plateform
stringclasses 1
value | repo_name
stringlengths 13
113
| name
stringlengths 3
74
| ext
stringclasses 1
value | path
stringlengths 12
229
| size
int64 23
843k
| source_encoding
stringclasses 9
values | md5
stringlengths 32
32
| text
stringlengths 23
843k
|
|---|---|---|---|---|---|---|---|---|
github
|
rising-turtle/slam_matlab-master
|
fitline3d.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/fitline3d.m
| 1,200 |
utf_8
|
3ca7a782577d8447e17c36ccf25f3d59
|
% FITLINE3D - Fits a line to a set of 3D points
%
% Usage: [L] = fitline3d(XYZ)
%
% Where: XYZ - 3xNpts array of XYZ coordinates
% [x1 x2 x3 ... xN;
% y1 y2 y3 ... yN;
% z1 z2 z3 ... zN]
%
% Returns: L - 3x2 matrix consisting of the two endpoints of the line
% that fits the points. The line is centered about the
% mean of the points, and extends in the directions of the
% principal eigenvectors, with scale determined by the
% eigenvalues.
%
% Author: Felix Duvallet (CMU)
% August 2006
function L = fitline3d(XYZ)
% Since the covariance matrix should be 3x3 (not NxN), need
% to take the transpose of the points.
XYZ = XYZ';
% find mean of the points
mu = mean(XYZ, 1);
% covariance matrix
C = cov(XYZ);
% get the eigenvalues and eigenvectors
[V, D] = eig(C);
% largest eigenvector is in the last column
col = size(V, 2); %get the number of columns
% get the last eigenvector column and the last eigenvalue
eVec = V(:, col);
eVal = D(col, col);
% start point - center about mean and scale eVector by eValue
L(:, 1) = mu' - sqrt(eVal)*eVec;
% end point
L(:, 2) = mu' + sqrt(eVal)*eVec;
|
github
|
rising-turtle/slam_matlab-master
|
ransac_plane.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/ransac_plane.m
| 9,877 |
utf_8
|
6161d8cc1a602a9c2796433f55d7b6dd
|
% RANSAC - Robustly fits a model to data with the RANSAC algorithm
%
% Usage:
%
% [M, inliers] = ransac(x, fittingfn, distfn, degenfn s, t, feedback, ...
% maxDataTrials, maxTrials)
%
% Arguments:
% x - Data sets to which we are seeking to fit a model M
% It is assumed that x is of size [d x Npts]
% where d is the dimensionality of the data and Npts is
% the number of data points.
%
% fittingfn - Handle to a function that fits a model to s
% data from x. It is assumed that the function is of the
% form:
% M = fittingfn(x)
% Note it is possible that the fitting function can return
% multiple models (for example up to 3 fundamental matrices
% can be fitted to 7 matched points). In this case it is
% assumed that the fitting function returns a cell array of
% models.
% If this function cannot fit a model it should return M as
% an empty matrix.
%
% distfn - Handle to a function that evaluates the
% distances from the model to data x.
% It is assumed that the function is of the form:
% [inliers, M] = distfn(M, x, t)
% This function must evaluate the distances between points
% and the model returning the indices of elements in x that
% are inliers, that is, the points that are within distance
% 't' of the model. Additionally, if M is a cell array of
% possible models 'distfn' will return the model that has the
% most inliers. If there is only one model this function
% must still copy the model to the output. After this call M
% will be a non-cell object representing only one model.
%
% degenfn - Handle to a function that determines whether a
% set of datapoints will produce a degenerate model.
% This is used to discard random samples that do not
% result in useful models.
% It is assumed that degenfn is a boolean function of
% the form:
% r = degenfn(x)
% It may be that you cannot devise a test for degeneracy in
% which case you should write a dummy function that always
% returns a value of 1 (true) and rely on 'fittingfn' to return
% an empty model should the data set be degenerate.
%
% s - The minimum number of samples from x required by
% fittingfn to fit a model.
%
% t - The distance threshold between a data point and the model
% used to decide whether the point is an inlier or not.
%
% feedback - An optional flag 0/1. If set to one the trial count and the
% estimated total number of trials required is printed out at
% each step. Defaults to 0.
%
% maxDataTrials - Maximum number of attempts to select a non-degenerate
% data set. This parameter is optional and defaults to 100.
%
% maxTrials - Maximum number of iterations. This parameter is optional and
% defaults to 1000.
%
% Returns:
% M - The model having the greatest number of inliers.
% inliers - An array of indices of the elements of x that were
% the inliers for the best model.
%
% For an example of the use of this function see RANSACFITHOMOGRAPHY or
% RANSACFITPLANE
% References:
% M.A. Fishler and R.C. Boles. "Random sample concensus: A paradigm
% for model fitting with applications to image analysis and automated
% cartography". Comm. Assoc. Comp, Mach., Vol 24, No 6, pp 381-395, 1981
%
% Richard Hartley and Andrew Zisserman. "Multiple View Geometry in
% Computer Vision". pp 101-113. Cambridge University Press, 2001
% Copyright (c) 2003-2006 Peter Kovesi
% School of Computer Science & Software Engineering
% The University of Western Australia
% pk at csse uwa edu au
% http://www.csse.uwa.edu.au/~pk
%
% Permission is hereby granted, free of charge, to any person obtaining a copy
% of this software and associated documentation files (the "Software"), to deal
% in the Software without restriction, subject to the following conditions:
%
% The above copyright notice and this permission notice shall be included in
% all copies or substantial portions of the Software.
%
% The Software is provided "as is", without warranty of any kind.
%
% May 2003 - Original version
% February 2004 - Tidied up.
% August 2005 - Specification of distfn changed to allow model fitter to
% return multiple models from which the best must be selected
% Sept 2006 - Random selection of data points changed to ensure duplicate
% points are not selected.
% February 2007 - Jordi Ferrer: Arranged warning printout.
% Allow maximum trials as optional parameters.
% Patch the problem when non-generated data
% set is not given in the first iteration.
% August 2008 - 'feedback' parameter restored to argument list and other
% breaks in code introduced in last update fixed.
% December 2008 - Octave compatibility mods
% June 2009 - Argument 'MaxTrials' corrected to 'maxTrials'!
function [M, inliers] = ransac(x, fittingfn, distfn, degenfn, s, t, feedback, ...
maxDataTrials, maxTrials)
Octave = exist('OCTAVE_VERSION') ~= 0;
% Test number of parameters
error ( nargchk ( 6, 9, nargin ) );
if nargin < 9; maxTrials = 1000; end;
if nargin < 8; maxDataTrials = 100; end;
if nargin < 7; feedback = 0; end;
[rows, npts] = size(x);
p = 0.99; % Desired probability of choosing at least one sample
% free from outliers
bestM = NaN; % Sentinel value allowing detection of solution failure.
trialcount = 0;
bestscore = 0;
N = 1; % Dummy initialisation for number of trials.
while N > trialcount
% Select at random s datapoints to form a trial model, M.
% In selecting these points we have to check that they are not in
% a degenerate configuration.
degenerate = 1;
count = 1;
while degenerate
% Generate s random indicies in the range 1..npts
% (If you do not have the statistics toolbox, or are using Octave,
% use the function RANDOMSAMPLE from my webpage)
if Octave | ~exist('randsample.m')
ind = randomsample(npts, s);
else
ind = randsample(npts, s);
end
% Test that these points are not a degenerate configuration.
degenerate = feval(degenfn, x(:,ind));
if ~degenerate
% Fit model to this random selection of data points.
% Note that M may represent a set of models that fit the data in
% this case M will be a cell array of models
M = feval(fittingfn, x(:,ind));
% Depending on your problem it might be that the only way you
% can determine whether a data set is degenerate or not is to
% try to fit a model and see if it succeeds. If it fails we
% reset degenerate to true.
if isempty(M)
degenerate = 1;
end
end
% Safeguard against being stuck in this loop forever
count = count + 1;
if count > maxDataTrials
warning('Unable to select a nondegenerate data set');
break
end
end
% Once we are out here we should have some kind of model...
% Evaluate distances between points and model returning the indices
% of elements in x that are inliers. Additionally, if M is a cell
% array of possible models 'distfn' will return the model that has
% the most inliers. After this call M will be a non-cell object
% representing only one model.
[inliers, M] = feval(distfn, M, x, t);
% Find the number of inliers to this model.
ninliers = length(inliers);
if ninliers > bestscore % Largest set of inliers so far...
bestscore = ninliers; % Record data for this model
bestinliers = inliers;
bestM = M;
% Update estimate of N, the number of trials to ensure we pick,
% with probability p, a data set with no outliers.
fracinliers = ninliers/npts;
pNoOutliers = 1 - fracinliers^s;
pNoOutliers = max(eps, pNoOutliers); % Avoid division by -Inf
pNoOutliers = min(1-eps, pNoOutliers);% Avoid division by 0.
N = log(1-p)/log(pNoOutliers);
end
trialcount = trialcount+1;
if feedback
fprintf('trial %d out of %d \r',trialcount, ceil(N));
end
% Safeguard against being stuck in this loop forever
if trialcount > maxTrials
warning( ...
sprintf('ransac reached the maximum number of %d trials',...
maxTrials));
break
end
end
fprintf('\n');
if ~isnan(bestM) % We got a solution
M = bestM;
inliers = bestinliers;
else
M = [];
inliers = [];
error('ransac was unable to find a useful solution');
end
|
github
|
rising-turtle/slam_matlab-master
|
Kabsch.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/Kabsch.m
| 22,003 |
utf_8
|
d9ea79a024f3c2dd185841ba44642f29
|
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<html lang="en">
<head>
<meta http-equiv="X-UA-Compatible" content="IE=8" />
<script type="text/javascript" src="/includes_content/nextgen/scripts/jquery/jquery-latest.js"></script>
<!-- START OF GLOBAL NAV -->
<link rel="stylesheet" href="/matlabcentral/css/sitewide.css" type="text/css" />
<link rel="stylesheet" href="/matlabcentral/css/mlc.css" type="text/css" />
<!--[if lt IE 7]>
<link href="/matlabcentral/css/ie6down.css" type="text/css" rel="stylesheet">
<![endif]-->
<meta http-equiv="content-type" content="text/html; charset=UTF-8" />
<meta name="keywords" content="file exchange, matlab answers, newsgroup access, link exchange, matlab blog, matlab central, simulink blog, matlab community, matlab and simulink community" />
<meta name="description" content="File exchange, MATLAB Answers, newsgroup access, Links, and Blogs for the MATLAB & Simulink user community" />
<link rel="stylesheet" href="/matlabcentral/css/fileexchange.css" type="text/css" />
<link rel="stylesheet" type="text/css" media="print" href="/matlabcentral/css/print.css" />
<title>Kabsch algorithm: Kabsch(P, Q, m) - File Exchange - MATLAB Central</title>
<script src="/matlabcentral/fileexchange/javascripts/prototype.js?1333472917" type="text/javascript"></script>
<script src="/matlabcentral/fileexchange/javascripts/effects.js?1333472917" type="text/javascript"></script>
<script src="/matlabcentral/fileexchange/javascripts/dragdrop.js?1333472917" type="text/javascript"></script>
<script src="/matlabcentral/fileexchange/javascripts/controls.js?1333472917" type="text/javascript"></script>
<script src="/matlabcentral/fileexchange/javascripts/application.js?1333472917" type="text/javascript"></script>
<script src="/matlabcentral/fileexchange/javascripts/searchfield.js?1333472917" type="text/javascript"></script>
<link href="/matlabcentral/fileexchange/stylesheets/application.css?1333472917" media="screen" rel="stylesheet" type="text/css" />
<link rel="search" type="application/opensearchdescription+xml" title="Search File Exchange" href="/matlabcentral/fileexchange/search.xml" />
</head>
<body>
<div id="header">
<div class="wrapper">
<!--put nothing in left div - only 11px wide shadow -->
<div class="main">
<div id="logo"><a href="/matlabcentral/" title="MATLAB Central Home"><img src="/matlabcentral/images/mlclogo-whitebgd.gif" alt="MATLAB Central" /></a></div>
<div id="headertools">
<script language="JavaScript1.3" type="text/javascript">
function submitForm(query){
choice = document.forms['searchForm'].elements['search_submit'].value;
if (choice == "entire1" || choice == "contest" || choice == "matlabcentral" || choice == "blogs"){
var newElem = document.createElement("input");
newElem.type = "hidden";
newElem.name = "q";
newElem.value = query.value;
document.forms['searchForm'].appendChild(newElem);
submit_action = '/searchresults/';
}
switch(choice){
case "matlabcentral":
var newElem = document.createElement("input");
newElem.type = "hidden";
newElem.name = "c[]";
newElem.value = "matlabcentral";
document.forms['searchForm'].appendChild(newElem);
selected_index = 0;
break
case "fileexchange":
var newElem = document.createElement("input");
newElem.type = "hidden";
newElem.name = "term";
newElem.value = query.value;
newElem.classname = "formelem";
document.forms['searchForm'].appendChild(newElem);
submit_action = "/matlabcentral/fileexchange/";
selected_index = 1;
break
case "answers":
var newElem = document.createElement("input");
newElem.type = "hidden";
newElem.name = "term";
newElem.value = query.value;
newElem.classname = "formelem";
document.forms['searchForm'].appendChild(newElem);
submit_action = "/matlabcentral/answers/";
selected_index = 2;
break
case "cssm":
var newElem = document.createElement("input");
newElem.type = "hidden";
newElem.name = "search_string";
newElem.value = query.value;
newElem.classname = "formelem";
document.forms['searchForm'].appendChild(newElem);
submit_action = "/matlabcentral/newsreader/search_results";
selected_index = 3;
break
case "linkexchange":
var newElem = document.createElement("input");
newElem.type = "hidden";
newElem.name = "term";
newElem.value = query.value;
newElem.classname = "formelem";
document.forms['searchForm'].appendChild(newElem);
submit_action = "/matlabcentral/linkexchange/";
selected_index = 4;
break
case "blogs":
var newElem = document.createElement("input");
newElem.type = "hidden";
newElem.name = "c[]";
newElem.value = "blogs";
document.forms['searchForm'].appendChild(newElem);
selected_index = 5;
break
case "trendy":
var newElem = document.createElement("input");
newElem.type = "hidden";
newElem.name = "search";
newElem.value = query.value;
newElem.classname = "formelem";
document.forms['searchForm'].appendChild(newElem);
submit_action = "/matlabcentral/trendy";
selected_index = 6;
break
case "cody":
var newElem = document.createElement("input");
newElem.type = "hidden";
newElem.name = "term";
newElem.value = query.value;
newElem.classname = "formelem";
document.forms['searchForm'].appendChild(newElem);
submit_action = "/matlabcentral/cody/";
selected_index = 7;
break
case "contest":
var newElem = document.createElement("input");
newElem.type = "hidden";
newElem.name = "c[]";
newElem.value = "contest";
document.forms['searchForm'].appendChild(newElem);
selected_index = 8;
break
case "entire1":
var newElem = document.createElement("input");
newElem.type = "hidden";
newElem.name = "c[]";
newElem.value = "entiresite";
document.forms['searchForm'].appendChild(newElem);
selected_index = 9;
break
default:
var newElem = document.createElement("input");
newElem.type = "hidden";
newElem.name = "c[]";
newElem.value = "entiresite";
document.forms['searchForm'].appendChild(newElem);
selected_index = 9;
break
}
document.forms['searchForm'].elements['search_submit'].selectedIndex = selected_index;
document.forms['searchForm'].elements['query'].value = query.value;
document.forms['searchForm'].action = submit_action;
}
</script>
<form name="searchForm" method="GET" action="" style="margin:0px; margin-top:5px; font-size:90%" onsubmit="submitForm(query)">
<label for="search">Search: </label>
<select name="search_submit" style="font-size:9px ">
<option value="matlabcentral">MATLAB Central</option>
<option value="fileexchange" selected> File Exchange</option>
<option value="answers"> Answers</option>
<option value="cssm"> Newsgroup</option>
<option value="linkexchange"> Link Exchange</option>
<option value="blogs"> Blogs</option>
<option value="trendy"> Trendy</option>
<option value="cody"> Cody</option>
<option value="contest"> Contest</option>
<option value="entire1">MathWorks.com</option>
</select>
<input type="text" name="query" size="10" class="formelem" value="" />
<input type="submit" value="Go" class="formelem gobutton" />
</form>
<ol id="access2">
<li class="first">
Amirhossein Tamjidi
</li>
<li><a href="/accesslogin/editCommunityProfile.do?uri=/matlabcentral/fileexchange/25746-kabsch-algorithm/content/Kabsch.m">Create Community Profile</a></li>
<li><a href="https://www.mathworks.com/accesslogin/logout.do?uri=http://www.mathworks.com/matlabcentral/fileexchange/25746-kabsch-algorithm/content/Kabsch.m">Log Out</a></li>
</ol>
</div>
<div id="globalnav">
<!-- from includes/global_nav.html -->
<ol>
<li class="active">
<a href="/matlabcentral/fileexchange/">File Exchange</a>
</li>
<li class=";">
<a href="/matlabcentral/answers/">Answers</a>
</li>
<li class=";">
<a href="/matlabcentral/newsreader/">Newsgroup</a>
</li>
<li class=";">
<a href="/matlabcentral/linkexchange/">Link Exchange</a>
</li>
<li class=";">
<a href="http://blogs.mathworks.com/">Blogs</a>
</li>
<li class=";">
<a href="/matlabcentral/trendy">Trendy</a>
</li>
<li class=";">
<a href="/matlabcentral/cody">Cody</a>
</li>
<li class=";">
<a href="/matlabcentral/contest/">Contest</a>
</li>
<li class="icon mathworks">
<a href="/">MathWorks.com</a>
</li>
</ol>
</div>
</div>
</div>
</div>
<div id="middle">
<div class="wrapper">
<div id="mainbody" class="columns2">
<div class="manifest">
<div class="ctaBtn ctaBlueBtn btnSmall">
<div class="btnCont">
<div class="btn download">
<a href="http://www.mathworks.com/matlabcentral/fileexchange/25746-kabsch-algorithm?controller=file_infos&download=true" title="Download Now">Download All</a>
</div>
</div>
</div>
<p class="license">
Code covered by the <a href="/matlabcentral/fileexchange/view_license?file_info_id=25746" onclick="window.open(this.href,'new_window','height=500,width=640,scrollbars=yes');return false;">BSD License</a>
<a href="/matlabcentral/fileexchange/help_license#bsd" class="info notext" onclick="window.open(this.href,'small','toolbar=no,resizable=yes,status=yes,menu=no,scrollbars=yes,width=600,height=550');return false;"> </a>
</p>
<h3 class="highlights_title">
Highlights from<br/>
<a href="http://www.mathworks.com/matlabcentral/fileexchange/25746-kabsch-algorithm" class="manifest_title">Kabsch algorithm</a>
</h3>
<ul class="manifest">
<li class="manifest">
<a href="http://www.mathworks.com/matlabcentral/fileexchange/25746-kabsch-algorithm/content/Kabsch.m" class="function" title="Function">Kabsch(P, Q, m)</a>
<span class="description">Find the Least Root Mean Square between two sets of N points in D dimensions</span>
</li>
<li class="manifest_allfiles">
<a href="http://www.mathworks.com/matlabcentral/fileexchange/25746-kabsch-algorithm/all_files">View all files</a></li>
</ul>
</div>
<table cellpadding="0" cellspacing="0" class="details file">
<tr>
<th class="maininfo">
<div id="details" class="content_type_details">
from
<a href="/matlabcentral/fileexchange/file_infos/25746-kabsch-algorithm" class="content_type_author">Kabsch algorithm</a></h2>
by <a href="/matlabcentral/fileexchange/authors/75295">Ehud Schreiber</a>
<br/>Find the rigid transformation & Least Root Mean Square distance between two paired sets of points
</p>
</div>
</th>
</tr>
<tr>
<td class="file">
<table cellpadding="0" cellspacing="0" border="0" class="fileview section">
<tr class="title">
<th><span class="heading">Kabsch(P, Q, m)</span></th>
</tr>
<tr>
<td>
<div class="codecontainer"><pre class="matlab-code">% Find the Least Root Mean Square between two sets of N points in D dimensions
% and the rigid transformation (i.e. translation and rotation)
% to employ in order to bring one set that close to the other,
% Using the Kabsch (1976) algorithm.
% Note that the points are paired, i.e. we know which point in one set
% should be compared to a given point in the other set.
%
% References:
% 1) Kabsch W. A solution for the best rotation to relate two sets of vectors. Acta Cryst A 1976;32:9223.
% 2) Kabsch W. A discussion of the solution for the best rotation to relate two sets of vectors. Acta Cryst A 1978;34:8278.
% 3) http://cnx.org/content/m11608/latest/
% 4) http://en.wikipedia.org/wiki/Kabsch_algorithm
%
% We slightly generalize, allowing weights given to the points.
% Those weights are determined a priori and do not depend on the distances.
%
% We work in the convention that points are column vectors;
% some use the convention where they are row vectors instead.
%
% Input variables:
% P : a D*N matrix where P(a,i) is the a-th coordinate of the i-th point
% in the 1st representation
% Q : a D*N matrix where Q(a,i) is the a-th coordinate of the i-th point
% in the 2nd representation
% m : (Optional) a row vector of length N giving the weights, i.e. m(i) is
% the weight to be assigned to the deviation of the i-th point.
% If not supplied, we take by default the unweighted (or equal weighted)
% m(i) = 1/N.
% The weights do not have to be normalized;
% we divide by the sum to ensure sum_{i=1}^N m(i) = 1.
% The weights must be non-negative with at least one positive entry.
% Output variables:
% U : a proper orthogonal D*D matrix, representing the rotation
% r : a D-dimensional column vector, representing the translation
% lrms: the Least Root Mean Square
%
% Details:
% If p_i, q_i are the i-th point (as a D-dimensional column vector)
% in the two representations, i.e. p_i = P(:,i) etc., and for
% p_i' = U p_i + r (' does not stand for transpose!)
% we have p_i' ~ q_i, that is,
% lrms = sqrt(sum_{i=1}^N m(i) (p_i' - q_i)^2)
% is the minimal rms when going over the possible U and r.
% (assuming the weights are already normalized).
%
function[U, r, lrms] = Kabsch(P, Q, m)
sz1 = size(P) ;
sz2 = size(Q) ;
if (length(sz1) ~= 2 || length(sz2) ~= 2)
error 'P and Q must be matrices' ;
end
if (any(sz1 ~= sz2))
error 'P and Q must be of same size' ;
end
D = sz1(1) ; % dimension of space
N = sz1(2) ; % number of points
if (nargin >= 3)
if (~isvector(m) || any(size(m) ~= [1 N]))
error 'm must be a row vector of length N' ;
end
if (any(m < 0))
error 'm must have non-negative entries' ;
end
msum = sum(m) ;
if (msum == 0)
error 'm must contain some positive entry' ;
end
m = m / msum ; % normalize so that weights sum to 1
else % m not supplied - use default
m = ones(1,N)/N ;
end
p0 = P*m' ; % the centroid of P
q0 = Q*m' ; % the centroid of Q
v1 = ones(1,N) ; % row vector of N ones
P = P - p0*v1 ; % translating P to center the origin
Q = Q - q0*v1 ; % translating Q to center the origin
% C is a covariance matrix of the coordinates
% C = P*diag(m)*Q'
% but this is inefficient, involving an N*N matrix, while typically D << N.
% so we use another way to compute Pdm = P*diag(m)
Pdm = zeros(D,N) ;
for i=1:N
Pdm(:,i) = m(i)*P(:,i) ;
end
C = Pdm*Q' ;
% C = P*Q' / N ; % (for the non-weighted case)
[V,S,W] = svd(C) ; % singular value decomposition
I = eye(D) ;
if (det(C) < 0)
I(D,D) = -1 ;
end
U = W*I*V' ;
r = q0 - U*p0 ;
Diff = U*P - Q ; % P, Q already centered
% lrms = sqrt(sum(sum(Diff.*Diff))/N) ; % (for the non-weighted case)
lrms = 0 ;
for i=1:N
lrms = lrms + m(i)*Diff(:,i)'*Diff(:,i) ;
end
lrms = sqrt(lrms) ;
end
</pre></div>
</td>
</tr>
</table>
</td>
</tr>
</table>
<p id="contactus">Contact us at <a href="mailto:[email protected]">[email protected]</a></p>
</div>
<div class="clearboth"> </div>
</div>
</div>
<!-- footer.html -->
<!-- START OF FOOTER -->
<div id="mlc-footer">
<script type="text/javascript">
function clickDynamic(obj, target_url, tracking_code) {
var pos=target_url.indexOf("?");
if (pos<=0) {
var linkComponents = target_url + tracking_code;
obj.href=linkComponents;
}
}
</script>
<div class="wrapper">
<div>
<ul id="matlabcentral">
<li class="copyright first">© 1994-2012 The MathWorks, Inc.</li>
<li class="first"><a href="/help.html" title="Site Help">Site Help</a></li>
<li><a href="/company/aboutus/policies_statements/patents.html" title="patents" rel="nofollow">Patents</a></li>
<li><a href="/company/aboutus/policies_statements/trademarks.html" title="trademarks" rel="nofollow">Trademarks</a></li>
<li><a href="/company/aboutus/policies_statements/" title="privacy policy" rel="nofollow">Privacy Policy</a></li>
<li><a href="/company/aboutus/policies_statements/piracy.html" title="preventing piracy" rel="nofollow">Preventing Piracy</a></li>
<li class="last"><a href="/matlabcentral/termsofuse.html" title="Terms of Use" rel="nofollow">Terms of Use</a></li>
<li class="icon"><a href="/company/rss/" title="RSS" class="rssfeed" rel="nofollow"><span class="text">RSS</span></a></li>
<li class="icon"><a href="http://www.mathworks.com/programs/bounce_hub_generic.html?s_cid=mlc_fbk&url=http://www.facebook.com/MATLAB" title="Facebook" class="facebook" rel="nofollow"><span class="text">Facebook</span></a></li>
<li class="last icon"><a href="http://www.mathworks.com/programs/bounce_hub_generic.html?s_cid=mlc_twt&url=http://www.twitter.com/MATLAB" title="Twitter" class="twitter" rel="nofollow"><span class="text">Twitter</span></a></li>
</ul>
<ul id="mathworks">
<li class="first sectionhead">Featured MathWorks.com Topics:</li>
<li class="first"><a href="/products/new_products/latest_features.html" onclick="clickDynamic(this, this.href, '?s_cid=MLC_new')">New Products</a></li>
<li><a href="/support/" title="support" onclick="clickDynamic(this, this.href, '?s_cid=MLC_support')">Support</a></li>
<li><a href="/help" title="documentation" onclick="clickDynamic(this, this.href, '?s_cid=MLC_doc')">Documentation</a></li>
<li><a href="/services/training/" title="training" onclick="clickDynamic(this, this.href, '?s_cid=MLC_training')">Training</a></li>
<li><a href="/company/events/webinars/" title="Webinars" onclick="clickDynamic(this, this.href, '?s_cid=MLC_webinars')">Webinars</a></li>
<li><a href="/company/newsletters/" title="newsletters" onclick="clickDynamic(this, this.href, '?s_cid=MLC_newsletters')">Newsletters</a></li>
<li><a href="/programs/trials/trial_request.html?prodcode=ML&s_cid=MLC_trials" title="MATLAB Trials">MATLAB Trials</a></li>
<li class="last"><a href="/company/jobs/opportunities/index_en_US.html" title="Careers" onclick="clickDynamic(this, this.href, '?s_cid=MLC_careers')">Careers</a></li>
</ul>
</div>
</div>
</div>
<!-- END OF FOOTER -->
<!-- SiteCatalyst code version: H.4. -->
<script language="JavaScript" type="text/javascript" src="/includes_content/jscript/omniture/s_code.js"></script>
<script language="JavaScript" type="text/javascript"><!--
s.pageName=document.title
s.eVar13='Version B'; //code for testing two versions of the intro text on the homepage products tab
/************* DO NOT ALTER ANYTHING BELOW THIS LINE ! **************/
var s_code=s.t();if(s_code)document.write(s_code)//--></script>
<script language="JavaScript" type="text/javascript"><!--
if(navigator.appVersion.indexOf('MSIE')>=0)document.write(unescape('%3C')+'\!-'+'-')
//--></script><!--/DO NOT REMOVE/-->
<!-- End SiteCatalyst code version: H.4. -->
<!--JF: Jai and I added the following code for GA to test mobile devices. 12/23/10 -->
<script type="text/javascript">
var _gaq = _gaq || [];
_gaq.push(['_setAccount', 'UA-7506924-7']);
_gaq.push(['_setDomainName', 'none']);
_gaq.push(['_setAllowLinker', true]);
var whichSection = location.href.split("/");
if ((whichSection[3]) && (whichSection[3].lastIndexOf(".html") == -1)) {
_gaq.push(['_setCustomVar', 1, 'Section1', whichSection[3], 3]);
// alert(whichSection[3]);
}
if ((whichSection[4]) && (whichSection[4].lastIndexOf(".html") == -1)) {
whichSectionCombined = whichSection[3] + "/" + whichSection[4];
_gaq.push(['_setCustomVar', 2, 'Section2', whichSectionCombined, 3]);
// alert(whichSectionCombined);
}
if ((whichSection[5]) && (whichSection[5].lastIndexOf(".html") == -1)) {
whichSectionCombined = whichSection[3] + "/" + whichSection[4] + "/" + whichSection[5];
_gaq.push(['_setCustomVar', 3, 'Section3', whichSectionCombined, 3]);
//alert(whichSectionCombined);
}
_gaq.push(['_trackPageview']);
(function() {
var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true;
ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js';
var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s);
})();
</script>
</body>
</html>
|
github
|
rising-turtle/slam_matlab-master
|
fitplane.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/fitplane.m
| 1,694 |
utf_8
|
a058a2453754cff9ea42776f1034436a
|
% FITPLANE - solves coefficients of plane fitted to 3 or more points
%
% Usage: B = fitplane(XYZ)
%
% Where: XYZ - 3xNpts array of xyz coordinates to fit plane to.
% If Npts is greater than 3 a least squares solution
% is generated.
%
% Returns: B - 4x1 array of plane coefficients in the form
% b(1)*X + b(2)*Y +b(3)*Z + b(4) = 0
% The magnitude of B is 1.
%
% See also: RANSACFITPLANE
% Copyright (c) 2003-2005 Peter Kovesi
% School of Computer Science & Software Engineering
% The University of Western Australia
% http://www.csse.uwa.edu.au/
%
% Permission is hereby granted, free of charge, to any person obtaining a copy
% of this software and associated documentation files (the "Software"), to deal
% in the Software without restriction, subject to the following conditions:
%
% The above copyright notice and this permission notice shall be included in
% all copies or substantial portions of the Software.
%
% The Software is provided "as is", without warranty of any kind.
% June 2003
function B = fitplane(XYZ)
[rows,npts] = size(XYZ);
if rows ~=3
error('data is not 3D');
end
if npts < 3
error('too few points to fit plane');
end
% Set up constraint equations of the form AB = 0,
% where B is a column vector of the plane coefficients
% in the form b(1)*X + b(2)*Y +b(3)*Z + b(4) = 0.
A = [XYZ' ones(npts,1)]; % Build constraint matrix
if npts == 3 % Pad A with zeros
A = [A; zeros(1,4)];
end
[u d v] = svd(A); % Singular value decomposition.
B = v(:,4); % Solution is last column of v.
|
github
|
rising-turtle/slam_matlab-master
|
iscolinear.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/iscolinear.m
| 2,318 |
utf_8
|
65025b7413f8f6b4cb16dd1689a5900f
|
% ISCOLINEAR - are 3 points colinear
%
% Usage: r = iscolinear(p1, p2, p3, flag)
%
% Arguments:
% p1, p2, p3 - Points in 2D or 3D.
% flag - An optional parameter set to 'h' or 'homog'
% indicating that p1, p2, p3 are homogneeous
% coordinates with arbitrary scale. If this is
% omitted it is assumed that the points are
% inhomogeneous, or that they are homogeneous with
% equal scale.
%
% Returns:
% r = 1 if points are co-linear, 0 otherwise
% Copyright (c) 2004-2005 Peter Kovesi
% School of Computer Science & Software Engineering
% The University of Western Australia
% http://www.csse.uwa.edu.au/
%
% Permission is hereby granted, free of charge, to any person obtaining a copy
% of this software and associated documentation files (the "Software"), to deal
% in the Software without restriction, subject to the following conditions:
%
% The above copyright notice and this permission notice shall be included in
% all copies or substantial portions of the Software.
%
% The Software is provided "as is", without warranty of any kind.
% February 2004
% January 2005 - modified to allow for homogeneous points of arbitrary
% scale (thanks to Michael Kirchhof)
function r = iscolinear(p1, p2, p3, flag)
if nargin == 3 % Assume inhomogeneous coords
flag = 'inhomog';
end
if ~all(size(p1)==size(p2)) | ~all(size(p1)==size(p3)) | ...
~(length(p1)==2 | length(p1)==3)
error('points must have the same dimension of 2 or 3');
end
% If data is 2D, assume they are 2D inhomogeneous coords. Make them
% homogeneous with scale 1.
if length(p1) == 2
p1(3) = 1; p2(3) = 1; p3(3) = 1;
end
if flag(1) == 'h'
% Apply test that allows for homogeneous coords with arbitrary
% scale. p1 X p2 generates a normal vector to plane defined by
% origin, p1 and p2. If the dot product of this normal with p3
% is zero then p3 also lies in the plane, hence co-linear.
r = abs(dot(cross(p1, p2),p3)) < eps;
else
% Assume inhomogeneous coords, or homogeneous coords with equal
% scale.
r = norm(cross(p2-p1, p3-p1)) < eps;
end
|
github
|
rising-turtle/slam_matlab-master
|
testfitplane.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/testfitplane.m
| 3,367 |
utf_8
|
422156cc86a4f56b427ec295e9306b63
|
% TESTFITPLANE - demonstrates RANSAC plane fitting
%
% Usage: testfitplane(outliers, sigma, t, feedback)
%
% Arguments:
% outliers - Fraction specifying how many points are to be
% outliers.
% sigma - Standard deviation of inlying points from the
% true plane.
% t - Distance threshold to be used by the RANSAC
% algorithm for deciding whether a point is an
% inlier.
% feedback - Optional flag 0 or 1 to turn on RANSAC feedback
% information.
%
% Try using: testfitplane(0.3, 0.05, 0.05)
%
% See also: RANSACFITPLANE, FITPLANE
% Copyright (c) 2003-2005 Peter Kovesi
% School of Computer Science & Software Engineering
% The University of Western Australia
% http://www.csse.uwa.edu.au/
%
% Permission is hereby granted, free of charge, to any person obtaining a copy
% of this software and associated documentation files (the "Software"), to deal
% in the Software without restriction, subject to the following conditions:
%
% The above copyright notice and this permission notice shall be included in
% all copies or substantial portions of the Software.
%
% The Software is provided "as is", without warranty of any kind.
% June 2003
function testfitplane(outliers, sigma, t, feedback)
if nargin == 3
feedback = 0;
end
% Hard wire some constants - vary these as you wish
npts = 100; % Number of 3D data points
% Define a plane ax + by + cz + d = 0
a = 10; b = -3; c = 5; d = 1;
B = [a b c d]';
B = B/norm(B);
outsigma = 30*sigma; % outlying points have a distribution that is
% 30 times as spread as the inlying points
vpts = round((1-outliers)*npts); % No of valid points
opts = npts - vpts; % No of outlying points
% Generate npts points in the plane
X = rand(1,npts);
Y = rand(1,npts);
Z = (-a*X -b*Y -d)/c;
XYZ = [X
Y
Z];
% Add uniform noise of +/-sigma
XYZ = XYZ + (2*rand(size(XYZ))-1)*sigma;
% Generate opts random outliers
n = length(XYZ);
ind = randperm(n); % get a random set of point indices
ind = ind(1:opts); % ... of length opts
% Add uniform noise of outsigma to the points chosen to be outliers.
% XYZ(:,ind) = XYZ(:,ind) + (2*rand(3,opts)-1)*outsigma;
XYZ(:,ind) = XYZ(:,ind) + sign(rand(3,opts)-.5).*(rand(3,opts)+1)*outsigma;
% Display the cloud of points
figure(1), clf, plot3(XYZ(1,:),XYZ(2,:),XYZ(3,:), 'r*');
% Perform RANSAC fitting of the plane
[Bfitted, P, inliers] = ransacfitplane(XYZ, t, feedback);
fprintf('Original plane coefficients: ');
fprintf('%8.3f ',B);
fprintf('\nFitted plane coefficients: ');
fprintf('%8.3f ',Bfitted);
fprintf('\n');
% Display the triangular patch formed by the 3 points that gave the
% plane of maximum consensus
patch(P(1,:), P(2,:), P(3,:), 'g')
box('on'), grid('on'), rotate3d('on')
fprintf('\nRotate image so that planar patch is seen edge on\n');
fprintf('If the fit has been successful the inlying points should\n');
fprintf('form a line\n\n');
|
github
|
rising-turtle/slam_matlab-master
|
ransacfitplane.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/ransacfitplane.m
| 4,285 |
utf_8
|
3a8b352c171a4ca6acb11d45a6c10caf
|
% RANSACFITPLANE - fits plane to 3D array of points using RANSAC
%
% Usage [B, P, inliers] = ransacfitplane(XYZ, t, feedback)
%
% This function uses the RANSAC algorithm to robustly fit a plane
% to a set of 3D data points.
%
% Arguments:
% XYZ - 3xNpts array of xyz coordinates to fit plane to.
% t - The distance threshold between data point and the plane
% used to decide whether a point is an inlier or not.
% feedback - Optional flag 0 or 1 to turn on RANSAC feedback
% information.
%
% Returns:
% B - 4x1 array of plane coefficients in the form
% b(1)*X + b(2)*Y +b(3)*Z + b(4) = 0
% The magnitude of B is 1.
% This plane is obtained by a least squares fit to all the
% points that were considered to be inliers, hence this
% plane will be slightly different to that defined by P below.
% P - The three points in the data set that were found to
% define a plane having the most number of inliers.
% The three columns of P defining the three points.
% inliers - The indices of the points that were considered
% inliers to the fitted plane.
%
% See also: RANSAC, FITPLANE
% Copyright (c) 2003-2008 Peter Kovesi
% School of Computer Science & Software Engineering
% The University of Western Australia
% http://www.csse.uwa.edu.au/
%
% Permission is hereby granted, free of charge, to any person obtaining a copy
% of this software and associated documentation files (the "Software"), to deal
% in the Software without restriction, subject to the following conditions:
%
% The above copyright notice and this permission notice shall be included in
% all copies or substantial portions of the Software.
%
% The Software is provided "as is", without warranty of any kind.
% June 2003 - Original version.
% Feb 2004 - Modified to use separate ransac function
% Aug 2005 - planeptdist modified to fit new ransac specification
% Dec 2008 - Much faster distance calculation in planeptdist (thanks to
% Alastair Harrison)
function [B, P, inliers] = ransacfitplane(XYZ, t, feedback)
if nargin == 2
feedback = 0;
end
[rows, npts] = size(XYZ);
if rows ~=3
error('data is not 3D');
end
if npts < 3
error('too few points to fit plane');
end
s = 3; % Minimum No of points needed to fit a plane.
fittingfn = @defineplane;
distfn = @planeptdist;
degenfn = @isdegenerate;
[P, inliers] = ransac_plane(XYZ, fittingfn, distfn, degenfn, s, t, feedback);
% Perform least squares fit to the inlying points
B = fitplane(XYZ(:,inliers));
%------------------------------------------------------------------------
% Function to define a plane given 3 data points as required by
% RANSAC. In our case we use the 3 points directly to define the plane.
function P = defineplane(X);
P = X;
%------------------------------------------------------------------------
% Function to calculate distances between a plane and a an array of points.
% The plane is defined by a 3x3 matrix, P. The three columns of P defining
% three points that are within the plane.
function [inliers, P] = planeptdist(P, X, t)
n = cross(P(:,2)-P(:,1), P(:,3)-P(:,1)); % Plane normal.
n = n/norm(n); % Make it a unit vector.
npts = length(X);
d = zeros(npts,1); % d will be an array of distance values.
% The following loop builds up the dot product between a vector from P(:,1)
% to every X(:,i) with the unit plane normal. This will be the
% perpendicular distance from the plane for each point
for i=1:3
d = d + (X(i,:)'-P(i,1))*n(i);
end
inliers = find(abs(d) < t);
%------------------------------------------------------------------------
% Function to determine whether a set of 3 points are in a degenerate
% configuration for fitting a plane as required by RANSAC. In this case
% they are degenerate if they are colinear.
function r = isdegenerate(X)
% The three columns of X are the coords of the 3 points.
r = iscolinear(X(:,1),X(:,2),X(:,3));
|
github
|
rising-turtle/slam_matlab-master
|
evaluate_result.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/all/evaluate_result.m
| 1,659 |
utf_8
|
e5ce33520dfb03572fe195820d79fe85
|
function evaluate_result
% de_y = get_y_error();
% fe = get_final_error();
[de_y,fe,fe_norm,fe_perecent,ye_percent,path] = get_error()
end
function [de_y,fe,fe_norm,fe_perecent,ye_percent,path] = get_error()
legend('off')
grid off
axis off
h = gcf;
axesObjs = get(h, 'Children');
dataObjs = get(axesObjs(1), 'Children');
objTypes = get(dataObjs(1), 'Type');
xdata1 = get(dataObjs(1), 'XData');
ydata1 = get(dataObjs(1), 'YData');
zdata1 = get(dataObjs(1), 'ZData');
path1 = 0;
for i =1:length(xdata1)-1
path1 = path1 + norm([xdata1(i+1);ydata1(i+1);zdata1(i+1)]-[xdata1(i);ydata1(i);zdata1(i)] );
end
xdata2 = get(dataObjs(2), 'XData');
ydata2 = get(dataObjs(2), 'YData');
zdata2 = get(dataObjs(2), 'ZData');
path2 = 0;
for i =1:length(xdata2)-1
path2 = path2 + norm([xdata2(i+1);ydata2(i+1);zdata2(i+1)]-[xdata2(i);ydata2(i);zdata2(i)] );
end
% figure;plot(abs(ydata1),'r');hold on;plot(abs(ydata2),'b')
de_y(1) = mean(abs(ydata1));
de_y(2) = mean(abs(ydata2));
fe(1,:) = [xdata1(end);ydata1(end);zdata1(end)]';
fe(2,:) = [xdata2(end);ydata2(end);zdata2(end)]';
fe_norm(1) = norm([xdata1(end);ydata1(end);zdata1(end)]);
fe_norm(2) = norm([xdata2(end);ydata2(end);zdata2(end)]);
fe_perecent(1) = fe_norm(1)/min(path1,path2)*100;
fe_perecent(2) = fe_norm(2)/min(path1,path2)*100;
ye_percent(1) = de_y(1)/path1*100;
ye_percent(2) = de_y(1)/path2*100;
path(1)=path1;
path(2)=path2;
end
% function fe = get_final_error()
% h = gcf;
% axesObjs = get(h, 'Children');
% dataObjs = get(axesObjs, 'Children');
% objTypes = get(dataObjs, 'Type');
% xdata = get(dataObjs, 'XData');
% ydata = get(dataObjs, 'YData');
% zdata = get(dataObjs, 'ZData');
% end
|
github
|
rising-turtle/slam_matlab-master
|
test_the_effect_of_orinetation_.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/all/test_the_effect_of_orinetation_.m
| 3,784 |
utf_8
|
080623a148f01b9929a71b510bfa1461
|
function [angle1,angle2,angle3,xx,trajectory,cov_info]=test_the_effect_of_orinetation_(snapshot_step,h_figure)
config_file_for_orientation
global myCONFIG
% create_step_in_source(2500)
if nargin==0
close all
snapshot_step = 700
h_figure = figure
end
VRO_result_file = [myCONFIG.PATH.SOURCE_FOLDER,'VRO_result/VRO_result.mat'];
PM_result_file = [myCONFIG.PATH.SOURCE_FOLDER,'PM_result/VRO_result.mat'];
nFiles = data_file_counting(myCONFIG.PATH.SOURCE_FOLDER,'d1')-1;
if ~exist(VRO_result_file,'file')
[xx,varargout] = Test_RANSAC_dead_reckoning_dr_ye(1,nFiles)
save(VRO_result_file,'xx')
else
load(VRO_result_file)
end
if ~exist(PM_result_file,'file')
i=5;
while (i)<nFiles
[x_k_k,p_k_k,dT1,dq1,q_expected] = read_snapshot(i);
trajectory(:,i) = x_k_k;
cov_info(:,:,i) = p_k_k;
i = i+1
end
save(PM_result_file,'trajectory','cov_info');
else
load(PM_result_file)
end
% % % % % % % % % % figure;plot3(trajectory(1,4:snapshot_step-1),...
% % % % % % % % % % trajectory(2,4:snapshot_step-1),...
% % % % % % % % % % trajectory(3,4:snapshot_step-1),'b')
% % % % % % % % % % hold on;
% % % % % % % % % % plot3(xx(1,4:snapshot_step-1),...
% % % % % % % % % % xx(2,4:snapshot_step-1),...
% % % % % % % % % % xx(3,4:snapshot_step-1),'r')
% % % % % % % % % % axis equal
% % % % % % % % % % grid on
% % % % % % % % % % legend('proposed method','VRO')
% % % % % % % % % % figure(h_figure)
% draw_camera( [trajectory(1:3,snapshot_step); trajectory(4:7,snapshot_step)], 'r' );hold on;
% draw_camera( [xx(1:3,snapshot_step); xx(4:7,snapshot_step)], 'b' );
% % % % % % % % % draw_camera( [[0;0;0]; trajectory(4:7,snapshot_step-1)], 'b' );hold on;
% % % % % % % % % draw_camera( [[0;0;0]; xx(4:7,snapshot_step-1)], 'r' );
% draw_camera( [[0;0;0]; [1;0;0;0]], 'g' );
% % % % % % % % % % axis equal
% % % % % % % % % % grid on
% legend('a','b')
% % % % % % % % % % legend('PM','VRO')
% first plane
R_PM = q2r(trajectory(4:7,snapshot_step-1));
p1_1 = [0 0 0];p2_1= [1 0 0];p3_1= [0 0 1];
[ a_0, b_0, c_0, d_0 ] = plane_exp2imp_3d ( p1_1, p2_1, p3_1 );
%%% normalize the representation
norm1 = norm([a_0,b_0,c_0]);
a_0 = a_0/norm1; b_0 = b_0/norm1; c_0 = c_0/norm1; d_0 = d_0/norm1;
f1 = R_PM(1,3);
g1 = R_PM(2,3);
h1 = R_PM(3,3);
angle1 = planes_imp_angle_line_3d ( a_0, b_0, c_0, d_0, f1, g1, h1 )*(180/pi);
%%%
R_VRO = q2r(xx(4:7,snapshot_step-1));
f2 = R_VRO(1,3);
g2 = R_VRO(2,3);
h2 = R_VRO(3,3);
angle2 = planes_imp_angle_line_3d ( a_0, b_0, c_0, d_0, f2, g2, h2 )*(180/pi);
[R_PF,T_PF] = plane_fit_to_data_save(snapshot_step);
R_PF = R_PF';
f3 = R_PF(1,3);
g3 = R_PF(2,3);
h3 = R_PF(3,3);
OUTPUT=SpinCalc('DCMtoEA231',R_PM,0.01,1);
temp = (180/pi)*normilize_angle_(OUTPUT*(pi/180))
R_compensate = e2r([0 OUTPUT(1)*pi/180 0]);
angle3 = planes_imp_angle_line_3d ( a_0, b_0, c_0, d_0, f3, g3, h3 )*(180/pi);
% % % % % % % % % % drawSpan(R_compensate'*R_PF(:,[1,3]), 'g');
disp('-------------')
disp(['proposed method = ',num2str(angle1)])
disp(['VRO = ',num2str(angle2)])
disp(['Plane fitting = ',num2str(angle3)])
% OUTPUT=SpinCalc('DCMtoEA123',R_VRO,0.01,1)
% OUTPUT=SpinCalc('DCMtoEA231',R_VRO,0.01,1)
% OUTPUT=SpinCalc('DCMtoEA213',R_VRO,0.01,1)
% OUTPUT=SpinCalc('DCMtoEA321',R_VRO,0.01,1)
end
function p = normilize_angle_(p)
for i=1:length(p)
if p(i)>pi
p(i) = p(i) - 2*pi;
end
end
end
% [cov_pose_shift,q_dpose,T_pose] = bootstrap_cov_calc(100,103);
% [d_euler,E] = q2eG(q_dpose,cov_pose_shift(4:7,4:7));
% euler_cov = (sqrt(diag(E))*180/pi);
% t_cov = sqrt(diag(cov_pose_shift(1:3,1:3)));
% disp((180/pi)*d_euler')
% disp(euler_cov')
%
% disp(T_pose')
% disp(t_cov')
% close all;test_the_effect_of_orinetation(120,figure)
|
github
|
rising-turtle/slam_matlab-master
|
scrollplot.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/all/scrollplot.m
| 59,195 |
utf_8
|
224671d57c562973a15b560590589614
|
function scrollHandles = scrollplot(varargin)
%SCROLLPLOT add scroll sub-window to the supplied plot handle
%
% scrollplot adds a scroll sub-window to any supplied plot handle(s).
% The user may specify initial view window parameters or use defaults.
% Dragging the side-bars or central patch modifies the respective parent
% axes limits interactively. Conversely, modifying the parent axes
% limits (with zoom, pan or programatically) modifies the corresponding
% scroll patch(es) accordingly. Works ok with log and reverse axes.
% Both X & Y scrolling are possible.
%
% Syntax:
% scrollHandles = scrollplot(plotHandles, propName,propValue,...)
%
% scrollplot(plotHandles) adds a scroll sub-window to the supplied
% plotHandles using default property values (see below).
% plotHandles may be any combination of axes and line/data handles.
% If plotHandles is not supplied then the current axes (<a href="matlab:help gca">gca</a>) is used.
%
% scrollplot(..., propName,propValue, ...) sets the property value(s)
% for the initial scroll view window. Property specification order does
% not matter. The following properties are supported (case-insensitive):
% - 'Axis' : string (default = 'X'; accepted values: 'X','Y','XY')
% - 'Min' : number (default = minimal value of actual plot data)
% sets the same value for both 'MinX' & 'MinY'
% - 'Max' : number (default = maximal value of actual plot data)
% sets the same value for both 'MaxX' & 'MaxY'
% - 'MinX','MinY': number (same as 'Min', but only for X or Y axis)
% - 'MaxX','MaxY': number (same as 'Max', but only for X or Y axis)
% - 'WindowSize' : number (default = entire range of actual plot data)
% sets the same value for 'WindowSizeX' & 'WindowSizeY'
% - 'WindowSizeX': number (same as 'WindowSize' but only for X axis)
% - 'WindowSizeY': number (same as 'WindowSize' but only for Y axis)
%
% scrollHandles = scrollplot(...) returns handle(s) to the scroll axes.
% The returned handles are regular axes with a few additional read-only
% properties:
% - 'ScrollSideBarHandles' - array of 2 handles to the scroll side-bars
% - 'ScrollPatchHandle' - handle to the central scroll patch
% - 'ScrollAxesHandle' - handle to the scroll axes =double(scrollHandles)
% - 'ParentAxesHandle' - handle to the parent axes
% - 'ScrollMin' - number
% - 'ScrollMax' - number
%
% Examples:
% scrollplot; % add scroll sub-window to the current axes (gca)
% scrollplot(plot(xdata,ydata), 'WindowSize',50); % plot with initial zoom
% scrollplot('Min',20, 'windowsize',70); % add x-scroll to current axes
% scrollplot([h1,h2], 'axis','xy'); % scroll both X&Y of 2 plot axes
% scrollplot('axis','xy', 'minx',20, 'miny',10); % separate scroll minima
%
% Notes:
% 1. Matlab 5: scrollplot might NOT work on Matlab versions earlier than 6 (R12)
% 2. Matlab 6: scrollplot is not interactive in zoom mode (ok in Matlab 7+)
% 3. Matlab 6: warnings are disabled as a side-effect (not in Matlab 7+)
% 4. scrollplot modifies the figure's WindowButtonMotionFcn callback
% 5. scrollplot works on 3D plots, but only X & Y axis are scrollable
%
% Warning:
% This code relies in [small] part on undocumented and unsupported
% Matlab functionality. It works on Matlab 6+, but use at your own risk!
%
% Bugs and suggestions:
% Please send to Yair Altman (altmany at gmail dot com)
%
% Change log:
% 2007-Jun-14: Enabled image exploration per suggestion by Joe Lotz; improved log axis-scaling behavior per suggestion by Fredric Moisy; added scroll visibility & deletion handlers; fixed minor error handling bug
% 2007-May-15: Added 'MinX' etc. params; clarified error msgs; added 'ParentAxesHandle' special prop; fixed 'xy' bugs
% 2007-May-14: Set focus on parent axes after scroll-axes creation; added special scroll props; allowed 'Axis'='xy'
% 2007-May-13: First version posted on MathWorks file exchange: <a href="http://www.mathworks.com/matlabcentral/fileexchange/loadFile.do?objectId=14984">http://www.mathworks.com/matlabcentral/fileexchange/loadFile.do?objectId=14984</a>
%
% See also:
% plot, gca
% Programming notes:
% 1. Listeners are set on parent axes's properties so that whenever
% any of them (xlim,ylim,parent,units,position) is modified, then
% so are the corresponding scroll axes properties.
% 2. To bypass the mode managers' (zoom, pan, ...) "hijack" of the
% WindowButtonUpFcn callback, we use the non-supported JavaFrame's
% AxisComponent MouseReleasedCallback: this doesn't work in Matlab 6
% so scrollplot is non-interactive in Matlab 6 during zoom mode.
% 3. To bypass the mode managers over the scroll axes (to ignore zoom/
% pan), we use the little-known 'ButtonDownFilter' mode property.
% 4. The special read-only properties in the returned scrollHandles were
% added using the undocumented schema.prop interface. Since these
% properties are not viewable (only accessible) in the regular axes
% handle but only in the handle(scrollHandles), the latter form is
% returned. If you need the regular (numeric) form, use either
% double(scrollHandles) or the new 'ScrollAxesHandle' read-only prop.
% License to use and modify this code is granted freely without warranty to all, as long as the original author is
% referenced and attributed as such. The original author maintains the right to be solely associated with this work.
% Programmed and Copyright by Yair M. Altman: altmany(at)gmail.com
% $Revision: 1.2 $ $Date: 2007/05/15 12:28:27 $
try
% Note: on some systems with Matlab 6, an OpenGL warning is displayed due to semi-
% transparent scroll patch. This may be safely ignored. Unfortunately, specific warning
% disabling was not yet available in Matlab 6 so we must turn off all warnings...
v = version;
if v(1)<='6'
warning off; %#ok for Matlab 6 compatibility
else
% Temporarily turn off log-axis warnings
oldWarn = warning('off','MATLAB:Axes:NegativeDataInLogAxis');
end
% Args check
[plotHandles, pvPairs] = parseparams(varargin);
if iscell(plotHandles)
plotHandles = [plotHandles{:}]; % cell2mat is not supported on old Matlab versions...
end
if isempty(plotHandles)
plotHandles = gca;
else
plotHandles = plotHandles(:); % ensure 1-D array of handles
end
% Ensure that all supplied handles are valid HG handles
if isempty(plotHandles) | ~all(ishandle(plotHandles)) %#ok for Matlab 6 compatibility (note that Matlab 6 did not have ishghandle())
myError('YMA:scrollplot:invalidHandle','invalid plot handle(s) passed to scrollplot');
end
% Get the list of axes handles (supplied handles may be axes or axes children)
validHandles = [];
try
for hIdx = 1 : length(plotHandles)
thisHandle = plotHandles(hIdx);
if ~strcmpi(get(thisHandle,'type'),'axes')
thisHandle = get(thisHandle,'Parent'); % old Matlab versions don't have ancestor()...
end
if ~strcmpi(get(thisHandle,'type'),'axes')
myError('YMA:scrollplot:invalidHandle','invalid plot handle passed to scrollplot - must be an axes or line/data handle');
end
validHandles = [validHandles, thisHandle]; %#ok mlint - preallocate
end
validHandles = unique(validHandles);
catch
% Probably not a valid axes/line, without a 'type' property (see isprop)
myError('YMA:scrollplot:invalidHandle','invalid plot handle(s) passed to scrollplot - must be an axes or line/data handle');
end
% Pre-process args necessary for creating the scroll-plots, if supplied
[pvPairs, axName] = preProcessArgs(pvPairs);
% For each unique axes, add the relevant scroll sub-plot
%try
scrollplotHandles = handle([]);
% Loop over all specified/inferred parent axes
for hIdx = 1 : length(validHandles)
% Loop over all requested scroll axes (x,y, or x&y) for this parent axes
hAx = validHandles(hIdx);
for axisIdx = 1 : length(axName)
% Add the new scroll plot axes
scrollplotHandles(end+1) = addScrollPlot(hAx,axName(axisIdx)); %#ok mlint - preallocate
% Process args, if supplied
processArgs(pvPairs,scrollplotHandles(end));
end
% Set the focus on the parent axes
axes(hAx);
end
%catch
% Probably not a valid axes handle
%myError('YMA:scrollplot:invalidHandle','invalid plot handle(s) passed to scrollplot - must be an axes or line/data handle');
%end
% If return scrollHandles was requested
if nargout
% Return the list of all scroll handles
scrollHandles = scrollplotHandles;
end
catch
v = version;
if v(1)<='6'
err.message = lasterr; % no lasterror function...
else
err = lasterror;
end
try
err.message = regexprep(err.message,'Error using ==> [^\n]+\n','');
catch
try
% Another approach, used in Matlab 6 (where regexprep is unavailable)
startIdx = findstr(err.message,'Error using ==> ');
stopIdx = findstr(err.message,char(10));
for idx = length(startIdx) : -1 : 1
idx2 = min(find(stopIdx > startIdx(idx))); %#ok ML6
err.message(startIdx(idx):stopIdx(idx2)) = [];
end
catch
% never mind...
end
end
if isempty(findstr(mfilename,err.message))
% Indicate error origin, if not already stated within the error message
err.message = [mfilename ': ' err.message];
end
if v(1)<='6'
while err.message(end)==char(10)
err.message(end) = []; % strip excessive Matlab 6 newlines
end
error(err.message);
else
rethrow(err);
end
end
% Restore original warnings (if available/possible)
try
warning(oldWarn);
catch
% never mind...
end
%end % scrollplot %#ok for Matlab 6 compatibility
%% Set-up a new scroll sub-plot window to the supplied axes handle
function hScroll = addScrollPlot(hAx,axName)
% Before modifying the original axes position, we must fix the labels (esp. xlabel)
hLabel = get(hAx, [axName 'Label']);
set(hLabel, 'units','normalized');
% Set a new scroll sub-plot in the bottom 10% of the original axes height
axPos = get(hAx,'position');
axVis = get(hAx,'visible');
axUnits = get(hAx,'units');
scaleStr = [axName 'Scale'];
dirStr = [axName 'Dir'];
limStr = [axName 'Lim'];
if strcmpi(axName,'x')
newScrollPos = axPos .* [1, 1, 1, 0.10];
newPlotPos = axPos .* [1, 1, 1, 0.80] + [0, 0.20*axPos(4), 0, 0];
specialStr = {'YTick',[]};
colStr = 'YColor'; % =axis line to hide
rotation = 0;
else % Y scroll
newScrollPos = [axPos(1)+axPos(3)*0.85, axPos(2), axPos(3)*0.1, axPos(4)];
newPlotPos = axPos .* [1, 1, 0.80, 1];
specialStr = {'XTick',[], 'YAxisLocation','right'};
colStr = 'XColor'; % =axis line to hide
rotation = 90;
end
hScroll = axes('units',axUnits, 'position',newScrollPos, 'visible',axVis, scaleStr,get(hAx,scaleStr), dirStr,get(hAx,dirStr), 'NextPlot','add', 'Box','off', specialStr{:}, 'FontSize',7, 'Tag','scrollAx', 'UserData',axName, 'DeleteFcn',@deleteScrollAx);
GRAY = 0.8 * [1,1,1];
try
bgColor = get(get(hAx,'parent'),'Color');
set(hScroll, colStr,bgColor);
catch
% Maybe the axes is contained in something without a 'Color' property
set(hScroll, colStr,GRAY);
end
%axis(hScroll, 'off');
set(hAx, 'position',newPlotPos);
% Store the parent axes in the scroll axes's appdata
setappdata(hScroll, 'parent',hAx);
% Set the scroll limits & data based on the original axes's data
% Note: use any axes child xdata to set the scroll limits, but
% ^^^^ only plot line children (not scatter/bar/polar etc)
axLines = get(hAx,'children');
%lim = [Inf, -Inf];
lim = get(hAx,limStr);
if isinf(lim(1)), lim(1)=+inf; end
if isinf(lim(2)), lim(2)=-inf; end
for lineIdx = 1 : length(axLines)
try
hLine = axLines(lineIdx);
xdata = get(hLine,'XData');
ydata = get(hLine,'YData');
try
name = get(hLine,'DisplayName');
catch
% Matlab 6 did not have 'DisplayName' property (used by legend) - never mind...
name = '';
end
if strcmpi(axName,'x'), data=xdata(1,:); else data=ydata(1,:); end
lim = [min(lim(1),min(data)), max(lim(2),max(data))];
linType = get(hLine,'type');
if strcmpi(linType,'line')
% Add plot line child if and only if it's a line
lineColor = GRAY; %=get(hLine,'color'); %orig color looks bad in scroll axes
hLine2 = plot(xdata, ydata, 'Parent',hScroll, 'color',lineColor, 'tag','scrollDataLine', 'HitTest','off');
if ~isempty(name)
set(hLine2, 'DisplayName',name);
end
% 2007-Jun-14: Enabled image exploration per suggestion by Joe Lotz
elseif strcmpi(linType,'image')
% Add miniature version of the main image
hLine2 = image(get(hLine,'CData'), 'Parent',hScroll); %#ok hLine2 used for debug
set(hScroll,'YDir','Reverse','XLim',get(hLine,'XData'),'YLim',get(hLine,'YData'));
end
catch
% Probably some axes child without data - skip it...
end
end
if lim(1) > lim(2)
curLim = get(hScroll, limStr);
if isinf(lim(1)), lim(1) = curLim(1); end
if isinf(lim(2)), lim(2) = curLim(2); end
end
if ~isempty(axLines) & lim(1) < lim(2) %#ok for Matlab 6 compatibility
set(hScroll, limStr,lim);
end
% Get the figure handle
hFig = ancestor(hAx,'figure');
% Prevent flicker on axes update
set(hFig, 'DoubleBuffer','on');
% Ensure that the axis component has a handle with callbacks
% Note: this is determined by the first invocation, so ensure we're the first...
axisComponent = getAxisComponent(hFig); %#ok unused
% Set the scroll handle-bars
xlim = get(hScroll, 'XLim');
ylim = get(hScroll, 'YLim');
hPatch = patch(xlim([1,1,2,2]), ylim([1,2,2,1]), 'b', 'FaceAlpha',.15, 'EdgeColor','w', 'EdgeAlpha',.15, 'ButtonDownFcn',@mouseDownCallback, 'tag','scrollPatch', 'userdata',axName); %Note: FaceAlpha causes an OpenGL warning in Matlab 6
commonProps = {'Parent',hScroll, 'LineWidth',3, 'ButtonDownFcn',@mouseDownCallback, 'tag','scrollBar'};
smallDelta = 0.01 * diff(lim); % don't use eps
if strcmpi(axName,'x')
hBars(1) = plot(xlim([1,1]), ylim, '-b', commonProps{:});
hBars(2) = plot(xlim([2,2]), ylim, '-b', commonProps{:});
else % Y scroll
hBars(1) = plot(xlim, ylim([1,1])+smallDelta, '-b', commonProps{:});
hBars(2) = plot(xlim, ylim([2,2])-smallDelta, '-b', commonProps{:});
end
try
set(hBars(1), 'DisplayName','Min');
set(hBars(2), 'DisplayName','Max');
catch
% Matlab 6 did not have 'DisplayName' property (used by legend) - never mind...
end
% TODO: maybe add a blue diamond or a visual handle in center of hBars?
set(hScroll, limStr,lim+smallDelta*[-1.2,1.2]);
% Help messages
msg = {'drag blue side-bars to zoom', 'drag central patch to pan'};
xText = getCenterCoord(hScroll, 'x');
yText = getCenterCoord(hScroll, 'y');
hText = text(xText,yText,msg, 'Color','r', 'Rotation',rotation, 'HorizontalAlignment','center', 'FontSize',9, 'FontWeight','bold', 'EraseMode','xor', 'HitTest','off', 'tag','scrollHelp'); %#ok ret val used for debug
hMenu = uicontextmenu;
set(hScroll, 'UIContextMenu',hMenu);
uimenu(hMenu, 'Label',msg{1}, 'Callback',@moveCursor, 'UserData',hBars(2));
uimenu(hMenu, 'Label',msg{2}, 'Callback',@moveCursor, 'UserData',hPatch);
% Set the mouse callbacks
winFcn = get(hFig,'WindowButtonMotionFcn');
if ~isempty(winFcn) & ~isequal(winFcn,@mouseMoveCallback) & (~iscell(winFcn) | ~isequal(winFcn{1},@mouseMoveCallback)) %#ok for Matlab 6 compatibility
setappdata(hFig, 'scrollplot_oldButtonMotionFcn',winFcn);
end
set(hFig,'WindowButtonMotionFcn',@mouseMoveCallback);
% Fix label position(s)
oldPos = get(hLabel, 'position');
if strcmpi(axName,'x')
if ~isempty(oldPos) & oldPos(2)<0 %#ok for Matlab 6 compatibility
% Only fix if the X label is on the bottom (usually yes)
set(hLabel, 'position',oldPos-[0,.20/.80,0]);
end
else % Y scroll
if ~isempty(oldPos) & oldPos(1)>0 %#ok for Matlab 6 compatibility
% Only fix if the Y label is on the right side (usually not)
set(hLabel, 'position',oldPos+[.20/.80,0,0]);
end
end
% Add property listeners
listenedPropNames = {'XLim','YLim','XDir','YDir','XScale','YScale','Position','Units','Parent'};
listeners = addPropListeners(hFig, hAx, hScroll, hPatch, hBars, listenedPropNames);
setappdata(hScroll, 'scrollplot_listeners',listeners); % These will be destroyed with hScroll so no need to un-listen upon hScroll deletion
% Add special properties
addSpecialProps(hAx, hScroll, hPatch, hBars, axName);
% Convert to handle object, so that the special properties become visible
hScroll = handle(hScroll);
return; % debug point
%end % addScrollPlot %#ok for Matlab 6 compatibility
%% Add parent axes listener
function listeners = addPropListeners(hFig, hAx, hScroll, hPatch, hBars, propNames)
% Listeners on parent axes properties
hhAx = handle(hAx);
for propIdx = 1 : length(propNames)
callback = {@parentAxesChanged, hFig, hAx, hPatch, hBars, propNames{propIdx}};
prop = findprop(hhAx, propNames{propIdx});
listeners(propIdx) = handle.listener(hhAx, prop, 'PropertyPostSet', callback); %#ok mlint - preallocate
end
% Listeners on scroll axes properties
hhScroll = handle(hScroll);
prop = findprop(hhScroll, 'Visible');
listeners(end+1) = handle.listener(hhScroll, prop, 'PropertyPostSet', {@updateParentPos,hScroll});
%end % addPropListeners %#ok for Matlab 6 compatibility
%% Add special scrollplot properties to the hScroll axes
function addSpecialProps(hAx, hScroll, hPatch, hBars, axName)
try
hhScroll = handle(hScroll);
% Read-only props
addNewProp(hhScroll,'ParentAxesHandle', hAx,1);
addNewProp(hhScroll,'ScrollAxesHandle', double(hScroll),1);
addNewProp(hhScroll,'ScrollPatchHandle', hPatch,1);
addNewProp(hhScroll,'ScrollSideBarHandles',hBars, 1);
% Note: setting the property's GetFunction is much cleaner but doesn't work in Matlab 6...
dataStr = [axName,'Data'];
addNewProp(hhScroll,'ScrollMin',unique(get(hBars(1),dataStr)),1); %,{@getBarVal,hBars(1),dataStr});
addNewProp(hhScroll,'ScrollMax',unique(get(hBars(2),dataStr)),1); %,{@getBarVal,hBars(2),dataStr});
catch
% Never mind...
end
%end % addSpecialProps %#ok for Matlab 6 compatibility
%% Add new property to supplied handle
function addNewProp(hndl,propName,initialValue,readOnlyFlag,getFunc,setFunc)
sp = schema.prop(hndl,propName,'mxArray');
set(hndl,propName,initialValue);
if nargin>3 & ~isempty(readOnlyFlag) & readOnlyFlag %#ok for Matlab 6 compatibility
set(sp,'AccessFlags.PublicSet','off'); % default='on'
end
if nargin>4 & ~isempty(getFunc) %#ok for Matlab 6 compatibility
set(sp,'GetFunction',getFunc); % unsupported in Matlab 6
end
if nargin>5 & ~isempty(setFunc) %#ok for Matlab 6 compatibility
set(sp,'SetFunction',setFunc); % unsupported in Matlab 6
end
%end % addNewProp %#ok for Matlab 6 compatibility
%% Callback for getting side-bar value
function propValue = getBarVal(object,propValue,varargin) %#ok object & propValue are unused
propValue = unique(get(varargin{:}));
%end % getBarVal %#ok for Matlab 6 compatibility
%% Pre-process args necessary for creating the scroll-plots, if supplied
function [pvPairs, axName] = preProcessArgs(pvPairs)
% Default axes is 'X'
axName = 'x';
% Special check for invalid format
if ~isempty(pvPairs) & ischar(pvPairs{end}) & any(strcmpi(pvPairs{end},{'axis','axes'})) %#ok for Matlab 6 compatibility
myError('YMA:scrollplot:invalidProperty','No data specified for scrollplot property ''Axis''');
end
% Loop over all supplied P-V pairs to pre-process the parameters
idx = 1;
while idx < length(pvPairs)
paramName = pvPairs{idx};
if ~ischar(paramName), idx=idx+1; continue; end
switch lower(paramName)
% Get the last axes requested by the user (if any)
% Check for 'axis' or 'axes' ('axes' is a typical typo of 'axis')
case {'axis','axes'}
axName = pvPairs{idx+1};
% Ensure we got a valid axis name: 'x','y','xy' or 'yx'
if ~ischar(axName) | ~any(strcmpi(axName,{'x','y','xy','yx'})) %#ok for Matlab 6 compatibility
myError('YMA:scrollplot:invalidProperty','Invalid scrollplot ''Axis'' property value: only ''x'',''y'' & ''xy'' are accepted');
end
% Remove from the PV pairs list and move on
axName = lower(axName);
pvPairs(idx:idx+1) = [];
% Placeholder for possible future pre-processed args
otherwise
% Skip...
idx = idx + 1;
end
end
%end % preProcessArgs %#ok for Matlab 6 compatibility
%% Process P-V argument pairs
function processArgs(pvPairs,hScroll)
try
minLim = [];
maxLim = [];
hScroll = double(hScroll); % Matlab 6 could not use findall with handle objects...
axName = get(hScroll, 'userdata');
if strcmpi(axName,'x')
otherAxName = 'y'; %#ok mlint mistaken warning - used below
else % Y scroll
otherAxName = 'x'; %#ok mlint mistaken warning - used below
end
dataStr = [axName 'Data'];
limStr = [axName 'Lim'];
while ~isempty(pvPairs)
% Ensure basic format is valid
paramName = '';
if ~ischar(pvPairs{1})
myError('YMA:scrollplot:invalidProperty','Invalid property passed to scrollplot');
elseif length(pvPairs) == 1
myError('YMA:scrollplot:invalidProperty',['No data specified for property ''' pvPairs{1} '''']);
end
% Process parameter values
paramName = pvPairs{1};
paramValue = pvPairs{2};
pvPairs(1:2) = [];
hScrollBars = unique(findall(hScroll, 'tag','scrollBar'));
hScrollPatches = unique(findall(hScroll, 'tag','scrollPatch'));
switch lower(paramName)
case {'min',['min' axName]}
set(hScrollBars(1:2:end), dataStr,paramValue([1,1]));
for patchIdx = 1 : length(hScrollPatches)
thisPatch = hScrollPatches(patchIdx);
data = get(thisPatch, dataStr);
if strcmpi(axName,'x')
set(thisPatch, dataStr,[paramValue([1;1]); data([4,4])]);
else % Y scroll
set(thisPatch, dataStr,[paramValue(1); data([2,2]); paramValue(1)]);
end
% Update the parent axes with the new limit
hAx = getappdata(get(thisPatch,'Parent'), 'parent');
lim = get(hAx, limStr);
set(hAx, limStr,[paramValue,lim(2)]);
end
minLim = paramValue;
case {'max',['max' axName]}
set(hScrollBars(2:2:end), dataStr,paramValue([1,1]));
for patchIdx = 1 : length(hScrollPatches)
thisPatch = hScrollPatches(patchIdx);
data = get(thisPatch, dataStr);
if strcmpi(axName,'x')
set(thisPatch, dataStr,[data([1,1]); paramValue([1;1])]);
else % Y scroll
set(thisPatch, dataStr,[data(1); paramValue([1;1]); data(1)]);
end
% Update the parent axes with the new limit
hAx = getappdata(get(thisPatch,'Parent'), 'parent');
lim = get(hAx, limStr);
set(hAx, limStr,[lim(1),paramValue]);
end
maxLim = paramValue;
case {'windowsize',['windowsize' axName]}
if isempty(pvPairs)
% No min,max after this param, so act based on data so far
if ~isempty(minLim)
if ~isempty(maxLim) & abs(maxLim-minLim-paramValue(1))>eps %#ok for Matlab 6 compatibility
myError('YMA:scrollplot:invalidWindowSize','Specified WindowSize value conflicts with earlier values specified for Min,Max');
end
pvPairs = {'Max', minLim+paramValue(1), pvPairs{:}}; % note: can't do [...,pvPairs] because of a Matlab6 bug when pvPairs={}
elseif ~isempty(maxLim) % Only max was specified...
pvPairs = {'Min', maxLim-paramValue(1), pvPairs{:}};
else
% No min,max: act based on actual min for each axes seperately
for scrollIdx = 1 : length(hScroll)
% Update the right side bar
thisScroll = hScroll(scrollIdx);
hScrollBars = unique(findall(thisScroll, 'tag','scrollBar'));
maxLim = get(hScrollBars(1), dataStr) + paramValue(1);
set(hScrollBars(2), dataStr,maxLim);
% Now update the patch
thisPatch = unique(findall(thisScroll, 'tag','scrollPatch'));
data = get(thisPatch, dataStr);
if strcmpi(axName,'x')
set(thisPatch, dataStr,[data([1,1]); maxLim']);
else % Y scroll
set(thisPatch, dataStr,[data(1); maxLim'; data(1)]);
end
% Finally, update the parent axes with the new limit
hAx = getappdata(thisScroll, 'parent');
lim = get(hAx, limStr);
set(hAx, limStr,[lim(1),maxLim(1)]);
end
end
else
% Push this P-V pair to the end of the params list (after min,max)
pvPairs = {pvPairs{:}, paramName, paramValue(1)};
end
% Not a good idea to let users play with position so easily...
%case 'position'
% set(hScroll, 'position',paramValue);
case {'axis','axes'}
% Do nothing (should never get here: should have been stripped by preProcessArgs()!)
case {['min' otherAxName], ['max' otherAxName], ['windowsize' otherAxName]}
% Do nothing (pass to other axes for processing)
otherwise
myError('YMA:scrollplot:invalidProperty','Unsupported property');
end % switch paramName
end % loop pvPairs
catch
if ~isempty(paramName), paramName = [' ''' paramName '''']; end
myError('YMA:scrollplot:invalidHandle',['Error setting scrollplot property' paramName ':' char(10) lasterr]);
end
%end % processArgs %#ok for Matlab 6 compatibility
%% Internal error processing
function myError(id,msg)
v = version;
if (v(1) >= '7')
error(id,msg);
else
% Old Matlab versions do not have the error(id,msg) syntax...
error(msg);
end
%end % myError %#ok for Matlab 6 compatibility
%% Get ancestor figure - used for old Matlab versions that don't have a built-in ancestor()
function hObj = ancestor(hObj,type)
if ~isempty(hObj) & ishandle(hObj) %#ok for Matlab 6 compatibility
%if ~isa(handle(hObj),type) % this is best but always returns 0 in Matlab 6!
if ~strcmpi(get(hObj,'type'),type)
hObj = ancestor(get(handle(hObj),'parent'),type);
end
end
%end % ancestor %#ok for Matlab 6 compatibility
%% Helper function to extract first data value(s) from an array
function data = getFirstVals(vals)
if isempty(vals)
data = [];
elseif iscell(vals)
for idx = 1 : length(vals)
thisVal = vals{idx};
data(idx) = thisVal(1); %#ok mlint - preallocate
end
else
data = vals(:,1);
end
%end % getFirstVal %#ok for Matlab 6 compatibility
%% Mouse movement outside the scroll patch area
function mouseOutsidePatch(hFig,inDragMode,hAx) %#ok Hax is unused
try
% Restore the original figure pointer (probably 'arrow', but not necessarily)
% On second thought, it should always be 'arrow' since zoom/pan etc. are disabled within hScroll
%if ~isempty(hAx)
% Only modify this within hScroll (outside the patch area) - not in other axes
set(hFig, 'Pointer','arrow');
%end
oldPointer = getappdata(hFig, 'scrollplot_oldPointer');
if ~isempty(oldPointer)
%set(hFig, oldPointer{:}); % see comment above
drawnow;
rmappdataIfExists(hFig, 'scrollplot_oldPointer');
if isappdata(hFig, 'scrollplot_mouseUpPointer')
setappdata(hFig, 'scrollplot_mouseUpPointer',oldPointer);
end
end
% Restore the original ButtonUpFcn callback
if isappdata(hFig, 'scrollplot_oldButtonUpFcn')
oldButtonUpFcn = getappdata(hFig, 'scrollplot_oldButtonUpFcn');
axisComponent = getappdata(hFig, 'scrollplot_oldButtonUpObj');
if ~isempty(axisComponent)
set(axisComponent, 'MouseReleasedCallback',oldButtonUpFcn);
else
set(hFig, 'WindowButtonUpFcn',oldButtonUpFcn);
end
rmappdataIfExists(hFig, 'scrollplot_oldButtonUpFcn');
end
% Additional cleanup
rmappdataIfExists(hFig, 'scrollplot_mouseDownPointer');
if ~inDragMode
rmappdataIfExists(hFig, 'scrollplot_originalX');
rmappdataIfExists(hFig, 'scrollplot_originalLimits');
end
catch
% never mind...
disp(lasterr);
end
%end % outsideScrollCleanup %#ok for Matlab 6 compatibility
%% Mouse movement within the scroll patch area
function mouseWithinPatch(hFig,inDragMode,hAx,scrollPatch,cx,isOverBar)
try
% Separate actions for X,Y scrolling
axName = get(hAx, 'userdata');
if strcmpi(axName,'x')
shapeStr = 'lrdrag';
else
shapeStr = 'uddrag';
end
dataStr = [axName 'Data'];
limStr = [axName 'Lim'];
% If we have entered the scroll patch area for the first time
axisComponent = getAxisComponent(hFig);
if ~isempty(axisComponent)
winUpFcn = get(axisComponent,'MouseReleasedCallback');
else
winUpFcn = get(hFig,'WindowButtonUpFcn');
end
if isempty(winUpFcn) | (~isequal(winUpFcn,@mouseUpCallback) & (~iscell(winUpFcn) | ~isequal(winUpFcn{1},@mouseUpCallback))) %#ok for Matlab 6 compatibility
% Set the ButtonUpFcn callbacks
if ~isempty(winUpFcn)
setappdata(hFig, 'scrollplot_oldButtonUpFcn',winUpFcn);
setappdata(hFig, 'scrollplot_oldButtonUpObj',axisComponent);
end
if ~isempty(axisComponent)
set(axisComponent, 'MouseReleasedCallback',{@mouseUpCallback,hFig});
else
set(hFig, 'WindowButtonUpFcn',@mouseUpCallback);
end
% Clear up potential junk that might confuse us later
rmappdataIfExists(hFig, 'scrollplot_clickedBarIdx');
end
% If this is a drag movement (i.e., mouse button is clicked)
if inDragMode
% Act according to the dragged object
if isempty(scrollPatch)
scrollPatch = findobj(hAx, 'tag','scrollPatch');
end
scrollBarIdx = getappdata(hFig, 'scrollplot_clickedBarIdx');
scrollBars = sort(findobj(hAx, 'tag','scrollBar'));
%barsXs = cellfun(@(c)c(1),get(scrollBars,dataStr)); % cellfun is very limited on Matlab 6...
barsXs = getFirstVals(get(scrollBars,dataStr));
if barsXs(1)>barsXs(2) % happens after dragging one bar beyond the other
scrollBarIdx = 3 - scrollBarIdx; % []=>[], 1=>2, 2=>1
scrollBars = scrollBars([2,1]);
end
oldPatchXs = get(scrollPatch, dataStr);
axLimits = get(hAx, limStr);
cx = min(max(cx,axLimits(1)),axLimits(2));
if isempty(scrollBarIdx) % patch drag
originalX = getappdata(hFig, 'scrollplot_originalX');
originalLimits = getappdata(hFig, 'scrollplot_originalLimits');
if ~isempty(originalLimits)
allowedDelta = [min(0,axLimits(1)-originalLimits(1)), max(0,axLimits(2)-originalLimits(2))];
deltaX = min(max(cx-originalX, allowedDelta(1)), allowedDelta(2));
if strcmpi(get(hAx,[axName 'Scale']), 'log')
newLimits = 10.^(log10(originalLimits) + deltaX);
else % linear axis scale
newLimits = originalLimits + deltaX;
end
%fprintf('%.3f ',[cx-originalX, deltaX, originalLimits(1), newLimits(1), allowedDelta])
%fprintf('\n');
if strcmpi(axName,'x')
set(scrollPatch, dataStr,newLimits([1,1,2,2]));
else
set(scrollPatch, dataStr,newLimits([1,2,2,1]));
end
set(scrollBars(1), dataStr,newLimits([1,1]));
set(scrollBars(2), dataStr,newLimits([2,2]));
setappdata(hFig, 'scrollplot_originalLimits', newLimits);
setappdata(hFig, 'scrollplot_originalX', cx);
if deltaX ~= 0
delete(findall(0,'tag','scrollHelp'));
end
end
elseif (scrollBarIdx == 1) % left/bottom bar drag
set(scrollBars(scrollBarIdx), dataStr,[cx,cx]);
if strcmpi(axName,'x')
set(scrollPatch, dataStr,[cx,cx, max(oldPatchXs)*[1,1]]);
else
set(scrollPatch, dataStr,[cx, max(oldPatchXs)*[1,1], cx]);
end
delete(findall(0,'tag','scrollHelp'));
else % right/top bar drag
set(scrollBars(scrollBarIdx), dataStr,[cx,cx]);
if strcmpi(axName,'x')
set(scrollPatch, dataStr,[min(oldPatchXs)*[1,1], cx,cx]);
else
set(scrollPatch, dataStr,[cx, min(oldPatchXs)*[1,1], cx]);
end
delete(findall(0,'tag','scrollHelp'));
end
% Modify the parent axes accordingly
parentAx = getappdata(hAx, 'parent');
newXLim = unique(get(scrollPatch,dataStr));
if length(newXLim) == 2 % might be otherwise if bars merge!
set(parentAx, limStr,newXLim);
end
% Mode managers (zoom/pan etc.) modify the cursor shape, so we need to force ours...
newPtr = getappdata(hFig, 'scrollplot_mouseDownPointer');
if ~isempty(newPtr)
setptr(hFig, newPtr);
end
else % Normal mouse movement (no drag)
% Modify the cursor shape
oldPointer = getappdata(hFig, 'scrollplot_oldPointer');
if isempty(oldPointer)
% Preserve original pointer shape for future use
setappdata(hFig, 'scrollplot_oldPointer',getptr(hFig));
end
if isOverBar
setptr(hFig, shapeStr);
setappdata(hFig, 'scrollplot_mouseDownPointer',shapeStr);
else
setptr(hFig, 'hand');
setappdata(hFig, 'scrollplot_mouseDownPointer','closedhand');
end
end
drawnow;
catch
% never mind...
disp(lasterr);
end
%end % mouseWithinPatch %#ok for Matlab 6 compatibility
%% Mouse movement callback function
function mouseMoveCallback(varargin)
try
try
% Temporarily turn off log-axis warnings
oldWarn = warning('off','MATLAB:Axes:NegativeDataInLogAxis');
catch
% never mind...
end
% Get the figure's current axes
hFig = gcbf;
if isempty(hFig) | ~ishandle(hFig), return; end %#ok just in case..
%hAx = get(hFig,'currentAxes');
hAx = getCurrentScrollAx(hFig);
inDragMode = isappdata(hFig, 'scrollplot_clickedBarIdx');
% Exit if already in progress - don't want to mess everything...
if isappdata(hFig,'scrollBar_inProgress'), return; end
% Fix case of Mode Managers (pan, zoom, ...)
try
modeMgr = get(hFig,'ModeManager');
hMode = modeMgr.CurrentMode;
set(hMode,'ButtonDownFilter',@shouldModeBeInactiveFcn);
catch
% Never mind - either an old Matlab (no mode managers) or no mode currently active
end
% If mouse pointer is not currently over any scroll axes
if isempty(hAx) %& ~inDragMode %#ok for Matlab 6 compatibility
% Perform cleanup
mouseOutsidePatch(hFig,inDragMode,hAx);
else
% Check whether the curser is over any side bar
scrollPatch = findobj(hAx, 'tag','scrollPatch');
isOverBar = 0;
cx = [];
if ~isempty(scrollPatch)
scrollPatch = scrollPatch(1);
axName = get(hAx,'userdata');
cp = get(hAx,'CurrentPoint');
cx = cp(1,1);
cy = cp(1,2);
xlim = get(hAx,'Xlim');
ylim = get(hAx,'Ylim');
limits = get(hAx,[axName 'Lim']);
barXs = unique(get(scrollPatch,[axName 'Data']));
if strcmpi(get(hAx,[axName 'Scale']), 'log')
fuzz = 0.01 * diff(log(abs(limits))); % tolerances (1%) in axes units
barXs = log10(barXs);
if strcmpi(axName,'x')
cx = log10(cx);
else
cy = log10(cy);
end
else
fuzz = 0.01 * diff(limits); % tolerances (1%) in axes units
end
if isempty(barXs), return; end
%disp(abs(cy-barXs)')
if strcmpi(axName,'x')
inXTest = any(barXs-fuzz < cx) & any(cx < barXs+fuzz);
inYTest = (ylim(1) < cy) & (cy < ylim(2));
isOverBar = any(abs(cx-barXs)<fuzz); %(barXs-fuzz < cx) & (cx < barXs+fuzz));
else % Y scroll
inXTest = (xlim(1) < cx) & (cx < xlim(2));
inYTest = any(barXs-fuzz < cy) & any(cy < barXs+fuzz);
isOverBar = any(abs(cy-barXs)<fuzz); %(barXs-fuzz < cy) & (cy < barXs+fuzz));
cx = cy; % for use in mouseWithinPatch below
end
scrollPatch = scrollPatch(inXTest & inYTest);
if strcmpi(get(hAx,[axName 'Scale']), 'log')
cx = 10^cx; % used below
end
end
% From this moment on, don't allow any interruptions
setappdata(hFig,'scrollBar_inProgress',1);
% If we're within the scroll patch area
if ~isempty(scrollPatch) | inDragMode %#ok for Matlab 6 compatibility
mouseWithinPatch(hFig,inDragMode,hAx,scrollPatch,cx,isOverBar);
else
% Perform cleanup
mouseOutsidePatch(hFig,inDragMode,hAx);
end
end
% Try to chain the original WindowButtonMotionFcn (if available)
try
hgfeval(getappdata(hFig, 'scrollplot_oldButtonMotionFcn'));
catch
% Never mind...
end
catch
% Never mind...
disp(lasterr);
end
rmappdataIfExists(hFig,'scrollBar_inProgress');
% Restore original warnings (if available/possible)
try
warning(oldWarn);
catch
% never mind...
end
%end % mouseMoveCallback %#ok for Matlab 6 compatibility
%% Mouse click down callback function
function mouseDownCallback(varargin)
try
% Modify the cursor shape (close hand)
hFig = gcbf; %varargin{3};
if isempty(hFig) & ~isempty(varargin) %#ok for Matlab 6 compatibility
hFig = ancestor(varargin{1},'figure');
end
if isempty(hFig) | ~ishandle(hFig), return; end %#ok just in case..
setappdata(hFig, 'scrollplot_mouseUpPointer',getptr(hFig));
newPtr = getappdata(hFig, 'scrollplot_mouseDownPointer');
if ~isempty(newPtr)
setptr(hFig, newPtr);
end
% Determine the clicked object: patch, left bar or right bar
hAx = get(hFig,'currentAxes');
if isempty(hAx), return; end
axName = get(hAx,'userdata');
limits = get(hAx,[axName 'Lim']);
cp = get(hAx,'CurrentPoint');
% Check whether the curser is over any side bar
barXs = [-inf,inf];
scrollBarIdx = [];
scrollPatch = findobj(hAx, 'tag','scrollPatch');
if ~isempty(scrollPatch)
scrollPatch = scrollPatch(1);
dataStr = [axName 'Data'];
barXs = unique(get(scrollPatch,dataStr));
if isempty(barXs), return; end
if strcmpi(axName,'x')
cx = cp(1,1);
else % Y scroll
cx = cp(1,2); % actually, this gets the y value...
end
if strcmpi(get(hAx,[axName 'Scale']), 'log')
fuzz = 0.01 * diff(log(abs(limits))); % tolerances (1%) in axes units
barXs = log10(barXs);
cx = log10(cx);
else
fuzz = 0.01 * diff(limits); % tolerances (1%) in axes units
end
inTest = abs(cx-barXs)<fuzz; %(barXs-fuzz < cx) & (cx < barXs+fuzz);
scrollBarIdx = find(inTest);
scrollBarIdx = scrollBarIdx(min(1:end)); %#ok - find(x,1) is unsupported on Matlab 6!
if strcmpi(get(hAx,[axName 'Scale']), 'log')
cx = 10^cx; % used below
barXs = 10.^barXs; % used below
end
% Re-sort side bars (might have been dragged one over the other...)
scrollBars = sort(findobj(hAx, 'tag','scrollBar'));
%barsXs = cellfun(@(c)c(1),get(scrollBars,'xdata')); % cellfun is very limited on Matlab 6...
barsXs = getFirstVals(get(scrollBars,dataStr));
if barsXs(1)>barsXs(2) % happens after dragging one bar beyond the other
set(scrollBars(1), dataStr,barsXs(2)*[1,1]);
set(scrollBars(2), dataStr,barsXs(1)*[1,1]);
end
end
setappdata(hFig, 'scrollplot_clickedBarIdx',scrollBarIdx);
setappdata(hFig, 'scrollplot_originalX',cx);
setappdata(hFig, 'scrollplot_originalLimits',barXs);
catch
% Never mind...
disp(lasterr);
end
%end % mouseDownCallback %#ok for Matlab 6 compatibility
%% Mouse click up callback function
function mouseUpCallback(varargin)
try
% Restore the previous (pre-click) cursor shape
hFig = gcbf; %varargin{3};
if isempty(hFig) & ~isempty(varargin) %#ok for Matlab 6 compatibility
hFig = varargin{3};
if isempty(hFig)
hFig = ancestor(varargin{1},'figure');
end
end
if isempty(hFig) | ~ishandle(hFig), return; end %#ok just in case..
if isappdata(hFig, 'scrollplot_mouseUpPointer')
mouseUpPointer = getappdata(hFig, 'scrollplot_mouseUpPointer');
set(hFig,mouseUpPointer{:});
rmappdata(hFig, 'scrollplot_mouseUpPointer');
end
% Cleanup data no longer needed
rmappdataIfExists(hFig, 'scrollplot_clickedBarIdx');
rmappdataIfExists(hFig, 'scrollplot_originalX');
rmappdataIfExists(hFig, 'scrollplot_originalLimits');
% Try to chain the original WindowButtonUpFcn (if available)
oldFcn = getappdata(hFig, 'scrollplot_oldButtonUpFcn');
if ~isempty(oldFcn) & ~isequal(oldFcn,@mouseUpCallback) & (~iscell(oldFcn) | ~isequal(oldFcn{1},@mouseUpCallback)) %#ok for Matlab 6 compatibility
hgfeval(oldFcn);
end
catch
% Never mind...
disp(lasterr);
end
%end % mouseUpCallback %#ok for Matlab 6 compatibility
%% Remove appdata if available
function rmappdataIfExists(handle, name)
if isappdata(handle, name)
rmappdata(handle, name)
end
%end % rmappdataIfExists %#ok for Matlab 6 compatibility
%% Get the figure's java axis component
function axisComponent = getAxisComponent(hFig)
try
if isappdata(hFig, 'scrollplot_axisComponent')
axisComponent = getappdata(hFig, 'scrollplot_axisComponent');
else
axisComponent = [];
javaFrame = get(hFig,'JavaFrame');
axisComponent = get(javaFrame,'AxisComponent');
axisComponent = handle(axisComponent, 'CallbackProperties');
if ~isprop(axisComponent,'MouseReleasedCallback')
axisComponent = []; % wrong axisComponent...
else
setappdata(hFig, 'scrollplot_axisComponent',axisComponent);
end
end
catch
% never mind...
end
%end % getAxisComponent %#ok for Matlab 6 compatibility
%% Get the scroll axes that the mouse is currently over
function hAx = getCurrentScrollAx(hFig)
try
hAx = [];
scrollAxes = findall(hFig, 'tag','scrollAx');
if isempty(scrollAxes), return; end % should never happen...
for axIdx = 1 : length(scrollAxes)
scrollPos(axIdx,:) = getPixelPos(scrollAxes(axIdx)); %#ok mlint - preallocate
end
cp = get(hFig, 'CurrentPoint'); % in Matlab pixels
inXTest = (scrollPos(:,1) <= cp(1)) & (cp(1) <= scrollPos(:,1)+scrollPos(:,3));
inYTest = (scrollPos(:,2) <= cp(2)) & (cp(2) <= scrollPos(:,2)+scrollPos(:,4));
hAx = scrollAxes(inXTest & inYTest);
hAx = hAx(min(1:end)); % ensure we return no more than asingle hAx!
catch
% never mind...
disp(lasterr);
end
%end % getCurrentScrollAx %#ok for Matlab 6 compatibility
%% Get pixel position of an HG object
function pos = getPixelPos(hObj)
try
% getpixelposition is unvectorized unfortunately!
pos = getpixelposition(hObj);
catch
% Matlab 6 did not have getpixelposition nor hgconvertunits so use the old way...
pos = getPos(hObj,'pixels');
end
%end % getPixelPos %#ok for Matlab 6 compatibility
%% Get position of an HG object in specified units
function pos = getPos(hObj,units)
% Matlab 6 did not have hgconvertunits so use the old way...
oldUnits = get(hObj,'units');
if strcmpi(oldUnits,units) % don't modify units unless we must!
pos = get(hObj,'pos');
else
set(hObj,'units',units);
pos = get(hObj,'pos');
set(hObj,'units',oldUnits);
end
%end % getPos %#ok for Matlab 6 compatibility
%% Temporary setting property value for a read-only property
function setOnce(hndl,propName,propValue)
try
prop = findprop(hndl,propName);
oldSetState = get(prop,'AccessFlags.PublicSet');
set(prop,'AccessFlags.PublicSet','on');
set(hndl,propName,propValue);
set(prop,'AccessFlags.PublicSet',oldSetState);
catch
% Never mind...
end
%end % setOnce %#ok for Matlab 6 compatibility
%% Callback for parent axes property changes
function parentAxesChanged(schemaProp, eventData, hFig, hAx, hScrollPatch, hScrollBars, propName) %#ok - first 2 are unused
try
if isempty(hFig) | ~ishandle(hFig), return; end %#ok just in case..
newPropVal = get(hAx,propName);
hScroll = get(hScrollPatch, 'Parent');
axName = get(hScroll, 'userdata');
if isappdata(hFig,'scrollBar_inProgress')
% Update the special prop values
if strcmpi(propName,[axName,'Lim'])
setOnce(handle(hScroll),'ScrollMin',newPropVal(1));
setOnce(handle(hScroll),'ScrollMax',newPropVal(2));
end
return;
end
switch propName
case 'XLim'
if strcmpi(axName,'x')
set(hScrollPatch, 'XData',newPropVal([1,1,2,2]));
set(hScrollBars(1), 'Xdata',newPropVal([1,1]));
set(hScrollBars(2), 'Xdata',newPropVal([2,2]));
setOnce(handle(hScroll),'ScrollMin',newPropVal(1));
setOnce(handle(hScroll),'ScrollMax',newPropVal(2));
end
case 'YLim'
if strcmpi(axName,'y')
set(hScrollPatch, 'YData',newPropVal([1,2,2,1]));
set(hScrollBars(1), 'Ydata',newPropVal([1,1]));
set(hScrollBars(2), 'Ydata',newPropVal([2,2]));
setOnce(handle(hScroll),'ScrollMin',newPropVal(1));
setOnce(handle(hScroll),'ScrollMax',newPropVal(2));
end
case 'Position'
if strcmpi(axName,'x')
newScrollPos = newPropVal .* [1, 1, 1, 0.10/0.80];
newScrollPos = newScrollPos - [0, 0.20/0.80*newPropVal(4), 0, 0];
else % Y scroll
newScrollPos = newPropVal .* [1, 1, 0.10/0.80, 1];
newScrollPos = newScrollPos + [(1+0.05/0.80)*newPropVal(3), 0, 0, 0];
end
axUnits = get(hAx, 'Units'); % units might be modified by Mode Managers bypassing listeners!
set(hScroll, 'Units',axUnits, 'Position',newScrollPos);
case {'Units','Parent','XDir','YDir','XScale','YScale'}
set(hScroll, propName,newPropVal);
otherwise
% Do nothing...
end
catch
% never mind...
disp(lasterr);
end
%end % parentAxesChanged %#ok for Matlab 6 compatibility
%% Determine whether a current mode manager should be active or not (filtered)
function shouldModeBeInactive = shouldModeBeInactiveFcn(hObj, eventData) %#ok - eventData is unused
try
shouldModeBeInactive = 0;
hFig = ancestor(hObj,'figure');
hScrollAx = getCurrentScrollAx(hFig);
shouldModeBeInactive = ~isempty(hScrollAx);
catch
% never mind...
disp(lasterr);
end
%end % shouldModeBeActiveFcn %#ok for Matlab 6 compatibility
%% hgfeval replacement for Matlab 6 compatibility
function hgfeval(fcn,varargin)
if isempty(fcn), return; end
if iscell(fcn)
feval(fcn{1},varargin{:},fcn{2:end});
elseif ischar(fcn)
evalin('base', fcn);
else
feval(fcn,varargin{:});
end
%end % hgfeval %#ok for Matlab 6 compatibility
%% Axis to screen coordinate transformation
function T = axis2Screen(ax)
% computes a coordinate transformation T = [xo,yo,rx,ry] that
% relates the normalized axes coordinates [xa,ya] of point [xo,yo]
% to its screen coordinate [xs,ys] (in the root units) by:
% xs = xo + rx * xa
% ys = yo + ry * ya
%
% See also SISOTOOL
%
% Note: this is a modified internal function within moveptr()
% Get axes normalized position in figure
T = getPos(ax,'normalized');
% Loop all the way up the hierarchy to the root
% Note: this fixes a bug in Matlab 7's moveptr implementation
parent = get(ax,'Parent');
while ~isempty(parent)
% Transform norm. axis coord -> parent coord.
if isequal(parent,0)
parentPos = get(0,'ScreenSize'); % Preserve screen units
else
parentPos = getPos(parent, 'normalized'); % Normalized units
end
T(1:2) = parentPos(1:2) + parentPos(3:4) .* T(1:2);
T(3:4) = parentPos(3:4) .* T(3:4);
parent = get(parent,'Parent');
end
%end % axis2Screen %#ok for Matlab 6 compatibility
%% Get centran axis location
function axisCoord = getCenterCoord(hAx, axName)
limits = get(hAx, [axName 'Lim']);
if strcmpi(get(hAx,[axName 'Scale']), 'log')
axisCoord = sqrt(abs(prod(limits))); %=10^mean(log10(abs(limits)));
else
axisCoord = mean(limits);
end
%end %getCenterCoord %#ok for Matlab 6 compatibility
%% Get normalized axis coordinates
function normCoord = getNormCoord(hAx, axName, curPos)
limits = get(hAx, [axName 'Lim']);
if strcmpi(get(hAx,[axName 'Scale']), 'log')
normCoord = (log2(curPos) - log2(limits(1))) / diff(log2(limits));
else
normCoord = (curPos-limits(1)) / diff(limits);
end
%end % getNormCoord %#ok for Matlab 6 compatibility
%% moveptr replacement for Matlab 6 compatibility
function moveptr(hAx, x, y)
% Compute normalized axis coordinates
NormX = getNormCoord(hAx, 'x', x);
NormY = getNormCoord(hAx, 'y', y);
% Compute the new coordinates in screen units
Transform = axis2Screen(hAx);
NewLoc = Transform(1:2) + Transform(3:4) .* [NormX NormY];
% Move the pointer
set(0,'PointerLocation',NewLoc);
%end % moveptr %#ok for Matlab 6 compatibility
%% UiContextMenu callback - Move cursor to center of requested element
function moveCursor(varargin)
try
% Get the x,y location of the center of the requested object
hScroll = handle(gca);
hObj = get(gcbo,'UserData');
x = mean(get(hObj,'XData'));
y = mean(get(hObj,'YData'));
% Move the mouse pointer to that location
% Note: Matlab 6 did not have moveptr() so we use a local version above
%moveptr(hScroll, 'init');
%moveptr(hScroll, 'move', x, y);
moveptr(hScroll, x, y);
% Call mouseMoveCallback to update the pointer shape
mouseMoveCallback;
drawnow;
catch
% Never mind...
disp(lasterr);
end
%end % moveCursor %#ok for Matlab 6 compatibility
%% Callback when scroll axes are deleted
function deleteScrollAx(varargin)
try
% Update the parent Axes position
hScroll = varargin{1};
updateParentPos([],[],hScroll,'off');
% Note: no need to remove hAx listeners since these are destroyed along with hScroll
catch
% Never mind - continue deletion process...
end
%end % deleteScrollAx %#ok for Matlab 6 compatibility
%% Update parent figure position based on scroll axes visibility
function updateParentPos(schemaProp, eventData, hScroll,scrollVisibility) %#ok first 2 params are unused
try
if nargin<4
scrollVisibility = get(hScroll, 'visible');
end
% Update the parent Axes position
hAx = get(hScroll, 'ParentAxesHandle');
axPos = get(hAx,'position');
axName = get(hScroll, 'userdata');
hLabel = get(hAx, [axName 'Label']);
set(hLabel, 'units','normalized');
oldPos = get(hLabel, 'position'); % Get the label position before the axes change
if strcmpi(scrollVisibility,'off')
ax_dy1 = 1/0.80;
ax_dy2 = -1/0.80;
label_delta = 0.20/0.80;
else %'on'
ax_dy1 = 0.80;
ax_dy2 = 1;
label_delta = -0.20/0.80;
end
if strcmpi(axName,'x')
newPlotPos = axPos .* [1, 1, 1, ax_dy1] + [0, 0.20*axPos(4)*ax_dy2, 0, 0];
else % Y scroll
newPlotPos = axPos .* [1, 1, ax_dy1, 1];
end
set(hAx, 'Position',newPlotPos);
% Fix label position(s)
if strcmpi(axName,'x')
if ~isempty(oldPos) & oldPos(2)<0 %#ok for Matlab 6 compatibility
% Only fix if the X label is on the bottom (usually yes)
set(hLabel, 'position',oldPos+[0,label_delta,0]);
end
else % Y scroll
if ~isempty(oldPos) & oldPos(1)>0 %#ok for Matlab 6 compatibility
% Only fix if the Y label is on the right side (usually not)
set(hLabel, 'position',oldPos-[label_delta,0,0]);
end
end
% Show/hide all the axes children (scroll patch, side-bars, text)
set(findall(hScroll), 'Visible',scrollVisibility);
% axisComponent gets re-created, so clear the cache
hFig = ancestor(hScroll,'figure');
rmappdata(hFig, 'scrollplot_axisComponent');
catch
% Never mind...
end
%end % updateParentPos %#ok for Matlab 6 compatibility
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% TODO %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% - maybe add a blue diamond or a visual handle in center of side-bars?
% - fix or bypass Matlab 6 OpenGL warning due to patch FaceAlpha property
|
github
|
rising-turtle/slam_matlab-master
|
plotframe.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/all/plotframe.m
| 1,677 |
utf_8
|
40c9bb130ec46ca33912f804e87f2eac
|
% PLOTFRAME - plots a coordinate frame specified by a homogeneous transform
%
% Usage: function plotframe(T, len, label)
%
% Arguments:
% T - 4x4 homogeneous transform
% len - length of axis arms to plot (defaults to 1)
% label - text string to append to x,y,z labels on axes
%
% len and label are optional and default to 1 and '' respectively
%
% See also: ROTX, ROTY, ROTZ, TRANS, INVHT
% Copyright (c) 2001 Peter Kovesi
% School of Computer Science & Software Engineering
% The University of Western Australia
% pk at csse uwa edu au
% http://www.csse.uwa.edu.au/
function plotframe(T, len, label, colr)
if ~all(size(T) == [4,4])
error('plotframe: matrix is not 4x4')
end
if ~exist('len','var')
len = 1;
end
if ~exist('label','var')
label = '';
end
if ~exist('colr','var')
colr = [0 0 1];
end
% Assume scale specified by T(4,4) == 1
origin = T(1:3, 4); % 1st three elements of 4th column
X = origin + len*T(1:3, 1); % point 'len' units out along x axis
Y = origin + len*T(1:3, 2); % point 'len' units out along y axis
Z = origin + len*T(1:3, 3); % point 'len' units out along z axis
line([origin(1),X(1)], [origin(2), X(2)], [origin(3), X(3)], 'color', colr);
line([origin(1),Y(1)], [origin(2), Y(2)], [origin(3), Y(3)], 'color', colr);
line([origin(1),Z(1)], [origin(2), Z(2)], [origin(3), Z(3)], 'color', colr);
text(X(1), X(2), X(3), ['x' label], 'color', colr);
text(Y(1), Y(2), Y(3), ['y' label], 'color', colr);
text(Z(1), Z(2), Z(3), ['z' label], 'color', colr);
|
github
|
rising-turtle/slam_matlab-master
|
compensate_badpixel_soonhac.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/all/compensate_badpixel_soonhac.m
| 2,255 |
utf_8
|
c10f2a65debeca406e3c75b955e4e84c
|
% Compenstate the bad pixel with low confidence by median filter
% Date : 3/13/12
% Author : Soonhac Hong ([email protected])
function [img, x, y, z, c] = compensate_badpixel_soonhac(img, x, y, z, c, confidence_cut_off)
e_index = c < confidence_cut_off;
x_median=medfilt2(x, [3 3],'symmetric');
y_median=medfilt2(y, [3 3],'symmetric');
z_median=medfilt2(z, [3 3],'symmetric');
image_median=medfilt2(img, [3 3],'symmetric');
z(e_index) = z_median(e_index);
x(e_index) = x_median(e_index);
y(e_index) = y_median(e_index);
img(e_index) = image_median(e_index);
%
%
%
%
%
% for i = 1:size(img,1) % row
% for j=1:size(img,2) % column
% if e_index(i,j) == 1
% start_i = i-1;
% end_i = i+1;
% start_j = j-1;
% end_j = j+1;
% point_i = 2;
% point_j = 2;
% if i == 1
% start_i = i;
% point_i = 1;
% if j == 1
% point_j = 1;
% end
% end
% if i == size(img,1)
% end_i = i;
% if j == 1
% point_j = 1;
% end
% end
% if j == 1
% start_j = j;
% if i == 1
% point_i = 1;
% end
% end
% if j == size(img,2)
% end_j = j;
% if i == 1
% point_i = 1;
% end
% end
% img_unit=medfilt2(img(start_i:end_i,start_j:end_j), [3 3],'symmetric');
% x_unit=medfilt2(x(start_i:end_i,start_j:end_j), [3 3],'symmetric');
% y_unit=medfilt2(y(start_i:end_i,start_j:end_j), [3 3],'symmetric');
% z_unit=medfilt2(z(start_i:end_i,start_j:end_j), [3 3],'symmetric');
% img(i,j) = img_unit(point_i,point_j);
% x(i,j) = x_unit(point_i,point_j);
% y(i,j) = y_unit(point_i,point_j);
% z(i,j) = z_unit(point_i,point_j);
% end
% end
% end
% end
|
github
|
rising-turtle/slam_matlab-master
|
fn_structdisp.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/all/fn_structdisp.m
| 2,279 |
utf_8
|
7abee2d7e63975b016be8f923567bc05
|
function fn_structdisp(Xname)
% function fn_structdisp Xname
% function fn_structdisp(X)
%---
% Recursively display the content of a structure and its sub-structures
%
% Input:
% - Xname/X one can give as argument either the structure to display or
% or a string (the name in the current workspace of the
% structure to display)
%
% A few parameters can be adjusted inside the m file to determine when
% arrays and cell should be displayed completely or not
% Thomas Deneux
% Copyright 2005-2012
if ischar(Xname)
X = evalin('caller',Xname);
else
X = Xname;
Xname = inputname(1);
end
if ~isstruct(X), error('argument should be a structure or the name of a structure'), end
rec_structdisp(Xname,X)
%---------------------------------
function rec_structdisp(Xname,X)
%---
%-- PARAMETERS (Edit this) --%
ARRAYMAXROWS = 10;
ARRAYMAXCOLS = 10;
ARRAYMAXELEMS = 30;
CELLMAXROWS = 10;
CELLMAXCOLS = 10;
CELLMAXELEMS = 30;
CELLRECURSIVE = false;
%----- PARAMETERS END -------%
disp([Xname ':'])
disp(X)
%fprintf('\b')
if isstruct(X) || isobject(X)
F = fieldnames(X);
nsub = length(F);
Y = cell(1,nsub);
subnames = cell(1,nsub);
for i=1:nsub
f = F{i};
Y{i} = X.(f);
subnames{i} = [Xname '.' f];
end
elseif CELLRECURSIVE && iscell(X)
nsub = numel(X);
s = size(X);
Y = X(:);
subnames = cell(1,nsub);
for i=1:nsub
inds = s;
globind = i-1;
for k=1:length(s)
inds(k) = 1+mod(globind,s(k));
globind = floor(globind/s(k));
end
subnames{i} = [Xname '{' num2str(inds,'%i,')];
subnames{i}(end) = '}';
end
else
return
end
for i=1:nsub
a = Y{i};
if isstruct(a) || isobject(a)
if length(a)==1
rec_structdisp(subnames{i},a)
else
for k=1:length(a)
rec_structdisp([subnames{i} '(' num2str(k) ')'],a(k))
end
end
elseif iscell(a)
if size(a,1)<=CELLMAXROWS && size(a,2)<=CELLMAXCOLS && numel(a)<=CELLMAXELEMS
rec_structdisp(subnames{i},a)
end
elseif size(a,1)<=ARRAYMAXROWS && size(a,2)<=ARRAYMAXCOLS && numel(a)<=ARRAYMAXELEMS
disp([subnames{i} ':'])
disp(a)
end
end
|
github
|
rising-turtle/slam_matlab-master
|
cov_pose_shift_calc.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/all/cov_pose_shift_calc.m
| 1,718 |
utf_8
|
6f51efd75828bc2370303048442bade9
|
function cov_pose_shift = cov_pose_shift_calc(Ya,Yb,R,T)
q_ab = R2q(R);
cov_pose_shift = zeros(7,7);
for i=1:size(Ya,2)
f_p = Ya(:,i);
f_n = Yb(:,i);
% cov_f_p = calc_cov_point(f_p);
% cov_f_n = calc_cov_point(f_n);
d2E_dX2_value = d2E_dX2_4cov(f_p,f_n,q_ab,T);
d2E_dFdX_value = [d2E_dfp_dX_4cov(f_p,f_n,q_ab,T) d2E_dfn_dX_4cov(f_p,f_n,q_ab,T)];
[cov_cart_f_p,tt,ttt] = calc_point_cov_jacobian(f_p);
[cov_cart_f_n,tt,ttt] = calc_point_cov_jacobian(f_n);
try
J_X2F = -pinv(d2E_dX2_value)*d2E_dFdX_value;
catch
disp('J_X2F calculation failed ! \n')
end
cov_pose_shift = cov_pose_shift + J_X2F*blkdiag(cov_cart_f_p,cov_cart_f_n)*J_X2F';
end
end
function [cov_cart,J_sph2cart,cov_spher] = calc_point_cov_jacobian(point)
[azimuth,elevation,r] = cart2sph(point(1),point(2),point(3));
x = r .* cos(elevation) .* cos(azimuth);
y = r .* cos(elevation) .* sin(azimuth);
z = r .* sin(elevation);
%%% check correctness
if norm([x,y,z]-[point(1),point(2),point(3)])>0.000000001
disp('ERROR OCCURRED IN SHP<-->CART CONVERISON')
end
J_sph2cart = [cos(elevation) * cos(azimuth) -r * cos(elevation) * sin(azimuth) -r * sin(elevation) * cos(azimuth);...
cos(elevation) * sin(azimuth) r * cos(elevation) * cos(azimuth) -r * sin(elevation) * sin(azimuth);...
sin(elevation) 0 r .* cos(elevation) ];
cov_spher = diag( [ (0.01/3)^2 , (0.24*pi/180/10)^2 .* [ 1 1 ] ] ); %%% uncertaint of the range is 1cm and I assume the uncertsainty of the azimuth and elevation is equal
%%% to sensor's angular resolution (0.24 degrees)
cov_cart = J_sph2cart*cov_spher*J_sph2cart';
end
|
github
|
rising-turtle/slam_matlab-master
|
test_orientation_observation.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/all/test_orientation_observation.m
| 3,049 |
utf_8
|
ce9f7def042bcd152cd6df91afcdd3b9
|
function com_error = test_orientation_observation(snap_step)
global myCONFIG
config_file
% snap_step =1;
[x_k_k,p_k_k,q_expected] = read_snapshot(snap_step)
[T,q,R,varargout]=Calculate_V_Omega_RANSAC_dr_ye(snap_step-1,snap_step)
com_error = calc_orientation_error(x_k_k,snap_step,q_expected);
R_expected = q2R(q_expected);
R_estimated = q2R(x_k_k(4:7));
[m_, a_] = find_angle_bw_2_vecs(R_expected(:,2), R_estimated(:,2));
display_angles(m_, a_);
end
function [x_k_k,p_k_k,q_expected] = read_snapshot(snap_step)
global myCONFIG
load([myCONFIG.PATH.DATA_FOLDER,'DataSnapshots/','snapshot',num2str(snap_step),'.mat'])
features_info = eval(['snapshot',num2str(snap_step),'.features_info']);
filter = eval(['snapshot',num2str(snap_step),'.filter']);
[R,T] = plane_fit_to_data(snap_step);
x_k_k = get_x_k_k(filter);
p_k_k = get_p_k_k(filter);
q_expected = R2q(R');
x_k_k = x_k_k(1:7);
p_k_k = p_k_k (1:7,1:7);
end
function com_error = calc_orientation_error(x_k_k,snap_step,q_expected)
vect_est = q2R(x_k_k(4:7))*[0; 0 ;1];
vect_gt = q2R(q_expected)*[0; 0 ;1];
% com_error = euler_vec' - [zeros(size(gt_pan),1),gt_pan,zeros(size(gt_pan),1)];
[m_, a_] = find_angle_bw_2_vecs(vect_est, vect_gt);
display_angles(m_, a_);
h3 = mArrow3([0;0;0],vect_est,'color',[1 0 0]);
h4 = mArrow3([0;0;0],vect_gt,'color',[0 1 0]);
% if i~=size(euler_vec,2)
delete(h3)
delete(h4)
% end
% com_error(i,:) = R2e(e2R(euler_vec(:,i))*(e2R([0 gt_pan(i) 0]))');
com_error = a_(7) ;
end
% load([myCONFIG.PATH.DATA_FOLDER,'DataSnapshots/','snapshot',num2str(myCONFIG.STEP.END-1),'.mat'])
% features_info = eval(['snapshot',num2str(myCONFIG.STEP.END-1),'.features_info']);
% filter = eval(['snapshot',num2str(myCONFIG.STEP.END-1),'.filter']);
% if step==initIm+1
% [R,T] = plane_fit_to_data(myCONFIG.STEP.END-1);
% R=eye(3);
% end
% x_k_k_temp(1:7)
% x_k_k_temp = get_x_k_k(filter);
% stacked_x_k_k(:,step) = [R'*x_k_k_temp(1:3); R2q(R'*q2R( x_k_k_temp(4:7))) ];
% % [V,q]=calc_gt_in_1pointRANSAC(1,step);
% % [V,q,time_gt]=get_gt_time(initIm,step);
% % q=q';%% temperorily
% % GroundTruth(:,step - initIm) = [V;q'];
% trajectory(:,step - initIm) = [R'*x_k_k_temp(1:3); R2q(R'*q2R( x_k_k_temp(4:7)))];
% % time_vector(step - initIm)=time_gt;
% p_k_k_temp = get_p_k_k(filter);
% stacked_p_k_k(:,:,step) = p_k_k_temp(1:7,1:7);
% step
% % NormError(step - initIm)=norm(x_k_k_temp(1:3)-V);
% %
% for i=1:size(euler_vec,2)
% vect_est = q2R(xx(4:7,i))*[1; 1 ;1];
% vect_gt = e2R([0 gt_pan(i)*pi/180 0])*[0; 0 ;1];
%
%
% % com_error = euler_vec' - [zeros(size(gt_pan),1),gt_pan,zeros(size(gt_pan),1)];
% [m_, a_] = find_angle_bw_2_vecs(vect_est, vect_gt);
% display_angles(m_, a_);
%
% h3 = mArrow3([0;0;0],vect_est,'color',[1 0 0]);
%
% h4 = mArrow3([0;0;0],vect_gt,'color',[0 1 0]);
% if i~=size(euler_vec,2)
% delete(h3)
% delete(h4)
% end
% % com_error(i,:) = R2e(e2R(euler_vec(:,i))*(e2R([0 gt_pan(i) 0]))');
% com_error(i,:) = a_(7) ;
% end
|
github
|
rising-turtle/slam_matlab-master
|
dispMEq.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/all/drawLA/dispMEq.m
| 8,277 |
utf_8
|
1e3bbf518742830f3c041b1d1c5c171b
|
function dispMEq(eq, varargin)
% Formatted display of a matrix equation
%
% Usage: dispEq(eq, arg)
% dispEq(eq, arg1, arg2, ...)
% dispEq(eq, arg1, arg2, ..., 'PropertyName', PropertyValue)
%
% INPUT:
% eq - character string defining a matrix equation, e.g., 'A*x=b'
% allowed arithmetical operations are: [+, -, *, / , \, =, ;]
% arg{1,2,...} - matrices corresponding to their symbolic definition in the equation,
% can be NaN (see the last 4 examples)
%
% (optional parameters)
% PropertyName: PropertyValue
% 'format' - a cell array with the formats for the text output. May be of 4 types:
% 1) a string specifying the num2str conversion format,
% see help to num2str for details;
% 2) 'elems' - output is symbolic, elementwise (default);
% 3) 'cols' - output is symbolic, grouped by colums;
% 4) 'rows' - output is symbolic, grouped by rows.
%
% OUTPUT:
% none
%
% Examples:
% A = rand(5); figure(1);clf; dispMEq('A', A);
% A = rand(5); figure(1);clf; dispMEq('A', A, 'format', {'%1.1f'});
% A = randn(5,3); x=rand(3,1); figure(1); clf; dispMEq('A*x=b', A, x, A*x);
% A = rand(5); [L,U]=lu(A); figure(1);clf; dispMEq('A=L*U',A,L,U);
% A = nan(5); figure(1);clf; dispMEq('A',A);
% A = nan(5); figure(1);clf; dispMEq('A',A,'format',{'cols'});
% A = nan(3,2); B = nan(2,4); figure(1);clf; dispMEq('A*B',A,B,'format',{'rows','cols'});
% y = {'\alpha' '\beta' '\gamma';'bla' 'bla' 'bla'}; figure(1);clf; dispMEq('y',y, 'format',{'rows'});
%
% See also: .
% Copyright (c) 2009, Dr. Vladimir Bondarenko <http://sites.google.com/site/bondsite>
% Check input:
error(nargchk(2,15,nargin));
if ~ischar(eq), error('The equation argument must be a string.'); end;
% Defaults:
global maxEntry minEntry
txtFormat = '%-5.2f'; % Default text format
% Parse Input:
regexprep(eq, ' ', ''); % remove white spaces
% ix = regexp(eq, '[=,+,\-,*,\\,/,;]'); % find arithmetic operations
ix = regexp(eq, '[=,+,\-,*,/,;]'); % find arithmetic operations
nOp = length(ix);
ix = [0 ix length(eq)+1];
if nargin<nOp+2, error('Number of numeric arguments is less then defined in the equation.');end;
for ii=1:nOp+1
if ~isnumeric(varargin{ii})
if iscell(varargin{ii})
A(ii).data = nan(size(varargin{ii}));
A(ii).max = nan;
A(ii).min = nan;
A(ii).userText = varargin{ii};
A(ii).txtFormat = txtFormat;
A(ii).m = size(A(ii).data,1);
A(ii).n = size(A(ii).data,2);
A(ii).symName = eq(ix(ii)+1:ix(ii+1)-1);
else
error('Error in the assignment of numeric arguments.');
end
else
A(ii).data = varargin{ii};
A(ii).max = max(A(ii).data(:));
A(ii).min = min(A(ii).data(:));
A(ii).m = size(A(ii).data,1);
A(ii).n = size(A(ii).data,2);
A(ii).symName = eq(ix(ii)+1:ix(ii+1)-1);
A(ii).txtFormat = txtFormat;
A(ii).userText = [];
end
end
maxEntry = max([A(:).max]);
minEntry = min([A(:).min]);
% Parse optional arguments:
if nargin > nOp+2
for ii=nOp+2:2:nargin-1
switch lower(varargin{ii})
case 'format'
txtFormat = varargin{ii+1};
if ~iscell(txtFormat), error('Format parameter must be a cell array.');end
dum = length(txtFormat);
if (dum>1)&&(dum~=nOp+1), error('Number of formats must either 1 or equal to the number of arguments.');end
for jj=1:nOp+1
if ~isempty(txtFormat{ (dum>1)*jj + (dum==1) })
A(jj).txtFormat = txtFormat{(dum>1)*jj + (dum==1)};
end
end
otherwise
error(['Unknown input parameter: ' varargin{ii}]);
end
end
end
% MAIN
sc = 3; % column scaling factor for the subplots
dum1 = sum([A(:).n])*sc + nOp; % overall number of subplots
dum2 = cumsum([A(:).n]*sc) + (0:nOp); % overall number of columns
% Display matrices
for ii=1:nOp+1
hsp(ii) = subplot(1,dum1,[(dum2(ii)-A(ii).n*sc+1) dum2(ii)]);
if max(A(ii).m,A(ii).n)<=20 % Display text if max dimension does not exceed 20.
dispMatrix(A(ii).data, A(ii).txtFormat);
dispText(A(ii));
else
imagesc(A(ii).data(1:min(A(ii).m,30),1:min(A(ii).n,30)), [minEntry maxEntry]);
axis equal tight off
title(A(ii).symName, 'FontSize', 16);
end
end
% Display arithmetic operations
for ii=1:nOp
op = eq(ix(ii+1));
% if strcmp(op, '*'), op='\times'; end;
if strcmp(op, '*'), op='x'; end;
% subplot(1,dum1, dum2(ii)+1);
pos1 = get(hsp(ii) , 'position');
pos2 = get(hsp(ii+1), 'position');
w = pos2(1)-(pos1(1)+pos1(3));
h = pos1(4);
subplot('position',[pos1(1)+pos1(3) pos1(2) w h])
text(.3,.5,op, 'FontUnits', 'normalized', 'FontSize', 0.05);
axis equal tight off
end
set(gcf, 'color', 'w');
end
%% Subfunctions:
%%
function dispMatrix(A,txtFormat)
% Plot matrix array to the current plot
global maxEntry minEntry
[m, n] = size(A);
if isnan(A)
switch lower(txtFormat)
case 'cols'
dum = 1:n;
A = dum(ones(m,1),:);
minEntry = 0;
maxEntry = n;
case 'rows'
dum = (1:m)';
A = dum(:,ones(n,1));
minEntry = 0;
maxEntry = m;
case 'elem'
A = reshape(1:m*n, m, n);
minEntry = 0;
maxEntry = m*n;
case 'user'
A = nan(m,n);
otherwise
end
end
% Display matrix
if isnan(A), minEntry = 0; maxEntry = 1;end;
imagesc(A, [minEntry maxEntry]);
cmap = colormap(jet(128));
colormap(cmap(30:end-10,:));
axis equal tight
set(gca, 'XTick',[], 'YTick', []);
end
%%
function dispText(A)
[m, n] = size(A.data);
m1 = m; n1 = n;
% Convert matrix to string
if ~isempty(A.userText)
txt = char(A.userText);
fcolor = 'k';
else
[txt, fcolor] = matrix2str(A.data, A.symName, A.txtFormat);
end
% Display text
title(A.symName, 'FontSize', 16);
axPos = get(gca, 'position');
[ii,jj] = meshgrid(1:n,1:m);
fsz = .45*max([axPos(3) axPos(4)])/max([m1 n1]);
ii = ii - .4*(m1/m);
jj = jj + .1;
text(ii(:),jj(:),txt, 'FontUnits', 'normalized',...
'FontSize', fsz,...
'Color', fcolor);
end
%%
function [txt,fcolor] = matrix2str(M,symName,txtFormat)
[m, n] = size(M);
[ii,jj] = meshgrid(1:n,1:m);
switch lower(txtFormat)
case 'cols'
if isscalar(M)==1
txt = symName; % No indices for 1-by-1 matrices, i.e., for scalars.
else
if n==1
dum = lower(symName);
txt = dum(ones(m,1),:);
else
txt = strcat(lower(symName), '_{', num2str(ii(:)), '}');
end
dum1 = ceil(m/2); dum2 = reshape(1:m*n,m,n);
ix = dum2([1:dum1-1,dum1+1:end],:);
txt(ix(:),:) = ' ';
end
fcolor = 'k';
case 'rows'
if isscalar(M)
txt = symName; % No indices for 1-by-1 matrices, i.e., for scalars.
else
if m==1
dum = lower(symName);
txt = dum(ones(n,1),:);
else
txt = strcat(lower(symName), '^{', num2str(jj(:)), '}');
end
dum1 = ceil(n/2); dum2 = reshape(1:m*n,m,n);
ix = dum2(:,[1:dum1-1,dum1+1:end]);
txt(ix(:),:) = ' ';
end
fcolor = 'k';
otherwise
if isnan(M) % Non-numeric input
fcolor = 'w';
if isscalar(M)
txt = symName; % No indices for 1-by-1 matrices, i.e., for scalars.
else
txt = strcat(symName, '_{', num2str(jj(:)), num2str(ii(:)), '}');
end
else % Numerical input
fcolor = 'k';
if abs(M-round(M))<realmin, txtFormat = '%-5.0f'; end % No fractional part for integers.
txt = num2str(M(:), txtFormat);
end
end
end
|
github
|
rising-turtle/slam_matlab-master
|
drawCircle.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/all/drawLA/drawCircle.m
| 3,103 |
utf_8
|
488a37088332d5ce21c6b8df98889e42
|
function drawCircle(varargin)
% Draw circle(s) in the xy-plane.
%
% Usage: drawCircle
% drawCircle(xc,yc,r)
% drawCircle(xc,yc,r, 'lType')
%
% INPUT:
% (optional)
% xc,yc - n-by-m matrices with (x,y) coordinates of the center(s)
% of n*m circles. Default: (0, 0).
% r - either a scalar or an n-by-m matrix of circle(s) radii.
% Default: 1.
% 'lType' - a string defining the line style and width (e.g., '2r-.').
% Default: 'b-'.
%
% OUTPUT:
% none
%
% Examples in 2D:
% clf; drawCircle;
% clf; drawCircle(3, 2, '2r-.');
% clf; x = randn(2,3); y = randn(2,3); r = rand(2,3); drawCircle(x,y,r,'gs');
% Examples in 3D:
% clf; drawVector([1 1 2]); hold on; drawCircle([0 1], [0 -1], [1,.5], '2r-'); hold off;
%
% See also: drawSphere, drawPlane, drawVector, drawLine, drawXLine, drawYLine.
% Copyright (c) 2009, Dr. Vladimir Bondarenko <http://sites.google.com/site/bondsite>
% Check input:
error(nargchk(0,4,nargin));
% Defaults:
N = 100; k = 0:N; % Number of points on the circle(s)
xc = 0; yc = 0; r = 1; % Center coordinates and radius.
lType = 'k-';
% Parse input:
for ii=1:nargin
if ischar(varargin{ii})
lType = varargin{ii};
elseif isscalar(varargin{ii})
if nargin==1
r = varargin{ii};
elseif ii==1
xc = varargin{ii};
elseif ii==2
yc = varargin{ii};
end
else
if ii==1
xc = varargin{ii};
elseif ii==2
yc = varargin{ii};
else
r = varargin{ii};
end
end
end
% Check input
if ~all(size(xc)==size(yc)), error('Dimensions of xc and yc must be equal.'); end;
if ~isscalar(r)&&~all(size(xc)==size(r))
error('Wrong dimensions of r. Must be scalar or matrix with dimensins of xc and yc');
end;
% Parse the line parameters
[lStyle,lWidth,lColor, lMarker] = parseLineType(lType);
% MAIN:
holdon = get(gca, 'NextPlot'); % Capture the NextPlot property
for ii=1:size(xc,1)
for jj = 1:size(xc,2)
xy = r(ii,jj)*exp(2*pi*1i*k./N) + xc(ii,jj) + 1i*yc(ii,jj);
x = real(xy); y = imag(xy);
line(x, y, 'LineStyle', lStyle, ...
'LineWidth', lWidth, ...
'Color' , lColor, ...
'Marker' , lMarker );
hold on;
end
end
axis equal
set(gca, 'NextPlot', holdon); % restore the NextPlot property
function [lStyle,lWidth,lColor, lMarker] = parseLineType(lType)
% Parse the line type
% get line style
lStyles = '--|:|-\.|-';
[dum1,dum2,dum3, lStyle] = regexp(lType, lStyles, 'once');
if isempty(lStyle), lStyle = 'none'; end
% get width
[dum1,dum2,dum3, lWidth] = regexp(lType, '\d*', 'once');
if isempty(lWidth), lWidth = 1; else lWidth = str2double(lWidth); end
% get color
lColors = 'y|m|c|r|g|b|w|k';
[dum1,dum2,dum3, lColor] = regexp(lType, lColors, 'once');
if isempty(lColor), lColor = 'k'; end
% get marker
lMarkers = '\+|o|\*|\.|x|s|d|\^|>|<|v|p|h|';
[dum1,dum2,dum3, lMarker] = regexp(lType, lMarkers, 'once');
if isempty(lMarker), lMarker = 'none'; end
|
github
|
rising-turtle/slam_matlab-master
|
efficient_pnp_gauss.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/all/EPnP_matlab/EPnP/efficient_pnp_gauss.m
| 7,985 |
utf_8
|
361396729ae9c9291df547c60c062e74
|
function [R,T,Xc,best_solution,opt]=efficient_pnp_gauss(x3d_h,x2d_h,A)
% EFFICIENT_PNP_GAUSS Main Function to solve the PnP problem
% as described in:
%
% Francesc Moreno-Noguer, Vincent Lepetit, Pascal Fua.
% Accurate Non-Iterative O(n) Solution to the PnP Problem.
% In Proceedings of ICCV, 2007.
%
% Note: In this version of the software we perform a final
% optimization using Gauss-Newton,which is not described in the
% paper.
%
% x3d_h: homogeneous coordinates of the points in world reference
% x2d_h: homogeneous position of the points in the image plane
% A: intrincic camera parameters
% R: Rotation of the camera system wrt world reference
% T: Translation of the camera system wrt world reference
% Xc: Position of the points in the camera reference
% best solution: dimension of the kernel for the best solution
% (before applying Gauss Newton).
% opt: some parameters of the optimization process
%
% Copyright (C) <2007> <Francesc Moreno-Noguer, Vincent Lepetit, Pascal Fua>
%
% This program is free software: you can redistribute it and/or modify
% it under the terms of the version 3 of the GNU General Public License
% as published by the Free Software Foundation.
%
% This program is distributed in the hope that it will be useful, but
% WITHOUT ANY WARRANTY; without even the implied warranty of
% MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
% General Public License for more details.
% You should have received a copy of the GNU General Public License
% along with this program. If not, see <http://www.gnu.org/licenses/>.
%
% Francesc Moreno-Noguer, CVLab-EPFL, October 2007.
% [email protected], http://cvlab.epfl.ch/~fmoreno/
Xw=x3d_h(:,1:3);
U=x2d_h(:,1:2);
THRESHOLD_REPROJECTION_ERROR=20;%error in degrees of the basis formed by the control points.
%If we have a larger error, we will compute the solution using a larger
%number of vectors in the kernel
%define control points in a world coordinate system (centered on the 3d
%points centroid)
Cw=define_control_points();
%compute alphas (linear combination of the control points to represent the 3d
%points)
Alph=compute_alphas(Xw,Cw);
%Compute M
M=compute_M_ver2(U,Alph,A);
%Compute kernel M
Km=kernel_noise(M,4); %in matlab we have directly the funcion km=null(M);
%1.-Solve assuming dim(ker(M))=1. X=[Km_end];------------------------------
dim_kerM=1;
X1=Km(:,end);
[Cc,Xc,sc]=compute_norm_sign_scaling_factor(X1,Cw,Alph,Xw);
[R,T]=getrotT(Xw,Xc); %solve exterior orientation
err(1)=reprojection_error_usingRT(Xw,U,R,T,A);
sol(1).Xc=Xc;
sol(1).Cc=Cc;
sol(1).R=R;
sol(1).T=T;
sol(1).error=err(1);
sol(1).betas=[1];
sol(1).sc=sc;
sol(1).Kernel=X1;
%2.-Solve assuming dim(ker(M))=2------------------------------------------
Km1=Km(:,end-1);
Km2=Km(:,end);
%control points distance constraint
D=compute_constraint_distance_2param_6eq_3unk(Km1,Km2);
dsq=define_distances_btw_control_points();
betas_=inv(D'*D)*D'*dsq;
beta1=sqrt(abs(betas_(1)));
beta2=sqrt(abs(betas_(3)))*sign(betas_(2))*sign(betas_(1));
X2=beta1*Km1+beta2*Km2;
[Cc,Xc,sc]=compute_norm_sign_scaling_factor(X2,Cw,Alph,Xw);
[R,T]=getrotT(Xw,Xc); %solve exterior orientation
err(2)=reprojection_error_usingRT(Xw,U,R,T,A);
sol(2).Xc=Xc;
sol(2).Cc=Cc;
sol(2).R=R;
sol(2).T=T;
sol(2).error=err(2);
sol(2).betas=[beta1,beta2];
sol(2).sc=sc;
sol(2).Kernel=[Km1,Km2];
%3.-Solve assuming dim(ker(M))=3------------------------------------------
if min(err)>THRESHOLD_REPROJECTION_ERROR %just compute if we do not have good solution in the previus cases
Km1=Km(:,end-2);
Km2=Km(:,end-1);
Km3=Km(:,end);
%control points distance constraint
D=compute_constraint_distance_3param_6eq_6unk(Km1,Km2,Km3);
dsq=define_distances_btw_control_points();
betas_=inv(D)*dsq;
beta1=sqrt(abs(betas_(1)));
beta2=sqrt(abs(betas_(4)))*sign(betas_(2))*sign(betas_(1));
beta3=sqrt(abs(betas_(6)))*sign(betas_(3))*sign(betas_(1));
X3=beta1*Km1+beta2*Km2+beta3*Km3;
[Cc,Xc,sc]=compute_norm_sign_scaling_factor(X3,Cw,Alph,Xw);
[R,T]=getrotT(Xw,Xc); %solve exterior orientation
err(3)=reprojection_error_usingRT(Xw,U,R,T,A);
sol(3).Xc=Xc;
sol(3).Cc=Cc;
sol(3).R=R;
sol(3).T=T;
sol(3).error=err(3);
sol(3).betas=[beta1,beta2,beta3];
sol(3).sc=sc;
sol(3).Kernel=[Km1,Km2,Km3];
end
%4.-Solve assuming dim(ker(M))=4------------------------------------------
if min(err)>THRESHOLD_REPROJECTION_ERROR %just compute if we do not have good solution in the previus cases
Km1=Km(:,end-3);
Km2=Km(:,end-2);
Km3=Km(:,end-1);
Km4=Km(:,end);
D=compute_constraint_distance_orthog_4param_9eq_10unk(Km1,Km2,Km3,Km4);
dsq=define_distances_btw_control_points();
lastcolumn=[-dsq',0,0,0]';
D_=[D,lastcolumn];
Kd=null(D_);
P=compute_permutation_constraint4(Kd);
lambdas_=kernel_noise(P,1);
lambda(1)=sqrt(abs(lambdas_(1)));
lambda(2)=sqrt(abs(lambdas_(6)))*sign(lambdas_(2))*sign(lambdas_(1));
lambda(3)=sqrt(abs(lambdas_(10)))*sign(lambdas_(3))*sign(lambdas_(1));
lambda(4)=sqrt(abs(lambdas_(13)))*sign(lambdas_(4))*sign(lambdas_(1));
lambda(5)=sqrt(abs(lambdas_(15)))*sign(lambdas_(5))*sign(lambdas_(1));
betass_=lambda(1)*Kd(:,1)+lambda(2)*Kd(:,2)+lambda(3)*Kd(:,3)+lambda(4)*Kd(:,4)+lambda(5)*Kd(:,5);
beta1=sqrt(abs(betass_(1)));
beta2=sqrt(abs(betass_(5)))*sign(betass_(2));
beta3=sqrt(abs(betass_(8)))*sign(betass_(3));
beta4=sqrt(abs(betass_(10)))*sign(betass_(4));
X4=beta1*Km1+beta2*Km2+beta3*Km3+beta4*Km4;
[Cc,Xc,sc]=compute_norm_sign_scaling_factor(X4,Cw,Alph,Xw);
[R,T]=getrotT(Xw,Xc); %solve exterior orientation
err(4)=reprojection_error_usingRT(Xw,U,R,T,A);
sol(4).Xc=Xc;
sol(4).Cc=Cc;
sol(4).R=R;
sol(4).T=T;
sol(4).error=err(4);
sol(4).betas=[beta1,beta2,beta3,beta4];
sol(4).sc=sc;
sol(4).Kernel=[Km1,Km2,Km3,Km4];
end
%5.-Gauss Newton Optimization------------------------------------------------------
[min_err,best_solution]=min(err);
Xc=sol(best_solution).Xc;
R=sol(best_solution).R;
T=sol(best_solution).T;
Betas=sol(best_solution).betas;
sc=sol(best_solution).sc;
Kernel=sol(best_solution).Kernel;
if best_solution==1
Betas=[0,0,0,Betas];
elseif best_solution==2
Betas=[0,0,Betas];
elseif best_solution==3
Betas=[0,Betas];
end
Km1=Km(:,end-3);
Km2=Km(:,end-2);
Km3=Km(:,end-1);
Km4=Km(:,end);
Kernel=[Km1,Km2,Km3,Km4];
%refine the solution iterating over the betas
Beta0=Betas/sc;
[Xc_opt,R_opt,T_opt,err_opt,iter]=optimize_betas_gauss_newton(Kernel,Cw,Beta0,Alph,Xw,U,A);
%Just update R,T,Xc if Gauss Newton improves results (which is almost
%always)
if err_opt<min_err
R=R_opt;
T=T_opt;
Xc=Xc_opt;
end
opt.Beta0=Beta0;
opt.Kernel=Kernel;
opt.iter=iter;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
function [R, T]=getrotT(wpts,cpts)
% This routine solves the exterior orientation problem for a point cloud
% given in both camera and world coordinates.
% wpts = 3D points in arbitrary reference frame
% cpts = 3D points in camera reference frame
n=size(wpts,1);
M=zeros(3);
ccent=mean(cpts);
wcent=mean(wpts);
for i=1:3
cpts(:,i)=cpts(:,i)-ccent(i)*ones(n,1);
wpts(:,i)=wpts(:,i)-wcent(i)*ones(n,1);
end
for i=1:n
M=M+cpts(i,:)'*wpts(i,:);
end
[U S V]=svd(M);
R=U*V';
if det(R)<0
R=-R;
end
T=ccent'-R*wcent';
%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
function [err,Urep]=reprojection_error_usingRT(Xw,U,R,T,A)
%clear all; close all; load reprojection_error_usingRT;
n=size(Xw,1);
P=A*[R,T];
Xw_h=[Xw,ones(n,1)];
Urep_=(P*Xw_h')';
%project reference points into the image plane
Urep=zeros(n,2);
Urep(:,1)=Urep_(:,1)./Urep_(:,3);
Urep(:,2)=Urep_(:,2)./Urep_(:,3);
%reprojection error
err_=sqrt((U(:,1)-Urep(:,1)).^2+(U(:,2)-Urep(:,2)).^2);
err=sum(err_)/n;
|
github
|
rising-turtle/slam_matlab-master
|
efficient_pnp.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/all/EPnP_matlab/EPnP/efficient_pnp.m
| 6,550 |
utf_8
|
b1b02989deb052da7480f60d06be7010
|
function [R,T,Xc,best_solution]=efficient_pnp(x3d_h,x2d_h,A)
% EFFICIENT_PNP Main Function to solve the PnP problem
% as described in:
%
% Francesc Moreno-Noguer, Vincent Lepetit, Pascal Fua.
% Accurate Non-Iterative O(n) Solution to the PnP Problem.
% In Proceedings of ICCV, 2007.
%
% x3d_h: homogeneous coordinates of the points in world reference
% x2d_h: homogeneous position of the points in the image plane
% A: intrincic camera parameters
% R: Rotation of the camera system wrt world reference
% T: Translation of the camera system wrt world reference
% Xc: Position of the points in the camera reference
%
% Copyright (C) <2007> <Francesc Moreno-Noguer, Vincent Lepetit, Pascal Fua>
%
% This program is free software: you can redistribute it and/or modify
% it under the terms of the version 3 of the GNU General Public License
% as published by the Free Software Foundation.
%
% This program is distributed in the hope that it will be useful, but
% WITHOUT ANY WARRANTY; without even the implied warranty of
% MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
% General Public License for more details.
% You should have received a copy of the GNU General Public License
% along with this program. If not, see <http://www.gnu.org/licenses/>.
%
% Francesc Moreno-Noguer, CVLab-EPFL, September 2007.
% [email protected], http://cvlab.epfl.ch/~fmoreno/
Xw=x3d_h(:,1:3);
U=x2d_h(:,1:2);
THRESHOLD_REPROJECTION_ERROR=20;%error in degrees of the basis formed by the control points.
%If we have a larger error, we will compute the solution using a larger
%number of vectors in the kernel
%define control points in a world coordinate system (centered on the 3d
%points centroid)
Cw=define_control_points();
%compute alphas (linear combination of the control points to represent the 3d
%points)
Alph=compute_alphas(Xw,Cw);
%Compute M
M=compute_M_ver2(U,Alph,A);
%Compute kernel M
Km=kernel_noise(M,4); %in matlab we have directly the funcion km=null(M);
%1.-Solve assuming dim(ker(M))=1. X=[Km_end];------------------------------
dim_kerM=1;
X1=Km(:,end);
[Cc,Xc]=compute_norm_sign_scaling_factor(X1,Cw,Alph,Xw);
[R,T]=getrotT(Xw,Xc); %solve exterior orientation
err(1)=reprojection_error_usingRT(Xw,U,R,T,A);
sol(1).Xc=Xc;
sol(1).Cc=Cc;
sol(1).R=R;
sol(1).T=T;
sol(1).error=err(1);
%2.-Solve assuming dim(ker(M))=2------------------------------------------
Km1=Km(:,end-1);
Km2=Km(:,end);
%control points distance constraint
D=compute_constraint_distance_2param_6eq_3unk(Km1,Km2);
dsq=define_distances_btw_control_points();
betas_=inv(D'*D)*D'*dsq;
beta1=sqrt(abs(betas_(1)));
beta2=sqrt(abs(betas_(3)))*sign(betas_(2))*sign(betas_(1));
X2=beta1*Km1+beta2*Km2;
[Cc,Xc]=compute_norm_sign_scaling_factor(X2,Cw,Alph,Xw);
[R,T]=getrotT(Xw,Xc); %solve exterior orientation
err(2)=reprojection_error_usingRT(Xw,U,R,T,A);
sol(2).Xc=Xc;
sol(2).Cc=Cc;
sol(2).R=R;
sol(2).T=T;
sol(2).error=err(2);
%3.-Solve assuming dim(ker(M))=3------------------------------------------
if min(err)>THRESHOLD_REPROJECTION_ERROR %just compute if we do not have good solution in the previus cases
Km1=Km(:,end-2);
Km2=Km(:,end-1);
Km3=Km(:,end);
%control points distance constraint
D=compute_constraint_distance_3param_6eq_6unk(Km1,Km2,Km3);
dsq=define_distances_btw_control_points();
betas_=inv(D)*dsq;
beta1=sqrt(abs(betas_(1)));
beta2=sqrt(abs(betas_(4)))*sign(betas_(2))*sign(betas_(1));
beta3=sqrt(abs(betas_(6)))*sign(betas_(3))*sign(betas_(1));
X3=beta1*Km1+beta2*Km2+beta3*Km3;
[Cc,Xc]=compute_norm_sign_scaling_factor(X3,Cw,Alph,Xw);
[R,T]=getrotT(Xw,Xc); %solve exterior orientation
err(3)=reprojection_error_usingRT(Xw,U,R,T,A);
sol(3).Xc=Xc;
sol(3).Cc=Cc;
sol(3).R=R;
sol(3).T=T;
sol(3).error=err(3);
end
%4.-Solve assuming dim(ker(M))=4------------------------------------------
if min(err)>THRESHOLD_REPROJECTION_ERROR %just compute if we do not have good solution in the previus cases
Km1=Km(:,end-3);
Km2=Km(:,end-2);
Km3=Km(:,end-1);
Km4=Km(:,end);
D=compute_constraint_distance_orthog_4param_9eq_10unk(Km1,Km2,Km3,Km4);
dsq=define_distances_btw_control_points();
lastcolumn=[-dsq',0,0,0]';
D_=[D,lastcolumn];
Kd=null(D_);
P=compute_permutation_constraint4(Kd);
lambdas_=kernel_noise(P,1);
lambda(1)=sqrt(abs(lambdas_(1)));
lambda(2)=sqrt(abs(lambdas_(6)))*sign(lambdas_(2))*sign(lambdas_(1));
lambda(3)=sqrt(abs(lambdas_(10)))*sign(lambdas_(3))*sign(lambdas_(1));
lambda(4)=sqrt(abs(lambdas_(13)))*sign(lambdas_(4))*sign(lambdas_(1));
lambda(5)=sqrt(abs(lambdas_(15)))*sign(lambdas_(5))*sign(lambdas_(1));
betass_=lambda(1)*Kd(:,1)+lambda(2)*Kd(:,2)+lambda(3)*Kd(:,3)+lambda(4)*Kd(:,4)+lambda(5)*Kd(:,5);
beta1=sqrt(abs(betass_(1)));
beta2=sqrt(abs(betass_(5)))*sign(betass_(2));
beta3=sqrt(abs(betass_(8)))*sign(betass_(3));
beta4=sqrt(abs(betass_(10)))*sign(betass_(4));
X4=beta1*Km1+beta2*Km2+beta3*Km3+beta4*Km4;
[Cc,Xc]=compute_norm_sign_scaling_factor(X4,Cw,Alph,Xw);
[R,T]=getrotT(Xw,Xc); %solve exterior orientation
err(4)=reprojection_error_usingRT(Xw,U,R,T,A);
sol(4).Xc=Xc;
sol(4).Cc=Cc;
sol(4).R=R;
sol(4).T=T;
sol(4).error=err(4);
end
[min_err,best_solution]=min(err);
Xc=sol(best_solution).Xc;
R=sol(best_solution).R;
T=sol(best_solution).T;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
function [R, T]=getrotT(wpts,cpts)
% This routine solves the exterior orientation problem for a point cloud
% given in both camera and world coordinates.
% wpts = 3D points in arbitrary reference frame
% cpts = 3D points in camera reference frame
n=size(wpts,1);
M=zeros(3);
ccent=mean(cpts);
wcent=mean(wpts);
for i=1:3
cpts(:,i)=cpts(:,i)-ccent(i)*ones(n,1);
wpts(:,i)=wpts(:,i)-wcent(i)*ones(n,1);
end
for i=1:n
M=M+cpts(i,:)'*wpts(i,:);
end
[U S V]=svd(M);
R=U*V';
if det(R)<0
R=-R;
end
T=ccent'-R*wcent';
%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
function [err,Urep]=reprojection_error_usingRT(Xw,U,R,T,A)
%clear all; close all; load reprojection_error_usingRT;
n=size(Xw,1);
P=A*[R,T];
Xw_h=[Xw,ones(n,1)];
Urep_=(P*Xw_h')';
%project reference points into the image plane
Urep=zeros(n,2);
Urep(:,1)=Urep_(:,1)./Urep_(:,3);
Urep(:,2)=Urep_(:,2)./Urep_(:,3);
%reprojection error
err_=sqrt((U(:,1)-Urep(:,1)).^2+(U(:,2)-Urep(:,2)).^2);
err=sum(err_)/n;
|
github
|
rising-turtle/slam_matlab-master
|
robust_dls_pnp.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/all/dls_pnp_matlab/robust_dls_pnp.m
| 956 |
utf_8
|
884e74eace41c8ba00c468ea725dca02
|
function [C_est, t_est, cost, flag] = robust_dls_pnp(p, z)
R = cat(3, rotx(pi/2), roty(pi/2), rotz(pi/2));
t = mean(p,2);
cost = inf;
for i = 1:3
% Make a random rotation
pp = R(:,:,i) * (p - repmat(t, 1, size(p,2)));
[C_est_i, t_est_i, cost_i, flag_i] = dls_pnp(pp, z);
for j = 1:length(cost_i)
t_est_i(:,j) = t_est_i(:,j) - C_est_i(:,:,j) * R(:,:,i) * t;
C_est_i(:,:,j) = C_est_i(:,:,j) * R(:,:,i);
end
if min(cost_i) < min(cost)
C_est = C_est_i;
t_est = t_est_i;
cost = cost_i;
flag = flag_i;
end
end
end
function r = rotx(t)
ct = cos(t);
st = sin(t);
r = [1 0 0;
0 ct -st;
0 st ct];
end
function r = roty(t)
% roty: rotation about y-axi-
ct = cos(t);
st = sin(t);
r = [ct 0 st;
0 1 0;
-st 0 ct];
end
function r = rotz(t)
% rotz: rotation about z-axis
ct = cos(t);
st = sin(t);
r = [ct -st 0
st ct 0
0 0 1];
end
|
github
|
rising-turtle/slam_matlab-master
|
rws.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/all/dls_pnp_matlab/rws.m
| 1,946 |
utf_8
|
24dabc05143cbf2c126ab352eb83b32f
|
function [C, t, p, z] = rws(N, sigma)
% this function generates a random camera pose, along with N random points,
% and also the perspective projections of those points.
% Generate a random global-to-camera rotation. This is the orientation of
% the global frame expressed in the camera frame of refence.
angle = 15;
C = rotx(angle * randn * pi/180 ) * roty( angle * randn * pi/180 );
% Generate a random global-to-camera translation. This is the origin of the
% global frame expressed in the camera frame.
t = randn(3,1);
% Create random 3D points within a 45 deg FOV (vertical and horizontal) of
% the camera. The points are between 0.5 and 5.5 meters from the camera.
Psens = zeros(3,N);
theta = (rand(N,1)*45 - 22.5) * pi/180;
phi = (rand(N,1)*45 - 22.5) * pi/180;
for i = 1:N
psens_unit = rotx(theta(i)) * roty(phi(i)) * [0;0;1];
alpha = rand * 5 + 0.5;
Psens(:,i) = alpha * psens_unit;
end
% Express the points in the global frame of reference
p = C' *(Psens - repmat(t,1,N));
% Construct the vector of perspective projections (i.e., image
% measurements) of the points,
z = zeros(2,N);
for i = 1:N
% create an instance of 2x1 pixel noise
noise = sigma * randn(2,1);
% You can uncomment the following lines in order to limit the noise to +/-
% 3 sigma
%
%
% if abs(noise(1)) > 3 * sigma
% noise(1) = sign(noise(1)) * 3 * sigma;
% end
% if abs(noise(2)) > 3 * sigma
% noise(2) = sign(noise(2)) * 3 * sigma;
% end
% Create the image measurement using the standard pinhole camera model
z(:,i) = [ Psens(1,i) / Psens(3,i) ; Psens(2,i) / Psens(3,i)] + noise;
end
end
function r = rotx(t)
%rotx: rotation around the x-axis
ct = cos(t);
st = sin(t);
r = [1 0 0;
0 ct -st;
0 st ct];
end
function r = roty(t)
% roty: rotation about y-axis
ct = cos(t);
st = sin(t);
r = [ct 0 st;
0 1 0;
-st 0 ct];
end
|
github
|
rising-turtle/slam_matlab-master
|
dls_pnp.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/all/dls_pnp_matlab/dls_pnp.m
| 52,158 |
utf_8
|
d6e96e3b74cc29c6f5ddd01bbdb2e1eb
|
function [C_est, t_est, cost, flag] = dls_pnp(p, z)
% DLS-PnP:
%
% This function performs the DLS-PnP method introduced at ICCV 2011
% Joel A. Hesch and Stergios I. Roumeliotis. "A direct least-squares (dls)
% solution for PnP". In Proc. of the Int. Conf. on Computer Vision,
% Barcelona, Spain, November 6-13, 2011.
%
% inputs:
% p: 3xN vector of 3D known point features
% z: 2xN vector of correpsonding image measurements (calibrated)
% Check the inputs
if size(z,1) > size(z,2) || size(p,1) > size(p,2)
fprintf('Usage: dls_pnp(p,z) \n p: 3xN matrix of 3D points \n z: 2xN matrix of corresponding 2D image measurements (normalized pixel coordinates)')
end
% make z into unit vectors from normalized pixel coords
z = [z; ones(1,size(z,2))];
z = z./ repmat(sqrt(sum(z.*z,1)),3,1);
% some preliminaries
flag = 0;
N = size(z,2);
% build coeff matrix
% An intermediate matrix, the inverse of what is called "H" in the paper
% (see eq. 25)
H = zeros(3);
for i = 1:N
H = H + eye(3) - z(:,i)*z(:,i)';
end
A = zeros(3,9);
for i = 1:N
A = A + (z(:,i)*z(:,i)' - eye(3)) * LeftMultVec(p(:,i));
end
A = H\A;
D = zeros(9);
for i = 1:N
D = D + (LeftMultVec(p(:,i)) + A)' * (eye(3) - z(:,i)*z(:,i)') * (LeftMultVec(p(:,i)) + A);
end
f1coeff = [2*D(1,6) - 2*D(1,8) + 2*D(5,6) - 2*D(5,8) + 2*D(6,1) + 2*D(6,5) + 2*D(6,9) - 2*D(8,1) - 2*D(8,5) - 2*D(8,9) + 2*D(9,6) - 2*D(9,8); % constant term
(6*D(1,2) + 6*D(1,4) + 6*D(2,1) - 6*D(2,5) - 6*D(2,9) + 6*D(4,1) - 6*D(4,5) - 6*D(4,9) - 6*D(5,2) - 6*D(5,4) - 6*D(9,2) - 6*D(9,4)); % s1^2 * s2
(4*D(1,7) - 4*D(1,3) + 8*D(2,6) - 8*D(2,8) - 4*D(3,1) + 4*D(3,5) + 4*D(3,9) + 8*D(4,6) - 8*D(4,8) + 4*D(5,3) - 4*D(5,7) + 8*D(6,2) + 8*D(6,4) + 4*D(7,1) - 4*D(7,5) - 4*D(7,9) - 8*D(8,2) - 8*D(8,4) + 4*D(9,3) - 4*D(9,7)); % s1 * s2
(4*D(1,2) - 4*D(1,4) + 4*D(2,1) - 4*D(2,5) - 4*D(2,9) + 8*D(3,6) - 8*D(3,8) - 4*D(4,1) + 4*D(4,5) + 4*D(4,9) - 4*D(5,2) + 4*D(5,4) + 8*D(6,3) + 8*D(6,7) + 8*D(7,6) - 8*D(7,8) - 8*D(8,3) - 8*D(8,7) - 4*D(9,2) + 4*D(9,4)); % s1 * s3
(8*D(2,2) - 8*D(3,3) - 8*D(4,4) + 8*D(6,6) + 8*D(7,7) - 8*D(8,8)); % s2 * s3
(4*D(2,6) - 2*D(1,7) - 2*D(1,3) + 4*D(2,8) - 2*D(3,1) + 2*D(3,5) - 2*D(3,9) + 4*D(4,6) + 4*D(4,8) + 2*D(5,3) + 2*D(5,7) + 4*D(6,2) + 4*D(6,4) - 2*D(7,1) + 2*D(7,5) - 2*D(7,9) + 4*D(8,2) + 4*D(8,4) - 2*D(9,3) - 2*D(9,7)); % s2^2 * s3
(2*D(2,5) - 2*D(1,4) - 2*D(2,1) - 2*D(1,2) - 2*D(2,9) - 2*D(4,1) + 2*D(4,5) - 2*D(4,9) + 2*D(5,2) + 2*D(5,4) - 2*D(9,2) - 2*D(9,4)); % s2^3
(4*D(1,9) - 4*D(1,1) + 8*D(3,3) + 8*D(3,7) + 4*D(5,5) + 8*D(7,3) + 8*D(7,7) + 4*D(9,1) - 4*D(9,9)); % s1 * s3^2
(4*D(1,1) - 4*D(5,5) - 4*D(5,9) + 8*D(6,6) - 8*D(6,8) - 8*D(8,6) + 8*D(8,8) - 4*D(9,5) - 4*D(9,9)); % s1
(2*D(1,3) + 2*D(1,7) + 4*D(2,6) - 4*D(2,8) + 2*D(3,1) + 2*D(3,5) + 2*D(3,9) - 4*D(4,6) + 4*D(4,8) + 2*D(5,3) + 2*D(5,7) + 4*D(6,2) - 4*D(6,4) + 2*D(7,1) + 2*D(7,5) + 2*D(7,9) - 4*D(8,2) + 4*D(8,4) + 2*D(9,3) + 2*D(9,7)); % s3
(2*D(1,2) + 2*D(1,4) + 2*D(2,1) + 2*D(2,5) + 2*D(2,9) - 4*D(3,6) + 4*D(3,8) + 2*D(4,1) + 2*D(4,5) + 2*D(4,9) + 2*D(5,2) + 2*D(5,4) - 4*D(6,3) + 4*D(6,7) + 4*D(7,6) - 4*D(7,8) + 4*D(8,3) - 4*D(8,7) + 2*D(9,2) + 2*D(9,4)); % s2
(2*D(2,9) - 2*D(1,4) - 2*D(2,1) - 2*D(2,5) - 2*D(1,2) + 4*D(3,6) + 4*D(3,8) - 2*D(4,1) - 2*D(4,5) + 2*D(4,9) - 2*D(5,2) - 2*D(5,4) + 4*D(6,3) + 4*D(6,7) + 4*D(7,6) + 4*D(7,8) + 4*D(8,3) + 4*D(8,7) + 2*D(9,2) + 2*D(9,4)); % s2 * s3^2
(6*D(1,6) - 6*D(1,8) - 6*D(5,6) + 6*D(5,8) + 6*D(6,1) - 6*D(6,5) - 6*D(6,9) - 6*D(8,1) + 6*D(8,5) + 6*D(8,9) - 6*D(9,6) + 6*D(9,8)); % s1^2
(2*D(1,8) - 2*D(1,6) + 4*D(2,3) + 4*D(2,7) + 4*D(3,2) - 4*D(3,4) - 4*D(4,3) - 4*D(4,7) - 2*D(5,6) + 2*D(5,8) - 2*D(6,1) - 2*D(6,5) + 2*D(6,9) + 4*D(7,2) - 4*D(7,4) + 2*D(8,1) + 2*D(8,5) - 2*D(8,9) + 2*D(9,6) - 2*D(9,8)); % s3^2
(2*D(1,8) - 2*D(1,6) - 4*D(2,3) + 4*D(2,7) - 4*D(3,2) - 4*D(3,4) - 4*D(4,3) + 4*D(4,7) + 2*D(5,6) - 2*D(5,8) - 2*D(6,1) + 2*D(6,5) - 2*D(6,9) + 4*D(7,2) + 4*D(7,4) + 2*D(8,1) - 2*D(8,5) + 2*D(8,9) - 2*D(9,6) + 2*D(9,8)); % s2^2
(2*D(3,9) - 2*D(1,7) - 2*D(3,1) - 2*D(3,5) - 2*D(1,3) - 2*D(5,3) - 2*D(5,7) - 2*D(7,1) - 2*D(7,5) + 2*D(7,9) + 2*D(9,3) + 2*D(9,7)); % s3^3
(4*D(1,6) + 4*D(1,8) + 8*D(2,3) + 8*D(2,7) + 8*D(3,2) + 8*D(3,4) + 8*D(4,3) + 8*D(4,7) - 4*D(5,6) - 4*D(5,8) + 4*D(6,1) - 4*D(6,5) - 4*D(6,9) + 8*D(7,2) + 8*D(7,4) + 4*D(8,1) - 4*D(8,5) - 4*D(8,9) - 4*D(9,6) - 4*D(9,8)); % s1 * s2 * s3
(4*D(1,5) - 4*D(1,1) + 8*D(2,2) + 8*D(2,4) + 8*D(4,2) + 8*D(4,4) + 4*D(5,1) - 4*D(5,5) + 4*D(9,9)); % s1 * s2^2
(6*D(1,3) + 6*D(1,7) + 6*D(3,1) - 6*D(3,5) - 6*D(3,9) - 6*D(5,3) - 6*D(5,7) + 6*D(7,1) - 6*D(7,5) - 6*D(7,9) - 6*D(9,3) - 6*D(9,7)); % s1^2 * s3
(4*D(1,1) - 4*D(1,5) - 4*D(1,9) - 4*D(5,1) + 4*D(5,5) + 4*D(5,9) - 4*D(9,1) + 4*D(9,5) + 4*D(9,9))]; % s1^3
f2coeff = [- 2*D(1,3) + 2*D(1,7) - 2*D(3,1) - 2*D(3,5) - 2*D(3,9) - 2*D(5,3) + 2*D(5,7) + 2*D(7,1) + 2*D(7,5) + 2*D(7,9) - 2*D(9,3) + 2*D(9,7); % constant term
(4*D(1,5) - 4*D(1,1) + 8*D(2,2) + 8*D(2,4) + 8*D(4,2) + 8*D(4,4) + 4*D(5,1) - 4*D(5,5) + 4*D(9,9)); % s1^2 * s2
(4*D(1,8) - 4*D(1,6) - 8*D(2,3) + 8*D(2,7) - 8*D(3,2) - 8*D(3,4) - 8*D(4,3) + 8*D(4,7) + 4*D(5,6) - 4*D(5,8) - 4*D(6,1) + 4*D(6,5) - 4*D(6,9) + 8*D(7,2) + 8*D(7,4) + 4*D(8,1) - 4*D(8,5) + 4*D(8,9) - 4*D(9,6) + 4*D(9,8)); % s1 * s2
(8*D(2,2) - 8*D(3,3) - 8*D(4,4) + 8*D(6,6) + 8*D(7,7) - 8*D(8,8)); % s1 * s3
(4*D(1,4) - 4*D(1,2) - 4*D(2,1) + 4*D(2,5) - 4*D(2,9) - 8*D(3,6) - 8*D(3,8) + 4*D(4,1) - 4*D(4,5) + 4*D(4,9) + 4*D(5,2) - 4*D(5,4) - 8*D(6,3) + 8*D(6,7) + 8*D(7,6) + 8*D(7,8) - 8*D(8,3) + 8*D(8,7) - 4*D(9,2) + 4*D(9,4)); % s2 * s3
(6*D(5,6) - 6*D(1,8) - 6*D(1,6) + 6*D(5,8) - 6*D(6,1) + 6*D(6,5) - 6*D(6,9) - 6*D(8,1) + 6*D(8,5) - 6*D(8,9) - 6*D(9,6) - 6*D(9,8)); % s2^2 * s3
(4*D(1,1) - 4*D(1,5) + 4*D(1,9) - 4*D(5,1) + 4*D(5,5) - 4*D(5,9) + 4*D(9,1) - 4*D(9,5) + 4*D(9,9)); % s2^3
(2*D(2,9) - 2*D(1,4) - 2*D(2,1) - 2*D(2,5) - 2*D(1,2) + 4*D(3,6) + 4*D(3,8) - 2*D(4,1) - 2*D(4,5) + 2*D(4,9) - 2*D(5,2) - 2*D(5,4) + 4*D(6,3) + 4*D(6,7) + 4*D(7,6) + 4*D(7,8) + 4*D(8,3) + 4*D(8,7) + 2*D(9,2) + 2*D(9,4)); % s1 * s3^2
(2*D(1,2) + 2*D(1,4) + 2*D(2,1) + 2*D(2,5) + 2*D(2,9) - 4*D(3,6) + 4*D(3,8) + 2*D(4,1) + 2*D(4,5) + 2*D(4,9) + 2*D(5,2) + 2*D(5,4) - 4*D(6,3) + 4*D(6,7) + 4*D(7,6) - 4*D(7,8) + 4*D(8,3) - 4*D(8,7) + 2*D(9,2) + 2*D(9,4)); % s1
(2*D(1,6) + 2*D(1,8) - 4*D(2,3) + 4*D(2,7) - 4*D(3,2) + 4*D(3,4) + 4*D(4,3) - 4*D(4,7) + 2*D(5,6) + 2*D(5,8) + 2*D(6,1) + 2*D(6,5) + 2*D(6,9) + 4*D(7,2) - 4*D(7,4) + 2*D(8,1) + 2*D(8,5) + 2*D(8,9) + 2*D(9,6) + 2*D(9,8)); % s3
(8*D(3,3) - 4*D(1,9) - 4*D(1,1) - 8*D(3,7) + 4*D(5,5) - 8*D(7,3) + 8*D(7,7) - 4*D(9,1) - 4*D(9,9)); % s2
(4*D(1,1) - 4*D(5,5) + 4*D(5,9) + 8*D(6,6) + 8*D(6,8) + 8*D(8,6) + 8*D(8,8) + 4*D(9,5) - 4*D(9,9)); % s2 * s3^2
(2*D(1,7) - 2*D(1,3) + 4*D(2,6) - 4*D(2,8) - 2*D(3,1) + 2*D(3,5) + 2*D(3,9) + 4*D(4,6) - 4*D(4,8) + 2*D(5,3) - 2*D(5,7) + 4*D(6,2) + 4*D(6,4) + 2*D(7,1) - 2*D(7,5) - 2*D(7,9) - 4*D(8,2) - 4*D(8,4) + 2*D(9,3) - 2*D(9,7)); % s1^2
(2*D(1,3) - 2*D(1,7) + 4*D(2,6) + 4*D(2,8) + 2*D(3,1) + 2*D(3,5) - 2*D(3,9) - 4*D(4,6) - 4*D(4,8) + 2*D(5,3) - 2*D(5,7) + 4*D(6,2) - 4*D(6,4) - 2*D(7,1) - 2*D(7,5) + 2*D(7,9) + 4*D(8,2) - 4*D(8,4) - 2*D(9,3) + 2*D(9,7)); % s3^2
(6*D(1,3) - 6*D(1,7) + 6*D(3,1) - 6*D(3,5) + 6*D(3,9) - 6*D(5,3) + 6*D(5,7) - 6*D(7,1) + 6*D(7,5) - 6*D(7,9) + 6*D(9,3) - 6*D(9,7)); % s2^2
(2*D(6,9) - 2*D(1,8) - 2*D(5,6) - 2*D(5,8) - 2*D(6,1) - 2*D(6,5) - 2*D(1,6) - 2*D(8,1) - 2*D(8,5) + 2*D(8,9) + 2*D(9,6) + 2*D(9,8)); % s3^3
(8*D(2,6) - 4*D(1,7) - 4*D(1,3) + 8*D(2,8) - 4*D(3,1) + 4*D(3,5) - 4*D(3,9) + 8*D(4,6) + 8*D(4,8) + 4*D(5,3) + 4*D(5,7) + 8*D(6,2) + 8*D(6,4) - 4*D(7,1) + 4*D(7,5) - 4*D(7,9) + 8*D(8,2) + 8*D(8,4) - 4*D(9,3) - 4*D(9,7)); % s1 * s2 * s3
(6*D(2,5) - 6*D(1,4) - 6*D(2,1) - 6*D(1,2) - 6*D(2,9) - 6*D(4,1) + 6*D(4,5) - 6*D(4,9) + 6*D(5,2) + 6*D(5,4) - 6*D(9,2) - 6*D(9,4)); % s1 * s2^2
(2*D(1,6) + 2*D(1,8) + 4*D(2,3) + 4*D(2,7) + 4*D(3,2) + 4*D(3,4) + 4*D(4,3) + 4*D(4,7) - 2*D(5,6) - 2*D(5,8) + 2*D(6,1) - 2*D(6,5) - 2*D(6,9) + 4*D(7,2) + 4*D(7,4) + 2*D(8,1) - 2*D(8,5) - 2*D(8,9) - 2*D(9,6) - 2*D(9,8)); % s1^2 * s3
(2*D(1,2) + 2*D(1,4) + 2*D(2,1) - 2*D(2,5) - 2*D(2,9) + 2*D(4,1) - 2*D(4,5) - 2*D(4,9) - 2*D(5,2) - 2*D(5,4) - 2*D(9,2) - 2*D(9,4))]; % s1^3
f3coeff = [2*D(1,2) - 2*D(1,4) + 2*D(2,1) + 2*D(2,5) + 2*D(2,9) - 2*D(4,1) - 2*D(4,5) - 2*D(4,9) + 2*D(5,2) - 2*D(5,4) + 2*D(9,2) - 2*D(9,4); % constant term
(2*D(1,6) + 2*D(1,8) + 4*D(2,3) + 4*D(2,7) + 4*D(3,2) + 4*D(3,4) + 4*D(4,3) + 4*D(4,7) - 2*D(5,6) - 2*D(5,8) + 2*D(6,1) - 2*D(6,5) - 2*D(6,9) + 4*D(7,2) + 4*D(7,4) + 2*D(8,1) - 2*D(8,5) - 2*D(8,9) - 2*D(9,6) - 2*D(9,8)); % s1^2 * s2
(8*D(2,2) - 8*D(3,3) - 8*D(4,4) + 8*D(6,6) + 8*D(7,7) - 8*D(8,8)); % s1 * s2
(4*D(1,8) - 4*D(1,6) + 8*D(2,3) + 8*D(2,7) + 8*D(3,2) - 8*D(3,4) - 8*D(4,3) - 8*D(4,7) - 4*D(5,6) + 4*D(5,8) - 4*D(6,1) - 4*D(6,5) + 4*D(6,9) + 8*D(7,2) - 8*D(7,4) + 4*D(8,1) + 4*D(8,5) - 4*D(8,9) + 4*D(9,6) - 4*D(9,8)); % s1 * s3
(4*D(1,3) - 4*D(1,7) + 8*D(2,6) + 8*D(2,8) + 4*D(3,1) + 4*D(3,5) - 4*D(3,9) - 8*D(4,6) - 8*D(4,8) + 4*D(5,3) - 4*D(5,7) + 8*D(6,2) - 8*D(6,4) - 4*D(7,1) - 4*D(7,5) + 4*D(7,9) + 8*D(8,2) - 8*D(8,4) - 4*D(9,3) + 4*D(9,7)); % s2 * s3
(4*D(1,1) - 4*D(5,5) + 4*D(5,9) + 8*D(6,6) + 8*D(6,8) + 8*D(8,6) + 8*D(8,8) + 4*D(9,5) - 4*D(9,9)); % s2^2 * s3
(2*D(5,6) - 2*D(1,8) - 2*D(1,6) + 2*D(5,8) - 2*D(6,1) + 2*D(6,5) - 2*D(6,9) - 2*D(8,1) + 2*D(8,5) - 2*D(8,9) - 2*D(9,6) - 2*D(9,8)); % s2^3
(6*D(3,9) - 6*D(1,7) - 6*D(3,1) - 6*D(3,5) - 6*D(1,3) - 6*D(5,3) - 6*D(5,7) - 6*D(7,1) - 6*D(7,5) + 6*D(7,9) + 6*D(9,3) + 6*D(9,7)); % s1 * s3^2
(2*D(1,3) + 2*D(1,7) + 4*D(2,6) - 4*D(2,8) + 2*D(3,1) + 2*D(3,5) + 2*D(3,9) - 4*D(4,6) + 4*D(4,8) + 2*D(5,3) + 2*D(5,7) + 4*D(6,2) - 4*D(6,4) + 2*D(7,1) + 2*D(7,5) + 2*D(7,9) - 4*D(8,2) + 4*D(8,4) + 2*D(9,3) + 2*D(9,7)); % s1
(8*D(2,2) - 4*D(1,5) - 4*D(1,1) - 8*D(2,4) - 8*D(4,2) + 8*D(4,4) - 4*D(5,1) - 4*D(5,5) + 4*D(9,9)); % s3
(2*D(1,6) + 2*D(1,8) - 4*D(2,3) + 4*D(2,7) - 4*D(3,2) + 4*D(3,4) + 4*D(4,3) - 4*D(4,7) + 2*D(5,6) + 2*D(5,8) + 2*D(6,1) + 2*D(6,5) + 2*D(6,9) + 4*D(7,2) - 4*D(7,4) + 2*D(8,1) + 2*D(8,5) + 2*D(8,9) + 2*D(9,6) + 2*D(9,8)); % s2
(6*D(6,9) - 6*D(1,8) - 6*D(5,6) - 6*D(5,8) - 6*D(6,1) - 6*D(6,5) - 6*D(1,6) - 6*D(8,1) - 6*D(8,5) + 6*D(8,9) + 6*D(9,6) + 6*D(9,8)); % s2 * s3^2
(2*D(1,2) - 2*D(1,4) + 2*D(2,1) - 2*D(2,5) - 2*D(2,9) + 4*D(3,6) - 4*D(3,8) - 2*D(4,1) + 2*D(4,5) + 2*D(4,9) - 2*D(5,2) + 2*D(5,4) + 4*D(6,3) + 4*D(6,7) + 4*D(7,6) - 4*D(7,8) - 4*D(8,3) - 4*D(8,7) - 2*D(9,2) + 2*D(9,4)); % s1^2
(6*D(1,4) - 6*D(1,2) - 6*D(2,1) - 6*D(2,5) + 6*D(2,9) + 6*D(4,1) + 6*D(4,5) - 6*D(4,9) - 6*D(5,2) + 6*D(5,4) + 6*D(9,2) - 6*D(9,4)); % s3^2
(2*D(1,4) - 2*D(1,2) - 2*D(2,1) + 2*D(2,5) - 2*D(2,9) - 4*D(3,6) - 4*D(3,8) + 2*D(4,1) - 2*D(4,5) + 2*D(4,9) + 2*D(5,2) - 2*D(5,4) - 4*D(6,3) + 4*D(6,7) + 4*D(7,6) + 4*D(7,8) - 4*D(8,3) + 4*D(8,7) - 2*D(9,2) + 2*D(9,4)); % s2^2
(4*D(1,1) + 4*D(1,5) - 4*D(1,9) + 4*D(5,1) + 4*D(5,5) - 4*D(5,9) - 4*D(9,1) - 4*D(9,5) + 4*D(9,9)); % s3^3
(4*D(2,9) - 4*D(1,4) - 4*D(2,1) - 4*D(2,5) - 4*D(1,2) + 8*D(3,6) + 8*D(3,8) - 4*D(4,1) - 4*D(4,5) + 4*D(4,9) - 4*D(5,2) - 4*D(5,4) + 8*D(6,3) + 8*D(6,7) + 8*D(7,6) + 8*D(7,8) + 8*D(8,3) + 8*D(8,7) + 4*D(9,2) + 4*D(9,4)); % s1 * s2 * s3
(4*D(2,6) - 2*D(1,7) - 2*D(1,3) + 4*D(2,8) - 2*D(3,1) + 2*D(3,5) - 2*D(3,9) + 4*D(4,6) + 4*D(4,8) + 2*D(5,3) + 2*D(5,7) + 4*D(6,2) + 4*D(6,4) - 2*D(7,1) + 2*D(7,5) - 2*D(7,9) + 4*D(8,2) + 4*D(8,4) - 2*D(9,3) - 2*D(9,7)); % s1 * s2^2
(4*D(1,9) - 4*D(1,1) + 8*D(3,3) + 8*D(3,7) + 4*D(5,5) + 8*D(7,3) + 8*D(7,7) + 4*D(9,1) - 4*D(9,9)); % s1^2 * s3
(2*D(1,3) + 2*D(1,7) + 2*D(3,1) - 2*D(3,5) - 2*D(3,9) - 2*D(5,3) - 2*D(5,7) + 2*D(7,1) - 2*D(7,5) - 2*D(7,9) - 2*D(9,3) - 2*D(9,7))]; % s1^3
% Construct the Macaulay matrix
% u0 = round(randn(1)*100);
% u1 = round(randn(1)*100);
% u2 = round(randn(1)*100);
% u3 = round(randn(1)*100);
%M2 = cayley_LS_M(f1coeff, f2coeff, f3coeff,u0,u1,u2,u3);
u = round(randn(4,1) * 100);
M2 = cayley_LS_M(f1coeff, f2coeff, f3coeff,u);
% construct the multiplication matrix via schur compliment of the Macaulay
% matrix
Mtilde = M2(1:27,1:27) - M2(1:27,28:120)/M2(28:120,28:120)*M2(28:120,1:27);
[V,~] = eig(Mtilde);
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% Now check the solutions %%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% extract the optimal solutions from the eigen decomposition of the
% Multiplication matrix
sols = zeros(3,27);
cost = zeros(27,1);
i = 1;
for k = 1:27
V(:,k) = V(:,k)/V(1,k);
if (imag(V(2,k)) == 0)
stmp = V([10 4 2],k);
H = Hessian(f1coeff, f2coeff, f3coeff, stmp);
if eig(H) > 0
sols(:,i) = stmp;
Cbar = cayley2rotbar(stmp);
CbarVec = Cbar';
CbarVec = CbarVec(:);
cost(i) = CbarVec' * D * CbarVec;
i = i+1;
end
end
end
sols = sols(:,1:i-1);
cost = cost(1:i-1);
C_est = zeros(3,3,size(sols,2));
t_est = zeros(3, size(sols,2));
for j = 1:size(sols,2)
% recover the optimal orientation
C_est(:,:,j) = 1/(1 + sols(:,j)' * sols(:,j)) * cayley2rotbar(sols(:,j));
A2 = zeros(3);
for i = 1:N
A2 = A2 + eye(3) - z(:,i)*z(:,i)';
end
b2 = zeros(3,1);
for i = 1:N
b2 = b2 + (z(:,i)*z(:,i)' - eye(3)) * C_est(:,:,j) * p(:,i);
end
% recover the optimal translation
t_est(:,j) = A2\b2;
end
% check that the points are infront of the center of perspectivity
sols_valid = [];
for k = 1:size(sols,2)
cam_points = C_est(:,:,k) * p + repmat(t_est(:,k),1, length(p));
if isempty(find(cam_points(3,:) < 0))
sols_valid = [sols_valid; k];
end
end
t_est = t_est(:,sols_valid);
C_est = C_est(:,:,sols_valid);
cost = cost(sols_valid);
end
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% Some helper functions %%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
function C = cayley2rotbar(s)
C = ( (1-s'*s) * eye(3) + 2 * skewsymm(s) + 2 * (s * s'))';
end
function C = skewsymm(X1)
% generates skew symmetric matrix
C = [0 , -X1(3) , X1(2)
X1(3) , 0 , -X1(1)
-X1(2) , X1(1) , 0];
end
function M = LeftMultVec(v)
% R * p = LeftMultVec(p) * vec(R)
M = [v' zeros(1,6);
zeros(1,3) v' zeros(1,3);
zeros(1,6) v'];
end
function M = cayley_LS_M(a,b,c,u) %,u1,u2,u3)
% Construct the Macaulay resultant matrix
M = [u(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(1) 0; u(4) u(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(1) a(10) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(10) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(10) 0; 0 u(4) u(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(10) a(14) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(1) 0 0 b(10) 0 0 0 0 0 0 0 0 0 0 b(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(14) 0 0 0 0 0 c(1) 0 0 0 0 0 0 0 0 0 c(10) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(14) 0; u(3) 0 0 u(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(11) 0 0 0 0 0 0 0 0 0 0 0 0 0 a(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(1) 0 0 0 0 0 0 b(11) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(11) c(1); 0 u(3) 0 u(4) u(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(11) a(5) 0 0 0 0 0 0 0 a(1) 0 0 0 0 0 a(10) 0 0 0 0 b(11) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(1) 0 0 0 0 b(10) 0 0 0 0 0 0 b(5) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(11) c(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(5) c(10); 0 0 u(3) 0 u(4) u(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(5) a(12) 0 0 0 0 0 a(1) 0 a(10) 0 0 0 0 0 a(14) 0 a(11) 0 0 b(5) 0 0 0 0 0 0 0 b(1) 0 0 b(11) 0 0 0 0 0 b(10) 0 0 0 0 b(14) 0 0 0 0 0 0 b(12) 0 0 0 0 0 c(11) 0 0 0 0 0 0 0 0 0 c(5) c(10) 0 0 0 0 0 0 0 0 0 0 c(1) 0 0 0 0 0 0 c(12) c(14); 0 0 0 u(3) 0 0 u(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(1) 0 0 0 0 a(15) 0 0 0 0 0 0 0 0 0 0 0 0 0 a(11) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(1) b(11) 0 0 0 0 0 0 b(15) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(1) 0 0 0 0 0 0 0 0 0 0 c(15) c(11); 0 0 0 0 u(3) 0 u(4) u(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(10) 0 0 0 a(15) a(6) 0 0 0 0 0 0 0 a(11) 0 a(1) 0 0 0 a(5) 0 0 0 0 b(15) 0 0 0 0 0 0 0 0 b(1) 0 0 0 0 0 0 0 b(11) 0 0 0 b(10) b(5) 0 0 0 0 0 0 b(6) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(15) c(11) 0 0 0 0 0 0 c(10) 0 0 0 0 c(1) 0 0 0 0 0 c(6) c(5); 0 0 0 0 0 u(3) 0 u(4) u(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(14) 0 0 0 a(6) 0 0 0 0 0 0 a(11) 0 a(5) 0 a(10) a(1) 0 0 a(12) 0 a(15) 0 0 b(6) 0 0 0 0 0 0 0 b(11) b(10) 0 b(15) b(1) 0 0 0 0 b(5) 0 0 0 b(14) b(12) 0 0 0 0 0 0 0 0 0 0 0 0 c(15) 0 0 0 0 0 0 0 0 0 c(6) c(5) 0 c(1) 0 0 0 0 c(14) 0 0 0 c(11) c(10) 0 0 0 0 0 0 c(12); u(2) 0 0 0 0 0 0 0 0 u(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(9) a(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(9) b(1) 0 0 0 c(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(9) 0; 0 u(2) 0 0 0 0 0 0 0 u(4) u(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(1) a(9) a(4) a(10) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(9) 0 0 0 0 b(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(4) b(10) 0 0 0 c(10) 0 0 0 0 0 0 0 0 0 0 c(9) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(1) c(4) 0; 0 0 u(2) 0 0 0 0 0 0 0 u(4) u(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(1) 0 0 0 0 a(10) a(4) a(8) a(14) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(9) 0 0 b(4) 0 0 b(1) 0 b(10) 0 0 0 0 0 b(9) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(8) b(14) 0 0 0 c(14) c(9) 0 0 0 0 0 0 0 0 0 c(4) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(1) 0 0 c(10) c(8) 0; 0 0 0 u(2) 0 0 0 0 0 u(3) 0 0 u(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(3) a(11) 0 0 a(1) 0 0 0 0 0 0 0 0 0 a(9) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(9) 0 0 b(1) 0 0 0 b(3) b(11) 0 0 0 c(11) 0 0 0 0 0 0 0 c(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(3) c(9); 0 0 0 0 u(2) 0 0 0 0 0 u(3) 0 u(4) u(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(11) a(3) a(17) a(5) 0 0 a(10) 0 0 0 a(9) 0 0 0 a(1) 0 a(4) 0 0 0 0 b(3) 0 0 0 0 b(11) b(1) 0 0 0 0 0 0 0 0 0 0 b(9) 0 0 0 0 b(4) 0 0 b(10) 0 0 0 b(17) b(5) 0 0 0 c(5) 0 c(1) 0 0 0 0 0 c(10) 0 0 c(3) c(9) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(11) c(17) c(4); 0 0 0 0 0 u(2) 0 0 0 0 0 u(3) 0 u(4) u(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 a(11) 0 0 0 0 a(5) a(17) 0 a(12) 0 0 a(14) 0 a(9) a(1) a(4) 0 0 0 a(10) 0 a(8) 0 a(3) 0 0 b(17) 0 b(1) b(11) 0 b(5) b(10) 0 b(9) 0 0 b(3) 0 0 0 0 0 b(4) 0 0 0 0 b(8) 0 0 b(14) 0 0 0 0 b(12) 0 0 0 c(12) c(3) c(10) 0 0 0 0 0 c(14) 0 0 c(17) c(4) 0 0 0 0 0 c(1) 0 0 0 0 c(9) 0 0 c(11) 0 0 c(5) 0 c(8); 0 0 0 0 0 0 u(2) 0 0 0 0 0 u(3) 0 0 u(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 a(1) a(9) 0 0 0 0 a(18) a(15) 0 0 a(11) 0 0 0 0 0 0 0 0 0 a(3) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(1) b(9) b(3) 0 0 b(11) 0 0 0 b(18) b(15) 0 0 0 c(15) 0 0 0 0 0 c(1) 0 c(11) 0 0 0 0 0 0 0 0 0 0 c(9) 0 0 0 0 0 0 0 0 0 0 c(18) c(3); 0 0 0 0 0 0 0 u(2) 0 0 0 0 0 u(3) 0 u(4) u(1) 0 0 0 0 0 0 0 0 0 0 0 0 a(10) a(4) 0 0 a(15) a(18) 0 a(6) 0 0 a(5) 0 0 0 a(3) a(1) a(9) 0 a(11) 0 a(17) 0 0 0 0 b(18) 0 0 0 0 b(15) b(11) 0 0 b(9) 0 0 0 0 b(1) 0 0 b(3) 0 0 b(10) b(4) b(17) 0 0 b(5) 0 0 0 0 b(6) 0 0 0 c(6) 0 c(11) 0 0 0 c(10) 0 c(5) c(1) 0 c(18) c(3) 0 0 0 0 0 0 c(4) 0 0 0 0 c(9) 0 0 0 0 c(15) 0 c(17); 0 0 0 0 0 0 0 0 u(2) 0 0 0 0 0 u(3) 0 u(4) u(1) 0 0 0 0 0 0 0 0 0 0 a(15) a(14) a(8) 0 0 a(6) 0 0 0 0 0 a(12) 0 a(3) a(11) a(17) a(10) a(4) a(9) a(5) 0 0 0 a(18) 0 0 0 0 b(11) b(15) 0 b(6) b(5) 0 b(3) b(4) 0 b(18) b(9) 0 b(10) 0 0 b(17) 0 0 b(14) b(8) 0 0 0 b(12) 0 0 0 0 0 0 0 0 0 c(18) c(5) 0 0 0 c(14) 0 c(12) c(10) 0 0 c(17) 0 c(9) 0 0 0 c(11) c(8) 0 0 0 c(3) c(4) 0 c(15) 0 0 c(6) 0 0; 0 0 0 0 0 0 0 0 0 u(2) 0 0 0 0 0 0 0 0 u(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(13) a(9) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(1) b(13) b(9) 0 0 c(1) c(9) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(13) 0; 0 0 0 0 0 0 0 0 0 0 u(2) 0 0 0 0 0 0 0 u(4) u(1) 0 0 0 0 0 0 0 0 0 0 0 0 a(1) a(9) a(13) a(19) a(4) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(10) b(13) 0 0 0 0 b(9) 0 b(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(10) b(19) b(4) 0 0 c(10) c(4) 0 0 0 0 0 0 0 0 0 0 c(13) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(1) c(9) c(19) 0; 0 0 0 0 0 0 0 0 0 0 0 u(2) 0 0 0 0 0 0 0 u(4) u(1) 0 0 0 0 0 0 a(1) a(9) 0 0 0 a(10) a(4) a(19) 0 a(8) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(13) 0 a(14) b(19) b(1) 0 b(9) 0 b(4) 0 b(10) 0 0 0 b(13) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(14) 0 b(8) 0 0 c(14) c(8) c(13) 0 0 0 0 0 0 0 0 0 c(19) 0 0 0 0 0 0 0 0 0 0 0 0 0 c(1) c(9) 0 c(10) c(4) 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 u(2) 0 0 0 0 0 u(3) 0 0 u(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 a(2) a(3) 0 a(1) a(9) 0 0 0 0 0 0 0 0 0 a(13) 0 0 0 a(11) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(13) 0 b(1) b(9) 0 0 b(11) b(2) b(3) 0 0 c(11) c(3) 0 0 0 c(1) 0 0 0 c(9) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(2) c(13); 0 0 0 0 0 0 0 0 0 0 0 0 0 u(2) 0 0 0 0 0 u(3) 0 u(4) u(1) 0 0 0 0 0 0 0 0 0 a(11) a(3) a(2) 0 a(17) 0 a(10) a(4) a(1) 0 0 a(13) 0 0 0 a(9) 0 a(19) 0 0 0 a(5) b(2) 0 0 0 0 b(3) b(9) b(11) 0 0 0 0 0 0 0 0 0 b(13) b(1) 0 0 0 b(19) 0 b(10) b(4) 0 0 b(5) 0 b(17) 0 0 c(5) c(17) 0 c(9) 0 c(10) 0 0 c(1) c(4) 0 0 c(2) c(13) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(11) c(3) 0 c(19); 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(2) 0 0 0 0 0 u(3) 0 u(4) u(1) 0 0 0 a(11) a(3) 0 0 0 a(5) a(17) 0 0 0 0 a(14) a(8) a(10) a(13) a(9) a(19) 0 0 0 a(4) 0 0 0 a(2) 0 a(12) 0 b(11) b(9) b(3) 0 b(17) b(4) b(5) b(13) 0 0 b(2) 0 0 0 0 0 b(19) b(10) 0 0 0 0 0 b(14) b(8) 0 0 b(12) 0 0 0 0 c(12) 0 c(2) c(4) 0 c(14) 0 0 c(10) c(8) 0 0 0 c(19) 0 0 0 0 0 c(9) 0 0 0 0 c(13) 0 c(11) c(3) 0 c(5) c(17) 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(2) 0 0 0 0 0 u(3) 0 0 u(1) 0 0 0 0 a(9) a(13) 0 0 0 0 0 a(18) 0 a(11) a(3) 0 0 0 0 0 0 0 0 0 a(2) 0 0 a(1) a(15) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(1) b(9) b(13) b(2) 0 b(11) b(3) 0 0 b(15) 0 b(18) 0 0 c(15) c(18) 0 0 0 c(11) c(1) c(9) 0 c(3) 0 0 0 0 0 0 0 0 0 0 c(13) 0 0 0 0 0 0 0 0 0 0 0 c(2); 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(2) 0 0 0 0 0 u(3) 0 u(4) u(1) 0 0 0 a(4) a(19) 0 a(15) a(18) 0 0 0 0 a(5) a(17) a(11) 0 0 a(2) a(9) a(13) 0 a(3) 0 0 0 0 a(10) a(6) 0 0 0 0 0 b(18) b(3) b(15) 0 b(13) 0 0 0 0 b(9) 0 0 b(2) b(11) b(10) b(4) b(19) 0 0 b(5) b(17) 0 0 b(6) 0 0 0 0 c(6) 0 0 c(3) 0 c(5) c(10) c(4) c(11) c(17) c(9) 0 0 c(2) 0 0 0 0 0 0 c(19) 0 0 0 0 c(13) 0 0 0 c(15) c(18) 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(2) 0 0 0 0 0 u(3) 0 u(4) u(1) a(15) a(18) a(8) 0 0 a(6) 0 0 0 0 0 a(12) 0 a(5) a(2) a(3) 0 a(4) a(19) a(13) a(17) 0 0 0 0 a(14) 0 0 b(15) b(3) b(18) 0 0 b(17) b(6) b(2) b(19) 0 0 b(13) 0 b(4) 0 0 0 b(5) b(14) b(8) 0 0 0 b(12) 0 0 0 0 0 0 0 0 0 0 0 c(17) 0 c(12) c(14) c(8) c(5) 0 c(4) 0 0 0 0 c(13) 0 0 0 c(3) 0 0 0 0 c(2) c(19) c(15) c(18) 0 c(6) 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(3) 0 0 0 0 0 0 0 0 0 0 0 0 0 a(11) a(3) 0 0 0 0 0 a(7) 0 0 a(15) 0 0 0 0 0 0 0 0 0 a(18) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(9) b(1) 0 0 0 b(11) b(3) b(18) 0 0 b(15) 0 0 0 0 b(7) 0 0 0 c(7) 0 0 c(1) 0 0 c(11) 0 c(15) 0 0 0 0 0 0 0 0 0 0 c(3) 0 0 c(9) 0 0 0 0 0 0 0 0 c(18); 0 0 0 0 0 0 u(3) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(11) 0 0 0 0 a(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 a(15) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(1) 0 0 0 0 0 b(11) b(15) 0 0 0 0 0 0 b(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(11) 0 0 c(1) 0 0 0 0 0 0 0 c(7) c(15); 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(3) 0 0 0 0 a(3) a(2) 0 0 0 0 0 0 0 a(15) a(18) 0 0 0 0 0 0 0 0 0 0 0 0 a(11) a(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(13) b(9) 0 0 b(11) b(3) b(2) 0 0 b(15) b(18) 0 0 b(7) 0 0 0 0 c(7) 0 0 0 c(9) c(15) c(11) c(3) 0 c(18) 0 0 0 0 0 0 0 0 0 0 c(2) 0 0 c(13) 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(18) 0 0 0 0 0 0 0 0 0 0 a(7) 0 0 0 a(2) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(20) 0 0 b(2) 0 0 0 b(18) 0 0 0 b(7) 0 0 0 c(7) 0 0 0 0 0 c(20) 0 c(2) 0 0 0 0 c(18) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(2) 0 0 0 0 0 0 0 a(15) a(18) 0 0 0 0 0 0 0 0 0 0 0 a(7) 0 a(3) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(20) b(13) 0 0 b(3) b(2) 0 0 b(15) b(18) 0 b(7) 0 0 0 0 0 c(7) 0 0 0 0 c(13) c(18) c(3) c(2) 0 0 0 c(15) 0 0 0 0 0 0 0 0 0 0 0 c(20) 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(9) 0 a(13) 0 0 a(20) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(13) b(9) b(20) 0 0 c(9) c(13) c(20) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(15) a(18) 0 0 0 0 0 0 0 0 a(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(9) 0 0 0 0 b(3) b(11) 0 0 0 b(15) b(18) 0 0 0 b(7) 0 0 0 0 0 0 0 0 0 0 0 c(11) 0 0 c(15) 0 c(7) 0 0 0 0 0 0 c(9) 0 0 0 c(18) 0 0 c(3) 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(18) 0 0 0 0 0 0 0 0 a(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 a(15) 0 0 0 0 0 0 0 0 0 0 0 b(13) 0 0 0 0 b(2) b(3) 0 0 b(15) b(18) 0 0 0 b(7) 0 0 0 0 0 0 0 0 0 0 0 0 c(3) c(7) c(15) c(18) 0 0 0 0 0 0 0 0 c(13) 0 0 0 0 0 0 c(2) 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(18) 0 0 0 0 0 0 0 0 0 0 0 b(20) 0 0 0 0 0 b(2) 0 0 b(18) 0 0 0 b(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 c(2) 0 c(18) 0 0 0 0 c(7) 0 0 0 0 c(20) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 u(4) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(12) 0 0 0 0 0 0 a(10) 0 a(14) 0 0 0 0 0 a(16) 0 a(5) 0 0 b(12) 0 0 0 0 0 0 0 b(10) 0 0 b(5) 0 0 0 0 0 b(14) 0 0 0 0 b(16) 0 0 0 0 0 0 0 0 0 0 0 0 c(5) 0 0 0 0 0 0 0 0 0 c(12) c(14) c(1) 0 0 0 0 0 0 0 c(11) 0 c(10) 0 0 0 0 0 0 0 c(16); 0 0 u(4) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(14) a(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(10) 0 0 b(14) 0 0 0 0 0 0 0 0 0 0 b(10) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(16) 0 0 0 0 0 c(10) 0 0 0 0 0 0 0 0 0 c(14) 0 0 0 0 0 0 0 0 0 c(1) 0 0 0 0 0 0 0 0 c(16) 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(15) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(1) 0 0 0 0 b(11) 0 0 0 0 0 b(15) b(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(1) 0 0 0 c(15) 0 0 c(11) 0 0 0 0 0 0 0 0 c(7); 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(14) 0 0 0 0 a(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(8) 0 0 0 0 0 b(14) 0 b(16) 0 0 0 0 0 b(8) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(8) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(9) 0 0 c(10) c(4) 0 0 0 0 c(14) 0 0 c(16) 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(14) a(8) 0 0 0 a(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(14) 0 b(8) 0 0 0 b(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(13) 0 0 c(4) c(19) 0 0 0 c(14) c(8) 0 c(16) 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(11) 0 0 0 0 b(15) 0 0 0 0 0 b(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(11) 0 0 0 c(7) 0 0 c(15) 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(8) 0 0 0 a(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(8) 0 0 b(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(20) 0 0 c(19) 0 0 0 0 c(8) 0 c(16) 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(3) 0 0 0 0 b(18) b(15) 0 0 0 b(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(15) 0 0 c(7) 0 0 0 0 0 0 0 0 c(3) 0 0 0 0 0 0 c(18) 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(4) 0 0 c(14) c(8) 0 0 0 0 c(16) 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(19) 0 0 c(8) 0 0 0 0 c(16) 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(7) 0 0 0 0 0 0 0 0 0 0 0 b(2) 0 0 0 0 0 b(18) 0 0 b(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(18) 0 c(7) 0 0 0 0 0 0 0 0 0 c(2) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(7) 0 0 0 a(18) 0 0 0 0 0 0 0 a(6) 0 0 0 0 0 0 0 0 0 0 0 b(19) 0 0 b(2) b(18) 0 b(17) 0 b(7) b(6) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(17) 0 c(6) 0 c(7) 0 c(18) 0 0 0 0 0 c(19) c(2) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(6) 0 0 0 0 0 0 0 0 0 0 0 0 a(7) 0 a(15) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(15) b(10) 0 0 b(11) 0 b(5) 0 b(7) 0 0 0 b(6) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(7) 0 0 c(10) c(11) 0 0 c(6) 0 0 c(5) 0 c(15) 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(12) 0 0 0 a(16) a(14) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(12) b(16) 0 0 b(14) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(5) c(14) 0 0 c(15) 0 0 0 c(6) 0 c(12) c(16) 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(14) 0 0 0 c(5) 0 0 0 c(12) 0 c(16) 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(18) 0 0 0 0 0 b(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(7) 0 0 0 0 0 0 0 0 0 0 0 c(18) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(6) 0 0 0 a(12) a(5) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(6) b(12) 0 0 b(5) b(14) 0 b(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(15) c(5) 0 c(14) 0 0 0 0 c(7) c(16) c(6) c(12) 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(15) 0 0 0 0 b(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(15) 0 0 0 0 0 0 c(7) 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(14) 0 0 0 c(16) 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(17) 0 0 b(18) b(7) 0 b(6) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(6) 0 0 0 0 0 c(7) 0 0 0 0 0 c(17) c(18) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(8) 0 0 c(16) 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(6) 0 0 b(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(6) c(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(12) 0 b(7) b(6) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(7) c(12) c(6) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 u(4) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(16) 0 0 0 0 0 0 0 0 0 0 a(5) 0 a(12) 0 a(14) a(10) 0 0 0 0 a(6) 0 0 0 0 0 0 0 0 0 0 b(5) b(14) 0 b(6) b(10) 0 0 0 0 b(12) 0 0 0 b(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 c(6) 0 0 0 0 0 0 0 0 0 0 c(12) c(11) c(10) 0 0 0 0 c(16) 0 c(15) 0 c(5) c(14) 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 u(3) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(12) 0 0 0 0 0 0 0 0 0 0 a(15) 0 a(6) 0 a(5) a(11) 0 0 0 0 a(7) 0 0 0 0 0 0 0 0 0 0 b(15) b(5) 0 b(7) b(11) b(10) 0 b(14) 0 b(6) 0 0 0 b(12) 0 0 0 0 0 0 0 0 0 0 0 0 0 c(7) 0 0 0 0 0 0 0 0 0 0 c(6) 0 c(11) 0 c(10) 0 0 c(12) 0 0 c(14) c(15) c(5) 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 b(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(8) 0 0 0 c(17) c(16) 0 c(12) 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(12) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(12) b(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(6) c(12) 0 c(16) c(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(6) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(16) 0 b(6) b(12) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(7) c(6) c(16) c(12) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(12) c(16) 0 0 c(6) 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(16) 0 0 0 c(12) 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(14) 0 a(16) 0 0 0 0 0 0 0 a(12) 0 0 0 0 0 0 0 0 0 0 b(14) 0 0 b(12) 0 0 0 0 0 b(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(12) 0 0 0 0 0 0 0 0 0 0 c(16) c(10) 0 0 0 c(11) 0 0 0 c(5) 0 c(14) 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(2) 0 0 0 0 0 0 0 0 0 0 a(18) 0 0 0 a(20) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(20) 0 0 0 b(2) 0 0 0 b(18) 0 0 0 c(18) 0 0 0 0 0 0 0 c(20) 0 0 0 0 c(2) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(2) 0 0 0 0 0 0 a(9) a(13) 0 0 a(10) a(4) a(19) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(14) a(20) 0 a(8) 0 b(9) 0 b(13) b(10) b(19) 0 b(4) 0 0 0 b(20) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(14) 0 b(8) 0 0 0 c(14) c(8) 0 c(20) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(9) c(13) c(10) c(4) c(19) 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(7) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(7) b(5) 0 0 b(15) 0 b(6) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(5) c(15) 0 0 0 0 0 c(6) 0 c(7) 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(7) 0 0 0 a(6) a(15) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(7) b(6) b(14) 0 b(15) b(5) 0 b(12) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(15) c(14) c(5) 0 0 0 0 0 c(12) c(7) c(6) 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(13) 0 a(20) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(20) b(13) 0 0 0 c(13) c(20) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(3) 0 0 0 a(17) 0 0 a(7) 0 0 0 0 0 a(6) 0 a(15) 0 0 0 a(3) a(2) 0 a(18) 0 0 0 0 a(5) 0 0 0 0 0 0 0 b(18) b(7) 0 b(2) 0 0 0 b(13) b(3) b(19) b(4) 0 b(15) b(5) b(17) 0 0 0 b(6) 0 0 0 0 0 0 0 0 0 0 0 c(18) c(4) c(6) c(5) c(17) c(15) 0 c(3) 0 0 0 0 0 0 c(13) 0 0 0 0 0 c(19) 0 c(2) 0 0 0 c(7) 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(2) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(2) a(1) a(9) a(13) 0 0 0 0 0 0 0 0 0 a(20) a(11) 0 0 a(3) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(20) b(1) b(9) b(13) b(11) 0 b(3) 0 b(2) 0 c(11) c(3) c(2) 0 0 0 c(9) 0 0 0 c(13) 0 c(1) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(20); 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(20) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(20) 0 0 0 c(20) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(14) 0 0 b(16) 0 0 0 0 0 0 0 0 0 0 b(14) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(14) 0 0 0 0 0 0 0 0 0 c(16) 0 0 0 0 0 c(1) 0 0 0 c(10) 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(17) 0 0 0 a(12) 0 0 0 0 0 a(16) 0 0 a(8) 0 a(19) 0 0 0 0 0 0 0 0 0 0 0 0 b(17) b(19) 0 b(12) 0 0 0 0 0 0 0 0 0 0 0 0 0 b(8) 0 0 0 0 b(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(8) 0 0 c(16) 0 0 c(20) 0 0 0 0 c(19) 0 c(2) 0 0 0 0 c(17) 0 c(12) 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(12) 0 a(16) 0 a(8) 0 0 0 0 0 0 0 0 0 b(12) 0 0 0 0 0 0 0 0 0 b(8) 0 b(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(16) 0 0 0 c(17) c(8) 0 0 c(18) c(12) 0 c(6) 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(4) 0 0 0 0 0 0 0 0 0 0 0 0 0 a(5) 0 0 0 0 a(12) 0 0 0 0 0 a(16) 0 a(4) a(10) a(8) 0 0 0 a(14) 0 0 0 a(17) 0 0 0 0 b(10) b(5) 0 b(12) b(14) 0 b(4) 0 0 b(17) 0 0 0 0 0 b(8) 0 0 0 0 0 0 0 b(16) 0 0 0 0 0 0 0 0 0 c(17) c(14) 0 0 0 0 0 c(16) 0 0 0 c(8) c(9) 0 0 0 0 c(10) 0 c(11) c(3) 0 c(4) 0 0 c(5) 0 0 c(12) 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(3) a(2) 0 0 0 0 a(4) a(19) 0 a(13) 0 0 0 0 0 0 a(20) a(5) 0 a(17) 0 0 0 0 0 0 0 b(3) 0 b(20) b(2) 0 0 0 0 0 0 0 0 0 0 b(13) 0 0 0 0 b(4) b(19) 0 b(17) b(5) 0 0 0 c(5) c(17) 0 0 0 c(20) 0 c(19) 0 0 c(13) 0 0 c(4) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(3) c(2) 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(6) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(15) a(18) 0 a(7) 0 0 0 0 0 0 0 0 0 0 0 0 b(7) 0 0 b(18) b(4) 0 0 b(3) b(15) b(17) b(5) 0 0 0 b(6) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(7) c(5) 0 0 c(6) 0 0 c(15) 0 0 0 0 0 c(4) c(3) 0 0 0 0 0 c(17) 0 c(18) 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(2) a(18) 0 0 0 a(6) 0 0 0 0 0 a(12) 0 0 a(17) 0 a(2) 0 a(19) 0 a(20) 0 0 0 0 0 a(8) 0 0 b(18) b(2) 0 b(6) 0 0 0 0 0 0 0 b(20) 0 b(19) 0 0 0 b(17) b(8) 0 0 0 b(12) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(8) 0 c(17) 0 c(19) c(12) 0 0 0 c(20) 0 0 0 c(2) 0 0 0 0 0 0 c(18) 0 c(6) 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(3) 0 0 0 0 0 0 0 0 0 0 0 0 a(5) a(17) 0 0 a(7) 0 0 0 0 0 a(6) 0 0 0 a(18) a(11) a(3) 0 a(15) 0 0 0 0 0 0 0 0 0 0 0 b(7) b(15) 0 0 b(3) 0 0 0 b(9) b(11) b(4) b(10) b(18) 0 0 b(5) b(17) 0 0 0 b(6) 0 0 0 0 0 0 0 0 0 0 c(15) c(10) 0 0 c(5) 0 c(6) c(11) 0 0 c(18) 0 0 0 c(9) 0 0 c(17) 0 0 c(4) 0 c(3) 0 0 0 0 c(7) 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(2) 0 0 0 a(19) 0 a(15) a(18) 0 0 0 0 a(5) a(17) 0 a(3) 0 0 0 a(13) a(20) 0 a(2) 0 0 a(6) 0 a(4) 0 0 0 0 0 b(15) 0 b(2) b(18) 0 b(20) 0 0 0 0 b(13) 0 0 0 b(3) b(4) b(19) 0 0 b(5) b(17) 0 b(6) 0 0 0 0 0 c(6) 0 0 0 c(2) 0 c(17) c(4) c(19) c(3) 0 c(13) c(5) 0 0 0 0 0 0 0 0 0 0 0 0 0 c(20) 0 0 c(15) c(18) 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(18) 0 0 0 0 0 a(17) 0 0 a(2) 0 0 0 a(20) 0 0 0 a(6) 0 0 0 a(19) 0 0 0 0 0 b(18) 0 0 0 0 0 0 0 0 0 b(20) 0 0 0 b(2) b(19) 0 0 0 b(17) 0 0 0 b(6) 0 0 0 c(6) 0 0 0 0 0 0 0 c(19) 0 c(2) 0 c(20) c(17) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(18) 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(2) 0 0 0 0 0 0 0 0 a(11) a(3) a(2) 0 0 0 a(10) a(4) a(19) a(9) 0 0 a(20) 0 0 0 a(13) 0 0 a(5) 0 0 a(17) 0 0 0 0 b(11) b(2) b(13) b(3) 0 0 0 0 0 0 0 0 0 b(20) b(9) 0 0 0 0 b(10) b(4) b(19) b(5) 0 b(17) 0 0 0 c(5) c(17) 0 0 c(13) 0 c(4) 0 0 c(9) c(19) 0 c(10) 0 c(20) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(11) c(3) c(2) 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(2) 0 0 0 a(17) 0 0 0 0 0 a(8) 0 0 a(19) 0 a(20) 0 0 0 0 0 a(12) 0 0 0 0 0 0 b(2) b(20) 0 b(17) 0 0 0 0 0 0 0 0 0 0 0 0 0 b(19) 0 0 0 0 b(8) 0 0 0 b(12) 0 0 0 c(12) 0 0 0 0 0 0 0 0 0 c(19) 0 0 c(8) 0 0 0 0 0 0 0 c(20) 0 0 0 0 0 0 c(2) 0 c(17) 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(7) 0 0 0 0 0 a(6) 0 0 a(18) 0 0 0 a(2) 0 0 0 0 0 0 0 a(17) 0 0 0 0 0 b(7) 0 0 0 0 0 0 0 0 b(20) b(2) 0 b(19) 0 b(18) b(17) 0 0 0 b(6) 0 0 0 0 0 0 0 0 0 0 0 0 0 c(19) 0 c(17) 0 c(18) 0 c(2) c(6) 0 0 0 0 0 c(20) 0 0 0 0 0 0 0 0 0 0 c(7) 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(12) 0 0 0 0 0 0 0 0 0 0 0 0 a(16) 0 a(8) 0 0 0 0 0 0 0 0 0 0 0 0 b(12) b(8) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(16) 0 0 0 0 0 c(19) 0 0 0 c(2) c(8) 0 c(17) 0 0 0 0 c(12) 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(2) 0 0 0 a(3) a(2) 0 0 a(5) a(17) 0 0 0 0 a(14) a(8) 0 a(4) a(20) a(13) 0 0 0 0 a(19) 0 0 a(12) 0 0 0 0 b(3) b(13) b(2) b(5) 0 b(19) b(17) b(20) 0 0 0 0 0 0 0 0 0 b(4) 0 0 0 0 b(14) b(8) 0 b(12) 0 0 0 0 0 c(12) 0 0 0 c(19) 0 c(8) 0 0 c(4) 0 0 c(14) 0 0 0 0 0 0 0 c(13) 0 0 0 0 c(20) 0 c(3) c(2) c(5) c(17) 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(4) 0 0 0 0 0 0 0 0 0 0 a(6) a(16) 0 0 0 0 0 0 0 0 0 0 0 a(17) a(5) 0 a(14) a(8) a(4) a(12) 0 0 0 0 0 0 0 0 b(5) b(6) 0 0 b(12) 0 b(17) b(8) 0 0 b(4) 0 b(14) 0 0 0 0 0 b(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(12) 0 0 0 c(16) 0 0 c(14) 0 0 0 c(3) c(4) 0 0 0 c(5) 0 c(15) c(18) 0 c(17) c(8) 0 c(6) 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(3) 0 0 0 0 0 0 0 0 0 0 a(7) a(12) 0 0 0 0 0 0 0 0 0 0 0 a(18) a(15) 0 a(5) a(17) a(3) a(6) 0 0 0 0 0 0 0 0 b(15) b(7) 0 0 b(6) 0 b(18) b(17) 0 0 b(3) b(4) b(5) b(8) b(14) 0 0 0 b(12) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(6) c(14) 0 0 c(12) 0 0 c(5) 0 0 0 0 c(3) 0 c(4) 0 c(15) 0 0 0 c(8) c(18) c(17) 0 c(7) 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(2) 0 0 0 0 0 a(19) 0 0 a(20) 0 0 0 0 0 0 0 a(17) 0 0 0 0 0 0 0 0 0 b(2) 0 0 0 0 0 0 0 0 0 0 0 0 0 b(20) 0 0 0 0 b(19) 0 0 0 b(17) 0 0 0 c(17) 0 0 0 0 0 0 0 0 0 c(20) 0 0 c(19) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(2) 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(4) a(6) 0 0 0 0 0 0 0 0 0 0 0 0 a(12) 0 a(17) 0 a(8) 0 a(19) 0 0 0 0 0 a(16) 0 0 b(6) b(17) 0 0 0 0 0 0 0 0 0 b(19) 0 b(8) 0 0 0 b(12) b(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(16) 0 c(12) 0 c(8) 0 0 0 c(2) c(19) 0 0 0 c(17) 0 c(18) 0 0 0 0 c(6) 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(4) 0 0 0 a(5) a(17) 0 0 0 a(12) 0 0 0 0 0 a(16) 0 a(14) a(19) a(4) 0 0 0 0 a(8) 0 0 0 0 0 0 0 b(5) b(4) b(17) 0 0 b(8) b(12) b(19) 0 0 0 0 0 0 0 0 0 b(14) 0 0 0 0 0 b(16) 0 0 0 0 0 0 0 0 0 0 0 c(8) 0 c(16) 0 0 c(14) 0 0 0 0 0 c(13) 0 0 0 0 c(4) 0 c(3) c(2) 0 c(19) 0 c(5) c(17) 0 c(12) 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(6) 0 a(12) 0 a(17) 0 0 0 0 0 0 0 0 0 b(6) 0 0 0 0 0 0 0 0 0 b(17) b(8) b(12) 0 b(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(16) 0 0 0 0 0 c(12) 0 0 0 c(18) c(17) 0 c(8) 0 c(6) 0 c(7) 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(12) 0 0 0 0 0 0 0 0 0 0 0 0 a(8) a(14) 0 0 0 0 a(16) 0 0 0 0 0 0 0 0 b(14) b(12) 0 0 b(16) 0 b(8) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(16) 0 0 0 0 0 0 0 0 0 0 c(4) 0 0 0 c(3) c(14) 0 c(5) c(17) 0 c(8) 0 0 c(12) 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(7) 0 a(6) 0 a(18) 0 0 0 0 0 0 0 0 0 b(7) 0 0 0 0 0 0 0 b(8) 0 b(18) b(17) b(6) 0 b(12) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(12) 0 0 0 0 0 c(6) 0 0 0 0 c(18) c(8) c(17) 0 c(7) 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 b(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(10) 0 0 0 c(14) 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(9) a(13) a(20) 0 0 0 0 0 0 0 0 a(11) 0 a(3) 0 0 a(2) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(9) b(13) b(20) b(3) b(11) b(2) 0 0 c(11) c(3) c(2) 0 0 0 0 c(13) 0 0 0 c(20) 0 c(9) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(13) a(20) 0 0 0 0 0 0 0 0 0 a(3) 0 a(2) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(13) b(20) 0 b(2) b(3) 0 0 0 c(3) c(2) 0 0 0 0 0 c(20) 0 0 0 0 0 c(13) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(20) 0 0 0 0 0 0 0 0 0 0 a(2) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(20) 0 0 0 b(2) 0 0 0 c(2) 0 0 0 0 0 0 0 0 0 0 0 0 c(20) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 u(4) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(10) 0 0 0 0 a(14) a(8) 0 a(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(4) 0 0 b(8) 0 0 b(10) 0 b(14) 0 0 0 0 0 b(4) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(16) 0 0 0 c(16) c(4) 0 0 0 0 0 0 0 0 0 c(8) 0 0 0 0 0 0 0 0 c(1) c(9) 0 0 0 0 c(10) 0 0 c(14) 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(4) 0 0 0 0 0 0 a(10) a(4) 0 0 0 a(14) a(8) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(19) 0 a(16) 0 b(10) 0 b(4) 0 b(8) 0 b(14) 0 0 0 b(19) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(16) 0 0 0 0 c(16) 0 c(19) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(9) c(13) 0 0 0 c(10) c(4) 0 c(14) c(8) 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(4) a(19) 0 0 a(14) a(8) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(16) 0 0 0 0 b(4) 0 b(19) b(14) 0 0 b(8) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(16) 0 0 0 0 0 c(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(13) c(20) 0 0 0 c(4) c(19) c(14) c(8) 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(19) 0 0 0 a(8) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(16) 0 0 0 0 0 0 b(19) 0 0 b(8) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(16) 0 0 0 c(16) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(20) 0 0 0 0 c(19) 0 c(8) 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(20) 0 0 0 0 0 0 0 0 0 0 0 a(1) 0 a(9) 0 0 a(13) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(9) b(1) b(13) 0 b(20) c(1) c(9) c(13) c(20) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(3) a(7) 0 0 0 0 0 0 0 0 0 0 0 0 a(6) 0 a(18) 0 a(17) 0 a(2) 0 0 0 0 0 a(12) 0 0 b(7) b(18) 0 0 0 0 0 0 0 0 0 b(2) b(19) b(17) 0 b(8) 0 b(6) b(12) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(8) 0 c(12) 0 c(6) 0 c(17) 0 0 0 0 c(2) 0 c(19) 0 c(18) 0 0 0 0 0 0 c(7) 0 0 0 0 0 0; 0 0 0 0 0 0 0 u(3) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(5) 0 0 0 a(7) 0 0 0 0 0 0 0 0 a(15) 0 a(11) 0 0 0 a(6) 0 0 0 0 b(7) 0 0 0 0 0 0 0 0 b(11) 0 0 0 b(1) 0 b(10) 0 b(15) 0 0 0 b(5) b(6) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(7) c(15) 0 0 0 c(1) 0 0 c(5) 0 0 c(10) 0 c(11) 0 0 0 0 0 0 c(6); 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(2) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(20) a(13) 0 0 0 0 0 0 0 0 0 0 0 0 0 a(1) 0 0 a(9) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(1) 0 b(9) b(20) b(13) 0 c(1) c(9) c(13) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(20) 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(2) 0 0 0 0 0 0 0 0 0 0 0 a(1) a(9) a(13) a(20) 0 a(19) 0 0 0 0 0 0 0 0 0 0 0 0 0 a(10) 0 0 a(4) b(20) 0 0 0 b(1) b(13) 0 b(9) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(10) 0 b(4) 0 b(19) 0 c(10) c(4) c(19) 0 0 0 0 0 0 0 0 0 0 c(20) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(1) c(9) c(13) 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(9) a(13) a(20) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(10) 0 a(4) 0 0 a(19) 0 0 0 0 b(9) b(20) 0 b(13) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(4) b(10) b(19) 0 0 c(10) c(4) c(19) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(9) c(13) c(20) 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(13) a(20) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(4) 0 a(19) 0 0 0 0 0 0 0 b(13) 0 0 b(20) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(19) b(4) 0 0 0 c(4) c(19) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(13) c(20) 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(20) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(19) 0 0 0 0 0 0 0 0 0 b(20) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(19) 0 0 0 c(19) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(20) 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 u(2) 0 0 0 0 a(13) a(20) 0 0 0 0 0 0 a(11) a(3) a(2) 0 0 0 0 0 0 0 0 0 0 a(15) 0 a(9) a(18) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(9) b(13) b(20) 0 b(11) b(3) b(2) b(15) 0 b(18) 0 0 0 c(15) c(18) 0 0 0 0 c(3) c(9) c(13) 0 c(2) 0 c(11) 0 0 0 0 0 0 0 0 c(20) 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(20) 0 0 0 0 0 0 0 a(3) a(2) 0 0 0 0 0 0 0 0 0 a(15) 0 a(18) 0 a(13) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(13) b(20) 0 0 b(3) b(2) 0 b(18) b(15) 0 0 0 c(15) c(18) 0 0 0 0 0 c(2) c(13) c(20) 0 0 0 c(3) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(13) a(20) 0 0 a(4) a(19) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(14) 0 a(8) 0 0 0 0 b(13) 0 b(20) b(4) 0 0 b(19) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(8) b(14) 0 0 0 c(14) c(8) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(13) c(20) c(4) c(19) 0 0 0; 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(20) 0 0 0 a(19) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a(8) 0 0 0 0 0 0 b(20) 0 0 b(19) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b(8) 0 0 0 c(8) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c(20) 0 c(19) 0 0 0 0]';
end
function [H] = Hessian(f1coeff, f2coeff, f3coeff, s)
% the vector of monomials is
% m = [ const ; s1^2 * s2 ; s1 * s2 ; s1 * s3 ; s2 * s3 ; s2^2 * s3 ; s2^3 ; ...
% s1 * s3^2 ; s1 ; s3 ; s2 ; s2 * s3^2 ; s1^2 ; s3^2 ; s2^2 ; s3^3 ; ...
% s1 * s2 * s3 ; s1 * s2^2 ; s1^2 * s3 ; s1^3]
%
% deriv of m w.r.t. s1
Hs1 = [0 ; 2 * s(1) * s(2) ; s(2) ; s(3) ; 0 ; 0 ; 0 ; ...
s(3)^2 ; 1 ; 0 ; 0 ; 0 ; 2 * s(1) ; 0 ; 0 ; 0 ; ...
s(2) * s(3) ; s(2)^2 ; 2*s(1)*s(3); 3 * s(1)^2];
% deriv of m w.r.t. s2
Hs2 = [0 ; s(1)^2 ; s(1) ; 0 ; s(3) ; 2 * s(2) * s(3) ; 3 * s(2)^2 ; ...
0 ; 0 ; 0 ; 1 ; s(3)^2 ; 0 ; 0 ; 2 * s(2) ; 0 ; ...
s(1) * s(3) ; s(1) * 2 * s(2) ; 0 ; 0];
% deriv of m w.r.t. s3
Hs3 = [0 ; 0 ; 0 ; s(1) ; s(2) ; s(2)^2 ; 0 ; ...
s(1) * 2 * s(3) ; 0 ; 1 ; 0 ; s(2) * 2 * s(3) ; 0 ; 2 * s(3) ; 0 ; 3 * s(3)^2 ; ...
s(1) * s(2) ; 0 ; s(1)^2 ; 0];
H = [ f1coeff' * Hs1 , f1coeff' * Hs2 , f1coeff' * Hs3;
f2coeff' * Hs1 , f2coeff' * Hs2 , f2coeff' * Hs3;
f3coeff' * Hs1 , f3coeff' * Hs2 , f3coeff' * Hs3];
end
%function [C] = cayley2rot(s)
% C = ( (1-s'*s) * eye(3) + 2 * skewsymm(s) + 2 * s * s')' / ( 1 + s' * s);
%end
|
github
|
rising-turtle/slam_matlab-master
|
compute_error.m
|
.m
|
slam_matlab-master/libs_dir/plane_fitting_code/all/dls_pnp_matlab/compute_error.m
| 1,320 |
utf_8
|
9661477a57e8c1cd36f99c33566fb9ba
|
function [da, dt] = compute_error(C, t, Cm, tm)
% compute the error quaternion btw. the true and the estimated solutions
% (using JPL definition of quaternions)
q_del = rot2quat(C' * Cm);
% compute the tilt angle error
da = norm(q_del(1:3) * 2);
% compute the position error
dt = norm(t - tm);
end
function q = rot2quat(R)
% converts a rotational matrix to a unit quaternion, according to JPL
% procedure (Breckenridge Memo)
T = trace(R);
[dummy maxpivot] = max([R(1,1) R(2,2) R(3,3) T]); %#ok<ASGLU>
switch maxpivot
case 1
q(1) = sqrt((1+2*R(1,1)-T)/4);
q(2:4) = 1/(4*q(1)) * [R(1,2)+R(2,1);
R(1,3)+R(3,1);
R(2,3)-R(3,2) ];
case 2
q(2) = sqrt((1+2*R(2,2)-T)/4);
q([1 3 4]) = 1/(4*q(2)) * [R(1,2)+R(2,1);
R(2,3)+R(3,2);
R(3,1)-R(1,3) ];
case 3
q(3) = sqrt((1+2*R(3,3)-T)/4);
q([1 2 4]) = 1/(4*q(3)) * [R(1,3)+R(3,1);
R(2,3)+R(3,2);
R(1,2)-R(2,1) ];
case 4
q(4) = sqrt((1+T)/4);
q(1:3) = 1/(4*q(4)) * [R(2,3)-R(3,2);
R(3,1)-R(1,3);
R(1,2)-R(2,1) ];
end % switch
% make column vector
q = q(:);
% 4th element is always positive
if q(4)<0
q = -q;
end
% quaternion normalization
q = q/sqrt(q'*q);
end
|
github
|
rising-turtle/slam_matlab-master
|
gtsamExamples.m
|
.m
|
slam_matlab-master/libs_dir/gtsam-toolbox-2.3.0-win64/toolbox/gtsam_examples/gtsamExamples.m
| 5,664 |
utf_8
|
f2621b78fabdb370c4f63d5e0309b7e9
|
function varargout = gtsamExamples(varargin)
% GTSAMEXAMPLES MATLAB code for gtsamExamples.fig
% GTSAMEXAMPLES, by itself, creates a new GTSAMEXAMPLES or raises the existing
% singleton*.
%
% H = GTSAMEXAMPLES returns the handle to a new GTSAMEXAMPLES or the handle to
% the existing singleton*.
%
% GTSAMEXAMPLES('CALLBACK',hObject,eventData,handles,...) calls the local
% function named CALLBACK in GTSAMEXAMPLES.M with the given input arguments.
%
% GTSAMEXAMPLES('Property','Value',...) creates a new GTSAMEXAMPLES or raises the
% existing singleton*. Starting from the left, property value pairs are
% applied to the GUI before gtsamExamples_OpeningFcn gets called. An
% unrecognized property name or invalid value makes property application
% stop. All inputs are passed to gtsamExamples_OpeningFcn via varargin.
%
% *See GUI Options on GUIDE's Tools menu. Choose "GUI allows only one
% instance to run (singleton)".
%
% See also: GUIDE, GUIDATA, GUIHANDLES
% Edit the above text to modify the response to help gtsamExamples
% Last Modified by GUIDE v2.5 03-Sep-2012 13:34:13
% Begin initialization code - DO NOT EDIT
gui_Singleton = 1;
gui_State = struct('gui_Name', mfilename, ...
'gui_Singleton', gui_Singleton, ...
'gui_OpeningFcn', @gtsamExamples_OpeningFcn, ...
'gui_OutputFcn', @gtsamExamples_OutputFcn, ...
'gui_LayoutFcn', [] , ...
'gui_Callback', []);
if nargin && ischar(varargin{1})
gui_State.gui_Callback = str2func(varargin{1});
end
if nargout
[varargout{1:nargout}] = gui_mainfcn(gui_State, varargin{:});
else
gui_mainfcn(gui_State, varargin{:});
end
% End initialization code - DO NOT EDIT
% --- Executes just before gtsamExamples is made visible.
function gtsamExamples_OpeningFcn(hObject, eventdata, handles, varargin)
% This function has no output args, see OutputFcn.
% hObject handle to figure
% eventdata reserved - to be defined in a future version of MATLAB
% handles structure with handles and user data (see GUIDATA)
% varargin command line arguments to gtsamExamples (see VARARGIN)
% Choose default command line output for gtsamExamples
handles.output = hObject;
% Update handles structure
guidata(hObject, handles);
OdometryExample;
% --- Outputs from this function are returned to the command line.
function varargout = gtsamExamples_OutputFcn(hObject, eventdata, handles)
% varargout cell array for returning output args (see VARARGOUT);
% hObject handle to figure
% eventdata reserved - to be defined in a future version of MATLAB
% handles structure with handles and user data (see GUIDATA)
% Get default command line output from handles structure
varargout{1} = handles.output;
% --------------------------------------------------------------------
function CloseMenuItem_Callback(hObject, eventdata, handles)
% hObject handle to CloseMenuItem (see GCBO)
% eventdata reserved - to be defined in a future version of MATLAB
% handles structure with handles and user data (see GUIDATA)
selection = questdlg(['Close ' get(handles.figure1,'Name') '?'],...
['Close ' get(handles.figure1,'Name') '...'],...
'Yes','No','Yes');
if strcmp(selection,'No')
return;
end
delete(handles.figure1)
% --- Executes on button press in Odometry.
function Odometry_Callback(hObject, eventdata, handles)
axes(handles.axes3);
echo on
OdometryExample;
echo off
% --- Executes on button press in Localization.
function Localization_Callback(hObject, eventdata, handles)
axes(handles.axes3);
echo on
LocalizationExample;
echo off
% --- Executes on button press in Pose2SLAM.
function Pose2SLAM_Callback(hObject, eventdata, handles)
axes(handles.axes3);
echo on
Pose2SLAMExample
echo off
% --- Executes on button press in Pose2SLAMCircle.
function Pose2SLAMCircle_Callback(hObject, eventdata, handles)
axes(handles.axes3);
echo on
Pose2SLAMExample_circle
echo off
% --- Executes on button press in Pose2SLAMManhattan.
function Pose2SLAMManhattan_Callback(hObject, eventdata, handles)
axes(handles.axes3);
Pose2SLAMExample_graph
% --- Executes on button press in Pose3SLAM.
function Pose3SLAM_Callback(hObject, eventdata, handles)
axes(handles.axes3);
echo on
Pose3SLAMExample
echo off
% --- Executes on button press in Pose3SLAMSphere.
function Pose3SLAMSphere_Callback(hObject, eventdata, handles)
axes(handles.axes3);
echo on
Pose3SLAMExample_graph
echo off
% --- Executes on button press in PlanarSLAM.
function PlanarSLAM_Callback(hObject, eventdata, handles)
axes(handles.axes3);
echo on
PlanarSLAMExample
echo off
% --- Executes on button press in PlanarSLAMSampling.
function PlanarSLAMSampling_Callback(hObject, eventdata, handles)
axes(handles.axes3);
PlanarSLAMExample_sampling
% --- Executes on button press in PlanarSLAMGraph.
function PlanarSLAMGraph_Callback(hObject, eventdata, handles)
axes(handles.axes3);
echo on
PlanarSLAMExample_graph
echo off
% --- Executes on button press in SFM.
function SFM_Callback(hObject, eventdata, handles)
axes(handles.axes3);
echo on
SFMExample
echo off
% --- Executes on button press in VisualISAM.
function VisualISAM_Callback(hObject, eventdata, handles)
axes(handles.axes3);
echo on
VisualISAMExample
echo off
% --- Executes on button press in StereoVO.
function StereoVO_Callback(hObject, eventdata, handles)
axes(handles.axes3);
echo on
StereoVOExample
echo off
% --- Executes on button press in StereoVOLarge.
function StereoVOLarge_Callback(hObject, eventdata, handles)
axes(handles.axes3);
StereoVOExample_large
|
github
|
rising-turtle/slam_matlab-master
|
VisualISAM_gui.m
|
.m
|
slam_matlab-master/libs_dir/gtsam-toolbox-2.3.0-win64/toolbox/gtsam_examples/VisualISAM_gui.m
| 10,009 |
utf_8
|
ed501f5a7d855d179385d3bb29e65500
|
function varargout = VisualISAM_gui(varargin)
% VisualISAM_gui: runs VisualSLAM iSAM demo in GUI
% Interface is defined by VisualISAM_gui.fig
% You can run this file directly, but won't have access to globals
% By running ViusalISAMDemo, you see all variables in command prompt
% Authors: Duy Nguyen Ta
% Last Modified by GUIDE v2.5 13-Jun-2012 23:15:43
% Begin initialization code - DO NOT EDIT
gui_Singleton = 1;
gui_State = struct('gui_Name', mfilename, ...
'gui_Singleton', gui_Singleton, ...
'gui_OpeningFcn', @VisualISAM_gui_OpeningFcn, ...
'gui_OutputFcn', @VisualISAM_gui_OutputFcn, ...
'gui_LayoutFcn', [] , ...
'gui_Callback', []);
if nargin && ischar(varargin{1})
gui_State.gui_Callback = str2func(varargin{1});
end
if nargout
[varargout{1:nargout}] = gui_mainfcn(gui_State, varargin{:});
else
gui_mainfcn(gui_State, varargin{:});
end
% End initialization code - DO NOT EDIT
% --- Executes just before VisualISAM_gui is made visible.
function VisualISAM_gui_OpeningFcn(hObject, ~, handles, varargin)
% This function has no output args, see OutputFcn.
% varargin command line arguments to VisualISAM_gui (see VARARGIN)
% Choose default command line output for VisualISAM_gui
handles.output = hObject;
% Update handles structure
guidata(hObject, handles);
% --- Outputs from this function are returned to the command line.
function varargout = VisualISAM_gui_OutputFcn(hObject, ~, handles)
% varargout cell array for returning output args (see VARARGOUT);
% Get default command line output from handles structure
varargout{1} = handles.output;
%----------------------------------------------------------
% Convenient functions
%----------------------------------------------------------
function showFramei(hObject, handles)
global frame_i
set(handles.frameStatus, 'String', sprintf('Frame: %d',frame_i));
drawnow
guidata(hObject, handles);
function showWaiting(handles, status)
set(handles.waitingStatus,'String', status);
drawnow
guidata(handles.waitingStatus, handles);
function triangle = chooseDataset(handles)
str = cellstr(get(handles.dataset,'String'));
sel = get(handles.dataset,'Value');
switch str{sel}
case 'triangle'
triangle = true;
case 'cube'
triangle = false;
end
function initOptions(handles)
global options
% Data options
options.triangle = chooseDataset(handles);
options.nrCameras = str2num(get(handles.numCamEdit,'String'));
options.showImages = get(handles.showImagesCB,'Value');
% iSAM Options
options.hardConstraint = get(handles.hardConstraintCB,'Value');
options.pointPriors = get(handles.pointPriorsCB,'Value');
options.batchInitialization = get(handles.batchInitCB,'Value');
%options.reorderInterval = str2num(get(handles.reorderIntervalEdit,'String'));
options.alwaysRelinearize = get(handles.alwaysRelinearizeCB,'Value');
% Display Options
options.saveDotFile = get(handles.saveGraphCB,'Value');
options.printStats = get(handles.printStatsCB,'Value');
options.drawInterval = str2num(get(handles.drawInterval,'String'));
options.cameraInterval = str2num(get(handles.cameraIntervalEdit,'String'));
options.drawTruePoses = get(handles.drawTruePosesCB,'Value');
options.saveFigures = get(handles.saveFiguresCB,'Value');
options.saveDotFiles = get(handles.saveGraphsCB,'Value');
%----------------------------------------------------------
% Callback functions for GUI elements
%----------------------------------------------------------
% --- Executes during object creation, after setting all properties.
function dataset_CreateFcn(hObject, ~, handles)
% handles empty - handles not created until after all CreateFcns called
% Hint: popupmenu controls usually have a white background on Windows.
% See ISPC and COMPUTER.
if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor'))
set(hObject,'BackgroundColor','white');
end
% --- Executes on selection change in dataset.
function dataset_Callback(hObject, ~, handles)
% Hints: contents = cellstr(get(hObject,'String')) returns dataset contents as cell array
% contents{get(hObject,'Value')} returns selected item from dataset
% --- Executes during object creation, after setting all properties.
function numCamEdit_CreateFcn(hObject, ~, handles)
% Hint: edit controls usually have a white background on Windows.
if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor'))
set(hObject,'BackgroundColor','white');
end
function numCamEdit_Callback(hObject, ~, handles)
% Hints: get(hObject,'String') returns contents of numCamEdit as text
% str2double(get(hObject,'String')) returns contents of numCamEdit as a double
% --- Executes on button press in showImagesCB.
function showImagesCB_Callback(hObject, ~, handles)
% Hint: get(hObject,'Value') returns toggle state of showImagesCB
% --- Executes on button press in hardConstraintCB.
function hardConstraintCB_Callback(hObject, ~, handles)
% Hint: get(hObject,'Value') returns toggle state of hardConstraintCB
% --- Executes on button press in pointPriorsCB.
function pointPriorsCB_Callback(hObject, ~, handles)
% Hint: get(hObject,'Value') returns toggle state of pointPriorsCB
% --- Executes during object creation, after setting all properties.
function batchInitCB_CreateFcn(hObject, eventdata, handles)
set(hObject,'Value',1);
% --- Executes on button press in batchInitCB.
function batchInitCB_Callback(hObject, ~, handles)
% Hint: get(hObject,'Value') returns toggle state of batchInitCB
% --- Executes on button press in alwaysRelinearizeCB.
function alwaysRelinearizeCB_Callback(hObject, ~, handles)
% Hint: get(hObject,'Value') returns toggle state of alwaysRelinearizeCB
% --- Executes during object creation, after setting all properties.
function reorderIntervalText_CreateFcn(hObject, ~, handles)
% Hint: edit controls usually have a white background on Windows.
if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor'))
set(hObject,'BackgroundColor','white');
end
% --- Executes during object creation, after setting all properties.
function reorderIntervalEdit_CreateFcn(hObject, ~, handles)
% Hint: edit controls usually have a white background on Windows.
if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor'))
set(hObject,'BackgroundColor','white');
end
% --- Executes during object creation, after setting all properties.
function drawInterval_CreateFcn(hObject, ~, handles)
% Hint: edit controls usually have a white background on Windows.
if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor'))
set(hObject,'BackgroundColor','white');
end
function drawInterval_Callback(hObject, ~, handles)
% Hints: get(hObject,'String') returns contents of drawInterval as text
% str2double(get(hObject,'String')) returns contents of drawInterval as a double
function cameraIntervalEdit_Callback(hObject, ~, handles)
% Hints: get(hObject,'String') returns contents of cameraIntervalEdit as text
% str2double(get(hObject,'String')) returns contents of cameraIntervalEdit as a double
% --- Executes during object creation, after setting all properties.
function cameraIntervalEdit_CreateFcn(hObject, ~, handles)
% Hint: edit controls usually have a white background on Windows.
% See ISPC and COMPUTER.
if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor'))
set(hObject,'BackgroundColor','white');
end
% --- Executes on button press in saveGraphCB.
function saveGraphCB_Callback(hObject, ~, handles)
% Hint: get(hObject,'Value') returns toggle state of saveGraphCB
% --- Executes on button press in printStatsCB.
function printStatsCB_Callback(hObject, ~, handles)
% Hint: get(hObject,'Value') returns toggle state of printStatsCB
% --- Executes on button press in drawTruePosesCB.
function drawTruePosesCB_Callback(hObject, ~, handles)
% Hint: get(hObject,'Value') returns toggle state of drawTruePosesCB
% --- Executes on button press in saveFiguresCB.
function saveFiguresCB_Callback(hObject, ~, handles)
% Hint: get(hObject,'Value') returns toggle state of saveFiguresCB
% --- Executes on button press in saveGraphsCB.
function saveGraphsCB_Callback(hObject, ~, handles)
% Hint: get(hObject,'Value') returns toggle state of saveGraphsCB
% --- Executes on button press in intializeButton.
function intializeButton_Callback(hObject, ~, handles)
global frame_i truth data noiseModels isam result nextPoseIndex options
% initialize global options
initOptions(handles)
% Generate Data
[data,truth] = gtsam.VisualISAMGenerateData(options);
% Initialize and plot
[noiseModels,isam,result,nextPoseIndex] = gtsam.VisualISAMInitialize(data,truth,options);
cla
gtsam.VisualISAMPlot(truth, data, isam, result, options)
frame_i = 2;
showFramei(hObject, handles)
% --- Executes on button press in runButton.
function runButton_Callback(hObject, ~, handles)
global frame_i truth data noiseModels isam result nextPoseIndex options
while (frame_i<size(truth.cameras,2))
frame_i = frame_i+1;
showFramei(hObject, handles)
[isam,result,nextPoseIndex] = gtsam.VisualISAMStep(data,noiseModels,isam,result,truth,nextPoseIndex);
if mod(frame_i,options.drawInterval)==0
showWaiting(handles, 'Computing marginals...');
gtsam.VisualISAMPlot(truth, data, isam, result, options)
showWaiting(handles, '');
end
end
% --- Executes on button press in stepButton.
function stepButton_Callback(hObject, ~, handles)
global frame_i truth data noiseModels isam result nextPoseIndex options
if (frame_i<size(truth.cameras,2))
frame_i = frame_i+1;
showFramei(hObject, handles)
[isam,result,nextPoseIndex] = gtsam.VisualISAMStep(data,noiseModels,isam,result,truth,nextPoseIndex);
showWaiting(handles, 'Computing marginals...');
gtsam.VisualISAMPlot(truth, data, isam, result, options)
showWaiting(handles, '');
end
|
github
|
rising-turtle/slam_matlab-master
|
find_pair_tp.m
|
.m
|
slam_matlab-master/ground_truth_zh/find_pair_tp.m
| 4,069 |
utf_8
|
48cda0aca2320418530334c250c0cc24
|
% Find pairs between two 3D point sets by using T pattern
%
% Author : Soonhac Hong ([email protected])
% Date : 2/18/14
%
% Input : gt_total : [time_stamp [x y z]*5]
function [ gt_total_pair ] = find_pair_tp( gt_total )
%UNTITLED Summary of this function goes here
% Detailed explanation goes here
%addpath('..\icp');
%addpath('D:\Soonhac\SW\icp'); % original code
%addpath('C:\Yiming\ground_truth_soonhac\icp'); % added by Yimin Zhao on @06/12/2015$
% figure;
plot_colors={'b.','r.','g.','m.','c.'};
%gt_total_pair=gt_total(1,:); % copy the first data
for i=1:size(gt_total,1)
%new_gt = [];
gt_total_pair(i,1) = gt_total(i,1);
min_idx_prev=[];
prev=[];
cur=[];
distance=[];
for j=1:5
%prev = [prev; gt_total_pair(i-1,2+(j-1)*3:4+(j-1)*3)];
cur = [cur; gt_total(i,2+(j-1)*3:4+(j-1)*3)];
end
% Find T pattern
cur_idx =[];
cur_mean = mean(cur);
for k=1:5
distance(k) = norm(cur(k,:) - cur_mean);
end
[~, sort_idx] = sort(distance);
cur_idx(1:2,1) = [sort_idx(1);sort_idx(2)];
temp_cur_1 = cur(sort_idx(1),:);
temp_cur_2 = cur(sort_idx(2),:);
v1 = temp_cur_2-temp_cur_1;
for k=3:5
%three_markers=[temp_cur_1;temp_cur_2;cur(sort_idx(k),:)];
v=cur(sort_idx(k),:)- temp_cur_1;
theta(k-2) = asin(norm(cross(v,v1))/(norm(v)*norm(v1)));
%d(k-2)=det(cross(v1,v));
%[plane error]= fit([three_markers(:,1),three_markers(:,2)],three_markers(:,3),'poly11');
%fit_error(k-2) = error.sse;
end
[~, min_idx] = min(theta);
temp_cur_3 = cur(sort_idx(min_idx+2),:);
cur_idx(3,1) = sort_idx(min_idx+2);
if norm(diff([temp_cur_1;temp_cur_3])) > norm(diff([temp_cur_2;temp_cur_3]))
cur_idx(1:2,1) = sort_idx(1:2);
else
cur_idx(1,1) = sort_idx(2);
cur_idx(2,1) = sort_idx(1);
end
find_idx=4;
for n=1:5
t=find(cur_idx==n);
if isempty(t)
cur_idx(find_idx,1) = n;
find_idx=find_idx+1;
end
end
% % find temp_cur_4 and temp_cur_5
% cur45=cur;
% cur45(cur_idx,:) = [];
% v3=cur(cur_idx(1,1),:) - cur(cur_idx(3,1),:);
% v4=cur45(1,:)-cur(cur_idx(3,1),:);
% v5=cur45(2,:)-cur(cur_idx(3,1),:);
% %theta4 = asin(norm(cross(v3,v4))/(norm(v4)*norm(v3)));
% %theta5 = asin(norm(cross(v3,v5))/(norm(v5)*norm(v3)));
% if v4(1)*v4(3) < 0 && v5(1)*v5(1) > 0
% find_idx=4;
% for n=1:5
% t=find(cur_idx==n);
% if isempty(t)
% cur_idx(find_idx,1) = n;
% find_idx=find_idx+1;
% end
% end
% else
% find_idx=5;
% for n=1:5
% t=find(cur_idx==n);
% if isempty(t)
% cur_idx(find_idx,1) = n;
% find_idx=find_idx-1;
% end
% end
% end
% determin temp_cur1 and temp_cur2
% Run ICP
%[match, TR, TT, ER, t] = icp_match(prev', cur', 20);
%[match, TR, TT, ER, t] = icp_match(prev', cur', 20,'Minimize','lmaPoint'); % {point} | plane | lmaPoint
% diff_dist=[];
% for k=1:5
% pre = gt_total(i-1,2+(k-1)*3:4+(k-1)*3);
% diff_dist = [diff_dist; norm(pre - cur)];
% end
% diff_dist(min_idx_prev)=intmax;
% [~, min_idx] = min(diff_dist);
% min_idx_prev = [min_idx_prev; min_idx];
%gt_total_pair(i,2:end)=new_cur;
for k=1:5
gt_total_pair(i,2+(k-1)*3:4+(k-1)*3)=cur(cur_idx(k),:);
%degbug by showing plot
% plot3(cur(cur_idx(k),1),cur(cur_idx(k),2),cur(cur_idx(k),3),plot_colors{k});
% hold on;
end
%gt_total_pair = [gt_total_pair; new_gt];
end
end
|
github
|
rising-turtle/slam_matlab-master
|
generate_gt_wpattern_syn_zh.m
|
.m
|
slam_matlab-master/ground_truth_zh/generate_gt_wpattern_syn_zh.m
| 8,468 |
utf_8
|
341b6b99aacac024b3929ca4d0226dcf
|
% Generate ground truth from motion capture data of MOTIVE, using Time to
% synchorize not the movement, also synchorize the camera record time
% Assumption : Motion caputre data of MOTIVE(file format : *.csv) has at least 5 markers on SR4K
%
% Author : David Zhang ([email protected])
% Date : 10/23/15
function generate_gt_wpattern_syn_zh()
clear all
clc
clf
close all
%% add 3rd libraries
add_path_zh;
%% Load motion capture data
[data_file_name, camera_record_name] = get_file_name();
%% open and load the data
fid = fopen(data_file_name);
if fid < 0
error(['Cannot open file ' data_file_name]);
end
% [line_data, gt, gt_total] = scan_data (fid); % retrive data from file
[gt, gt_total] = synchorize_gt_record(data_file_name, camera_record_name);
%% T pattern, find matched points
% ---|--- 4/5 1 5/4
% | 2
% | 3
% find the T pattern in each frame, and find 5 matched points
[ gt_total_pair ] = find_pair_tp( gt_total );
[ gt_total_pair(:,11:16)] = find_pair_nn( gt_total_pair(:,11:16));
%% added by Yimin Zhao on @13/07/2015
marker=[];
for m=1:5
marker=[marker; [gt_total_pair(1,(m-1)*3+2),gt_total_pair(1,(m-1)*3+3),gt_total_pair(1,(m-1)*3+4)]];
end
%% transform all the points from the motion capture coordinate frame into local coordinate frame
% set the origin of motion capture system to the center of five LEDs in initial position
% origin = mean(marker,1);
% gt_total_pair(:,2:end) = gt_total_pair(:,2:end) - repmat(origin, size(gt_total_pair,1), 5); % some concern
% f_w = f_l*R_l2w, p^l = R_l2w * p^w
[R_l2w, t_l2w] = compute_initial_T(marker);
[ gt_total_pair(:,2:16)] = transform_pc( gt_total_pair(:,2:16), R_l2w, t_l2w);
[ gt(:, 2:4)] = transform_pc(gt(:, 2:4), R_l2w, t_l2w);
%% original of Soonhac
%gt_total_pair(:,2:end) = gt_total_pair(:,2:end) - repmat(gt_total_pair(1,2:4), size(gt_total_pair,1), 5);
%% generate pose
[gt_pose, gt_pose_euler, distance_total] = compute_transformation(gt_total_pair);
%% only convert with dataset_3
% gt_pose = [gt_pose(:,1), gt_pose(:,2), gt_pose(:,3), gt_pose(:,4:end)];
%% plot result
% plot_distance(distance_total);
plot_gt_pose(gt_pose_euler);
% plot_Rxyz(gt_pose_euler);
% plot_ground_truth1(gt);
plot_gt_pairs(gt_total_pair);
% plot_ground_truth2(gt_total);
% plot_TPattern(gt_total);
% plot_displacement(gt_total);
%% Save result
out_file_name=strrep(data_file_name, 'csv','dat_wp');
total_out_file_name=strrep(data_file_name, 'csv','dat_total_wp');
gt_pose_out_file_name=strrep(data_file_name, 'csv','dat_pose_wp');
dlmwrite(out_file_name,gt,' '); % [time_stamp x y z]
dlmwrite(total_out_file_name,gt_total_pair,' '); % [time_stamp [x y z]*5]
dlmwrite(gt_pose_out_file_name,gt_pose,' '); % [time_stamp [x y z q1 q2 q3 q4]
end
function [gt_pose, gt_pose_euler, distance_total] = add_new_trans(rot, trans, timestamp, ...
op_pset1, op_pset2, gt_pose,gt_pose_euler, distance_total)
q = R2q(rot);
gt_pose=[gt_pose; timestamp, trans' q'];
%
e = R2e(rot);
gt_pose_euler=[gt_pose_euler; timestamp, trans' e'];
% check relative distance b/w markers for rigid body
if i==1
for k=2:5
distance(k-1)=norm(op_pset1(k,:)-op_pset1(1,:));
end
distance_total=[distance_total; distance];
end
for k=2:5
distance(k-1)=norm(op_pset2(k,:)-op_pset2(1,:));
end
distance_total=[distance_total; distance];
end
function [gt_pose, gt_pose_euler, distance_total] = compute_transformation(gt_total_pair)
gt_pose=[gt_total_pair(1,1), 0,0,0,1,0,0,0];
gt_pose_euler=[gt_total_pair(1,1), 0,0,0,0,0,0];
distance_total=[];
for i=1:size(gt_total_pair,1)-1
op_pset1 = [];
op_pset2 = [];
for k=1:5
op_pset1 = [op_pset1; gt_total_pair(1,2+(k-1)*3:4+(k-1)*3)];
op_pset2 = [op_pset2; gt_total_pair(i+1,2+(k-1)*3:4+(k-1)*3)];
end
[rot, trans, sta] = find_transform_matrix(op_pset2', op_pset1');
% [rot, trans, sta] = find_transform_matrix(op_pset1', op_pset2');
if sta > 0
[gt_pose, gt_pose_euler, distance_total] = add_new_trans(rot, trans, gt_total_pair(i+1,1),...
op_pset1, op_pset2, gt_pose, gt_pose_euler, distance_total);
else
% sta;
% [rot, trans, valid] = computeT_with_previous(op_pset1, op_pset2, gt_pose_euler);
% if valid % find valid transformation
% [gt_pose, gt_pose_euler, distance_total] = add_new_trans(rot, trans, gt_total_pair(i+1,1), ...
% op_pset1, op_pset2, gt_pose, gt_pose_euler, distance_total);
% end
%% use this more robust function to compute [R, t]
[rot, trans] = eq_point(op_pset2', op_pset1');
% [rot, trans] = eq_point(op_pset1', op_pset2');
[gt_pose, gt_pose_euler, distance_total] = add_new_trans(rot, trans, gt_total_pair(i+1,1), ...
op_pset1, op_pset2, gt_pose, gt_pose_euler, distance_total);
end
end
end
function [rot, trans, valid] = computeT_with_previous(op_pset1, op_pset2, gt_pose_euler)
valid = 0;
j = size(gt_pose_euler,1);
for k=j-1:-1:1
p = gt_pose_euler(k, 2:end);
%% [R t]-1 = [R' -R't]
R = euler_to_rot(p(4), p(5), p(6));
t = p(1:3);
R_INV = R';
t_INV = -R'*t';
op_pset_prev = R_INV*op_pset2' + repmat(t_INV, 1, 5);
[rot, trans, sta] = find_transform_matrix(op_pset_prev, op_pset1');
if sta > 0 %%
rot = R * rot;
trans = R * trans + t;
valid = 1;
return ;
end
end
end
function plot_distance(distance_total)
%% show relative distance between marker for checking rigid body
figure;
plot_colors={'b.','r.','g.','m.','c.'};
for k=1:4
plot(distance_total(:,k),plot_colors{k});
hold on;
end
xlabel('Frame');
ylabel('Relative Distance');
grid;
legend('v^1_2','v^1_3','v^1_4','v^1_5');
hold off;
end
function plot_gt_pose(gt_pose_euler)
%% show gt_pose
figure;
plot3(gt_pose_euler(:,2),gt_pose_euler(:,3),gt_pose_euler(:,4),'.-');
hold on;
plot3(gt_pose_euler(1,2),gt_pose_euler(1,3),gt_pose_euler(1,4),'g*', 'MarkerSize', 10);
hold off;
axis equal;
grid;
xlabel('X');ylabel('Y');zlabel('Z');
title('Translaton');
end
function plot_Rxyz(gt_pose_euler)
plot_colors={'b.','r.','g.','m.','c.'};
figure;
title_list ={'Rx','Ry','Rz'};
for i=1:3
%plot(gt_pose(:,i+4),plot_colors{i});
%subplot(3,1,i);plot(gt_pose_euler(:,i),plot_colors{i});
subplot(3,1,i);plot(gt_pose_euler(:,i+4)*180/pi(),plot_colors{i});
title(title_list{i});grid;
%hold on;
end
xlabel('frame');
%ylabel('Orientation [quaternion]');
ylabel('Orientation [degree]');
%legend('Rx','Ry','Rz');
end
function plot_ground_truth1(gt)
%% show ground truth
figure;
plot3(gt(:,2),gt(:,3),gt(:,4),'.-');
hold on;
plot3(gt(1,2),gt(1,3),gt(1,4),'g*', 'MarkerSize', 10);
hold off;
axis equal;
grid;
xlabel('X');ylabel('Y');zlabel('Z');
end
function plot_gt_pairs(gt_total_pair)
%% show ground truth by pairs
plot_colors={'b.','r.','g.','m.','c.'};
figure;
%plot_colors={'b.-','r.-','g.-','m.-','c.-'};
for i=1:5
plot3(gt_total_pair(:,2+3*(i-1)),gt_total_pair(:,3+3*(i-1)),gt_total_pair(:,4+3*(i-1)),plot_colors{i});
hold on;
end
plot3(gt_total_pair(1,2),gt_total_pair(1,3),gt_total_pair(1,4),'g*', 'MarkerSize', 10);
hold off;
axis equal;
grid;
xlabel('X');ylabel('Y');zlabel('Z');
title('GT pairs');
end
function plot_TPattern(gt_total)
plot_colors={'b.','r.','g.','m.','c.'};
figure;
px = zeros(5, 1); py = zeros(5,1); pz = zeros(5,1);
for i=1:5
px(i) = gt_total(1,2+3*(i-1));
py(i) = gt_total(1,3+3*(i-1));
pz(i) = gt_total(1,4+3*(i-1));
hold on;
plot3(px(i), py(i), pz(i), plot_colors{i});
end
grid;
axis equal;
hold off;
end
function plot_ground_truth2(gt_total)
%% show ground truth
plot_colors={'b.','r.','g.','m.','c.'};
figure;
%plot_colors={'b.-','r.-','g.-','m.-','c.-'};
for i=1:5
plot3(gt_total(:,2+3*(i-1)),gt_total(:,3+3*(i-1)),gt_total(:,4+3*(i-1)),plot_colors{i});
hold on;
end
plot3(gt_total(1,2),gt_total(1,3),gt_total(1,4),'g*', 'MarkerSize', 10);
hold off;
axis equal;
grid;
xlabel('X');ylabel('Y');zlabel('Z');
end
function plot_displacement(gt)
%% show displacement
gt_diff = diff(gt(:,2:4),5,1);
%[~,gt_diff] = gradient(gt(:,2:4));
for i=1:size(gt_diff,1)
displacement(i,1) = norm(gt_diff(i,:));
end
figure;
plot(displacement);
xlabel('Frame');
ylabel('displacement [m]');
end
|
github
|
rising-turtle/slam_matlab-master
|
rmat2quat.m
|
.m
|
slam_matlab-master/ground_truth_zh/rmat2quat.m
| 795 |
utf_8
|
238c076762b7266d10f6726f613f660c
|
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Converts (orthogonal) rotation matrices R to (unit) quaternion
% representations
%
% Input: A 3x3xn matrix of rotation matrices
% Output: A 4xn matrix of n corresponding quaternions
%
% http://en.wikipedia.org/wiki/Rotation_matrix#Quaternion
function quaternion = rmat2quat(R)
Qxx = R(1,1,:);
Qxy = R(1,2,:);
Qxz = R(1,3,:);
Qyx = R(2,1,:);
Qyy = R(2,2,:);
Qyz = R(2,3,:);
Qzx = R(3,1,:);
Qzy = R(3,2,:);
Qzz = R(3,3,:);
w = 0.5 * sqrt(1+Qxx+Qyy+Qzz);
x = 0.5 * sign(Qzy-Qyz) .* sqrt(1+Qxx-Qyy-Qzz);
y = 0.5 * sign(Qxz-Qzx) .* sqrt(1-Qxx+Qyy-Qzz);
z = 0.5 * sign(Qyx-Qxy) .* sqrt(1-Qxx-Qyy+Qzz);
quaternion = reshape([w;x;y;z],4,[]);
end
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
github
|
rising-turtle/slam_matlab-master
|
quat2rmat.m
|
.m
|
slam_matlab-master/ground_truth_zh/quat2rmat.m
| 798 |
utf_8
|
fb153dbeaab428656e51bea7b133b87c
|
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Converts (unit) quaternion representations to (orthogonal) rotation matrices R
%
% Input: A 4xn matrix of n quaternions
% Output: A 3x3xn matrix of corresponding rotation matrices
%
% http://en.wikipedia.org/wiki/Quaternions_and_spatial_rotation#From_a_quaternion_to_an_orthogonal_matrix
function R = quat2rmat(quaternion)
q0(1,1,:) = quaternion(1,:);
qx(1,1,:) = quaternion(2,:);
qy(1,1,:) = quaternion(3,:);
qz(1,1,:) = quaternion(4,:);
R = [q0.^2+qx.^2-qy.^2-qz.^2 2*qx.*qy-2*q0.*qz 2*qx.*qz+2*q0.*qy;
2*qx.*qy+2*q0.*qz q0.^2-qx.^2+qy.^2-qz.^2 2*qy.*qz-2*q0.*qx;
2*qx.*qz-2*q0.*qy 2*qy.*qz+2*q0.*qx q0.^2-qx.^2-qy.^2+qz.^2];
end
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
github
|
rising-turtle/slam_matlab-master
|
compute_initial_Tnew.m
|
.m
|
slam_matlab-master/ground_truth_zh/compute_initial_Tnew.m
| 3,696 |
utf_8
|
8d867c8f8fc844654fc6bb156c2b5ec4
|
%
% new T-bar on the robocane for data collection
% find the initial transformation from world coordinate to camera
% coordinate, z
% the result shows that /
% T pattern, find matched points /
% x ---|--- 5 1 4 /-----> x
% | 2 |
% z| 3 | y
% p1 (0, 0, 0) |
% p2 (0, 0, -0.0838)
% p3 (0, 0, -0.1681)
% p4 (0.127, 0, 0)
% p5 (-0.134, 0, 0)
% Author : David Zhang ([email protected])
% Date : Jan. 30 2018
% from led to the front plate of sr4k, consider Tplate_cam later
% z axis offset 22.5 mm
% y axis offset 10.5 mm + (height=65/2) = 10.5 + 32.5 = 43 mm
function [rot, trans] = compute_initial_Tnew(p_world)
if nargin == 0
p_world = [1.6982 0.6840 2.4577;
1.6432 0.7339 2.4971;
1.5857 0.7820 2.5376;
1.7844 0.7082 2.5511;
1.6088 0.6648 2.3616];
end
z_shift = 0;
y_shift = 0;
%% for different cases, have to pay attention which case fit, by checking that whether the result of
% generate_gt_wpattern_syn_zh consistent in plot_gt_and_estimate
%% case 1
p_local = [0 0 0; 0 0 -0.0838; 0 0 -0.1681; 0.127, 0, 0; -0.134, 0, 0];
%% case 2
% p_local = [0 0 0; 0 0 -0.0838; 0 0 -0.1681; 0.1302 0 0; -0.1317 0 0 ];
p_local = p_local + repmat([0 y_shift z_shift], 5, 1);
%% compute the transfrom, pl = Tlc * pc, Tlc = find_transform_matrix(pl, pc);
[rot, trans, sta] = find_transform_matrix(p_local', p_world');
% [rot, trans] = eq_point(p_local', p_world');
tmp_p_l = rot * p_world' + repmat(trans, 1, 5);
end
function compute_initial_five_pts()
path_dir = '.';
addpath(strcat(path_dir, '\Localization'));
addpath(strcat(path_dir, '\slamtoolbox\slamToolbox_11_09_08\FrameTransforms\Rotations'));
%% Load motion capture data
data_file_name = '.\motion_capture_data\test_10_16_2015\Take 2015-10-16 04.49.46 PM.csv';
%% open and load the data
fid = fopen(data_file_name);
if fid < 0
error(['Cannot open file ' data_file_name]);
end
[line_data, gt, gt_total] = scan_data (fid); % retrive data from file
%% T pattern, find matched points
% ---|--- 4/5 1 5/4
% | 2
% | 3
% find the T pattern in each frame, and find 5 matched points
[ gt_total_pair ] = find_pair_tp( gt_total );
[ gt_total_pair(:,11:16)] = find_pair_nn( gt_total_pair(:,11:16));
rel_dis = compute_relative_dis(gt_total_pair(:, 2:16));
plot_gt_pairs(gt_total_pair);
end
function [rel_dis] = compute_relative_dis(pts)
pt_1 = pts(:, 1:3); pt_2 = pts(:, 4:6); pt_3 = pts(:, 7:9);
pt_4 = pts(:, 10:12); pt_5 = pts(:, 13:15);
pt_0 = pt_1;
pt_1 = pt_1 - pt_0; pt_2 = pt_2 - pt_0; pt_3 = pt_3 - pt_0;
pt_4 = pt_4 - pt_0; pt_5 = pt_5 - pt_0;
d1 = sqrt(diag(pt_1*pt_1'));
d2 = sqrt(diag(pt_2*pt_2'));
d3 = sqrt(diag(pt_3*pt_3'));
d4 = sqrt(diag(pt_4*pt_4'));
d5 = sqrt(diag(pt_5*pt_5'));
rel_dis = [d1 d2 d3 d4 d5];
end
function plot_gt_pairs(gt_total_pair)
%% show ground truth by pairs
plot_colors={'b.','r.','g.','y.','c.'};
figure;
%plot_colors={'b.-','r.-','g.-','m.-','c.-'};
for i=1:5
plot3(gt_total_pair(:,2+3*(i-1)),gt_total_pair(:,3+3*(i-1)),gt_total_pair(:,4+3*(i-1)),plot_colors{i});
hold on;
end
plot3(gt_total_pair(1,2),gt_total_pair(1,3),gt_total_pair(1,4),'g*', 'MarkerSize', 10);
hold off;
axis equal;
grid;
xlabel('X');ylabel('Y');zlabel('Z');
title('GT pairs');
end
|
github
|
rising-turtle/slam_matlab-master
|
extractTPattern.m
|
.m
|
slam_matlab-master/ground_truth_zh/extractTPattern.m
| 4,990 |
utf_8
|
d95b3060ae44175ad6d00c6f87ecdeca
|
%%
% Jan. 30 2018, He Zhang, [email protected]
% extract T Pattern from motion capture data, and store it into TPattern.log
% the T Pattern is described in compute_initial_Tnew.m
function extractTPattern(fname, ouf)
clear all
clc
clf
close all
global g_TN;
g_TN = 5;
global g_6;
g_6 = 0;
if nargin == 0
fname = './motion_capture_data/Dense_Slow_640x480_30_b.csv';
ouf = './motion_capture_data/Dense_Slow_640x480_30_b_TPattern.log';
end
%% read data from .csv file
pts = extract_TPoints(fname);
fprintf('pattern extracted given 6 points %d\n', g_6);
%% compute sum distance
sum_dis = compute_distance(pts(:, 2:end));
mu = mean(sum_dis);
sigma = std(sum_dis);
%% save it
dlmwrite(ouf,pts,' '); % [time_stamp pt1 pt2 pt3 pt4 pt5]
end
function pts = extract_TPoints(fname)
%% scan all lines into a cell array
global g_TN;
fid = fopen(fname);
columns=textscan(fid,'%s','delimiter','\n');
lines=columns{1};
N=size(lines,1);
pts=[];
cnt_b5 = 0;
for i=1:N
line_i=lines{i};
% line_data = textscan(line_i,'%s %d %f %f %f %f %f %f %d %s %f %f %f %d %s %f %f %f %d %s %f %f %f %d %s %f %f %f %d %s','delimiter',',');
l = textscan(line_i, '%s %d %f %f %f\n', 'delimiter', ',');
%% valid data
if strcmp(l{1}, 'frame')
if l{5} >= 5
%% read data
line_data = [];
if l{5} == 5
line_data = textscan(line_i,'%s %d %f %f %f %f %f %f %d %s %f %f %f %d %s %f %f %f %d %s %f %f %f %d %s %f %f %f %d %s\n','delimiter',',');
elseif l{5} == 6
line_data = textscan(line_i,'%s %d %f %f %f %f %f %f %d %s %f %f %f %d %s %f %f %f %d %s %f %f %f %d %s %f %f %f %d %s %f %f %f %d %s\n','delimiter',',');
else
fprintf('extractTPattern: not handle %d points\n', l{5});
continue;
end
time_stamp = line_data{3};
%% extract marker point
marker=[];
for m=1:line_data{5}
marker=[marker; [line_data{(m-1)*5+6},line_data{(m-1)*5+7},line_data{(m-1)*5+8}]];
end
%% extract TPattern
TP = get_T_pattern(marker);
if size(TP, 1) == g_TN
cnt_b5 = cnt_b5 + 1;
pts =[pts; time_stamp reshape(TP', 1, 15)];
end
end
end
end
fclose(fid);
fprintf('scan_data.m: cnt 5-points %d, \n', cnt_b5);
end
function TP = get_T_pattern(markers)
global g_TN
global g_6
TP = [];
if size(markers, 1) == g_TN
TP = TPattern(markers);
elseif size(markers, 1) == g_TN + 1
TP = TPatternPlus(markers);
if size(TP, 1) == g_TN
g_6 = g_6 + 1;
end
end
end
%% only work when size(markers,1) = g_TN + 1
function TP = TPatternPlus(markers)
global g_TN
TP = [];
for i=1:size(markers, 1)
pts = [];
for j = 1:size(markers,1)
if j == i
continue; % exclude point i
end
pts = [pts; markers(j,:)];
end
TP = TPattern(pts);
if size(TP, 1) == g_TN
break;
end
end
end
function TP = TPattern(markers)
global g_TN;
%% T pattern, details in compute_initial_Tnew.m
T_dis = [0.51, 0.48, 0.68, 0.75, 0.76];
pdis = compute_distance_pattern(markers);
TP = zeros(size(markers));
% first find point 1, 2, 3, then distinguish point 4, 5
seq = [0 0 0 0 0];
flag = [0 0 0 0 0];
thre = 0.01; % 1cm
cnt_45 = 0;
pre_j = 0;
pre_dj = 0;
sorted_dis = sort(pdis);
for i =1:g_TN
fprintf('%f ', sorted_dis(i));
end
fprintf('\n');
for i=1:g_TN
d = pdis(i);
for j = 1:g_TN
if abs(d - T_dis(j)) < thre
seq(i) = j;
flag(j) = 1;
end
end
%% distinguish 4 and 5. the TPattern design is so bad
if d > 0.75 && d < 0.79
if cnt_45 == 0
pre_j = i;
pre_dj = d;
else
if d < pre_dj
seq(i) = 4;
seq(pre_j) = 5;
else
seq(i) = 5;
seq(pre_j) = 4;
end
flag(4) = 1;
flag(5) = 1;
end
cnt_45 = cnt_45 + 1;
end
end
if sum(flag) ~= g_TN
TP = [];
else
for i=1:g_TN
if seq(i) == 0
TP = [];
break;
end
TP(seq(i),:) = markers(i, :);
end
end
end
function [sum_dis] = compute_distance_pattern(pts)
global g_TN;
sum_dis = zeros(g_TN, 1);
for i=1:g_TN
pti = pts(i, :);
dis_m = 0; % sum of point i to all other points
for m=1:g_TN
ptj = pts(m, :);
d_pt = ptj - pti;
dis_m = dis_m + sqrt(sum(d_pt.*d_pt));
end
sum_dis(i) = dis_m;
end
end
|
github
|
rising-turtle/slam_matlab-master
|
transform_pc.m
|
.m
|
slam_matlab-master/ground_truth_zh/transform_pc.m
| 333 |
utf_8
|
61bdbcf94164db5c26d374657b5b494e
|
%
% transform point cloud given [R t]
% Author : David Zhang ([email protected])
% Date : 10/23/15
function [pc] = transform_pc(pc, R, t)
[m, n] = size(pc);
loop_n = n/3;
translation = repmat(t, 1, m);
for i=1:loop_n
pc_T = pc(:,(i-1)*3+1:i*3)';
pc_T = R*pc_T + translation;
pc(:, (i-1)*3+1:i*3) = pc_T';
end
end
|
github
|
rising-turtle/slam_matlab-master
|
compare_error_lsd_rgbd_vo.m
|
.m
|
slam_matlab-master/ground_truth_zh/compare_error_lsd_rgbd_vo.m
| 2,666 |
utf_8
|
476dd1331f449491f422d3bfbb435ea0
|
function compare_error_lsd_rgbd_vo()
%%
% Author : David Zhang ([email protected])
% Date : 08/08/16
% compute the RMSE of the relative displacement error given the result
% from VO of LSDSLAM and of RGBDSLAM
%% comparison between dense-track and sparse-track, use dataset_3
gt_f = '.\motion_capture_data\test_10_16_2015\Take 2015-10-16 04.49.46 PM.dat_pose_wp';
es_f = '.\motion_capture_data\test_10_16_2015\dataset_3\compare_lsd_vo\key_frame_trajectory.log'; % lsd-vo
es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_3\compare_lsd_vo\trajectory_estimate.txt'; % rgbd-vo
es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_3\data3_vro_estimate.txt';
es_f = '.\motion_capture_data\test_10_16_2015\dataset_3\data3_plane_em_vro_4_estimate.txt';
add_path_zh;
%% compute RTE for rgbd's vo and lsd's vo respectively
E_lsd = computeRTE(gt_f, es_f);
E_rgbd = computeRTE(gt_f, es_f2);
%% draw them
st = 2;
et = min(size(E_rgbd,1), size(E_lsd,1));
e_lsd = E_lsd(st:et,1);
plot(e_lsd, 'b-*');
xlabel('Number of Keyframes #', 'FontSize', 20);
ylabel('Relative Translational Error [m]', 'FontSize', 20);
hold on;
e_rgbd = E_rgbd(st:et,1);
plot(e_rgbd, 'r-*');
set(gca, 'fontsize', 17);
h = legend('VO LSD-SLAM', 'VO RGBD-SLAM');
set(h, 'FontSize', 24);
grid on;
end
function [E_ab, E_sq] = computeRTE(gt_f, es_f)
gt = load(gt_f);
es = load(es_f);
% es2 = load(es_f2);
%% synchroization with time
[es_syn, gt_syn] = syn_time_with_gt(es, gt);
E_sq = zeros(size(es_syn,1), 2);
E_ab = zeros(size(es_syn,1), 2);
Te_1 = eye(4);
Tg_1 = eye(4);
j = 1;
for i=2:1:size(es_syn,1)
pe_2 = es_syn(i, 2:end);
pe_2_seq = pe_2(4:7); pe_2_seq(2:4) = pe_2(4:6); pe_2_seq(1) = pe_2(7);
pg_2 = gt_syn(i, 2:end);
Re_2 = quat2rmat(pe_2_seq');
te_2 = pe_2(1:3)';
Rg_2 = quat2rmat(pg_2(4:7)');
tg_2 = pg_2(1:3)';
Te_2 = combine(Re_2, te_2);
Tg_2 = combine(Rg_2, tg_2);
[t_sq, r_sq] = compute_squared_error(Te_1, Te_2, Tg_1, Tg_2);
E_sq(j,1) = t_sq; E_sq(j, 2) = r_sq;
E_ab(j,1) = sqrt(t_sq); E_ab(j,2) = sqrt(r_sq);
j = j+1;
Te_1 = Te_2;
Tg_1 = Tg_2;
end
E_sq(j:end,:) = [];
E_ab(j:end,:) = [];
end
function [T] = combine(R, t)
T = [R t; 0 0 0 1];
end
function [R, t] = decompose(T)
R = T(1:3, 1:3);
t = T(1:3, 4);
end
function [t_sq, r_sq] = compute_squared_error(Te_1, Te_2, Tg_1, Tg_2)
dTe12 = Te_1\Te_2; % inv(Te_1)*Te_2;
dTg12 = Tg_1\Tg_2; %inv(Tg_1)*Tg_2;
deltaT = dTe12\dTg12; % inv(dTe12)*dTg12;
[R, t] = decompose(deltaT);
e = R2e(R);
e = e.*180./pi;
t_sq = t'*t;
r_sq = e'*e;
end
|
github
|
rising-turtle/slam_matlab-master
|
transform_TR.m
|
.m
|
slam_matlab-master/ground_truth_zh/transform_TR.m
| 443 |
utf_8
|
a666dc39ac19c4ad53318391dd685460
|
%%
% transform coodinate system
% pay attention to quaternion sequence
function [pose] = transform_TR(pose, R, t)
T_w2l = combine(R, t);
for i = 1:size(pose, 1)
t_l2i = pose(i, 1:3);
q = pose(i, 4:7);
R_l2i = quat2rmat(q');
T_l2i = combine(R_l2i, t_l2i');
T_w2i = T_w2l * T_l2i;
[R_w2i, t_w2i] = decompose(T_w2i);
pose(i, 1:3) = t_w2i';
q = rmat2quat(R_w2i);
pose(i, 4:7) = [q(2:4); q(1)]';
end
end
|
github
|
rising-turtle/slam_matlab-master
|
eq_point.m
|
.m
|
slam_matlab-master/ground_truth_zh/eq_point.m
| 834 |
utf_8
|
e792beddfd466db4f709817ca9559e8e
|
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% this function is from icp function
function [R,T] = eq_point(q,p,weights)
if nargin <= 2
weights = ones(1,size(q,2));
end
m = size(p,2);
n = size(q,2);
% normalize weights
weights = weights ./ sum(weights);
% find data centroid and deviations from centroid
q_bar = q * transpose(weights);
q_mark = q - repmat(q_bar, 1, n);
% Apply weights
q_mark = q_mark .* repmat(weights, 3, 1);
% find data centroid and deviations from centroid
p_bar = p * transpose(weights);
p_mark = p - repmat(p_bar, 1, m);
% Apply weights
%p_mark = p_mark .* repmat(weights, 3, 1);
N = p_mark*transpose(q_mark); % taking points of q in matched order
[U,~,V] = svd(N); % singular value decomposition
R = V*diag([1 1 det(U*V')])*transpose(U);
T = q_bar - R*p_bar;
end
|
github
|
rising-turtle/slam_matlab-master
|
get_ground_truth_pose.m
|
.m
|
slam_matlab-master/ground_truth_zh/get_ground_truth_pose.m
| 1,938 |
utf_8
|
96628d57e5d78e6f7f222759ad420b01
|
function get_ground_truth_pose()
%
% Author : David Zhang ([email protected])
% Date : 08/08/16
% get the ground truth pose of the keyframes
%
gt_f = '.\motion_capture_data\test_10_16_2015\Take 2015-10-16 04.49.46 PM.dat_pose_wp';
es_f = '.\motion_capture_data\test_10_16_2015\dataset_3\compare_lsd_vo\key_frame_trajectory.log'; % lsd-vo
% es_f = '.\motion_capture_data\test_10_16_2015\dataset_3\compare_lsd_vo\trajectory_estimate.txt'; % rgbd-vo
add_path_zh;
gt = load(gt_f);
es = load(es_f);
%% synchroization with time
[es_syn, gt_syn] = syn_time_with_gt(es, gt);
% gt sequence: timestamp x y z qw qx qy qz => timestamp x y z qx qy qz qw
gt = [es_syn(:,1) gt_syn(:,2:4) gt_syn(:, 6:8) gt_syn(:, 5)];
% dlmwrite('.\tmp\data_3_rgbd_gt_pose.log', gt, 'delimiter', ' ');
dlmwrite('.\tmp\data_3_lsd_gt_pose.log', gt, 'delimiter', ' ');
sid = 1;
eid = 100; %size(gt,1);
plot3(gt(sid:eid,2), gt(sid:eid,3), gt(sid:eid,4), 'k-*');
hold on;
plot3(es_syn(sid:eid,2), es_syn(sid:eid,3), es_syn(sid:eid,4), 'b*-');
% test_gt(gt_syn(1:end,2:8));
end
%% test whether the gt's transformation works correctly
function test_gt(gt)
tmp_v = [0 0 0 1 0 0 0];
pre_gt = gt(1,:);
T_g1 = toTrans(pre_gt);
T_n1 = toTrans(tmp_v);
nv = tmp_v;
for i=2:size(gt,1)
cur_gt = gt(i,:);
T_g2 = toTrans(cur_gt);
% gt incremental transformation
inc_trans = T_g1\T_g2; % inv(Tg1)*Tg2
T_n2 = T_n1*inc_trans;
cur_v = toPose(T_n2);
nv = [nv; cur_v];
% next iteration
T_n1 = T_n2;
T_g1 = T_g2;
end
hold on;
plot3(nv(:,1), nv(:,2), nv(:,3), 'b*-');
end
function T = toTrans(p)
% R = quat2rmat(p(4), p(5), p(6), p(7));
q = p(4:7);
t = p(1:3);
R = quat2rmat(q');
T = [R t'; 0 0 0 1];
end
function p = toPose(T)
R = T(1:3,1:3);
q = rmat2quat(R);
t = T(1:3,4);
p(1:3) = t';
p(4:7) = q';
end
|
github
|
rising-turtle/slam_matlab-master
|
generate_gt_trajectory.m
|
.m
|
slam_matlab-master/ground_truth_zh/generate_gt_trajectory.m
| 8,560 |
utf_8
|
83821a42059115510139114d096fa98f
|
% input: inf: *.csv file
% Assumption : Motion caputre data of MOTIVE(file format : *.csv) has at
% least 5 markers on camera
%
% input: inf .log file include TPattern points extracted by extractTPattern.m
% output: ouf: .log file [timestamp x, y, z, qx, qy, qz, qw]
% Author : David Zhang ([email protected])
% Date : Jan. 28 2018
function generate_gt_trajectory(inf, ouf)
clear all
clc
clf
close all
if nargin == 0
% inf = './motion_capture_data/Dense_Slow_640x480_30_b.csv';
inf = './motion_capture_data/Dense_Slow_640x480_30_b_TPattern.log';
ouf = './motion_capture_data/Dense_Slow_640x480_30_b_trajectory.log';
ouf2 = './motion_capture_data/Dense_Slow_640x480_30_b_trajectory_g.log';
end
%% add 3rd libraries
add_path_zh;
%% read data
gt_total_pair = load(inf);
timestamp = gt_total_pair(:,1);
pts = gt_total_pair(:,2:end);
%% T pattern, find matched points
% ---|--- 4/5 1 5/4
% | 2
% | 3
% find the T pattern in each frame, and find 5 matched points
% [ gt_total_pair ] = find_pair_tp( gt_total );
% [ gt_total_pair(:,11:16)] = find_pair_nn( gt_total_pair(:,11:16));
%% added by Yimin Zhao on @13/07/2015
marker=[];
for m=1:5
marker=[marker; [gt_total_pair(1,(m-1)*3+2),gt_total_pair(1,(m-1)*3+3),gt_total_pair(1,(m-1)*3+4)]];
end
%% transform all the points from the motion capture coordinate frame into local coordinate frame
% set the origin of motion capture system to the center of five LEDs in initial position
% origin = mean(marker,1);
% gt_total_pair(:,2:end) = gt_total_pair(:,2:end) - repmat(origin, size(gt_total_pair,1), 5); % some concern
% f_w = f_l*R_l2w, p^l = R_l2w * p^w
[R_l2w, t_l2w] = compute_initial_Tnew(marker);
[ gt_total_pair(:,2:16)] = transform_pc( gt_total_pair(:,2:16), R_l2w, t_l2w);
% [ gt(:, 2:4)] = transform_pc(gt(:, 2:4), R_l2w, t_l2w);
%% original of Soonhac
%gt_total_pair(:,2:end) = gt_total_pair(:,2:end) - repmat(gt_total_pair(1,2:4), size(gt_total_pair,1), 5);
%% generate pose
[gt_pose, gt_pose_euler, distance_total] = compute_transformation(gt_total_pair);
%% transform back into global coordinates
R_w2l = R_l2w';
t_w2l = -R_w2l*t_l2w;
% e = R2e(R_w2l);
% e(1) = 0;
% R_w2l = e2R(e);
% t_w2l = [0 0 0]'; % only use
gt_pose_back = gt_pose;
% gt_pose_back(:, 2:4) = transform_pc(gt_pose_euler(:,2:4), R_w2l, t_w2l);
gt_pose_back(:, 2:8) = transform_TR(gt_pose(:,2:8), R_w2l, t_w2l);
% e(1:2) = 0; %% get rid of yaw component
% e(3) = -e(3);
% R_w2l = e2R(e);
% gt_pose_back(:, 2:4) = transform_pc(gt_pose_back(:,2:4), R_w2l, t_w2l);
%% only convert with dataset_3
% gt_pose = [gt_pose(:,1), gt_pose(:,2), gt_pose(:,3), gt_pose(:,4:end)];
%% plot result
% plot_distance(distance_total);
plot_gt_pose(gt_pose_euler);
% plot_Rxyz(gt_pose_euler);
% plot_ground_truth1(gt);
plot_gt_pairs(gt_total_pair);
plot_gt_pose(gt_pose_back);
% plot_ground_truth2(gt_total);
% plot_TPattern(gt_total);
% plot_displacement(gt_total);
%% Save result
% out_file_name=strrep(data_file_name, 'csv','dat_wp');
% total_out_file_name=strrep(data_file_name, 'csv','dat_total_wp');
% gt_pose_out_file_name=strrep(data_file_name, 'csv','dat_pose_wp');
% dlmwrite(out_file_name,gt,' '); % [time_stamp x y z]
% dlmwrite(total_out_file_name,gt_total_pair,' '); % [time_stamp [x y z]*5]
% dlmwrite(gt_pose_out_file_name,gt_pose,' '); % [time_stamp [x y z q1 q2 q3 q4]
dlmwrite(ouf, gt_pose,' '); % [time_stamp [x y z q1 q2 q3 q4]
dlmwrite(ouf2, gt_pose_back, ' '); % [time_stamp [x y z q1 q2 q3 q4]
end
function [gt_pose, gt_pose_euler, distance_total] = add_new_trans(rot, trans, timestamp, ...
op_pset1, op_pset2, gt_pose,gt_pose_euler, distance_total)
q = R2q(rot);
gt_pose=[gt_pose; timestamp, trans' q'];
%
e = R2e(rot);
gt_pose_euler=[gt_pose_euler; timestamp, trans' e'];
% check relative distance b/w markers for rigid body
if i==1
for k=2:5
distance(k-1)=norm(op_pset1(k,:)-op_pset1(1,:));
end
distance_total=[distance_total; distance];
end
for k=2:5
distance(k-1)=norm(op_pset2(k,:)-op_pset2(1,:));
end
distance_total=[distance_total; distance];
end
function [gt_pose, gt_pose_euler, distance_total] = compute_transformation(gt_total_pair)
gt_pose=[gt_total_pair(1,1), 0,0,0,1,0,0,0];
gt_pose_euler=[gt_total_pair(1,1), 0,0,0,0,0,0];
distance_total=[];
for i=1:size(gt_total_pair,1)-1
op_pset1 = [];
op_pset2 = [];
for k=1:5
op_pset1 = [op_pset1; gt_total_pair(1,2+(k-1)*3:4+(k-1)*3)];
op_pset2 = [op_pset2; gt_total_pair(i+1,2+(k-1)*3:4+(k-1)*3)];
end
[rot, trans, sta] = find_transform_matrix(op_pset2', op_pset1');
% [rot, trans, sta] = find_transform_matrix(op_pset1', op_pset2');
if sta > 0
[gt_pose, gt_pose_euler, distance_total] = add_new_trans(rot, trans, gt_total_pair(i+1,1),...
op_pset1, op_pset2, gt_pose, gt_pose_euler, distance_total);
else
% sta;
% [rot, trans, valid] = computeT_with_previous(op_pset1, op_pset2, gt_pose_euler);
% if valid % find valid transformation
% [gt_pose, gt_pose_euler, distance_total] = add_new_trans(rot, trans, gt_total_pair(i+1,1), ...
% op_pset1, op_pset2, gt_pose, gt_pose_euler, distance_total);
% end
%% use this more robust function to compute [R, t]
[rot, trans] = eq_point(op_pset2', op_pset1');
% [rot, trans] = eq_point(op_pset1', op_pset2');
[gt_pose, gt_pose_euler, distance_total] = add_new_trans(rot, trans, gt_total_pair(i+1,1), ...
op_pset1, op_pset2, gt_pose, gt_pose_euler, distance_total);
end
end
end
function plot_distance(distance_total)
%% show relative distance between marker for checking rigid body
figure;
plot_colors={'b.','r.','g.','m.','c.'};
for k=1:4
plot(distance_total(:,k),plot_colors{k});
hold on;
end
xlabel('Frame');
ylabel('Relative Distance');
grid;
legend('v^1_2','v^1_3','v^1_4','v^1_5');
hold off;
end
function plot_gt_pose(gt_pose_euler)
%% show gt_pose
figure;
plot3(gt_pose_euler(:,2),gt_pose_euler(:,3),gt_pose_euler(:,4),'.-');
hold on;
plot3(gt_pose_euler(1,2),gt_pose_euler(1,3),gt_pose_euler(1,4),'g*', 'MarkerSize', 20);
hold off;
axis equal;
grid;
xlabel('X');ylabel('Y');zlabel('Z');
title('Translaton');
end
function plot_Rxyz(gt_pose_euler)
plot_colors={'b.','r.','g.','m.','c.'};
figure;
title_list ={'Rx','Ry','Rz'};
for i=1:3
%plot(gt_pose(:,i+4),plot_colors{i});
%subplot(3,1,i);plot(gt_pose_euler(:,i),plot_colors{i});
subplot(3,1,i);plot(gt_pose_euler(:,i+4)*180/pi(),plot_colors{i});
title(title_list{i});grid;
%hold on;
end
xlabel('frame');
%ylabel('Orientation [quaternion]');
ylabel('Orientation [degree]');
%legend('Rx','Ry','Rz');
end
function plot_ground_truth1(gt)
%% show ground truth
figure;
plot3(gt(:,2),gt(:,3),gt(:,4),'.-');
hold on;
plot3(gt(1,2),gt(1,3),gt(1,4),'g*', 'MarkerSize', 10);
hold off;
axis equal;
grid;
xlabel('X');ylabel('Y');zlabel('Z');
end
function plot_gt_pairs(gt_total_pair)
%% show ground truth by pairs
plot_colors={'b.','r.','g.','m.','c.'};
figure;
%plot_colors={'b.-','r.-','g.-','m.-','c.-'};
for i=1:5
plot3(gt_total_pair(:,2+3*(i-1)),gt_total_pair(:,3+3*(i-1)),gt_total_pair(:,4+3*(i-1)),plot_colors{i});
hold on;
end
plot3(gt_total_pair(1,2),gt_total_pair(1,3),gt_total_pair(1,4),'g*', 'MarkerSize', 20);
hold off;
axis equal;
grid;
xlabel('X');ylabel('Y');zlabel('Z');
title('GT pairs');
end
function plot_TPattern(gt_total)
plot_colors={'b.','r.','g.','m.','c.'};
figure;
px = zeros(5, 1); py = zeros(5,1); pz = zeros(5,1);
for i=1:5
px(i) = gt_total(1,2+3*(i-1));
py(i) = gt_total(1,3+3*(i-1));
pz(i) = gt_total(1,4+3*(i-1));
hold on;
plot3(px(i), py(i), pz(i), plot_colors{i});
end
grid;
axis equal;
hold off;
end
function plot_ground_truth2(gt_total)
%% show ground truth
plot_colors={'b.','r.','g.','m.','c.'};
figure;
%plot_colors={'b.-','r.-','g.-','m.-','c.-'};
for i=1:5
plot3(gt_total(:,2+3*(i-1)),gt_total(:,3+3*(i-1)),gt_total(:,4+3*(i-1)),plot_colors{i});
hold on;
end
plot3(gt_total(1,2),gt_total(1,3),gt_total(1,4),'g*', 'MarkerSize', 10);
hold off;
axis equal;
grid;
xlabel('X');ylabel('Y');zlabel('Z');
end
function plot_displacement(gt)
%% show displacement
gt_diff = diff(gt(:,2:4),5,1);
%[~,gt_diff] = gradient(gt(:,2:4));
for i=1:size(gt_diff,1)
displacement(i,1) = norm(gt_diff(i,:));
end
figure;
plot(displacement);
xlabel('Frame');
ylabel('displacement [m]');
end
|
github
|
rising-turtle/slam_matlab-master
|
find_pair_nn.m
|
.m
|
slam_matlab-master/ground_truth_zh/find_pair_nn.m
| 816 |
utf_8
|
a1524f22e650404f71c7aa8e8853bd08
|
% Find pairs between two 3D point sets by Nearest Neighbors
%
% Author : Soonhac Hong ([email protected])
% Date : 2/21/14
%
% Input : gt_total : [[x y z]*n]
function [ gt_total_pair] = find_pair_nn( gt_total)
% Find pairs by nearest neighbor
gt_total_pair=gt_total(1,:);
for i=2:size(gt_total,1)
min_idx_prev=[];
for j=1:2
cur = gt_total(i,1+(j-1)*3:3+(j-1)*3);
diff_dist=[];
for k=1:2
pre = gt_total_pair(i-1,1+(k-1)*3:3+(k-1)*3);
diff_dist = [diff_dist; norm(pre - cur)];
end
diff_dist(min_idx_prev)=intmax;
[~, min_idx] = min(diff_dist);
min_idx_prev = [min_idx_prev; min_idx];
gt_total_pair(i,1+(min_idx-1)*3:3+(min_idx-1)*3)=cur;
end
end
end
|
github
|
rising-turtle/slam_matlab-master
|
plot_ate_translation_error.m
|
.m
|
slam_matlab-master/ground_truth_zh/plot_ate_translation_error.m
| 1,920 |
utf_8
|
ad162345ee6f3a0cbc13de7dd392e0da
|
% Author : David Zhang ([email protected])
% Date : 11/09/15
% plot the the ATE displacement error at each step,
function plot_ate_translation_error()
add_path_zh;
%% dataset_1
% gt_f = '.\motion_capture_data\test_10_16_2015\Take 2015-10-16 04.25.55 PM.dat_pose_wp';
% es_f = '.\motion_capture_data\test_10_16_2015\dataset_1\trajectory_vo_estimate.txt';
% es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_1\trajectory_vo_p_estimate.txt';
% es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_1\trajectory_vo_pem_estimate.txt';
%% dataset_2
gt_f = '.\motion_capture_data\test_10_16_2015\Take 2015-10-16 04.46.44 PM.dat_pose_wp';
es_f = '.\motion_capture_data\test_10_16_2015\dataset_2\trajectory_vo_estimate.txt';
% es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_2\trajectory_vo_p_estimate.txt';
es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_2\trajectory_vo_pem_estimate.txt';
%% dataset_3
% es_f = '.\motion_capture_data\test_10_16_2015\trajectory_estimate_04.49.46.txt';
% gt_f = '.\motion_capture_data\test_10_16_2015\Take 2015-10-16 04.49.46 PM.dat_pose_wp';
% es_f = '.\motion_capture_data\test_10_16_2015\dataset_3\trajectory_vo_estimate.txt';
% es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_3\trajectory_vo_p_estimate.txt';
% es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_3\trajectory_vo_pem_estimate.txt';
mode = 'ATE';
[rmse, E_sq, E_ab] = compute_error(gt_f, es_f, mode);
E1 = E_ab;
[rmse, E_sq, E_ab] = compute_error(gt_f, es_f2, mode);
E2 = E_ab;
plot_error(E1, E2);
end
function plot_error(E1, E2)
et1 = sqrt(E1(:,1));
et2 = sqrt(E2(:,1));
plot(et1, 'r+');
hold on;
plot(et2, 'b+');
xlabel('nth camera frame #');
ylabel('translational error [m]');
legend('VO', 'VO_p');
end
|
github
|
rising-turtle/slam_matlab-master
|
compute_initial_T.m
|
.m
|
slam_matlab-master/ground_truth_zh/compute_initial_T.m
| 3,809 |
utf_8
|
fd2892b25e171900ac986ea56e8b49a7
|
%
% find the initial transformation from world coordinate to camera
% coordinate, y z
% the result shows that | /
% T pattern, find matched points | /
% ---|--- 5/4 1 4/5 x ----|/
% | 2
% | 3
% p1 (0, 0, 0)
% p2 (0, 0, -0.0838)
% p3 (0, 0, -0.1681)
% p4/5 (-0.1317, 0, 0)
% p5/4 (0.1302, 0, 0)
% Author : David Zhang ([email protected])
% Date : 10/23/15
% from led to the front plate of sr4k
% z axis offset 22.5 mm
% y axis offset 10.5 mm + (height=65/2) = 10.5 + 32.5 = 43 mm
function [rot, trans] = compute_initial_T(p_world)
if nargin == 0
p_world = [3.2867 0.8038 -1.1012;
3.3616 0.8403 -1.0947;
3.4369 0.8772 -1.0855;
3.2812 0.7928 -0.9697;
3.2941 0.8155 -1.2305];
p_world = [3.3603 0.7747 -1.0545;
3.4357 0.8076 -1.039;
3.511 0.8418 -1.0212;
3.3872 0.7772 -1.182;
3.3349 0.7741 -0.9254];
end
z_shift = -0.0225;
y_shift = +0.043;
%% for different cases, have to pay attention which case fit, by checking that whether the result of
% generate_gt_wpattern_syn_zh consistent in plot_gt_and_estimate
%% case 1
p_local = [0 0 0; 0 0 -0.0838; 0 0 -0.1681; -0.1317 0 0; 0.1302 0 0];
%% case 2
% p_local = [0 0 0; 0 0 -0.0838; 0 0 -0.1681; 0.1302 0 0; -0.1317 0 0 ];
p_local = p_local + repmat([0 y_shift z_shift], 5, 1);
%% compute the transfrom, pl = Tlc * pc, Tlc = find_transform_matrix(pl, pc);
[rot, trans, sta] = find_transform_matrix(p_local', p_world');
% [rot, trans] = eq_point(p_local', p_world');
tmp_p_l = rot * p_world' + repmat(trans, 1, 5);
end
function compute_initial_five_pts()
path_dir = '.';
addpath(strcat(path_dir, '\Localization'));
addpath(strcat(path_dir, '\slamtoolbox\slamToolbox_11_09_08\FrameTransforms\Rotations'));
%% Load motion capture data
data_file_name = '.\motion_capture_data\test_10_16_2015\Take 2015-10-16 04.49.46 PM.csv';
%% open and load the data
fid = fopen(data_file_name);
if fid < 0
error(['Cannot open file ' data_file_name]);
end
[line_data, gt, gt_total] = scan_data (fid); % retrive data from file
%% T pattern, find matched points
% ---|--- 4/5 1 5/4
% | 2
% | 3
% find the T pattern in each frame, and find 5 matched points
[ gt_total_pair ] = find_pair_tp( gt_total );
[ gt_total_pair(:,11:16)] = find_pair_nn( gt_total_pair(:,11:16));
rel_dis = compute_relative_dis(gt_total_pair(:, 2:16));
plot_gt_pairs(gt_total_pair);
end
function [rel_dis] = compute_relative_dis(pts)
pt_1 = pts(:, 1:3); pt_2 = pts(:, 4:6); pt_3 = pts(:, 7:9);
pt_4 = pts(:, 10:12); pt_5 = pts(:, 13:15);
pt_0 = pt_1;
pt_1 = pt_1 - pt_0; pt_2 = pt_2 - pt_0; pt_3 = pt_3 - pt_0;
pt_4 = pt_4 - pt_0; pt_5 = pt_5 - pt_0;
d1 = sqrt(diag(pt_1*pt_1'));
d2 = sqrt(diag(pt_2*pt_2'));
d3 = sqrt(diag(pt_3*pt_3'));
d4 = sqrt(diag(pt_4*pt_4'));
d5 = sqrt(diag(pt_5*pt_5'));
rel_dis = [d1 d2 d3 d4 d5];
end
function plot_gt_pairs(gt_total_pair)
%% show ground truth by pairs
plot_colors={'b.','r.','g.','y.','c.'};
figure;
%plot_colors={'b.-','r.-','g.-','m.-','c.-'};
for i=1:5
plot3(gt_total_pair(:,2+3*(i-1)),gt_total_pair(:,3+3*(i-1)),gt_total_pair(:,4+3*(i-1)),plot_colors{i});
hold on;
end
plot3(gt_total_pair(1,2),gt_total_pair(1,3),gt_total_pair(1,4),'g*', 'MarkerSize', 10);
hold off;
axis equal;
grid;
xlabel('X');ylabel('Y');zlabel('Z');
title('GT pairs');
end
|
github
|
rising-turtle/slam_matlab-master
|
plot_gt_and_estimate.m
|
.m
|
slam_matlab-master/ground_truth_zh/plot_gt_and_estimate.m
| 5,255 |
utf_8
|
cdc1577b62ee7dcf19136ad6ee3972bd
|
% Author : David Zhang ([email protected])
% Date : 10/23/15
% plot the estimated trajectory and the ground truth
function plot_gt_and_estimate(gt_f, es_f)
if nargin == 0
%% dataset_1
% gt_f = '.\motion_capture_data\test_10_16_2015\Take 2015-10-16 04.25.55 PM.dat_pose_wp';
% es_f = '.\motion_capture_data\test_10_16_2015\dataset_1\trajectory_vo_estimate.txt';
% es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_1\trajectory_vo_p_estimate.txt';
% es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_1\trajectory_vo_pem_estimate.txt';
%% dataset_2
% gt_f = '.\motion_capture_data\test_10_16_2015\Take 2015-10-16 04.46.44 PM.dat_pose_wp';
% es_f = '.\motion_capture_data\test_10_16_2015\dataset_2\trajectory_vo_estimate.txt';
% es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_2\trajectory_vo_p_estimate.txt';
% es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_2\trajectory_vo_pem_estimate.txt';
%% dataset_3
% es_f = '.\motion_capture_data\test_10_16_2015\trajectory_estimate_04.49.46.txt';
% gt_f = '.\motion_capture_data\test_10_16_2015\Take 2015-10-16 04.49.46 PM.dat_pose_wp';
% es_f = '.\motion_capture_data\test_10_16_2015\dataset_3\trajectory_vo_estimate.txt';
% es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_3\trajectory_vo_p_estimate.txt';
% es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_3\trajectory_vo_pem_estimate.txt';
%% dataset_4
% gt_f = '.\motion_capture_data\test_10_16_2015\Take 2015-10-16 04.51.46 PM.dat_pose_wp';
% es_f = '.\motion_capture_data\test_10_16_2015\dataset_4\trajectory_vo_estimate.txt';
% es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_4\trajectory_vo_p_estimate.txt';
%% comparison between dense-track and sparse-track, use dataset_3
gt_f = '.\motion_capture_data\test_10_16_2015\Take 2015-10-16 04.49.46 PM.dat_pose_wp'; % ground truth
es_f2 ='.\motion_capture_data\test_10_16_2015\dataset_3\compare_lsd_vo\key_frame_trajectory.log'; % lsd-slam
es_f = '.\motion_capture_data\test_10_16_2015\dataset_3\compare_lsd_vo\trajectory_estimate.txt'; % rgbd-slam
es_f = '.\motion_capture_data\test_10_16_2015\dataset_3\data3_vro_estimate.txt';
es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_3\data3_plane_em_vro_estimate.txt';
end
gt = load(gt_f);
es = load(es_f);
es_p = load(es_f2);
% traj_len = compute_trajectory_length(es_p(:,2:4));
gt = syn_last_frame(gt, es(:,1));
% p = 0.513; %1;
% st = 1;
% et = size(gt,1)*p;
% sti = int32(size(gt,1)*p) ;
% eti = int32(size(gt,1)*0.763);
% plot_xyz(gt(sti:eti,2), gt(sti:eti,3), gt(sti:eti,4), '-k');
gt(:,2:8) = transform_traj(gt(:,2:8));
es(:,2:8) = change_order(es(:,2:8));
es_p(:,2:8) = change_order(es_p(:,2:8));
es(:,2:8) = transform_traj(es(:,2:8));
es_p(:,2:8) = transform_traj(es_p(:,2:8));
% plot_xyz(gt(:,2), gt(:,3), gt(:,4), 'k');
plot_xyz(gt(:,2), gt(:,3), gt(:,4), 'y');
hold on;
st = 1;% 1200;
et = size(es,1); % 1800
plot_xyz(es(st:et,2), es(st:et,3), es(st:et,4), '-r');
hold on;
plot_xyz(es_p(st:et,2), es_p(st:et,3), es_p(st:et,4),'-g');
% legend('GT','VO','VO_P');
h = legend('GroundTruth', 'VO RGBD-SLAM', 'VO LSD-SLAM');
set(h, 'FontSize', 14);
hold on;
plot3(gt(1,2), gt(1,3), gt(1,4), 'g*', 'MarkerSize', 15, 'LineWidth',2);
% plot3(gt(sti,2), gt(sti,3), gt(sti,4), 'k*', 'MarkerSize', 15, 'LineWidth',2);
% plot3(es(st,2), es(st,3), es(st,4), 'r*', 'MarkerSize', 15, 'LineWidth',2);
% plot3(es_p(st,2), es_p(st,3), es_p(st,4), 'b*', 'MarkerSize', 15, 'LineWidth',2);
%% try save meshlab
% save_meshlab('./tmp/trajectory_mesh_2.ply',gt(:,2:4), es(:, 2:4), es_p(:,2:4));
end
function np = change_order(p)
np = p ;
np(:,5:7) = p(:,4:6);
np(:,4) = p(:,7);
end
%% save into meshlab
function save_meshlab(f, gt, es1, es2)
% pts_gt = generate_mesh_pts(gt, [0 0 0]); % black for gt
% pts_gt = generate_mesh_pts(gt, [153 0 153]); % purple
pts_gt = generate_mesh_pts(gt, [255 255 0]); % yellow
pts_es1 = generate_mesh_pts(es1, [255 0 0]); % red for es1
pts_es2 = generate_mesh_pts(es2, [0 255 0]); % blue for es2
%% write them into a ply file
pts = [pts_gt; pts_es1; pts_es2];
n = size(pts, 1);
fid = fopen(f, 'w');
fprintf(fid, 'ply\nformat ascii 1.0\nelement vertex %d\nproperty float x\nproperty float y\nproperty float z\nproperty uchar red\nproperty uchar green\nproperty uchar blue\nend_header\n', n);
fclose(fid);
dlmwrite(f,pts,'-append', 'delimiter',' ');
end
function [pts] = generate_mesh_pts(t, c)
pts = t;
pc = repmat( c, size(t,1), 1);
pts = [pts pc];
end
function l = compute_trajectory_length(p)
l = 0;
for i=2:size(p,1)
dt = p(i,:) - p(i-1,:);
l = l + sqrt(dt*dt');
end
end
function gt = syn_last_frame(gt, es_t)
time_passed = es_t(end) - es_t(1);
index = size(gt,1);
while gt(index,1) - gt(1,1) > time_passed
index = index - 1;
end
gt(index+1:size(gt,1), :) = [];
end
function plot_xyz(x, y, z, c)
plot3(x, y, z, c, 'LineWidth', 2);
hold on;
hold off;
axis equal;
grid;
xlabel('X (m)');ylabel('Y (m)');zlabel('Z (m)');
end
|
github
|
rising-turtle/slam_matlab-master
|
transform_traj.m
|
.m
|
slam_matlab-master/ground_truth_zh/transform_traj.m
| 715 |
utf_8
|
b10b621acceb5b4ffe34ca4642d61bff
|
%%
% Aug. 16, 2016, David Z
% Transform a trajectory into a new reference
function tp = transform_traj(p, iniT)
if nargin < 2
iniT = [0.2 0.03 0.98 -0.43;
-0.98 -0.08 0.2 -0.29;
0.09 -1. 0.01 0.23;
0 0 0 1];
end
np = vect_T(iniT);
tp = np;
new_T = iniT;
last_T = trans_T(p(1,:));
for i =2:size(p,1)
cur_T = trans_T(p(i,:));
del_T = last_T\cur_T;
new_T = new_T*del_T;
np = vect_T(new_T);
tp = [tp; np];
last_T = cur_T;
end
end
function T = trans_T(p)
R = quat2rmat(p(4:7)');
t = p(1:3)';
T = [R t; 0 0 0 1];
end
function p = vect_T(T)
R = T(1:3,1:3);
t = T(1:3,4);
q = rmat2quat(R);
p = [t' q'];
end
|
github
|
rising-turtle/slam_matlab-master
|
getTPattern.m
|
.m
|
slam_matlab-master/ground_truth_zh/getTPattern.m
| 1,328 |
utf_8
|
9bc54a7788714c6f84971a73e5515dcd
|
%%
% Jan. 28 2018, He Zhang, [email protected]
% explore the T pattern
%
function getTPattern(fname)
global g_TN;
g_TN = 5;
if nargin == 0
fname = './motion_capture_data/Dense_Slow_640x480_30_b.csv';
end
%% read data from .csv file
pts = get_data(fname);
%% compute sum distance
sum_dis = compute_distance(pts);
mu = mean(sum_dis);
sigma = std(sum_dis);
end
function pts = get_data(fname)
%% scan all lines into a cell array
global g_TN;
fid = fopen(fname);
columns=textscan(fid,'%s','delimiter','\n');
lines=columns{1};
N=size(lines,1);
pts=[];
cnt_b5 = 0;
for i=1:N
line_i=lines{i};
line_data = textscan(line_i,'%s %d %f %f %f %f %f %f %d %s %f %f %f %d %s %f %f %f %d %s %f %f %f %d %s %f %f %f %d %s','delimiter',',');
if strcmp(line_data{1}, 'frame')
if ~isempty(line_data{6})
% time_stamp = line_data{3};
marker=[];
if line_data{5} == 5 % original
for m=1:line_data{5}
marker=[marker; [line_data{(m-1)*5+6},line_data{(m-1)*5+7},line_data{(m-1)*5+8}]];
end
cnt_b5 = cnt_b5 + 1;
pts =[pts; reshape(marker', 1, 15)];
end
end
end
end
fclose(fid);
fprintf('scan_data.m: cnt 5-points %d, \n', cnt_b5);
end
|
github
|
rising-turtle/slam_matlab-master
|
deconstruct.m
|
.m
|
slam_matlab-master/ground_truth_zh/deconstruct.m
| 116 |
utf_8
|
41b09cb36372256be5cdd13704898870
|
function [q, t] = deconstruct(T)
[R,t] = decompose(T);
q = rmat2quat(R);
q = [q(2) q(3) q(4) q(1)];
end
|
github
|
rising-turtle/slam_matlab-master
|
plot_rte_translation_error.m
|
.m
|
slam_matlab-master/ground_truth_zh/plot_rte_translation_error.m
| 2,679 |
utf_8
|
c6a2dedbcd822057858f9662f7e72390
|
% Author : David Zhang ([email protected])
% Date : 11/13/15
% plot the accumulated RTE displacement error at each step,
function plot_rte_translation_error()
add_path_zh;
%% dataset_1
gt_f = '.\motion_capture_data\test_10_16_2015\Take 2015-10-16 04.25.55 PM.dat_pose_wp';
es_f = '.\motion_capture_data\test_10_16_2015\dataset_1\trajectory_vo_estimate.txt';
% es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_1\trajectory_vo_p_estimate.txt';
es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_1\trajectory_vo_pem_estimate.txt';
%% dataset_2
% gt_f = '.\motion_capture_data\test_10_16_2015\Take 2015-10-16 04.46.44 PM.dat_pose_wp';
% es_f = '.\motion_capture_data\test_10_16_2015\dataset_2\trajectory_vo_estimate.txt';
% es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_2\trajectory_vo_p_estimate.txt';
% es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_2\trajectory_vo_pem_estimate.txt';
%% dataset_3
gt_f = '.\motion_capture_data\test_10_16_2015\Take 2015-10-16 04.49.46 PM.dat_pose_wp';
% es_f = '.\motion_capture_data\test_10_16_2015\dataset_3\trajectory_vo_estimate.txt';
es_f = '.\motion_capture_data\test_10_16_2015\dataset_3\data3_vro_estimate.txt';
% es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_3\trajectory_vo_p_estimate.txt';
% es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_3\trajectory_vo_pem_estimate.txt';
% es_f = '.\motion_capture_data\test_10_16_2015\dataset_3\data3_plane_em_vro_estimate.txt';
es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_3\data3_plane_em_vro_4_estimate.txt';
% es_f2 = '.\motion_capture_data\test_10_16_2015\dataset_3\compare_lsd_vo\trajectory_estimate.txt';
mode = 'RTE';
[rmse, E_sq, E_ab] = compute_error(gt_f, es_f, mode);
E1 = E_ab;
[rmse, E_sq, E_ab] = compute_error(gt_f, es_f2, mode);
E2 = E_ab;
plot_error(E1, E2);
end
function E = accumulate(E)
for i=2:size(E,1)
E(i) = E(i-1) + E(i);
end
end
function plot_error(E1, E2)
et1 = sqrt(E1(:,1));
et2 = sqrt(E2(:,1));
et1 = accumulate(et1);
et2 = accumulate(et2);
plot_error_impl(et1, et2, 'translational error [m]');
figure;
et1 = sqrt(E1(:,2));
et2 = sqrt(E2(:,2));
et1 = accumulate(et1);
et2 = accumulate(et2);
plot_error_impl(et1, et2, 'rotational error [m]');
end
function plot_error_impl(et1, et2, y_label)
plot(et1, 'r+');
hold on;
plot(et2, 'b+');
xlabel('nth camera frame #');
ylabel(y_label);
legend('VO', 'VO_p');
end
|
github
|
rising-turtle/slam_matlab-master
|
compute_distance.m
|
.m
|
slam_matlab-master/ground_truth_zh/compute_distance.m
| 662 |
utf_8
|
a959fb86cb7213760e2844ca61964ba3
|
%% compute sum of mutual distance between all points
function [sum_dis] = compute_distance(pts)
[row, col] = size(pts);
global g_TN;
sum_dis = zeros(row, col/3);
for i=1:row
pt_dis = zeros(1, g_TN);
for m=1:g_TN
pt = pts(i, (m-1)*3+1:(m*3));
dis_m = 0; % sum of point i to all other points
for n =1:g_TN
pt_j = pts(i, (n-1)*3+1:(n*3));
d_pt = pt_j - pt;
dis_m = dis_m + sqrt(sum(d_pt.*d_pt));
end
pt_dis(1, m) = dis_m;
end
sorted_dis = sort(pt_dis);
sum_dis(i,:) = sorted_dis;
end
end
|
github
|
rising-turtle/slam_matlab-master
|
synchorize_gt_record.m
|
.m
|
slam_matlab-master/ground_truth_zh/synchorize_gt_record.m
| 2,150 |
utf_8
|
57bbd95d8600f2e1b5c43fe5b69a9210
|
%
% synchorize the timestamp between motion capture and camera record
%
% Author : David Zhang ([email protected])
% Date : 10/28/15
function [gt, gt_total] = synchorize_gt_record(gt_fname, record_log)
clc
if nargin == 0
%% Load motion capture data
% gt_fname = '.\motion_capture_data\test_10_16_2015\Take 2015-10-16 04.49.46 PM.csv';
% record_log = '.\motion_capture_data\test_10_16_2015\timestamp_04.49.46.log';
gt_fname = './motion_capture_data/test_10_16_2015/Take 2015-10-16 04.25.55 PM.csv';
record_log = './motion_capture_data/test_10_16_2015/timestamp_04.25.55.log';
end
%% load gt data
fid = fopen(gt_fname);
if fid < 0
error(['Cannot open file ' data_file_name]);
end
[line_data, gt, gt_total] = scan_data(fid);
gt_time_seq = gt(:,1);
% gt_time_base = repmat(gt_time_seq(1), size(gt_time_seq,1), 1);
% gt_time_shift = gt_time_seq - gt_time_base;
%% load camera record data
% fid = fopen(record_log);
camera_record_data = load(record_log);
cr_time_seq = camera_record_data(:, 2);
cr_time_base = repmat(cr_time_seq(1), size(cr_time_seq,1), 1);
cr_time_seq = cr_time_seq - cr_time_base;
%% find the synchronize start point
global g_index
g_index = 1;
potential_syn_points = find(cr_time_seq > gt_time_seq(1));
for i=1:size(potential_syn_points)
cr_index = potential_syn_points(i);
gt_index = find_syn_index(cr_time_seq(cr_index), gt_time_seq);
if gt_index ~= -1
break;
end
end
%% delete the non-synchronize part
M = size(gt,1);
gt = gt(gt_index:M, :);
gt_total = gt_total(gt_index:M, :);
fprintf('synchorize_gt_record.m: gt_index %d, cr_index %d\n', gt_index, cr_index);
end
function index = find_syn_index(cr_elapse_time, gt_time_seq)
global g_index
%% gt has 120 hz, 1/120 ~ 0.0083
gt_time_span_upper = 0.009;
while g_index < size(gt_time_seq,1)
cur_t = gt_time_seq(g_index);
if abs(cur_t - cr_elapse_time) < gt_time_span_upper
index = g_index;
break;
end
if cur_t > cr_elapse_time
index = -1;
break;
end
g_index = g_index+1;
end
end
|
github
|
rising-turtle/slam_matlab-master
|
compute_error.m
|
.m
|
slam_matlab-master/ground_truth_zh/compute_error.m
| 6,020 |
utf_8
|
5435bdb9264dc70c0879cef7e7e930e5
|
function [rmse, E_sq, E_ab] = compute_error(gt_f, es_f, mode)
%
% Author : David Zhang ([email protected])
% Date : 11/06/15
% compute the RMSE of the relative displacement error, following
% kuemmerl09auro's approach, RTE, and ATE,
%
if nargin == 0
%% dataset_1
% gt_f = '.\motion_capture_data\test_10_16_2015\Take 2015-10-16 04.25.55 PM.dat_pose_wp';
% es_f = '.\motion_capture_data\test_10_16_2015\dataset_1\trajectory_vo_estimate.txt';
% es_f = '.\motion_capture_data\test_10_16_2015\dataset_1\trajectory_vo_p_estimate.txt';
% es_f = '.\motion_capture_data\test_10_16_2015\dataset_1\trajectory_vo_pem_estimate.txt';
%% dataset_2
% gt_f = '.\motion_capture_data\test_10_16_2015\Take 2015-10-16 04.46.44 PM.dat_pose_wp';
% es_f = '.\motion_capture_data\test_10_16_2015\dataset_2\trajectory_vo_estimate.txt';
% es_f = '.\motion_capture_data\test_10_16_2015\dataset_2\trajectory_vo_pem_estimate.txt';
% es_f = '.\motion_capture_data\test_10_16_2015\dataset_2\trajectory_vo_p_estimate.txt';
%% dataset_3
gt_f = '.\motion_capture_data\test_10_16_2015\Take 2015-10-16 04.49.46 PM.dat_pose_wp';
% es_f = '.\motion_capture_data\test_10_16_2015\dataset_3\trajectory_vo_estimate.txt';
% es_f = '.\motion_capture_data\test_10_16_2015\dataset_3\trajectory_vo_p_estimate.txt';
%es_f = '.\motion_capture_data\test_10_16_2015\dataset_3\trajectory_vo_pem_estimate.txt';
es_f = '.\motion_capture_data\test_10_16_2015\dataset_3\data3_vro_estimate.txt';
es_f = '.\motion_capture_data\test_10_16_2015\dataset_3\data3_plane_em_vro_2_estimate.txt';
%% comparison between dense-track and sparse-track, use dataset_3
% gt_f = '.\motion_capture_data\test_10_16_2015\Take 2015-10-16 04.49.46 PM.dat_pose_wp';
% es_f = '.\motion_capture_data\test_10_16_2015\dataset_3\compare_lsd_vo\key_frame_trajectory.log'; % lsd-vo
% es_f = '.\motion_capture_data\test_10_16_2015\dataset_3\compare_lsd_vo\trajectory_estimate.txt'; % rgbd-vo
mode = 'RTE'; % 'RTE'
end
if nargin == 2
mode = 'ATE'; % 'RTE'
end
add_path_zh;
gt = load(gt_f);
es = load(es_f);
% es2 = load(es_f2);
%% synchroization with time
[es_syn, gt_syn] = syn_time_with_gt(es, gt);
%% statistically collect error information
% compute the relative transformation between any two VO pairs, and then
% collect translation error and rotational error
% [es_syn2, gt_syn2] = syn_time_with_gt(es2, gt);
%
st = 120;
et = st + 30;
plot_xyz(es_syn(st:et, 2), es_syn(st:et, 3), es_syn(st:et, 4), 'b*-');
hold on;
plot_xyz(gt_syn(st:et, 2), gt_syn(st:et, 3), gt_syn(st:et, 4), 'k*-');
% hold on;
% plot_xyz(es_syn2(st:et, 2), es_syn2(st:et, 3), es_syn2(st:et, 4), '-r');
% hold on;
% plot_xyz(gt_syn2(st:et, 2), gt_syn2(st:et, 3), gt_syn2(st:et, 4), '-c');
% es_syn = es_syn2(st:et, :);
% gt_syn = gt_syn(st:et, :);
if strcmp(mode,'RTE')
[E_ab, E_sq] = computeRTE(es_syn, gt_syn);
else
[E_ab, E_sq] = computeATE(es_syn, gt_syn);
end
E = E_ab;
plot_error(E);
me = mean(E);
stde = std(E);
maxe = max(E);
fprintf('\n');
fprintf('E_abs mean: %f %f, std: %f %f \n', me(1), me(2), stde(1), stde(2));
fprintf('E_abs max: %f %f \n', maxe(1), maxe(2));
E = E_sq;
N = size(E,1);
rmse = sqrt(sum(E)./N);
me = mean(E);
stde = std(E);
fprintf('E_sqr mean: %f %f, std: %f %f rmse %f %f \n', me(1), me(2), stde(1), ...
stde(2), rmse(1), rmse(2));
end
function [E_ab, E_sq] = computeATE(es_syn, gt_syn)
E_sq = zeros(size(es_syn,1), 2);
E_ab = zeros(size(es_syn,1), 2);
for i=2:size(es_syn,1)
pe_2 = es_syn(i, 2:end);
pe_2_seq = pe_2(4:7); pe_2_seq(2:4) = pe_2(4:6); pe_2_seq(1) = pe_2(7);
pg_2 = gt_syn(i, 2:end);
Re_2 = quat2rmat(pe_2_seq');
te_2 = pe_2(1:3)';
Rg_2 = quat2rmat(pg_2(4:7)');
tg_2 = pg_2(1:3)';
Te_2 = combine(Re_2, te_2);
Tg_2 = combine(Rg_2, tg_2);
deltaT = Te_2\Tg_2;
[R, t] = decompose(deltaT);
e = R2e(R);
e = e.*180./pi;
t_sq = t'*t;
r_sq = e'*e;
E_sq(i,1) = t_sq; E_sq(i, 2) = r_sq;
E_ab(i,1) = sqrt(t_sq); E_ab(i,2) = sqrt(r_sq);
end
end
function [E_ab, E_sq] = computeRTE(es_syn, gt_syn)
E_sq = zeros(size(es_syn,1), 2);
E_ab = zeros(size(es_syn,1), 2);
Te_1 = eye(4);
Tg_1 = eye(4);
j = 1;
for i=2:1:size(es_syn,1)
pe_2 = es_syn(i, 2:end);
pe_2_seq = pe_2(4:7); pe_2_seq(2:4) = pe_2(4:6); pe_2_seq(1) = pe_2(7);
pg_2 = gt_syn(i, 2:end);
Re_2 = quat2rmat(pe_2_seq');
te_2 = pe_2(1:3)';
% pg_2(5:6) = pg_2(5:6)*-1; % why gt's qx qy is has different signs with vo's qx qy
Rg_2 = quat2rmat(pg_2(4:7)');
tg_2 = pg_2(1:3)';
Te_2 = combine(Re_2, te_2);
Tg_2 = combine(Rg_2, tg_2);
[t_sq, r_sq] = compute_squared_error(Te_1, Te_2, Tg_1, Tg_2);
E_sq(j,1) = t_sq; E_sq(j, 2) = r_sq;
E_ab(j,1) = sqrt(t_sq); E_ab(j,2) = sqrt(r_sq);
j = j+1;
Te_1 = Te_2;
Tg_1 = Tg_2;
end
E_sq(j:end,:) = [];
E_ab(j:end,:) = [];
end
function plot_error(E)
et = E(:,1);
er = E(:,2);
plot(et, 'r-*');
xlabel('relation #');
ylabel('translational error [m]');
% figure;
% plot(er, 'r+');
% xlabel('relation #');
% ylabel('angular error [deg]');
end
function [t_sq, r_sq] = compute_squared_error(Te_1, Te_2, Tg_1, Tg_2)
dTe12 = Te_1\Te_2; % inv(Te_1)*Te_2;
dTg12 = Tg_1\Tg_2; %inv(Tg_1)*Tg_2;
deltaT = dTe12\dTg12; % inv(dTe12)*dTg12;
[R, t] = decompose(deltaT);
e = R2e(R);
e = e.*180./pi;
t_sq = t'*t;
r_sq = e'*e;
end
function plot_xyz(x, y, z, c)
plot3(x, y, z, c, 'LineWidth', 2);
hold on;
hold off;
axis equal;
grid;
xlabel('X (m)');ylabel('Y (m)');zlabel('Z (m)');
end
function [T] = combine(R, t)
T = [R t; 0 0 0 1];
end
function [R, t] = decompose(T)
R = T(1:3, 1:3);
t = T(1:3, 4);
end
|
github
|
rising-turtle/slam_matlab-master
|
pose_optimization_example_landmark_2d.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/pose_optimization_example_landmark_2d.m
| 6,710 |
utf_8
|
0600bb720d4a8ddc710779bc6a61242b
|
% 2D Example of Pose Graph Optimization
% Data : 4/12/12
% Author : Soonhac Hong ([email protected])
function pose_optimization_example_landmark_2d()
% 2D case
pose_data=[1 1 0 0 0; 1 2 2 0 0 ; 2 3 2 0 pi/2];
landmark_data=[3 4 2 0 ; 1 5 2 2 ; 2 5 2 0 ]; % [first pose, second pose, constraint[x,y,theta]]
xinit = [0.5; 0.0; 0.2; 2.3; 0.1; -0.2; 4.1; 0.1; pi/2; 4.0; 2.0; 2.1; 2.1]
motion_noise = [0.3;0.3;0.1];
pose_num = size(pose_data,1);
variable_size = (size(pose_data,2)-2); % size of each variable [x, y, r]
total_num = size(unique(pose_data(:,1:2)),1)*variable_size + size(unique(landmark_data(:,2)),1)*2; % first position is not optimized.
Omega = zeros(total_num,total_num);
Xi = zeros(total_num,1);
Odometry=zeros(total_num,1);
%Odometry = zeros(pose_num,1);
% Fill the elements of Omega and Xi
% Initial point
for i=1:variable_size
Omega(i,i) = 1;
Xi(i) = pose_data(1,i+2);
Odometry(i) = pose_data(1,i+2);
end
%% Pose data
for i=2:pose_num
unit_data = pose_data(i,:);
current_index = unit_data(1);
next_index = unit_data(2);
movement = unit_data(3:2+variable_size);
%previous_theta = pose_data(i-1,2+variable_size);
% Adjust index according to the size of each variable
if current_index ~= 1
current_index = (current_index - 1) * variable_size + 1;
end
if next_index ~= -1
next_index = (next_index - 1) * variable_size + 1;
end
for j=0:variable_size-1
% Fill diagonal elements of Omega
switch j
case 0
diagonal_factor = cos(movement(3));
offdiagonal_factor = 1;
case 1
diagonal_factor = cos(movement(3));
offdiagonal_factor = -1;
case 2
diagonal_factor = 1;
offdiagonal_factor = -1;
end
Omega(current_index+j,current_index+j) = Omega(current_index+j,current_index+j) + diagonal_factor*motion_noise(j+1);
Omega(next_index+j,next_index+j) = Omega(next_index+j,next_index+j) + 1/motion_noise(j+1);
% Fill Off-diagonal elements of Omega
Omega(current_index+j,next_index+j) = Omega(current_index+j,next_index+j) + (-1)/motion_noise(j+1);
Omega(next_index+j,current_index+j) = Omega(next_index+j,current_index+j) + (-1)*diagonal_factor/motion_noise(j+1);
if j <= 1
Omega(current_index+j,current_index+j+offdiagonal_factor) = Omega(current_index+j,next_index+j+offdiagonal_factor) + (-1)*offdiagonal_factor*sin(movement(3))/motion_noise(j+1);
Omega(next_index+j,current_index+j+offdiagonal_factor) = Omega(next_index+j,current_index+j+offdiagonal_factor) + offdiagonal_factor*sin(movement(3))/motion_noise(j+1);
end
% Fill Xi
Xi(current_index+j) = Xi(current_index+j) + (-1)*movement(j+1)/motion_noise(j+1);
Xi(next_index+j) = Xi(next_index+j) + movement(j+1)/motion_noise(j+1);
end
% Update Odometry
if abs(current_index - next_index) == 3
translation=[0 0 1]';
for t=i:-1:2
unit_movement=pose_data(t,3:5);
translation = [cos(unit_movement(3)) -sin(unit_movement(3)) unit_movement(1); sin(unit_movement(3)) cos(unit_movement(3)) unit_movement(2); 0 0 1]*translation;
end
%translation = Odometry(current_index:current_index+1) + movement(1:2)';
Odometry(next_index:next_index+1) = translation(1:2);
orientation = Odometry(current_index+2) + movement(3);
if orientation > pi*2
orientation = orientation - pi*2;
end
Odometry(next_index+2) = orientation;
end
end
%% Landmark data
for i=1:size(landmark_data,1);
unit_data = landmark_data(i,:);
current_index = unit_data(1);
next_index = unit_data(2) - pose_num; % should be changed to number of unique pose
movement = unit_data(3:2+variable_size-1);
%previous_theta = pose_data(i-1,2+variable_size);
% Adjust index according to the size of each variable
%if current_index ~= 1
current_index = (current_index - 1) * variable_size +1;
%end
if next_index ~= -1
next_index = (next_index - 1) * variable_size-1 + variable_size*pose_num + 1;
end
for j=0:variable_size-2
% Fill diagonal elements of Omega
Omega(current_index+j,current_index+j) = Omega(current_index+j,current_index+j) + 1*motion_noise(j+1);
Omega(next_index+j,next_index+j) = Omega(next_index+j,next_index+j) + 1/motion_noise(j+1);
% Fill Off-diagonal elements of Omega
Omega(current_index+j,next_index+j) = Omega(current_index+j,next_index+j) + (-1)/motion_noise(j+1);
Omega(next_index+j,current_index+j) = Omega(next_index+j,current_index+j) + (-1)/motion_noise(j+1);
% Fill Xi
Xi(current_index+j) = Xi(current_index+j) + (-1)*movement(j+1)/motion_noise(j+1);
Xi(next_index+j) = Xi(next_index+j) + movement(j+1)/motion_noise(j+1);
end
% Update Odometry
%if abs(current_index - next_index) == 3
translation=[landmark_data(i,3:4)'; 1];
for t=unit_data(1):-1:2
unit_movement=pose_data(t,3:5);
translation = [cos(unit_movement(3)) -sin(unit_movement(3)) unit_movement(1); sin(unit_movement(3)) cos(unit_movement(3)) unit_movement(2); 0 0 1]*translation;
end
%translation = Odometry(current_index:current_index+1) + movement(1:2)';
Odometry(next_index:next_index+1) = translation(1:2);
%end
end
%Omega
%Xi
%mu = Omega^-1 * Xi
% Using LM
xdata = Omega;
ydata = Xi;
%myfun = @(x,xdata)Rot(x(1:3))*xdata+repmat(x(4:6),1,length(xdata));
myfun = @(x,xdata)xdata*x(1:size(xdata,1));
options = optimset('Algorithm', 'levenberg-marquardt');
%x = lsqcurvefit(myfun, zeros(6,1), p, q, [], [], options);
x = lsqcurvefit(myfun, xinit, xdata, ydata, [], [], options)
x_mat = vec2mat(x(1:pose_num*variable_size),3);
Odometry_mat = vec2mat(Odometry(1:pose_num*variable_size),3);
xinit_mat = vec2mat(xinit(1:pose_num*variable_size),3);
x_lm_mat = vec2mat(x(1+pose_num*variable_size:end),2);
Odometry_lm_mat = vec2mat(Odometry(1+pose_num*variable_size:end),2);
xinit_lm_mat = vec2mat(xinit(1+pose_num*variable_size,end),2);
plot(xinit_mat(:,1), xinit_mat(:,2) ,'bo-');
hold on;
plot(Odometry_mat(:,1), Odometry_mat(:,2) ,'go-');
plot(x_mat(:,1), x_mat(:,2) ,'ro-');
plot(x_lm_mat(:,1), x_lm_mat(:,2) ,'bd');
plot(Odometry_lm_mat(:,1), Odometry_lm_mat(:,2) ,'gd');
plot(xinit_lm_mat(:,1), xinit_lm_mat(:,2) ,'rd');
hold off;
legend('Initial','Odometry','Optimized','Initial LM','Odometry LM','Optimized LM');
%legend('Initial','Optimized');
end
|
github
|
rising-turtle/slam_matlab-master
|
generate_location_info.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/generate_location_info.m
| 2,639 |
utf_8
|
6d427ec892a4afe59429489cac5899bc
|
% Generate the location infomation for Text-To-Speech and plots
%
% Author : Soonhac Hong ([email protected])
% History :
% 8/27/14 : Created
%
function [location_info, location_file_index] = generate_location_info(gtsam_pose_result, plot_xyz_result, location_file_index)
import gtsam.*
% Extract the optimized pose from gtsam data structure if the optimized pose is not available
if isempty(plot_xyz_result)
keys = KeyVector(gtsam_pose_result.keys);
% isp_fd = fopen(file_name_pose, 'w');
initial_max_index = keys.size-1;
for i=0:initial_max_index
key = keys.at(i);
x = gtsam_pose_result.at(key);
T = x.matrix();
[ry, rx, rz] = rot_to_euler(T(1:3,1:3));
plot_xyz_result(i+1,:)=[x.x x.y x.z];
end
end
%% Generate the location information
% %ex1
% location_info_list={'Room 222 is your left;',...
% 'Room 230 is your right;',...
% 'Room 245 is your left;',...
% 'Room 300 is your right;'};
% short_location_info_list={'Rm 222',...
% 'Rm 230',...
% 'Rm 245',...
% 'Rm 300'};
% location_info_xy=[0, 1; -0.1, 5; -0.5, 10; -7, 5]; % unit : [m]
% etas 523 exp2
location_info_list={'Room 523 is your right;',...
'Room 526 is your left;',...
'Room 528 is your left;',...
'Room 527 is your right;',...
'Room 529 is your right;'};
short_location_info_list={'Rm 523',...
'Rm 526',...
'Rm 528',...
'Rm 527',...
'Rm 529'};
location_info_xy=[0, 2; -0.1, 5; 0.1, 8.5; 3.5, 8; 9, 7.7]; % unit : [m]
% location_info_list={'Room 507 is your left;',...
% 'Room 505 is your right;',...
% 'Room 504 is your right;',...
% 'Room 503 is your right;',...
% 'Room 501 is your right;'};
% short_location_info_list={'Rm 507',...
% 'Rm 505',...
% 'Rm 504',...
% 'Rm 503',...
% 'Rm 501'};
% location_info_xy=[0, 11; 0, 12.5; 0, 16.5; 0, 18; -1, 19.5]; % unit : [m]
location_threshold = 0.4; % [m]
location_info_file_name = sprintf('C:\\SC-DATA-TRANSFER\\location_info_%d.txt', location_file_index);
location_info=[];
current_position = [plot_xyz_result(end,1),plot_xyz_result(end,2),plot_xyz_result(end,3)]; % [x, y, z]
for i=location_file_index:size(location_info_xy,1)
distance = norm(current_position(1, 1:2) - location_info_xy(i,:));
if distance < location_threshold
location_info = short_location_info_list{i};
% Write the location information to the location_info.txt
fd = fopen(location_info_file_name, 'w');
fprintf(fd,'%s',location_info_list{i});
fclose(fd);
location_file_index = location_file_index + 1;
break;
end
end
end
|
github
|
rising-turtle/slam_matlab-master
|
plot_isam.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/plot_isam.m
| 6,643 |
utf_8
|
b50a91a7288a763a25eedeb7ae97e1f4
|
% Plot the result of graph SLAM, isam
%
% Author : Soonhac Hong ([email protected])
% Date : 2/22/12
function plot_isam(org_file_name, opt_file_name)
[org_poses] = load_graph_isp(org_file_name);
[opt_poses] = load_graph_isam(opt_file_name);
% Show the pose
start_index = 1;
%end_index = min(size(org_poses,1), size(opt_poses)); %80;
end_index = size(org_poses,1);
if size(org_poses,2) == 2 % SE2
figure; plot(org_poses(start_index:end_index,1), org_poses(start_index:end_index,2),'b*-');
%xlabel('X [m]');
%ylabel('Y [m]');
%grid;[e_t_pose e_o_pose] = convert_o2p(f_index, t_pose, o_pose)
%figure;
hold on;
plot(opt_poses(start_index:end_index,1), opt_poses(start_index:end_index,2),'ro-','LineWidth',2);
%plot_groundtruth();
xlabel('X [m]');
ylabel('Y [m]');
grid;
legend('vro','isam'); %,'eGT');
%xlim([-0.1 0.6]);
%ylim([-0.1 0.6]);
hold off;
elseif size(org_poses,2) == 3 % SE3:QUAT
figure; plot3(org_poses(:,1), org_poses(:,2), org_poses(:,3), 'b:', 'LineWidth',2);
%xlabel('X [m]');
%ylabel('Y [m]');
%zlabel('Z [m]');plot_g2o.m
%grid;[e_t_pose e_o_pose] = convert_o2p(f_index, t_pose, o_pose)
%figure;
hold on;
plot3(opt_poses(:,1), opt_poses(:,2), opt_poses(:,3), 'r-.', 'LineWidth',2);
%plot_groundtruth_3D();
%plot_gt_etas();
xlabel('X [m]');
ylabel('Y [m]');
zlabel('Z [m]');
legend('vro','isam','GT_e');
grid;
axis equal;
hold off;
end
end
function plot_groundtruth()
gt_x = [0 0 2135 2135 0];
gt_y = [0 1220 1220 0 0];
gt_x = gt_x / 1000; % [mm] -> [m]
gt_y = gt_y / 1000; % [mm] -> [m]
plot(gt_x,gt_y,'g-','LineWidth',2);
end
function plot_groundtruth_3D()
% inch2mm = 304.8; % 1 cube = 12 inch = 304.8 mm
% gt_x = [0 0 13*inch2mm 13*inch2mm 0];
% gt_y = [-1*inch2mm 6*inch2mm 6*inch2mm -1*inch2mm -1*inch2mm];
% %gt_x = [0 0 2135 2135 0];
% %gt_y = [0 1220 1220 0 0];
% gt_z = [0 0 0 0 0];
% gt_x = gt_x / 1000; % [mm] -> [m]
% gt_y = gt_y / 1000; % [mm] -> [m]
% gt_z = gt_z / 1000; % [mm] -> [m]
inch2m = 0.0254; % 1 inch = 0.0254 m
gt_x = [0 0 150 910 965 965 910 50 0 0];
gt_y = [0 24 172.5 172.5 122.5 -122.5 -162.5 -162.5 -24 0];
gt_x = [gt_x 0 0 60 60+138 60+138+40 60+138+40 60+138 60 0 0];
gt_y = [gt_y 0 24 38.5+40 38.5+40 38.5 -38.5 -38.5-40 -38.5-40 -24 0];
gt_x = gt_x * inch2m;
gt_y = gt_y * inch2m;
gt_z = zeros(length(gt_x),1);
plot3(gt_x,gt_y,gt_z,'g-','LineWidth',2);
end
% function [t_pose] = load_graph_sam(file_name)
% fid = fopen(file_name);
% data = textscan(fid, '%s %d %f %f %f '); % 2D format
% fclose(fid);
%
% % Convert data
% % Pose
% data_name = data{1};
% data_name_list = {'ODOMETRY','LANDMARK','EDGE3','VERTEX_SE2'};
% vertex_index = 1;
% edge_index = 1;
% for i = 1 : size(data_name,1)
% if strcmp(data_name{i}, data_name_list{4}) % VERTEX_SE2
% %unit_data =[];
% %for j=3:5
% % unit_data = [unit_data data{j}(i)];
% %end
% f_index(vertex_index,:) = data{2}(i);
% t_pose(vertex_index,:) = [data{3}(i) data{4}(i) 0];
% o_pose(vertex_index,:) = [ 0 0 data{5}(i)];
% vertex_index = vertex_index + 1;
% % elseif strcmp(data_name{i}, data_name_list{2}) % EDGE_SE2
% % unit_data =[gt_y = gt_y / [e_t_pose e_o_pose] = convert_o2p(f_index, t_pose, o_pose)1000; % [mm] -> [m]];
% % for j=4:12
% % unit_data = [unit_data data{j}(i)];
% % end% elseif strcmp(data_name{i}, data_name_list{2}) % EDGE_SE2
% % unit_data =[];
% % for j=4:12
% % unit_data= [unit_data data{j}(i)];
% % end
% % edges(edge_index,:) = unit_data;
% % edge_index = edge_index + 1;
%
% % edges(edge_index,:) = unit_data;
% % edge_index = edge_index + 1;
% % elseif strcmp(data_name{i}, data_name_list{3}) % VERTEX_SE3:QUAT
% % unit_data =[];
% % for j=3:9
% % unit_data = [unit_data data{j}(i)];
% % end
% % poses(vertex_index,:) = unit_data;
% % vertex_index = vertex_index + 1;
% % elseif strcmp(data_name{i}, data_name_list{4}) % EDGE_SE3:QUAT
% % unit_data =[];
% % for j=2:31
% % unit_data = [unit_data data{j}(i)];
% % end
% % edge[e_t_pose e_o_pose] = convert_o2p(f_index, t_pose, o_pose)s(edge_index,:) = unit_data;
% % edge_index = edge_index + 1;
% end
%
% end
% end
% function [poses] = load_graph_opt(file_name)
% fid = fopen(file_name);
%
%
% % Convert data
% % Pose
%
% data_name_list = {'Pose2d_Node','Pose2d_Pose2d_Factor'};
% vertex_index = 1;
% edge_index = 1;
% while ~feof(fid) %for i = 1 : size(data_name,1)
% header = textscan(fid, '%s',1); % 2D format
% data_name = header{1};
% if strcmp(data_name, data_name_list{1}) % VERTEX_SE2
% data = textscan(fid, '%f (%f,%f,%f)');
% unit_data =[];
% for j=1:4
% unit_data = [unit_data data{j}];
% end
% f_index(vertex_index, :) = unit_data(1);
% t_pose(vertex_index,:) = [unit_data(2:3) 0];
% o_pose(vertex_index,:) = [0 0 unit_data(4)];
% vertex_index = vertex_index + 1;
% % elseif strcmp(data_name{i}, data_name_list{2}) % EDGE_SE2
% % unit_data =[];
% % for j=4:12
% % unit_data= [unit_data data{j}(i)];
% % end
% % edges(edge_index,:) = unit_data;
% % edge_index = edge_index + 1;
% % elseif strcmp(data_name{i}, data_name_list{3}) % VERTEX_SE3:QUAT
% % unit_data =[];
% % for j=3:9
% % unit_data = [unit_data data{j}(i)];
% % end
% % poses(vertex_index,:) = unit_data;
% % vertex_index = vertex_index + 1;
% % elseif strcmp(data_name{i}, data_name_list{4}) % EDGE_SE3:QUAT
% % unit_data =[];
% % for j=2:31
% % unit_data = [unit_data data{j}(i)];
% % end
% % edges(edge_index,:) = unit_data;
% % edge_index = edge_index + 1;
% end
%
% end
% fclose(fid);
% poses = [t_pose(:,1:2) o_pose(:,3)];
% end
|
github
|
rising-turtle/slam_matlab-master
|
load_graph_g2o.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/load_graph_g2o.m
| 2,350 |
utf_8
|
90f4a5ad47e0b0c319d82f953af06196
|
% Load graph of vro
function [poses edges fpts_poses fpts_edges] = load_graph_g2o(file_name)
fid = fopen(file_name);
data = textscan(fid, '%s %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f'); % 2D format
fclose(fid);
% Convert data
% Pose
data_name = data{1};
data_name_list = {'VERTEX_SE2','EDGE_SE2','VERTEX_SE3:QUAT','EDGE_SE3:QUAT'};
fpts_name_list ={'VERTEX_XY','EDGE_SE2_XY'};
vertex_index = 1;
edge_index = 1;
fpts_pose_index =1;
fpts_edge_index =1;
fpts_poses =[];
fpts_edges =[];
for i = 1 : size(data_name,1)
if strcmp(data_name{i}, data_name_list{1}) % VERTEX_SE2
unit_data =[];
for j=3:5
unit_data = [unit_data data{j}(i)];
end
poses(vertex_index,:) = unit_data;
vertex_index = vertex_index + 1;
elseif strcmp(data_name{i}, data_name_list{2}) % EDGE_SE2
unit_data =[];
for j=2:12
unit_data = [unit_data data{j}(i)];
end
edges(edge_index,:) = unit_data;
edge_index = edge_index + 1;
elseif strcmp(data_name{i}, fpts_name_list{1}) % VERTEX_XY
unit_data =[];
for j=3:5
unit_data = [unit_data data{j}(i)];
end
fpts_poses(fpts_pose_index,:) = unit_data;
fpts_pose_index = fpts_pose_index + 1;
elseif strcmp(data_name{i}, fpts_name_list{2}) % EDGE_SE2_XY
unit_data =[];
for j=2:12
unit_data = [unit_data data{j}(i)];
end
fpts_edges(fpts_edge_index,:) = unit_data;
fpts_edge_index = fpts_edge_index + 1;
elseif strcmp(data_name{i}, data_name_list{3}) % VERTEX_SE3:QUAT
unit_data =[];
for j=3:9
unit_data = [unit_data data{j}(i)];
end
poses(vertex_index,:) = unit_data;
vertex_index = vertex_index + 1;
elseif strcmp(data_name{i}, data_name_list{4}) % EDGE_SE3:QUAT
unit_data =[];
for j=2:31
unit_data = [unit_data data{j}(i)];
end
edges(edge_index,:) = unit_data;
edge_index = edge_index + 1;
end
end
end
|
github
|
rising-turtle/slam_matlab-master
|
Convert_map.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/Convert_map.m
| 2,344 |
utf_8
|
43aaeae9934a5b80a0b75f7f680234ef
|
% Convert map based its locations
%
% Author : Soonhac Hong ([email protected])
% Date : 5/16/14
function Convert_map()
%input_file_name = 'D:\Soonhac\SW\GraphSLAM\results\isam\3d\revisiting9_10m_Replace_pose_zero_827_vro_gtsam_feature_total_v1.ply';
%output_file_name = 'D:\Soonhac\SW\GraphSLAM\results\isam\3d\revisiting9_10m_Replace_pose_zero_827_vro_gtsam_feature_total_v21.ply';
input_file_name = 'D:\Soonhac\SW\GraphSLAM\results\isam\3d\revisiting9_10m_Replace_pose_zero_827_vro_gtsam_feature_v1.ply';
output_file_name = 'D:\Soonhac\SW\GraphSLAM\results\isam\3d\revisiting9_10m_Replace_pose_zero_827_vro_gtsam_feature_v2.ply';
%% Load an input file
fid = fopen(input_file_name);
if fid < 0
error(['Cannot open file ' input_file_name]);
end
% scan all lines into a cell array
columns=textscan(fid,'%s','delimiter','\n');
lines=columns{1};
fclose(fid);
%% Convert the color of a trajectory
x_threshold_min = -4;
x_threshold_max = 0.5;
y_threshold_min = -3;
y_threshold_max = 7;
z_threshold = -0.67;
new_data=[];
new_data_index = 1;
for i=1:size(lines,1)
i
line_i=lines{i};
if i > 12 % check intensity data
v = textscan(line_i,'%f %f %f %d %d %d',1);
if v{1} < x_threshold_max && v{1} > x_threshold_min && v{2} < y_threshold_max && v{2} > y_threshold_min
if v{3} < z_threshold
new_data(new_data_index, :)=[v{1},v{2},v{3},double(v{4}),double(v{5}),double(v{6})];
new_data_index = new_data_index + 1;
%fprintf(fd,'%f %f %f %d %d %d\n',v{1},v{2},v{3}, target_color);
end
else
new_data(new_data_index, :)=[v{1},v{2},v{3},double(v{4}),double(v{5}),double(v{6})];
new_data_index = new_data_index + 1;
end
end
end
%% Write data
ply_headers={'ply','format ascii 1.0','comment Created with XYZRGB_to_PLY', 'element_vertex_dummy', 'property float x', 'property float y','property float z','property uchar red','property uchar green','property uchar blue','end_header'};
nply_data = size(new_data,1)
element_vertex_n = sprintf('element vertex %d',nply_data);
ply_headers{4} = element_vertex_n;
fd = fopen(output_file_name, 'w');
for i=1:size(ply_headers,2)
fprintf(fd,'%s\n',ply_headers{i});
end
for i=1:nply_data
fprintf(fd,'%f %f %f %d %d %d\n',new_data(i,:));
end
fclose(fd);
end
|
github
|
rising-turtle/slam_matlab-master
|
get_global_transformation_dataname.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/get_global_transformation_dataname.m
| 11,980 |
utf_8
|
8be9a4cd85dc54f2a37329cbcb602d99
|
% Get global transformation for each data set
%
% Author : Soonhac Hong ([email protected])
% Date : 10/22/12
function [h_global] = get_global_transformation_dataname(data_name, dynamic_index, isgframe)
addpath('..\slamtoolbox\slamToolbox_11_09_08\FrameTransforms\Rotations');
% rx, ry, rz : [degree]
% tx, ty, tz : [mm]
rx=0; ry=0; rz=0; tx=0; ty=0; tz=0;
switch data_name
case {'square'}
switch dynamic_index
case 16
%h_global = [euler_to_rot(0, -15.4, 0) [0 0 0]'; 0 0 0 1]; % square_700
rx = -15.4;
end
case {'etas'}
switch dynamic_index
case 1
%h_global = [euler_to_rot(0, -36.3, 0) [0 0 0]'; 0 0 0 1]; % etas
%rx = -36.3;
rx = -29.5089; ry = 1.1837; rz=3.6372;
case 3
rx = -28.9278; ry = 1.1894; rz=0.8985;
case 5
rx = -29.2181; ry = 1.1830; rz=1.8845;
end
case {'loops'}
switch dynamic_index
case 2
%h_global = [euler_to_rot(1.23, -25.6, 3.51) [0 0 0]'; 0 0 0 1]; % loops_2
rx=-25.6; ry=1.23; rz=3.51;
case 3
%h_global =[euler_to_rot(1.2670, -24.3316, 4.6656) [0 0 0]'; 0 0 0 1]; %loops_3
rx=-24.3316; ry=1.2670; rz=0; %4.6656;
case 13
rx=-25.6109; ry=1.2418; rx=3.8448
end
case {'kinect_tum'}
switch dynamic_index
case 2
% start image : 13050033527.670034
% Groundtruth : 1305033527.6662 1.4906 -1.1681 0.6610 0.8959 0.0713 -0.0460 -0.4361
init_qauterion = [0.8959, 0.0713, -0.046, -0.4361];
% temp_rot = q2R(init_qauterion);
% [r1, r2, r3] = rot_to_euler(temp_rot);
% r2d = 180 / pi;
% temp_rot = euler_to_rot(r2*r2d, r1*r2d, r3*r2d);
temp_r = q2e(init_qauterion) * 180 / pi;
%temp_rot = euler_to_rot(temp_r(2), temp_r(1), temp_r(3)); % degree
%temp_rot = euler_to_rot(45, -45, 90) * temp_rot ;
%[temp_r(2) temp_r(1) temp_r(3)] = rot_to_euler(temp_rot);
temp_trans = [1.4906 -1.1681 0.661]*1000;
%temp_r = temp_r * 180 / pi;
rx = temp_r(1); ry = temp_r(2); rz = temp_r(3);
%rx =0; ry =0; rz=0;
tx = temp_trans(1); ty = temp_trans(2); tz = temp_trans(3);
end
case {'loops2', 'amir_vro'}
switch dynamic_index
case 1
%h_global =[euler_to_rot(1.2670, -24.3316, 4.6656) [0 0 0]'; 0 0 0 1]; %loops_3
rx=-24.4985419798496; ry=1.26190320102629; rz=4.69088809909393; %4.6656;
case 2
rx=-23.3420; ry=1.2739; rz=4.3772;
case 3
rx=-24.5234; ry=1.2565; rz=4.2682;
case 4
rx=-24.2726; ry=1.2606; rz=2.7462;
case 5
rx=-26.6191; ry=1.2173; rz=5.0601; %2.7462;
case 6
rx=-25.7422; ry=1.2396; rz=3.0821;
case 7
rx=-24.6791; ry=1.2562; rz=4.1917; %3.0821;
case 8
%h_global =[euler_to_rot(1.2670, -24.3316, 4.6656) [0 0 0]'; 0 0 0 1]; %loops_3
rx=-23.8464; ry=1.2721; rz= 3.7964; %3.75437022813459;
%init_qauterion = [0.977813222516827,-0.206522692608165,0.00392574917797234,0.0348463456121639];
%temp_r = q2e(init_qauterion) * 180 / pi;
%rx = temp_r(1); ry = temp_r(2); rz = temp_r(3);
case 9
rx=-25.6806; ry=1.2383; rz=3.4203;
case 10
rx=-25.1216960394087; ry=1.25311495911881; rz=-0.909688742543562;
case 11
%h_global = [euler_to_rot(1.23, -25.6, 3.51) [0 0 0]'; 0 0 0 1]; % loops_2
%rx=-25.6; ry=1.23; rz=3.51;
rx=-33.7709; ry=1.0903; rz=3.0738;
case 12 % same as exp 7
rx=-24.6791; ry=1.2562; rz=4.1917; %3.0821;
end
case {'sparse_feature'}
switch dynamic_index
case 1
rx=-29.9267; ry=1.1978; rz=0; %-2.6385; %4.69088809909393; %4.6656;
case 2
rx=-28.7386; ry=1.2247; rz=0.8001; %-2.6385; %4.69088809909393; %4.6656;
case 3
rx=-31.1218; ry=1.1713; rz=-1.7365;
case 4
rx=-29.3799; ry=1.2133; rz=-2.4075;
case {5,6,7,8}
rx = -39.1930; ry = 1.0134; rz = -1.2122;
case {9, 10, 11}
rx = -37.1533; ry = 1.0455; rz = -3.0266;
case {12,16}
rx = -38.3570; ry = 1.0315; rz = -0.3114;
case 13
rx=-28.3021; ry=1.2357; rz=-1.3391;
case 14
rx=-29.1073; ry=1.2133; rz=-0.5520;
end
case {'swing','swing2'}
switch dynamic_index
case 1
rx=-32.2179; ry=1.1567; rz=-1.1344;
case 2
rx=-28.6379; ry=1.2207; rz=-1.5099;
case 3
rx=-33.3490; ry=1.1279; rz=-3.1225;
case 4
rx=-31.0073; ry=1.1770; rz=-0.6052;
case 5
rx=-31.7611; ry=1.1627; rz=-0.9156;
case 7
rx=-19.2560; ry=0.8450; rz=0.5809;
case 8
rx=-17.5219; ry=0.8546; rz=-4.5522;
case 9
rx=-17.2286; ry=0.8616; rz=-3.4105;
case 10
rx=-18.1945; ry=0.8501; rz=1.4029;
case 11
rx=-18.2734; ry=0.8556; rz=0.2481;
case 12
rx=-24.4356; ry=0.7918; rz=-0.9333;
case 13
rx=-17.2649; ry=0.8549; rz=-2.8848;
case 14
rx=-20.1519; ry=0.8284; rz=-1.3661;
case 15
rx=-20.5691; ry=0.8302; rz=-0.7713;
case 16
rx=-19.8892; ry=0.8389; rz=-1.3603;
case 17
rx=-24.4211; ry=0.7959; rz=-3.0972;
case 18
rx=-21.7207; ry=0.8163; rz=-4.3600;
case 19
rx=-19.3202; ry=0.8454; rz=-2.7863;
case 20
rx=-20.2662; ry=0.8354; rz=-2.5763;
case 21
rx=-19.5310; ry=0.8403; rz=-1.9310;
case 22
rx=-18.3081; ry=0.8536; rz=-4.6978;
case 23
rx=-16.7653; ry=0.8678; rz=-4.2483;
case 24
rx=-16.4797; ry=0.8697; rz=-2.3778;
case 25
rx=-17.3060; ry=0.8587; rz=-3.1753;
case 26
rx=-16.9306; ry=0.8576; rz=-2.6512;
end
case {'motive'}
switch dynamic_index
case 1
rx=-22.2924; ry=0.8198; rz=-4.3367;
case 2
rx=-23.9217; ry=0.7990; rz=-3.1450;
case 11
rx=-17.5025; ry=1.3989; rz=-1.5120;
case 12
rx=-17.1856; ry=1.4018; rz=-0.3795;
case 13
rx=-16.8428; ry=0.8657; rz=-4.1148;
case 15
rx=-20.3861; ry=0.8323; rz=-4.4565;
case 16
rx=-21.2226; ry=0.8261; rz=-6.5253;
case 17
rx=-19.2100; ry=1.3724; rz=-4.0953;
case 18
rx=-18.4614; ry=1.3822; rz=-2.2221;
case 19
rx=-14.2685; ry=0.8810; rz=-2.4514;
case 20
rx=-14.8030; ry=0.8806; rz=-3.6865;
case 21
rx=-18.6913; ry=0.8436; rz=-1.3773;
case 22
rx=-15.3041; ry=0.8815; rz=0.3623;
case 23
%rx=-14.6923; ry=0.8813; rz=0.5994;
rx=-15.8020; ry=0.8704; rz=-1.0940;
case 24
%rx=-13.9262; ry=0.8867; rz=2.9371;
rx=-13.4739; ry=0.8915; rz=-3.8607;
case 25
%rx=-12.5003; ry=0.8973; rz=-1.6229;
rx=-12.5852; ry=0.8966; rz=-1.7487;
case 26
%rx=-13.1931; ry=0.8922; rz=0.3750;
rx=-13.1931; ry=0.8922; rz=0.3750;
case 27
rx=-13.7715; ry=0.8868; rz=0.7914;
case 28
rx=-14.2207; ry=0.8836; rz=-1.8911;
case 29
rx=-13.4205; ry=0.8920; rz=-1.8912;
case 30
rx=-15.2320; ry=0.8793; rz=-0.8757;
case 31
rx=-15.8020; ry=0.8704; rz=-1.0940;
case 32
rx=-18.3082; ry=1.3832; rz=2.6882;
case 33
rx=-18.3023; ry=1.3870; rz=-0.4907;
case 34
rx=-16.3416; ry=1.4141; rz=-1.3016;
case 35
rx=-15.9030; ry=1.4182; rz=-2.7589;
case 36
rx=-18.8863; ry=1.3755; rz=-4.6635;
case 37
rx=-15.9639; ry=1.4180; rz=0.1582;
end
case {'object_recognition'}
switch dynamic_index
case 1
rx=-37.7656; ry=1.0435; rz=-5.2416;
case 2
rx=-9.4696; ry=0.9351; rz=8.0194;
case 3
rx=-49.0312; ry=0.4693; rz=17.9271;
case 4
rx = -102.3278; ry=-0.4342; rz=-4.9235;
case 5
rx=-36.4384; ry=1.0755; rz=22.0853;
case 6
rx=-36.8751; ry=1.0886; rz=24.7719;
case 7
rx = -26.0697; ry=1.2667; rz=3.3011;
case 8
rx = -34.3278; ry=0.6732; rz=3.7436;
case 9
rx = -28.8132; ry=0.7380; rz=10.7045;
case 10
rx = -35.7148; ry=0.6569; rz=-10.9565;
case 11
rx = -27.3853; ry=0.7568; rz=-7.5140;
end
case {'map'}
switch dynamic_index
case 2
rx=-20.9872; ry=0.8237; rz= -1.6569;
case 3
rx=-21.9661; ry=0.8124; rz=1.3211;
case 4
rx=-20.0099; ry=0.8344; rz=3.4412;
case 5
rx=-20.7988; ry=0.8285; rz=-1.4896;
case 6
%rx=-20.8341; ry=0.8281; rz=-1.5208;
rx=-17.5434; ry=1.4023; rz=-0.4495;
case 7
rx=-19.7766; ry=1.3672; rz=-2.9528;
case 8
rx=-20.7362; ry=1.3549; rz=-3.8060;
case 9
rx=-17.1903; ry=1.4010; rz=0.2855;
end
end
%h_global_tf = [euler_to_rot(0, 90, 0) [0 0 0]'; 0 0 0 1];
if strcmp(isgframe, 'gframe')
h_global = [e2R([rx*pi/180, ry*pi/180, rz*pi/180]) [tx ty tz]'; 0 0 0 1];
else
%h_global = [euler_to_rot(ry, rx, rz) [tx ty tz]'; 0 0 0 1];
h_global = [euler_to_rot(rz, rx, ry) [tx ty tz]'; 0 0 0 1];
end
%h_global = h_global * h_global_tf;
%h_global = [euler_to_rot(0, 0, 15) [tx ty tz]'; 0 0 0 1];
%h_global = [euler_to_rot(rz, rx, ry) [tx ty tz]'; 0 0 0 1];
%h_global = [euler_to_rot(0, 0, 0) [0 0 0]'; 0 0 0 1];
%h_global = [euler_to_rot(0, -23.8, 0) [0 0 0]'; 0 0 0 1]; % square_1000
%init_qauterion = [0.8959, 0.0695, -0.0461, -0.4364];
%temp_rot = q2R(init_qauterion);
%[r1, r2, r3] = rot_to_euler(temp_rot);
%r2d = 180 / pi;
%temp_rot = euler_to_rot(r2*r2d, r1*r2d, r3*r2d);
%temp_r = q2v(init_qauterion) * 180 / pi;
%temp_rot = euler_to_rot(temp_r(2), temp_r(1), temp_r(3)); % degree
%temp_trans = [1.4908 -1.1707 0.6603]*1000;
%h_global = [temp_rot temp_trans'; 0 0 0 1]; % kinect_tum
%h_global_temp = [euler_to_rot(90, 0, 90) [0 0 0]'; 0 0 0 1]; % ?????
%h_global_temp = [[1 0 0; 0 0 -1; 0 1 0] [0 0 0]'; 0 0 0 1];
%h_global = h_global * h_global_temp;
%h_global = [euler_to_rot(0, 0, 0) temp_trans'; 0 0 0 1];
end
|
github
|
rising-turtle/slam_matlab-master
|
Convert_map_culling.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/Convert_map_culling.m
| 1,973 |
utf_8
|
5f09a0f4dbe560f1581a95e4b29f0753
|
% Convert map based its locations
%
% Author : Soonhac Hong ([email protected])
% Date : 5/16/14
function Convert_map_culling()
%input_file_name = 'D:\Soonhac\SW\GraphSLAM\results\isam\3d\revisiting9_10m_Replace_pose_zero_827_vro_gtsam_feature_total_v1.ply';
%output_file_name = 'D:\Soonhac\SW\GraphSLAM\results\isam\3d\revisiting9_10m_Replace_pose_zero_827_vro_gtsam_feature_total_v21.ply';
input_file_name = 'D:\Soonhac\SW\GraphSLAM\results\isam\3d\revisiting9_10m_Replace_pose_zero_827_vro_gtsam_feature_v2.ply';
output_file_name = 'D:\Soonhac\SW\GraphSLAM\results\isam\3d\revisiting9_10m_Replace_pose_zero_827_vro_gtsam_feature_v3.ply';
%% Load an input file
fid = fopen(input_file_name);
if fid < 0
error(['Cannot open file ' input_file_name]);
end
% scan all lines into a cell array
columns=textscan(fid,'%s','delimiter','\n');
lines=columns{1};
fclose(fid);
%% Convert the color of a trajectory
x_threshold_min = -4;
x_threshold_max = 0.5;
y_threshold_min = -3;
y_threshold_max = 7;
z_threshold = -0.67;
new_data=[];
new_data_index = 1;
for i=1:size(lines,1)
i
line_i=lines{i};
if i > 12 % check intensity data
if mod(i,2) == 0
v = textscan(line_i,'%f %f %f %d %d %d',1);
new_data(new_data_index, :)=[v{1},v{2},v{3},double(v{4}),double(v{5}),double(v{6})];
new_data_index = new_data_index + 1;
end
end
end
%% Write data
ply_headers={'ply','format ascii 1.0','comment Created with XYZRGB_to_PLY', 'element_vertex_dummy', 'property float x', 'property float y','property float z','property uchar red','property uchar green','property uchar blue','end_header'};
nply_data = size(new_data,1)
element_vertex_n = sprintf('element vertex %d',nply_data);
ply_headers{4} = element_vertex_n;
fd = fopen(output_file_name, 'w');
for i=1:size(ply_headers,2)
fprintf(fd,'%s\n',ply_headers{i});
end
for i=1:nply_data
fprintf(fd,'%f %f %f %d %d %d\n',new_data(i,:));
end
fclose(fd);
end
|
github
|
rising-turtle/slam_matlab-master
|
load_graph_toro.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/load_graph_toro.m
| 2,468 |
utf_8
|
9ad77c44caa6bd902416b9677c82e351
|
% Plot the data of TORO
% Author : Soonhac Hong ([email protected])
% Date : 4/19/2012
function [poses edges fpts_poses fpts_edges] = load_graph_toro(file_name)
fid = fopen(file_name);
data = textscan(fid, '%s %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f'); % 2D format
fclose(fid);
% Convert data
% Pose
data_name = data{1};
data_name_list = {'VERTEX','EDGE','VERTEX2','EDGE2','VERTEX3','EDGE3'};
fpts_name_list ={'VERTEX_XY','EDGE_SE2_XY'};
vertex_index = 1;
edge_index = 1;
fpts_pose_index =1;
fpts_edge_index =1;
fpts_poses =[];
fpts_edges =[];
for i = 1 : size(data_name,1)
if strcmp(data_name{i}, data_name_list{1}) || strcmp(data_name{i}, data_name_list{3}) % VERTEX
unit_data =[];
for j=3:5
unit_data = [unit_data data{j}(i)];
end
poses(vertex_index,:) = unit_data;
vertex_index = vertex_index + 1;
elseif strcmp(data_name{i}, data_name_list{2}) || strcmp(data_name{i}, data_name_list{4}) % EDGE
unit_data =[];
for j=2:12
unit_data = [unit_data data{j}(i)];
end
edges(edge_index,:) = unit_data;
edge_index = edge_index + 1;
elseif strcmp(data_name{i}, fpts_name_list{1}) % VERTEX_XY
unit_data =[];
for j=3:5
unit_data = [unit_data data{j}(i)];
end
fpts_poses(fpts_pose_index,:) = unit_data;
fpts_pose_index = fpts_pose_index + 1;
elseif strcmp(data_name{i}, fpts_name_list{2}) % EDGE_SE2_XY
unit_data =[];
for j=2:12
unit_data = [unit_data data{j}(i)];
end
fpts_edges(fpts_edge_index,:) = unit_data;
fpts_edge_index = fpts_edge_index + 1;
elseif strcmp(data_name{i}, data_name_list{5}) % VERTEX3
unit_data =[];
for j=3:8
unit_data = [unit_data data{j}(i)];
end
poses(vertex_index,:) = unit_data;
vertex_index = vertex_index + 1;
elseif strcmp(data_name{i}, data_name_list{6}) % EDGE3
unit_data =[];
for j=2:30
unit_data = [unit_data data{j}(i)];
end
edges(edge_index,:) = unit_data;
edge_index = edge_index + 1;
end
end
|
github
|
rising-turtle/slam_matlab-master
|
graph_slam_1d.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/graph_slam_1d.m
| 2,026 |
utf_8
|
0855b87c8e04ca339c4e5517526e64ec
|
% 1-D graph SLAM
% Data : 4/12/12
% Author : Soonhac Hong ([email protected])
function graph_slam_1d()
% Data
%pose_data = [1 1 -3; 1 2 5; 2 3 3]; % first row : initial pose, second row : [1]= current index, [2] = next index, [3] = movement
%result = [-3; 2; 5];
% pose_data =[1 1 0; 1 2 3; 2 3 4; 1 3 6; 3 4 2];
%
% [-9.66914000000000,10.9800100000000,6.85198000000000;]
% [-12.9620100000000,25.6074800000000,18.3592900000000;]
% [-0.776060000000000,16.2314800000000,6.47385000000000;]
% [-4.67272000000000,36.1912600000000,18.4479100000000;]
% [0.930870000000000,38.1723100000000,17.0636200000000;]
pose_data =[1 1 0; 1 2 -9.66914000000000; 2 3 -12.9620100000000; 3 4 -0.776060000000000; 4 5 -4.67272000000000; 5 6 0.930870000000000];
motion_noise = 1.0;
pose_num = size(unique(pose_data(:,1:2)),1);
Omega = zeros(pose_num,pose_num);
Xi = zeros(pose_num,1);
Odometry = zeros(pose_num,1);
% Fill the elements of Omega and Xi
Omega(1,1) = 1;
Xi(1) = pose_data(1,3);
Odometry = pose_data(1,3);
for i=2:size(pose_data,1);
unit_data = pose_data(i,:);
current_index = unit_data(1);
next_index = unit_data(2);
movement = unit_data(3);
% Fill diagonal elements of Omega
Omega(current_index,current_index) = Omega(current_index,current_index) + 1/motion_noise;
Omega(next_index,next_index) = Omega(next_index,next_index) + 1/motion_noise;
% Fill Off-diagonal elements of Omega
Omega(current_index,next_index) = Omega(current_index,next_index) + (-1)/motion_noise;
Omega(next_index,current_index) = Omega(next_index,current_index) + (-1)/motion_noise;
% Fill Xi
Xi(current_index) = Xi(current_index) + (-1)*movement/motion_noise;
Xi(next_index) = Xi(next_index) + movement/motion_noise;
% Update Odometry
if abs(current_index - next_index) == 1
Odometry(next_index) = Odometry(next_index-1) + movement;
end
end
Omega
Xi
mu = Omega^-1 * Xi
plot(Odometry,'bd-');
hold on;
plot(mu,'ro-');
hold off;
legend('Odometry','Optimized');
end
|
github
|
rising-turtle/slam_matlab-master
|
plot_gtsam.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/plot_gtsam.m
| 8,191 |
utf_8
|
32b10ae3e271c52b77844abca2c62522
|
% Plot the result of graph SLAM, isam
%
% Author : Soonhac Hong ([email protected])
% Date : 2/22/12
function plot_gtsam(org_file_name, opt_file_name)
[org_poses] = load_graph_isp(org_file_name);
[opt_poses] = load_graph_isp(opt_file_name);
compute_trajectory_error(org_poses, opt_poses);
% Show the pose
start_index = 1;
%end_index = min(size(org_poses,1), size(opt_poses)); %80;
end_index = size(org_poses,1);
if size(org_poses,2) == 2 % SE2
figure; plot(org_poses(start_index:end_index,1), org_poses(start_index:end_index,2),'b*-');
%xlabel('X [m]');
%ylabel('Y [m]');
%grid;[e_t_pose e_o_pose] = convert_o2p(f_index, t_pose, o_pose)
%figure;
hold on;
plot(opt_poses(start_index:end_index,1), opt_poses(start_index:end_index,2),'ro-','LineWidth',2);
%plot_groundtruth();
xlabel('X [m]');
ylabel('Y [m]');
grid;
legend('vro_icp','gtsam'); %,'eGT');
%xlim([-0.1 0.6]);
%ylim([-0.1 0.6]);
hold off;
elseif size(org_poses,2) == 3 % SE3:QUAT
figure; plot3(org_poses(:,1), org_poses(:,2), org_poses(:,3), 'b:', 'LineWidth',2);
%xlabel('X [m]');
%ylabel('Y [m]');
%zlabel('Z [m]');plot_g2o.m
%grid;[e_t_pose e_o_pose] = convert_o2p(f_index, t_pose, o_pose)
%figure;
hold on;
plot3(opt_poses(:,1), opt_poses(:,2), opt_poses(:,3), 'r-.', 'LineWidth',2);
%plot_groundtruth_3D();
%plot_gt_etas();
xlabel('X [m]');
ylabel('Y [m]');
zlabel('Z [m]');
legend('vro icp','gtsam','GT_e');
grid;
axis equal;
hold off;
end
end
function compute_trajectory_error(org_poses, opt_poses)
%Compute trajectory percentage error
% org_poses(end,:)
% opt_poses(end,:)
org_poses_first_last_distance = sqrt(sum(abs(org_poses(1,:)-org_poses(end,:)).^2))
opt_poses_first_last_distance = sqrt(sum(abs(opt_poses(1,:)-opt_poses(end,:)).^2))
%Compute trajectory length
org_poses_trajectory_length = 0;
opt_poses_trajectory_length = 0;
for i=1:size(org_poses,1)-1
org_poses_trajectory_length = org_poses_trajectory_length + sqrt(sum(abs(org_poses(i,:)-org_poses(i+1,:)).^2));
end
for i=1:size(org_poses,1)-1
opt_poses_trajectory_length = opt_poses_trajectory_length + sqrt(sum(abs(opt_poses(i,:)-opt_poses(i+1,:)).^2));
end
org_poses_trajectory_length
opt_poses_trajectory_length
%Compute percentage error
org_poses_trajectory_error = org_poses_first_last_distance * 100 / org_poses_trajectory_length
opt_poses_trajectory_error = opt_poses_first_last_distance * 100 / opt_poses_trajectory_length
% %Compute height error
% org_poses_delta = abs(org_poses - repmat(org_poses(1,:), size(org_poses,1),1));
% opt_poses_delta = abs(opt_poses - repmat(opt_poses(1,:), size(opt_poses,1),1));
%
% mean_org_poses_delta_height = mean(org_poses_delta(:,3))
% mean_opt_poses_delta_height = mean(opt_poses_delta(:,3))
%
% % Show height error
% figure;
% plot(org_poses_delta(:,3),'b-');
% hold on;
% plot(opt_poses_delta(:,3),'r-');
% legend('vro','gtsam');
% grid;
% ylabel('height error [m]');
% xlabel('Step');
% hold off;
end
function plot_groundtruth()
gt_x = [0 0 2135 2135 0];
gt_y = [0 1220 1220 0 0];
gt_x = gt_x / 1000; % [mm] -> [m]
gt_y = gt_y / 1000; % [mm] -> [m]
plot(gt_x,gt_y,'g-','LineWidth',2);
end
function plot_groundtruth_3D()
% inch2mm = 304.8; % 1 cube = 12 inch = 304.8 mm
% gt_x = [0 0 13*inch2mm 13*inch2mm 0];
% gt_y = [-1*inch2mm 6*inch2mm 6*inch2mm -1*inch2mm -1*inch2mm];
% %gt_x = [0 0 2135 2135 0];
% %gt_y = [0 1220 1220 0 0];
% gt_z = [0 0 0 0 0];
% gt_x = gt_x / 1000; % [mm] -> [m]
% gt_y = gt_y / 1000; % [mm] -> [m]
% gt_z = gt_z / 1000; % [mm] -> [m]
inch2m = 0.0254; % 1 inch = 0.0254 m
gt_x = [0 0 150 910 965 965 910 50 0 0];
gt_y = [0 24 172.5 172.5 122.5 -122.5 -162.5 -162.5 -24 0];
gt_x = [gt_x 0 0 60 60+138 60+138+40 60+138+40 60+138 60 0 0];
gt_y = [gt_y 0 24 38.5+40 38.5+40 38.5 -38.5 -38.5-40 -38.5-40 -24 0];
gt_x = gt_x * inch2m;
gt_y = gt_y * inch2m;
gt_z = zeros(length(gt_x),1);
plot3(gt_x,gt_y,gt_z,'g-','LineWidth',2);
end
% function [t_pose] = load_graph_sam(file_name)
% fid = fopen(file_name);
% data = textscan(fid, '%s %d %f %f %f '); % 2D format
% fclose(fid);
%
% % Convert data
% % Pose
% data_name = data{1};
% data_name_list = {'ODOMETRY','LANDMARK','EDGE3','VERTEX_SE2'};
% vertex_index = 1;
% edge_index = 1;
% for i = 1 : size(data_name,1)
% if strcmp(data_name{i}, data_name_list{4}) % VERTEX_SE2
% %unit_data =[];
% %for j=3:5
% % unit_data = [unit_data data{j}(i)];
% %end
% f_index(vertex_index,:) = data{2}(i);
% t_pose(vertex_index,:) = [data{3}(i) data{4}(i) 0];
% o_pose(vertex_index,:) = [ 0 0 data{5}(i)];
% vertex_index = vertex_index + 1;
% % elseif strcmp(data_name{i}, data_name_list{2}) % EDGE_SE2
% % unit_data =[gt_y = gt_y / [e_t_pose e_o_pose] = convert_o2p(f_index, t_pose, o_pose)1000; % [mm] -> [m]];
% % for j=4:12
% % unit_data = [unit_data data{j}(i)];
% % end% elseif strcmp(data_name{i}, data_name_list{2}) % EDGE_SE2
% % unit_data =[];
% % for j=4:12
% % unit_data= [unit_data data{j}(i)];
% % end
% % edges(edge_index,:) = unit_data;
% % edge_index = edge_index + 1;
%
% % edges(edge_index,:) = unit_data;
% % edge_index = edge_index + 1;
% % elseif strcmp(data_name{i}, data_name_list{3}) % VERTEX_SE3:QUAT
% % unit_data =[];
% % for j=3:9
% % unit_data = [unit_data data{j}(i)];
% % end
% % poses(vertex_index,:) = unit_data;
% % vertex_index = vertex_index + 1;
% % elseif strcmp(data_name{i}, data_name_list{4}) % EDGE_SE3:QUAT
% % unit_data =[];
% % for j=2:31
% % unit_data = [unit_data data{j}(i)];
% % end
% % edge[e_t_pose e_o_pose] = convert_o2p(f_index, t_pose, o_pose)s(edge_index,:) = unit_data;
% % edge_index = edge_index + 1;
% end
%
% end
% end
% function [poses] = load_graph_opt(file_name)
% fid = fopen(file_name);
%
%
% % Convert data
% % Pose
%
% data_name_list = {'Pose2d_Node','Pose2d_Pose2d_Factor'};
% vertex_index = 1;
% edge_index = 1;
% while ~feof(fid) %for i = 1 : size(data_name,1)
% header = textscan(fid, '%s',1); % 2D format
% data_name = header{1};
% if strcmp(data_name, data_name_list{1}) % VERTEX_SE2
% data = textscan(fid, '%f (%f,%f,%f)');
% unit_data =[];
% for j=1:4
% unit_data = [unit_data data{j}];
% end
% f_index(vertex_index, :) = unit_data(1);
% t_pose(vertex_index,:) = [unit_data(2:3) 0];
% o_pose(vertex_index,:) = [0 0 unit_data(4)];
% vertex_index = vertex_index + 1;
% % elseif strcmp(data_name{i}, data_name_list{2}) % EDGE_SE2
% % unit_data =[];
% % for j=4:12
% % unit_data= [unit_data data{j}(i)];
% % end
% % edges(edge_index,:) = unit_data;
% % edge_index = edge_index + 1;
% % elseif strcmp(data_name{i}, data_name_list{3}) % VERTEX_SE3:QUAT
% % unit_data =[];
% % for j=3:9
% % unit_data = [unit_data data{j}(i)];
% % end
% % poses(vertex_index,:) = unit_data;
% % vertex_index = vertex_index + 1;
% % elseif strcmp(data_name{i}, data_name_list{4}) % EDGE_SE3:QUAT
% % unit_data =[];
% % for j=2:31
% % unit_data = [unit_data data{j}(i)];
% % end
% % edges(edge_index,:) = unit_data;
% % edge_index = edge_index + 1;
% end
%
% end
% fclose(fid);
% poses = [t_pose(:,1:2) o_pose(:,3)];
% end
|
github
|
rising-turtle/slam_matlab-master
|
convert_pc2ply_map_registration_v2.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/convert_pc2ply_map_registration_v2.m
| 5,723 |
utf_8
|
cdf00bc9983ff11fc3c1ea7f124934dc
|
% Write 3D point clouds to a ply file using map registration algorithm
% Conver isp to ply for meshlab
%
% Author : Soonhac Hong ([email protected])
% History :
% 1/8/14 : Created
% 2/27/14 : Update map by mean value of correspondence and include quality
% factor
function convert_pc2ply_map_registration_v2(ply_headers, ply_file_name, poses, data_index, dynamic_index, isgframe)
data_name_list=get_data_name_list(); %{'pitch', 'pan', 'roll','x2', 'y2', 'c1', 'c2','c3','c4','m','etas','loops2','kinect_tum','sparse_feature','motive',};
feature_ply_file_name=strrep(ply_file_name, '.ply','_feature_mag_reg_v2.ply')
feature_ct_file_name=strrep(ply_file_name, '.ply','_feature_mag_reg_v2.ct')
% Write headers
pose_interval = 1;
sample_interval = 2;
image_width = 176;
image_height = 144;
show_image_width = floor(image_width/sample_interval);
show_image_height = floor(image_height/sample_interval);
% nfeature = size(poses,1)*show_image_width*show_image_height;
% element_vertex_n = sprintf('element vertex %d',nfeature);
% ply_headers{4} = element_vertex_n;
% for i=1:size(ply_headers,2)
% fprintf(fd,'%s\n',ply_headers{i});
% end
% Write data
for i = 1:size(poses,1)
% if vro_cpp == 1
% h{i} = [euler_to_rot(vro_o_pose(i,2), vro_o_pose(i,1), vro_o_pose(i,3)) vro_t_pose(i,:)'; 0 0 0 1];
% else
if strcmp(isgframe, 'gframe')
h{i} = [e2R([poses(i,4), poses(i,5), poses(i,6)]) poses(i,1:3)'; 0 0 0 1]; % e2R([rx,ry,rz]) [radian]
else
h{i} = [euler_to_rot(poses(i,5)*180/pi, poses(i,4)*180/pi, poses(i,6)*180/pi) poses(i,1:3)'; 0 0 0 1]; % euler_to_rot(ry, rx, rz) [degree]
end
end
distance_threshold_max = 5; %8; %5;
distance_threshold_min = 0.8;
ply_data=[];
ply_data_index = 1;
map=[];
map_kd_tree=[];
map_quality=[];
last_pose = size(poses,1)
map_ct=[];
for i=1:pose_interval:last_pose
i
if check_stored_visual_feature(data_name_list{data_index+5}, dynamic_index, i, true, 'intensity') == 0
[img, x, y, z, c, elapsed_pre] = LoadSR_no_bpc(data_name_list{data_index+5}, 'gaussian', 0, dynamic_index, i, 1, 'int');
else
[frm, des, elapsed_sift, img, x, y, z, c, elapsed_pre] = load_visual_features(data_name_list{data_index+5}, dynamic_index, i, true, 'intensity');
end
confidence_threshold = floor(max(max(c))/2);
%confidence_threshold = 0; % for object_recognition
ct_start=tic;
unit_map=[];
for j=1:show_image_width
for k=1:show_image_height
col_idx = (j-1)*sample_interval + 1;
row_idx = (k-1)*sample_interval + 1;
%unit_pose = [x(row_idx,col_idx), y(row_idx, col_idx), z(row_idx, col_idx)];
unit_pose = [-x(row_idx,col_idx), z(row_idx, col_idx), y(row_idx, col_idx)];
unit_pose_distance = sqrt(sum(unit_pose.^2));
%if img(row_idx, col_idx) > 50
if unit_pose(3) <= -0.1 && img(row_idx, col_idx) < 200 % 50
if c(row_idx,col_idx) >= confidence_threshold && unit_pose_distance < distance_threshold_max && unit_pose_distance > distance_threshold_min
unit_pose_global = h{i}*[unit_pose, 1]';
[map_registration_flag, unit_map_quality, map]=check_map_registration(map_kd_tree, map, [unit_pose_global(1:3,1)', double(img(row_idx, col_idx))]);
if map_registration_flag
unit_color = [img(row_idx, col_idx),img(row_idx, col_idx),img(row_idx, col_idx)];
%fprintf(fd,'%f %f %f %d %d %d\n',unit_pose_global(1:3,1)', unit_color);
ply_data(ply_data_index,:) = [unit_pose_global(1:3,1)', double(unit_color)];
unit_map(ply_data_index,:) = [unit_pose_global(1:3,1)', double(unit_color(1)), 0, 1]; %[x,y,z,gray,radius,n]
ply_data_index = ply_data_index + 1;
else
if unit_map_quality >= 0
map_quality = [map_quality; unit_map_quality];
end
end
end
end
end
end
map = [map; unit_map];
%update k-d tree with new map
map_kd_tree = kdtree_build(map(:,1:4));
map_ct(i,1) = toc(ct_start);
end
% Write data
nply_data = size(ply_data,1)
element_vertex_n = sprintf('element vertex %d',nply_data);
ply_headers{4} = element_vertex_n;
fd = fopen(feature_ply_file_name, 'w');
for i=1:size(ply_headers,2)
fprintf(fd,'%s\n',ply_headers{i});
end
for i=1:nply_data
fprintf(fd,'%f %f %f %d %d %d\n',ply_data(i,:));
end
fclose(fd);
%save computational time
dlmwrite(feature_ct_file_name,map_ct,' ');
% Display map quality
mean_map_quality = mean(map_quality)
std_map_quality = std(map_quality)
end
function [registration_flag, map_quality, map] = check_map_registration(map_kd_tree, map, point)
registration_flag = true;
dist_th = 0.3; % [m]
gray_th = 255*0.01; % [gray level]
map_quality = -1;
if ~isempty(map_kd_tree)
% find closest point using kd-tree
temp_idx = kdtree_nearest_neighbor(map_kd_tree, point);
closest_point = map(temp_idx,:);
dist = norm(point(1:3) - closest_point(1:3));
gray_dist = abs(point(4) - closest_point(4));
% check distance less than threshold and determine the flag
if dist < dist_th && gray_dist < gray_th
registration_flag = false;
map(temp_idx,1:4) = sum([point;closest_point(1:4)*closest_point(6)])/(closest_point(6)+1);
map_quality = norm(map(temp_idx,1:3)-point(1:3));
map(temp_idx,5) = map_quality;
map(temp_idx,6) = map(temp_idx,6) + 1;
end
end
end
|
github
|
rising-turtle/slam_matlab-master
|
analyze_bundler_file.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/analyze_bundler_file.m
| 3,795 |
utf_8
|
d9454f3886b73ec5e3a3f2cb4db97ca5
|
% Analyze bundler file
%
% Author : Soonhac Hong ([email protected])
% Date : 2/13/13
function analyze_bundler_file()
graphslam_addpath;
addpath('D:\soonhac\Project\PNBD\SW\ASEE\slamtoolbox\slamToolbox_11_09_08\FrameTransforms\Rotations');
addpath('D:\soonhac\Project\PNBD\SW\ASEE\slamtoolbox\slamToolbox_11_09_08\DataManagement');
%file_name = 'data/ba/ros_sba/exp1_bus_door_straight_150_Replace_pose_feature_426_48_ros_sba.out';
file_name = 'results/ros_sba/exp1_bus_door_straight_150_Replace_pose_feature_426_48_ros_sba_result.out';
%file_name = 'data/ba/ros_sba/exp1_bus_door_straight_150_Replace_pose_feature_1614_101_ros_sba.out';
%file_name = 'results/ros_sba/exp1_bus_door_straight_150_Replace_pose_feature_1614_101_ros_sba_result.out';
%file_name = 'data/ba/ros_sba/sample_bundler_file.out';
%file_name = 'results/ros_sba/result.out';
%%load ROS-SBA output file
[camera_poses, camera_parameters, landmark_position, landmark_projection] = load_rossba(file_name);
show_camera_pose(camera_poses, true, 'no', 'no', 'on', 'k.-' );
hold on;
%plot3(landmark_position(:,1),landmark_position(:,2),landmark_position(:,3),'m.');
axis equal;
%% Anlayze landmark position
landmark_distance = sqrt(sum(landmark_position.^2,2));
%% Analyze projection
%cam = initialize_cam2();
cam = initialize_cam();
features_info=[];
u0 = cam.Cx;
v0 = cam.Cy;
%ros_sba_T = sr4k_p2T([0,0,0,pi/2,0,0]);
% temp_t = camera_poses(1,1:3)';
% temp_R = e2R(camera_poses(1,4:6)');
% reference_T = [temp_R, temp_t; 0 0 0 1];
%reference_T = inv(ros_sba_T)*reference_T;
for i=1:size(camera_poses,1)
temp_t = camera_poses(i,1:3)';
temp_R = e2R(camera_poses(i,4:6)');
%temp_R = eye(3);
% T = [temp_R, temp_t; 0 0 0 1];
% %T = bundler_T*T; % Convert Bundler frame to SBA frame.
% temp_t = T(1:3,4);
% temp_R = T(1:3,1:3);
projection_idx = find(landmark_projection(:,1) == (i-1));
estimated_uv=[];
measured_uv=[];
for j=1:size(projection_idx,1)-1
measured_uv(j,:) = landmark_projection(projection_idx(j), end-1:end);
measured_uv(j,:) = [measured_uv(j,1) + u0, -1 * measured_uv(j,2) + v0];
temp_p = landmark_position(landmark_projection(projection_idx(j), 2)+1,:)';
%temp_p = reference_T * [landmark_position(landmark_projection(projection_idx(j), 2)+1,:)'; 1]; % The reference frame of all landmark positions is the first camera coordinate !!!!!
%temp_p = bundler_T * [landmark_position(landmark_projection(projection_idx(j), 2)+1,:)'; 1]; % The reference frame of all landmark positions is the first camera coordinate !!!!!
estimated_uv(j,:) = hi_cartesian_test(temp_p(1:3), temp_t, temp_R, cam, features_info)';
%estimated_uv(j,:) = hi_cartesian_test(temp_p(1:3), [0;0;0], eye(3), cam, features_info )';
end
figure;
plot(estimated_uv(:,1), estimated_uv(:,2),'b+');
hold on;
plot(measured_uv(:,1),measured_uv(:,2),'ro');
legend('Estimated','Measured');
hold off;
projection_error = sqrt(sum((estimated_uv - measured_uv(:,1:2)).^2,2));
projection_error_mean_std(i,:) = [mean(projection_error), std(projection_error)];
end
figure;
errorbar(projection_error_mean_std(:,1), projection_error_mean_std(:,2),'b.');
end
function cam = initialize_cam2()
nRows = 480;
nCols = 640;
k1= 0; %-0.85016;
k2= 0; %0.56153;
cam.Cx = 320;
cam.Cy = 240;
cam.fd = 430;
cam.f = 430;
cam.k1 = k1;
cam.k2 = k2;
cam.nRows = nRows;
cam.nCols = nCols;
cam.d = 4/nCols*0.001;
cam.F = cam.f/cam.d;
% cam.f = f;
cam.dx = cam.d; %% dx and dy are actually slightly different dx =4/176 while dy = 3.17/144
cam.dy = cam.d;
cam.K = sparse( [ cam.fd 0 cam.Cx;
0 cam.fd cam.Cy;
0 0 1] );
cam.model = 'two_distortion_parameters';
end
|
github
|
rising-turtle/slam_matlab-master
|
generate_feature_index.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/generate_feature_index.m
| 2,078 |
utf_8
|
a2d297a437866c5e4fc478c009089384
|
% Generate feature index through all data
%
% Author : Soonhac Hong ([email protected])
% Date : 3/1/13
function [feature_points] = generate_feature_index(feature_points)
feature_index_list=[];
camera_index_list=[];
feature_index = 1;
for i=1:size(feature_points,1)
if feature_points(i,1) == 1 && feature_points(i,2) == 2 && feature_points(i,3) == 1
feature_index_list(i,1) = feature_index;
feature_index = feature_index + 1;
camera_index_list(i,1) = feature_points(i,3);
elseif feature_points(i,1) == 1 && feature_points(i,2) == 2 && feature_points(i,3) == 2
feature_index_list(i,1) = feature_index_list(i-feature_index+1,1);
camera_index_list(i,1) = feature_points(i,3);
else
[unit_feature_index, unit_camera_index]= find_similar_feature_index(feature_points(1:i,:), feature_index_list, camera_index_list);
% if i==411
% disp('debug');
% end
feature_index_list(i,1) = unit_feature_index;
if unit_feature_index > feature_index
feature_index = unit_feature_index;
end
camera_index_list(i,1) = unit_camera_index;
end
end
feature_points = [feature_points, feature_index_list, camera_index_list];
end
function [feature_index, camera_index] = find_similar_feature_index(feature_points, feature_index_list, camera_index_list)
if feature_points(end,3) == 1
camera_index = feature_points(end,1);
else
camera_index = feature_points(end,2);
end
same_camera_index_list = find(camera_index_list == camera_index);
% Find similarity by pixel index if feature points observed by same camera
if ~isempty(same_camera_index_list)
same_camera_index = find(feature_points(same_camera_index_list,7) == feature_points(end,7) & feature_points(same_camera_index_list,8) == feature_points(end,8));
if ~isempty(same_camera_index)
feature_index = feature_index_list(same_camera_index(1),1);
else
feature_index = max(feature_index_list) + 1;
end
else
feature_index = max(feature_index_list) + 1;
end
end
|
github
|
rising-turtle/slam_matlab-master
|
compensate_vro.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/compensate_vro.m
| 10,828 |
utf_8
|
75a630c580d4f470b446748ce1cce2d2
|
% Compensate the missing pose
% Author : Soonhac Hong ([email protected])
% Date : 4/19/20
function [new_f_index, new_t_pose, new_o_pose, new_fpts_index, new_pose_std] = compensate_vro(f_index, t_pose, o_pose, fpts_index, pose_std, pose_std_flag, cmp_option)
cur_index = f_index(1,1);
new_index = 1;
f_index_length = size(f_index,1);
new_fpts_index = [];
new_pose_std = [];
if strcmp(cmp_option,'Replace')
for i = 1:f_index_length
missing_index = [0 0];
if f_index(i,1) == cur_index
new_f_index(new_index,:) = f_index(i,:);
new_t_pose(new_index,:) = t_pose(i,:);
new_o_pose(new_index,:) = o_pose(i,:);
if pose_std_flag == 1
new_pose_std(new_index,:) = pose_std(i,:);
end
elseif f_index(i,1) == cur_index + 1
cur_index = f_index(i,1);
new_f_index(new_index,:) = f_index(i,:);
new_t_pose(new_index,:) = t_pose(i,:);
new_o_pose(new_index,:) = o_pose(i,:);
if pose_std_flag == 1
new_pose_std(new_index,:) = pose_std(i,:);
end
else % missing constraints
n_missing_index = f_index(i,1) - (cur_index + 1);
missing_index = missing_index + [n_missing_index n_missing_index];
% t_pose_compensation = (t_pose(i,:) + t_pose(i-1,:))/2;
% o_pose_compensation = (o_pose(i,:) + o_pose(i-1,:))/2;
% for k=1:n_missing_index
% new_f_index(new_index,:) = [cur_index+k, cur_index+k+1];
% new_t_pose(new_index,:) = t_pose_compensation;
% new_o_pose(new_index,:) = o_pose_compensation;
% new_index = new_index + 1
% end
cur_index = f_index(i,1) - missing_index(1);
new_f_index(new_index,:) = f_index(i,:) - missing_index;
new_t_pose(new_index,:) = t_pose(i,:);
new_o_pose(new_index,:) = o_pose(i,:);
if pose_std_flag == 1
new_pose_std(new_index,:) = pose_std(i,:);
end
f_index(i:f_index_length,:) = f_index(i:f_index_length,:) - repmat(missing_index, f_index_length-i+1, 1);
end
new_index = new_index + 1;
end
% check successiveness of second pose
index_diff = diff(new_f_index,1,1);
for i=1:size(index_diff,1);
if index_diff(i,2) ~= 1 && index_diff(i,1) == 0
new_f_index(i+1,2) = new_f_index(i,2) + 1;
end
end
% check consecutive of neighbor pose
index_diff = diff(new_f_index,1,1);
% Comment out for loop closure
for i=2:size(new_f_index,1);
if abs(new_f_index(i,1)-new_f_index(i,2)) >= 2 && abs(new_f_index(i,1)-new_f_index(i-1,1)) ~= 0
new_f_index(i,2) = new_f_index(i,1) + 1;
end
end
% Adjuste index of features
if ~isempty(fpts_index)
fpts_index_length = size(fpts_index,1);
new_index = 1;
cur_index = fpts_index(1,1);
for i = 1:fpts_index_length
if fpts_index(i,1) == cur_index
new_fpts_index(new_index,:) = fpts_index(i,:);
elseif fpts_index(i,1) == cur_index + 1
cur_index = fpts_index(i,1);
new_fpts_index(new_index,:) = fpts_index(i,:);
else % missing constraints
n_missing_index = fpts_index(i,1) - (cur_index + 1);
missing_index = [n_missing_index n_missing_index];
cur_index = fpts_index(i,1) - missing_index(1);
new_fpts_index(new_index,:) = fpts_index(i,:) - missing_index;
fpts_index(i:fpts_index_length,:) = fpts_index(i:fpts_index_length,:) - repmat(missing_index, fpts_index_length-i+1, 1);
end
new_index = new_index + 1;
end
% check successiveness of second pose
index_diff = diff(new_fpts_index,1,1);
for i=1:size(index_diff,1);
if index_diff(i,2) > 1 && index_diff(i,1) == 0
update_index = find(new_fpts_index(:,1) == new_fpts_index(i+1,1) & new_fpts_index(:,2) == new_fpts_index(i+1,2));
new_fpts_index(update_index,2) = new_fpts_index(i,2) + 1;
end
end
end
elseif strcmp(cmp_option,'Linear')
missing_index = 0;
for i = 1:f_index_length
if f_index(i,1) == cur_index
new_f_index(new_index,:) = f_index(i,:);
new_t_pose(new_index,:) = t_pose(i,:);
new_o_pose(new_index,:) = o_pose(i,:);
if pose_std_flag == 1
new_pose_std(new_index,:) = pose_std(i,:);
end
elseif f_index(i,1) == cur_index + 1
cur_index = f_index(i,1);
new_f_index(new_index,:) = f_index(i,:);
new_t_pose(new_index,:) = t_pose(i,:);
new_o_pose(new_index,:) = o_pose(i,:);
if pose_std_flag == 1
new_pose_std(new_index,:) = pose_std(i,:);
end
else % missing constraints
n_missing_index = f_index(i,1) - (cur_index + 1);
missing_index = missing_index + n_missing_index;
t_pose_compensation = (t_pose(i,:) + t_pose(i-1,:))/2;
o_pose_compensation = (o_pose(i,:) + o_pose(i-1,:))/2;
if pose_std_flag == 1
pose_std_compensation = (pose_std(i,:) + pose_std(i-1,:))/2;
end
for k=1:n_missing_index
new_f_index(new_index,:) = [cur_index+k, cur_index+k+1];
new_t_pose(new_index,:) = t_pose_compensation;
new_o_pose(new_index,:) = o_pose_compensation;
if pose_std_flag == 1
new_pose_std(new_index,:) = pose_std_compensation;
end
new_index = new_index + 1;
end
cur_index = f_index(i,1);
new_f_index(new_index,:) = f_index(i,:);
new_t_pose(new_index,:) = t_pose(i,:);
new_o_pose(new_index,:) = o_pose(i,:);
if pose_std_flag == 1
new_pose_std(new_index,:) = pose_std(i,:);
end
end
new_index = new_index + 1;
end
% check successiveness of second pose
tmp_new_f_index = new_f_index;
tmp_new_t_pose = new_t_pose;
tmp_new_o_pose = new_o_pose;
tmp_new_pose_std = new_pose_std;
new_f_index = [];
new_t_pose = [];
new_o_pose = [];
new_pose_std = [];
index_diff = diff(tmp_new_f_index,1,1);
new_index = 1;
for i=1:size(index_diff,1);
new_f_index(new_index,:) = tmp_new_f_index(i,:);
new_t_pose(new_index,:) = tmp_new_t_pose(i,:);
new_o_pose(new_index,:) = tmp_new_o_pose(i,:);
if pose_std_flag == 1
new_pose_std(new_index,:) = tmp_new_pose_std(i,:);
end
new_index = new_index + 1;
if index_diff(i,2) ~= 1 && index_diff(i,1) == 0
%if (index_diff(i,2) - index_diff(i,1)) >= 1
t_pose_compensation = (tmp_new_t_pose(i,:) + tmp_new_t_pose(i-1,:))/2;
o_pose_compensation = (tmp_new_o_pose(i,:) + tmp_new_o_pose(i-1,:))/2;
if pose_std_flag == 1
pose_std_compensation = (tmp_new_pose_std(i,:) + tmp_new_pose_std(i-1,:))/2;
end
for k=1:index_diff(i,2)-1
new_f_index(new_index,:) = [tmp_new_f_index(i,1), tmp_new_f_index(i,2)+k];
new_t_pose(new_index,:) = t_pose_compensation;
new_o_pose(new_index,:) = o_pose_compensation;
if pose_std_flag == 1
new_pose_std(new_index,:) = pose_std_compensation;
end
new_index = new_index + 1;
end
end
end
new_f_index(new_index,:) = tmp_new_f_index(i+1,:);
new_t_pose(new_index,:) = tmp_new_t_pose(i+1,:);
new_o_pose(new_index,:) = tmp_new_o_pose(i+1,:);
if pose_std_flag == 1
new_pose_std(new_index,:) = tmp_new_pose_std(i+1,:);
end
% check consecutive of neighbor pose
tmp_new_f_index = new_f_index;
tmp_new_t_pose = new_t_pose;
tmp_new_o_pose = new_o_pose;
if pose_std_flag == 1
tmp_new_pose_std = new_pose_std;
end
new_f_index = tmp_new_f_index(1,:);
new_t_pose = tmp_new_t_pose(1,:);
new_o_pose = tmp_new_o_pose(1,:);
if pose_std_flag == 1
new_pose_std = tmp_new_pose_std(1,:);
end
new_index = 2;
for i=2:size(tmp_new_f_index,1);
if abs(tmp_new_f_index(i,1) - tmp_new_f_index(i,2)) >= 2 && abs(tmp_new_f_index(i,1) - tmp_new_f_index(i-1,1)) ~= 0
%new_f_index(i,2) = new_f_index(i,1) + 1;
t_pose_compensation = (tmp_new_t_pose(i,:) + tmp_new_t_pose(i-1,:))/2;
o_pose_compensation = (tmp_new_o_pose(i,:) + tmp_new_o_pose(i-1,:))/2;
if pose_std_flag == 1
pose_std_compensation = (tmp_new_pose_std(i,:) + tmp_new_pose_std(i-1,:))/2;
end
interval = abs(tmp_new_f_index(i,1)-tmp_new_f_index(i,2)) - 1;
for k=1:interval
new_f_index(new_index,:) = [tmp_new_f_index(i,1), tmp_new_f_index(i,1)+k];
new_t_pose(new_index,:) = t_pose_compensation;
new_o_pose(new_index,:) = o_pose_compensation;
if pose_std_flag == 1
new_pose_std(new_index,:) = pose_std_compensation;
end
new_index = new_index + 1;
end
end
new_f_index(new_index,:) = tmp_new_f_index(i,:);
new_t_pose(new_index,:) = tmp_new_t_pose(i,:);
new_o_pose(new_index,:) = tmp_new_o_pose(i,:);
if pose_std_flag == 1
new_pose_std(new_index,:) = tmp_new_pose_std(i,:);
end
new_index = new_index + 1;
end
end
% Reduce the constraints
% tmp_new_f_index = new_f_index;
% tmp_new_t_pose = new_t_pose;
% tmp_new_o_pose = new_o_pose;
% new_f_index = tmp_new_f_index(1,:);
% new_t_pose = tmp_new_t_pose(1,:);
% new_o_pose = tmp_new_o_pose(1,:);
% new_index = 2;
%
% min_edge = 1;
% unit_edge_cnt = 1;
%
% for i=2:size(tmp_new_f_index,1)
% if tmp_new_f_index(i,1) ~= tmp_new_f_index(i-1,1)
% new_f_index(new_index,:) = tmp_new_f_index(i,:);
% new_t_pose(new_index,:) = tmp_new_t_pose(i,:);
% new_o_pose(new_index,:) = tmp_new_o_pose(i,:);
% new_index = new_index + 1;
% unit_edge_cnt = 1;
% elseif unit_edge_cnt < min_edge
% new_f_index(new_index,:) = tmp_new_f_index(i,:);
% new_t_pose(new_index,:) = tmp_new_t_pose(i,:);
% new_o_pose(new_index,:) = tmp_new_o_pose(i,:);
% new_index = new_index + 1;
% unit_edge_cnt = unit_edge_cnt + 1;
% end
% end
end
|
github
|
rising-turtle/slam_matlab-master
|
run_pose_graph_optimization_v2.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/run_pose_graph_optimization_v2.m
| 1,322 |
utf_8
|
d28a8104e89fa69ed5a18db6501d5382
|
% Run pose graph optimization
%
% Author : Soonhac Hong ([email protected])
% History :
% 3/26/14 : Created
function [result, graph, initial, h_global, location_file_index, location_info_history]=run_pose_graph_optimization_v2(vro_result, vro_pose_std, graph, initial, h_global, dis, location_flag, location_file_index, location_info_history, lcd_found)
import gtsam.*
t = gtsam.Point3(0, 0, 0);
if isempty(h_global)
h_global = get_global_transformation_single('smart_cane');
rot = h_global(1:3,1:3);
R = gtsam.Rot3(rot);
origin= gtsam.Pose3(R,t);
initial.insert(0,origin);
end
pgc_t=tic;
[graph,initial] = construct_pose_graph(vro_result, vro_pose_std, graph, initial);
first = initial.at(0);
pgc_ct =toc(pgc_t)
graph.add(NonlinearEqualityPose3(0, first));
gtsam_t=tic;
optimizer = LevenbergMarquardtOptimizer(graph, initial);
result = optimizer.optimizeSafely();
gtsam_ct =toc(gtsam_t)
% Show the results in the plot and generate location information
if dis==true
[location_file_index, location_info_history]=plot_graph_initial_result_v2(initial, result, location_flag, location_file_index, location_info_history, lcd_found);
elseif location_flag == true && lcd_found == true
[location_info, location_file_index] = generate_location_info_v2(result,[], location_file_index);
end
end
|
github
|
rising-turtle/slam_matlab-master
|
convert_o2p_adaptive.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/convert_o2p_adaptive.m
| 5,408 |
utf_8
|
0c1a9b2d0a86628ce2495a472c1f80ae
|
% Conver the odometery to the global pose
% t_pose [mm]
% o_pose [degree]
function [pose_index, e_t_pose e_o_pose fpts_h] = convert_o2p_adaptive(data_index, dynamic_index, f_index, t_pose, o_pose, feature_points, dense_index, sparse_interval)
% Generate the vertex using VRO
disp('Generate the vertex using VRO.');
previous_index = -1;
if size(f_index,1) > 0
vro_index = 1;
if dense_index == 1
for i=1:size(f_index,1)
relative_max_index = max(find(f_index(:,1) == f_index(i,1)))
if f_index(i,1) > previous_index && (i == relative_max_index || check_orientation(o_pose(i,:)) == 1) %abs(f_index(i,1) - f_index(i,2)) == 1
vro_t_pose(vro_index,:) = t_pose(i,:);
vro_o_pose(vro_index,:) = o_pose(i,:);
pose_index(vro_index) = f_index(i,1);
vro_index = vro_index + 1;
previous_index = f_index(i,2)-1;
end
end
else % Sparse pose
% Generate the vertex using VRO with the only maximum constraints
%sparse_interval = 2;
next_index = min(f_index(:,1));
max_index = max(f_index(:,1));
while next_index <= max_index
temp_index=find(f_index(:,1) == next_index);
if isempty(temp_index)
next_index = next_index + 1;
else
%max_temp_index = max(temp_index);
if length(temp_index) > sparse_interval
max_temp_index = min(temp_index)+sparse_interval;
else
max_temp_index = max(temp_index);
end
vro_t_pose(vro_index,:) = t_pose(max_temp_index,:);
vro_o_pose(vro_index,:) = o_pose(max_temp_index,:);
pose_index(vro_index) = f_index(max_temp_index,1);
vro_index = vro_index + 1;
next_index = f_index(max_temp_index,2);
end
end
end
else
vro_t_pose = t_pose;
vro_o_pose = o_pose;
end
% Generate the vertex from feature points at each pose
disp('Generate the vertex from feature points at each pose.');
if ~isempty(feature_points)
fpts = cell(size(pose_index,2),1);
%current_pose_index = 1;
for i = 1:size(feature_points,1)
if feature_points(i,3) == 1
current_pose_index = feature_points(i,1);
else%pose_index
current_pose_index = feature_points(i,2);
end
cell_index = find(pose_index == current_pose_index);
if ~isempty(cell_index)
fpts{cell_index,1} = [fpts{cell_index,1}; current_pose_index feature_points(i,4:6)];
end
end
% Eliminate duplicated points
fpts = eliminate_duplication(fpts);
end
% Calculate Homogenous Transformation in 3D
disp('Calculate Homogenous Transformation in 3D.');
e_t_pose = zeros(size(vro_t_pose,1), 4);
h_global = get_global_transformation(data_index, dynamic_index);
for i = 1:size(vro_t_pose,1)
h{i} = [euler_to_rot(vro_o_pose(i,1), vro_o_pose(i,2), vro_o_pose(i,3)) vro_t_pose(i,:)'; 0 0 0 1];
end
disp('Convert poses w.r.t the global frame.');
fpts_h = cell(size(pose_index,2),1);
for k = 2:size(e_t_pose,1)
for j = k-1 : -1: 1
if j == k-1
temp_pose = h{j}*[ 0 0 0 1]';
if ~isempty(feature_points)
unit_data = fpts{k,1};
for f = 1:size(unit_data,1)
temp_fpts(:,f) = h{j}*[unit_data(f,2:4) 1]';
end
unit_data=[];
end
temp_h = h{j};
else
temp_pose = h{j}*temp_pose;
if ~isempty(feature_points)
unit_data = fpts{k,1};
for f = 1:size(unit_data,1)
temp_fpts(:,f) = h{j}* temp_fpts(:,f);
end
unit_data=[];
end
temp_h = h{j}*temp_h;
end
end
e_t_pose(k,:) = [h_global * temp_pose]';
if ~isempty(feature_points)
unit_data_h=[];
for f = 1:size(fpts{k,1},1)
unit_data_h(f,:) = [h_global* temp_fpts(:,f)]';
end
%temp_data = fpts{k,1};
fpts_h{k,1}= [fpts{k,1}(:,1) unit_data_h];
end
temp_h = h_global * temp_h;
[e_o_pose(k,1) e_o_pose(k,2) e_o_pose(k,3)] = rot_to_euler(temp_h(1:3,1:3));
end
e_t_pose(1,:) = [h_global * [0 0 0 1]']';
if ~isempty(feature_points)
for f = 1:size(fpts{1,1},1)
fpts_h{1,1}(f,:) = [fpts{1,1}(f,1) [h_global* [fpts{1,1}(f,2:4) 1]']'];
end
else
fpts_h={};
end
temp_h = h_global;
[e_o_pose(1,1) e_o_pose(1,2) e_o_pose(1,3)] = rot_to_euler(temp_h(1:3,1:3));
end
function [new_ftps] = eliminate_duplication(fpts)
new_fpts = cell(size(fpts));
for i = 1:size(fpts,1)
unit_cell = fpts{i,1};
new_unit_cell = [];
for j=1:size(unit_cell,1)
[duplication_index, duplication_flag] = check_duplication(new_unit_cell, unit_cell(j,:));
if duplication_flag == 0
new_unit_cell = [new_unit_cell; unit_cell(j,:)];
end
end
new_ftps{i,1} = new_unit_cell;
end
end
function [orientation_dominant] = check_orientation(unit_o_pose)
o_threshold = 3.0 ; %0.24; %[degree]
idx = find(abs(unit_o_pose) > o_threshold);
if isempty(idx)
orientation_dominant = 0
else
orientation_dominant = 1
end
end
|
github
|
rising-turtle/slam_matlab-master
|
kinect_relative_pose_gt.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/kinect_relative_pose_gt.m
| 2,205 |
utf_8
|
99fd8e96694f4792837b8a2ac1a0d976
|
% Compute relative pose b/w two images from ground truth in kinect_tum
%
% Author : Soonhac Hong ([email protected])
% Date : 12/20/12
function kinect_relative_pose_gt(dir_index, depth_file_index_1, depth_file_index_2)
format LONGG;
addpath('D:\soonhac\Project\PNBD\SW\ASEE\slamtoolbox\slamToolbox_11_09_08\FrameTransforms\Rotations');
addpath('D:\soonhac\Project\PNBD\SW\ASEE\slamtoolbox\slamToolbox_11_09_08\DataManagement');
% Get time stamp of color files
fisrt_depth_file_time_stamps = get_timestamp_kinect_tum(dir_index,depth_file_index_1)
second_depth_file_time_stamps = get_timestamp_kinect_tum(dir_index,depth_file_index_2)
% Load ground truth
[gt rgbdslam rtime] = Load_kinect_gt(dir_index);
% Find first time stamp
%first_timestamp = get_timestamp_kinect_tum(dir_index,file_index_1);
gt_first_index = find(gt(:,1) > fisrt_depth_file_time_stamps, 1);
if gt(gt_first_index,1) - fisrt_depth_file_time_stamps > fisrt_depth_file_time_stamps - gt(gt_first_index-1,1)
gt_first_index = gt_first_index - 1;
end
% Find second time stamp
%second_timestamp = get_timestamp_kinect_tum(dir_index,file_index_2);
gt_second_index = find(gt(:,1) > second_depth_file_time_stamps, 1);
if gt(gt_second_index,1) - second_depth_file_time_stamps > second_depth_file_time_stamps - gt(gt_second_index-1,1)
gt_second_index = gt_second_index - 1;
end
gt_first_tf =[q2R([gt(gt_first_index,5:8)]) [gt(gt_first_index,2:4)]'; 0 0 0 1];
gt_second_tf =[q2R([gt(gt_second_index,5:8)]) [gt(gt_second_index,2:4)]'; 0 0 0 1];
rgbdslam_first_tf =[q2R([rgbdslam(gt_first_index,5:8)]) [rgbdslam(gt_first_index,2:4)]'; 0 0 0 1];
rgbdslam_second_tf =[q2R([rgbdslam(gt_second_index,5:8)]) [rgbdslam(gt_second_index,2:4)]'; 0 0 0 1];
gt_relative_tf = inv(gt_first_tf) * gt_second_tf;
gt_translation = [gt_relative_tf(1,4) gt_relative_tf(2,4) gt_relative_tf(3,4)];
gt_euler = R2e(gt_relative_tf(1:3,1:3));
rgbdslam_relative_tf = inv(rgbdslam_first_tf) * rgbdslam_second_tf;
rgbdslam_translation = [rgbdslam_relative_tf(1,4) rgbdslam_relative_tf(2,4) rgbdslam_relative_tf(3,4)];
rgbdslam_euler = R2e(rgbdslam_relative_tf(1:3,1:3));
figure;
plot(gt_translation,'g*');
hold on;
plot(rgbdslam_translation,'rd');
hold off;
end
|
github
|
rising-turtle/slam_matlab-master
|
plot3_gtsam.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/plot3_gtsam.m
| 1,456 |
utf_8
|
5e536d0602db5d998c7d610d2e8c0e25
|
% This file was modified from gtsam/toolbox/+gtsam/plot3DTreajectory.m
function plot3_gtsam(values,linespec,frames,scale,marginals)
% plot3DTrajectory plots a 3D trajectory
% plot3DTrajectory(values,linespec,frames,scale,marginals)
if ~exist('scale','var') || isempty(scale), scale=1; end
if ~exist('frames','var'), scale=[]; end
import gtsam.*
haveMarginals = exist('marginals', 'var');
keys = KeyVector(values.keys);
holdstate = ishold;
hold on
% Plot poses and covariance matrices
lastIndex = [];
for i = 0:keys.size-1
key = keys.at(i);
x = values.at(key);
if isa(x, 'gtsam.Pose3')
if ~isempty(lastIndex)
% Draw line from last pose then covariance ellipse on top of
% last pose.
lastKey = keys.at(lastIndex);
lastPose = values.at(lastKey);
plot3([ x.x; lastPose.x ], [ x.y; lastPose.y ], [ x.z; lastPose.z ], linespec);
if haveMarginals
P = marginals.marginalCovariance(lastKey);
else
P = [];
end
%gtsam.plotPose3(lastPose, P, scale);
end
lastIndex = i;
end
end
% Draw final pose
if ~isempty(lastIndex)
lastKey = keys.at(lastIndex);
lastPose = values.at(lastKey);
if haveMarginals
P = marginals.marginalCovariance(lastKey);
else
P = [];
end
%gtsam.plotPose3(lastPose, P, scale);
end
if ~holdstate
hold off
end
grid;
end
|
github
|
rising-turtle/slam_matlab-master
|
construct_pose_graph.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/construct_pose_graph.m
| 2,474 |
utf_8
|
1d5fc7f3f5bebe8be1aa17d55d8bca3a
|
% Construct a pose graph for Graph SLAM
%
% Author : Soonhac Hong ([email protected])
% History :
% 3/26/14 : Created
function [graph,initial] = construct_pose_graph(vro_result, vro_pose_std, graph, initial)
import gtsam.*
n=size(vro_result,1);
N=n;
pose_size=max(vro_result(:,2));
first_index_offset = 1;
step_threshold = 2;
sigma_level = 1;
node_count=0;
edge_count=0;
for i=1:n
e=vro_result(i,:);
i1=e(1)-first_index_offset;
i2=e(2)-first_index_offset;
t = gtsam.Point3(e(6), e(7), e(8));
R = gtsam.Rot3.Ypr(e(5), e(3), e(4));
dpose = gtsam.Pose3(R,t);
pose_std = vro_pose_std(i,3:end)'; %[ry rx rz tx ty tz]
pose_std =[pose_std(4:6); pose_std(2); pose_std(1); pose_std(3)];
if (abs(i2-i1)==1) || check_reliability_static7(pose_std, sigma_level) == 1
pose_noise_model = noiseModel.Diagonal.Sigmas(pose_std);
graph.add(BetweenFactorPose3(i1, i2, dpose, pose_noise_model));
edge_count = edge_count + 1;
if abs(i2-i1)==1
if i2>i1
initial.insert(i2,initial.at(i1).compose(dpose));
else
initial.insert(i1,initial.at(i2).compose(dpose.inverse));
end
end
end
end
end
function [reliability_flag] = check_reliability_static6(pose_std, sigma_level)
reliability_flag = 1;
translation_sigma = 0.007; %14; %[m]
orientation_sigma_rx = 0.12*pi/180;%0.57*pi/180; %[radian]
orientation_sigma_ry = 0.12*pi/180;%0.11*pi/180; %[radian]
orientation_sigma_rz = 0.12*pi/180;%0.30*pi/180; %[radian]
std_pose_std = [orientation_sigma_ry,orientation_sigma_rx,orientation_sigma_rz,translation_sigma,translation_sigma,translation_sigma];
for i=1:6
if pose_std(i) > (sigma_level * std_pose_std(i))
reliability_flag = 0;
break;
end
end
end
function [reliability_flag] = check_reliability_static7(pose_std, sigma_level)
reliability_flag = 1;
translation_sigma = 0.056; %14; %[m]
orientation_sigma_rx = 0.96*pi/180;%0.57*pi/180; %[radian]
orientation_sigma_ry = 0.96*pi/180;%0.11*pi/180; %[radian]
orientation_sigma_rz = 0.96*pi/180;%0.30*pi/180; %[radian]
std_pose_std = [orientation_sigma_ry,orientation_sigma_rx,orientation_sigma_rz,translation_sigma,translation_sigma,translation_sigma];
for i=1:6
if pose_std(i) > (sigma_level * std_pose_std(i))
reliability_flag = 0;
break;
end
end
end
|
github
|
rising-turtle/slam_matlab-master
|
plot_graph_initial_result.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/plot_graph_initial_result.m
| 2,568 |
utf_8
|
48f6abd5162643f1f8d96e7f6012a8ec
|
% Show plots of inital pose and optimized pose
%
% Author : Soonhac Hong ([email protected])
% History :
% 3/27/14 : Created
%
function [location_file_index, location_info_history] = plot_graph_initial_result(gtsam_pose_initial, gtsam_pose_result, location_flag, location_file_index, location_info_history)
import gtsam.*
keys = KeyVector(gtsam_pose_initial.keys);
initial_max_index = keys.size-1;
for i=0:initial_max_index
key = keys.at(i);
x = gtsam_pose_initial.at(key);
T = x.matrix();
[ry, rx, rz] = rot_to_euler(T(1:3,1:3));
plot_xyz_initial(i+1,:)=[x.x x.y x.z];
end
keys = KeyVector(gtsam_pose_result.keys);
initial_max_index = keys.size-1;
for i=0:initial_max_index
key = keys.at(i);
x = gtsam_pose_result.at(key);
T = x.matrix();
[ry, rx, rz] = rot_to_euler(T(1:3,1:3));
plot_xyz_result(i+1,:)=[x.x x.y x.z];
end
location_info=[];
if location_flag == true
[location_info, location_file_index] = generate_location_info([], plot_xyz_result, location_file_index); % for finding rooms
end
figure(1);
subplot(1,2,2);
plot(plot_xyz_result(:,1),plot_xyz_result(:,2),'r-', 'LineWidth', 2);
hold on;
plot(plot_xyz_result(1,1),plot_xyz_result(1,2),'ko', 'LineWidth', 3,'MarkerSize', 3);
text(plot_xyz_result(1,1)-1.5,plot_xyz_result(1,2)-1.5,'Start','Color',[0 0 0]);
xlabel('X');ylabel('Y');
% Modify size of x in the graph
%xlim([-7 15]); % for etas523_exp2
%xlim([-15 5]); % for Amir's exp1
xlim([-10 10]); % for etas523_exp2_lefthallway
ylim([-5 15]);
% Display location information
text_offset=[1,0; -6.5,0; -1,2; -1,-1.5; 0,-1.5]; % etas etas 523 exp2
%text_offset=[-6,0; 1,0; 1,0; 1,0; 0,2]; % etas etas 523 lefthallway
if ~isempty(location_info_history)
for i=1:size(location_info_history,1)
text_xy=[location_info_history{i,1}+text_offset(i,1),location_info_history{i,2}+text_offset(i,2)];
text(text_xy(1), text_xy(2),location_info_history(i,3),'Color',[0 0 1]);
plot(location_info_history{i,1}, location_info_history{i,2},'bd', 'LineWidth', 2, 'MarkerSize', 5);
end
end
if ~isempty(location_info)
text_xy=[plot_xyz_result(end,1)+text_offset(location_file_index-1,1), plot_xyz_result(end,2)+text_offset(location_file_index-1,2)];
text(text_xy(1), text_xy(2),location_info, 'Color',[0 0 1]);
location_info_history(location_file_index-1,:) ={plot_xyz_result(end,1),plot_xyz_result(end,2),location_info};
plot(plot_xyz_result(end,1), plot_xyz_result(end,2),'bd', 'LineWidth', 2, 'MarkerSize', 5);
end
hold off;
grid;
legend('PGO');
axis equal;
drawnow;
end
|
github
|
rising-turtle/slam_matlab-master
|
pose_optimization_example_2d.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/pose_optimization_example_2d.m
| 4,277 |
utf_8
|
75b4771c3e2e5f8defec05bbcc23317e
|
% 2D Example of Pose Graph Optimization
% Data : 4/12/12
% Author : Soonhac Hong ([email protected])
function pose_optimization_example_2d()
% 2D case
pose_data=[1 1 0 0 0; 1 2 2 0 0 ; 2 3 2 0 pi/2; 3 4 2 0 pi/2; 4 5 2 0 pi/2 ; 5 2 2 0 pi/2]; % [first pose, second pose, constraint[x,y,theta]]
xinit = [0.5; 0.0; 0.2; 2.3; 0.1; -0.2; 4.1; 0.1; pi/2; 4.0; 2.0; pi; 2.1; 2.1; -pi/2]
motion_noise = [0.3;0.3;0.1];
variable_size = (size(pose_data,2)-2); % size of each variable [x, y, r]
pose_num = size(unique(pose_data(:,1:2)),1)*variable_size; % first position is not optimized.
Omega = zeros(pose_num,pose_num);
Xi = zeros(pose_num,1);
Odometry=zeros(pose_num,1);
%Odometry = zeros(pose_num,1);
% Fill the elements of Omega and Xi
% Initial point
for i=1:variable_size
Omega(i,i) = 1;
Xi(i) = pose_data(1,i+2);
Odometry(i) = pose_data(1,i+2);
end
for i=2:size(pose_data,1)
unit_data = pose_data(i,:);
current_index = unit_data(1);
next_index = unit_data(2);
movement = unit_data(3:2+variable_size);
previous_theta = pose_data(i-1,2+variable_size);
% Adjust index according to the size of each variable
if current_index ~= 1
current_index = (current_index - 1) * variable_size + 1;
end
if next_index ~= -1
next_index = (next_index - 1) * variable_size + 1;
end
for j=0:variable_size-1
% Fill diagonal elements of Omega
switch j
case 0
diagonal_factor = cos(movement(3));
offdiagonal_factor = 1;
case 1
diagonal_factor = cos(movement(3));
offdiagonal_factor = -1;
case 2
diagonal_factor = 1;
offdiagonal_factor = -1;
end
Omega(current_index+j,current_index+j) = Omega(current_index+j,current_index+j) + diagonal_factor*motion_noise(j+1);
Omega(next_index+j,next_index+j) = Omega(next_index+j,next_index+j) + 1/motion_noise(j+1);
% Fill Off-diagonal elements of Omega
Omega(current_index+j,next_index+j) = Omega(current_index+j,next_index+j) + (-1)/motion_noise(j+1);
Omega(next_index+j,current_index+j) = Omega(next_index+j,current_index+j) + (-1)*diagonal_factor/motion_noise(j+1);
if j <= 1
Omega(current_index+j,current_index+j+offdiagonal_factor) = Omega(current_index+j,next_index+j+offdiagonal_factor) + (-1)*offdiagonal_factor*sin(movement(3))/motion_noise(j+1);
Omega(next_index+j,current_index+j+offdiagonal_factor) = Omega(next_index+j,current_index+j+offdiagonal_factor) + offdiagonal_factor*sin(movement(3))/motion_noise(j+1);
end
% Fill Xi
Xi(current_index+j) = Xi(current_index+j) + (-1)*movement(j+1)/motion_noise(j+1);
Xi(next_index+j) = Xi(next_index+j) + movement(j+1)/motion_noise(j+1);
end
% Update Odometry
if abs(current_index - next_index) == 3
translation=[0 0 1]';
for t=i:-1:2
unit_movement=pose_data(t,3:5);
translation = [cos(unit_movement(3)) -sin(unit_movement(3)) unit_movement(1); sin(unit_movement(3)) cos(unit_movement(3)) unit_movement(2); 0 0 1]*translation;
end
%translation = Odometry(current_index:current_index+1) + movement(1:2)';
Odometry(next_index:next_index+1) = translation(1:2);
orientation = Odometry(current_index+2) + movement(3);
if orientation > pi*2
orientation = orientation - pi*2;
end
Odometry(next_index+2) = orientation;
end
end
%Omega
%Xi
%mu = Omega^-1 * Xi
% Using LM
xdata = Omega;
ydata = Xi;
%myfun = @(x,xdata)Rot(x(1:3))*xdata+repmat(x(4:6),1,length(xdata));
myfun = @(x,xdata)xdata*x(1:size(xdata,1));
options = optimset('Algorithm', 'levenberg-marquardt');
%x = lsqcurvefit(myfun, zeros(6,1), p, q, [], [], options);
x = lsqcurvefit(myfun, xinit, xdata, ydata, [], [], options)
x_mat = vec2mat(x,3);
Odometry_mat = vec2mat(Odometry,3);
xinit_mat = vec2mat(xinit,3);
plot(xinit_mat(:,1), xinit_mat(:,2) ,'bd-');
hold on;
plot(Odometry_mat(:,1), Odometry_mat(:,2) ,'gd-');
plot(x_mat(:,1), x_mat(:,2) ,'ro-');
hold off;
legend('Initial','Odometry','Optimized');
%legend('Initial','Optimized');
end
|
github
|
rising-turtle/slam_matlab-master
|
plot_g2o.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/plot_g2o.m
| 3,805 |
utf_8
|
7e0fcebfeb6496b1ba99eb889d50d7af
|
% Plot the result of graph SLAM, g2o
%
% Author : Soonhac Hong ([email protected])
% Date : 2/22/12
function plot_g2o(org_file_name, opt_file_name, feature_flag)
[ org_poses org_edges org_fpts_poses org_fpts_edges] = load_graph_g2o(org_file_name);
[ opt_poses opt_edges opt_fpts_poses opt_fpts_edges] = load_graph_g2o(opt_file_name);
% Show the pose
start_index = 1;
end_index = min(size(org_poses,1), size(opt_poses)); %80;
if size(org_poses,2) == 3 % SE2
figure;
%plot(org_poses(start_index:end_index,1), org_poses(start_index:end_index,2),'b.-','LineWidth',2);
plot(org_poses(:,1), org_poses(:,2),'b.-','LineWidth',2);
if ~isempty(org_fpts_poses) && feature_flag == 1
hold on;
plot(org_fpts_poses(:,1), org_fpts_poses(:,2), 'bd');
end
%xlabel('X [m]');
%ylabel('Y [m]');
%grid;
%figure;
hold on;
%plot(opt_poses(start_index:end_index,1), opt_poses(start_index:end_index,2),'r.-','LineWidth',2);
plot(opt_poses(:,1), opt_poses(:,2),'r.-','LineWidth',2);
if ~isempty(opt_fpts_poses) && feature_flag == 1
plot(opt_fpts_poses(:,1), opt_fpts_poses(:,2), 'rd');
end
%plot_groundtruth();
xlabel('X [m]');
ylabel('Y [m]');
grid;
%legend('vro','vro fpts', 'g2o', 'g2o fpts','eGT');
legend('vro','g2o');
%legend('vro','g2o');
%xlim([-0.1 0.2]);
%ylim([-0.1 0.4]);
axis equal;
hold off;
elseif size(org_poses,2) == 7 % SE3:QUAT
figure; plot3(org_poses(start_index:end_index,1), org_poses(start_index:end_index,2), org_poses(start_index:end_index,3), 'b-');
%xlabel('X [m]');
%ylabel('Y [m]');
%zlabel('Z [m]');
%grid;
%figure;
hold on;
plot3(opt_poses(start_index:end_index,1), opt_poses(start_index:end_index,2), opt_poses(start_index:end_index,3), 'r-');
xlabel('X [m]');
ylabel('Y [m]');
zlabel('Z [m]');
grid;
axis equal;
hold off;
end
end
function plot_groundtruth()
gt_x = [0 0 2135 2135 0];
gt_y = [0 1220 1220 0 0];
gt_x = gt_x / 1000; % [mm] -> [m]
gt_y = gt_y / 1000; % [mm] -> [m]
plot(gt_x,gt_y,'g-','LineWidth',2);
end
% function [poses edges] = load_graph(file_name)
% fid = fopen(file_name);
% data = textscan(fid, '%s %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f'); % 2D format
% fclose(fid);
%
% % Convert data
% % Pose
% data_name = data{1};
% data_name_list = {'VERTEX_SE2','EDGE_SE2','VERTEX_SE3:QUAT','EDGE_SE3:QUAT'};
% vertex_index = 1;
% edge_index = 1;
% for i = 1 : size(data_name,1)
% if strcmp(data_name{i}, data_name_list{1}) % VERTEX_SE2
% unit_data =[];
% for j=3:5
% unit_data = [unit_data data{j}(i)];
% end
% poses(vertex_index,:) = unit_data;
% vertex_index = vertex_index + 1;
% elseif strcmp(data_name{i}, data_name_list{2}) % EDGE_SE2
% unit_data =[];
% for j=2:12
% unit_data = [unit_data data{j}(i)];
% end
% edges(edge_index,:) = unit_data;
% edge_index = edge_index + 1;
% elseif strcmp(data_name{i}, data_name_list{3}) % VERTEX_SE3:QUAT
% unit_data =[];
% for j=3:9
% unit_data = [unit_data data{j}(i)];
% end
% poses(vertex_index,:) = unit_data;
% vertex_index = vertex_index + 1;
% elseif strcmp(data_name{i}, data_name_list{4}) % EDGE_SE3:QUAT
% unit_data =[];
% for j=2:31
% unit_data = [unit_data data{j}(i)];
% end
% edges(edge_index,:) = unit_data;
% edge_index = edge_index + 1;
% end
%
% end
% end
|
github
|
rising-turtle/slam_matlab-master
|
generate_location_info_v2.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/generate_location_info_v2.m
| 1,803 |
utf_8
|
636d5ba368cd57e0ff5b309c1a1ab788
|
% Generate the location infomation for Text-To-Speech and plots
%
% Author : Soonhac Hong ([email protected])
% History :
% 8/27/14 : Created
%
function [location_info, location_file_index] = generate_location_info_v2(gtsam_pose_result, plot_xyz_result, location_file_index)
import gtsam.*
location_info=[];
% Extract the optimized pose from gtsam data structure if the optimized pose is not available
if isempty(plot_xyz_result)
keys = KeyVector(gtsam_pose_result.keys);
% isp_fd = fopen(file_name_pose, 'w');
initial_max_index = keys.size-1;
for i=0:initial_max_index
key = keys.at(i);
x = gtsam_pose_result.at(key);
T = x.matrix();
[ry, rx, rz] = rot_to_euler(T(1:3,1:3));
plot_xyz_result(i+1,:)=[x.x x.y x.z];
end
end
if size(plot_xyz_result,1) < 50 % Avoid adjacency of first n steps with respect to the starting point
return;
end
%% Generate the location information
location_info_list={'You return to the starting point;'};
short_location_info_list={'Start'};
location_info_xy=[0, 0]; % unit : [m]
location_threshold = 0.4; % [m]
location_info_file_name = sprintf('C:\\SC-DATA-TRANSFER\\location_info_%d.txt', location_file_index);
current_position = [plot_xyz_result(end,1),plot_xyz_result(end,2),plot_xyz_result(end,3)]; % [x, y, z]
for i=location_file_index:size(location_info_xy,1)
distance = norm(current_position(1, 1:2) - location_info_xy(i,:));
if distance < location_threshold
location_info = short_location_info_list{i};
% Write the location information to the location_info.txt
fd = fopen(location_info_file_name, 'w');
fprintf(fd,'%s',location_info_list{i});
fclose(fd);
location_file_index = location_file_index + 1;
break;
end
end
end
|
github
|
rising-turtle/slam_matlab-master
|
convert_isp2ply.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/convert_isp2ply.m
| 1,477 |
utf_8
|
156009055f11b3be476e3513612330c7
|
% Conver isp to ply for meshlab
%
% Author : Soonhac Hong ([email protected])
% Date : 4/1/13
function convert_isp2ply(isp_file_name, file_index, dynamic_index, isgframe, vro_name)
[poses] = load_graph_isp(isp_file_name);
isp_finder = strfind(isp_file_name,'isp');
if(isempty(isp_finder))
ply_file_name=strrep(isp_file_name, 'opt','ply')
if strcmp(vro_name, 'vro') == 1
color=[255, 0, 0]; % red
else
color=[0, 0, 255]; % blue
end
is_opt_file = 1;
else
ply_file_name=strrep(isp_file_name, 'isp','ply')
color=[0, 255, 0]; % green
is_opt_file = 0;
end
fd = fopen(ply_file_name, 'w');
% Write header
element_vertex_n = sprintf('element vertex %d',size(poses,1));
ply_headers={'ply','format ascii 1.0','comment Created with XYZRGB_to_PLY', element_vertex_n, 'property float x', 'property float y','property float z','property uchar red','property uchar green','property uchar blue','end_header'};
for i=1:size(ply_headers,2)
fprintf(fd,'%s\n',ply_headers{i});
end
% Write data
for j=1:size(poses,1)
fprintf(fd,'%f %f %f %d %d %d\n',poses(j,1:3), color);
end
fclose(fd)
if(is_opt_file == 1)
convert_pc2ply(ply_headers, ply_file_name, poses, file_index, dynamic_index, isgframe);
%convert_pc2ply_map_registration(ply_headers, ply_file_name, poses, file_index, dynamic_index, isgframe);
%convert_pc2ply_map_registration_v2(ply_headers, ply_file_name, poses, file_index, dynamic_index, isgframe);
end
end
|
github
|
rising-turtle/slam_matlab-master
|
gslam_plot_comparison.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/gslam_plot_comparison.m
| 10,628 |
utf_8
|
ab0569c4aae179bf191b516d557a3428
|
% plot comparison of g2o and isam
function gslam_plot_comparison()
% file_index = 5; % 5 = whitecane
% dynamic_index = 16;
file_index = 9; % 5 = whitecane 6 = etas
dynamic_index = 1; % 15:square_500, 16:square_700, 17:square_swing
etas_nFrame_list = [979 1479 979 1979 1889]; %[3rd_straight, 3rd_swing, 4th_straigth, 4th_swing, 5th_straight, 5th_swing]
loops_nFrame_list = [0 1359 2498 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2669];
kinect_tum_nFrame_list = [0 98 0 0];
m_nFrame_list = [0 5414 0 729];
compensation_option ={'Replace','Linear'};
feature_flag = 0;
etas_vro_size_list = [1260 0 1665 0 9490 0]; %1974
loops_vro_size_list = [0 13098 12476 0 0 0 0 0 0 0 0 0 0 0 0 0 0 26613];
kinect_tum_vro_size_list = [0 98 0 0];
vro_name_list={'vro','vro_icp','vro_icp_ch','icp'};
vro_name_index = 3;
vro_name=vro_name_list{vro_name_index};
loops2_nFrame_list = [582 398 398 832]; %2498 578
loops2_vro_size_list = [2891 0 0 5754]; %4145
loops2_pose_size_list = [582 0 0 832];
switch file_index
case 5
nFrame = m_nFrame_list(dynamic_index - 14);
vro_size = 6992; %5382; %46; %5365; %5169;
case 6
nFrame = etas_nFrame_list(dynamic_index); %289; 5414; %46; %5468; %296; %46; %86; %580; %3920;
vro_size = 1260; %etas_vro_size_list(dynamic_index); %1951; 1942; %5382; %46; %5365; %5169;
case 7
nFrame = loops_nFrame_list(dynamic_index);
vro_size = loops_vro_size_list(dynamic_index);
case 8
nFrame = kinect_tum_nFrame_list(dynamic_index);
vro_size = kinect_tum_vro_size_list(dynamic_index);
case 9
nFrame = loops2_nFrame_list(dynamic_index);
vro_size = loops2_vro_size_list(dynamic_index);
pose_size = loops2_pose_size_list(dynamic_index);
end
if feature_flag == 1
feature_pose_name = 'pose_feature';
else
feature_pose_name = 'pose_zero';
end
[g2o_result_dir_name, isam_result_dir_name, vro_dir_name, dynamic_dir_name, toro_dir_name] = get_file_names(file_index, dynamic_index);
% switch dynamic_index
% case 15
% vro_size = 3616;
% case 16
% vro_size = 5169;
% case 11
% vro_size = 84;
% end
%vro_file_name = sprintf('%s%s_%s_%s_%d.g2o', vro_dir_name, dynamic_dir_name, compensation_option{2}, feature_pose_name, vro_size);
%g2o_result_file_name = sprintf('%s%s_%d.opt', g2o_result_dir_name, dynamic_dir_name, vro_size);
%isam_result_file_name = sprintf('%s%s_%s_%s_zero_%d_isam.opt', isam_result_dir_name, dynamic_dir_name, compensation_option{1}, feature_pose_name, vro_size);
%isp_file_name = sprintf('%s%s_%s_%s_zero_%d.isp', vro_dir_name, dynamic_dir_name, compensation_option{2}, feature_pose_name, 1974);
%isp_file_name_plus1 = sprintf('%s%s_%s_%s_zero_%d_329.isp', vro_dir_name, dynamic_dir_name, compensation_option{1}, feature_pose_name, vro_size);
%isp_file_name_plus2 = sprintf('%s%s_%s_%s_zero_%d_610.isp', vro_dir_name, dynamic_dir_name, compensation_option{1}, feature_pose_name, vro_size);
%isp_file_name_plus3 = sprintf('%s%s_%s_%s_zero_%d_325.isp', vro_dir_name, dynamic_dir_name, compensation_option{1}, feature_pose_name, vro_size);
%isp_file_name_plus1_minlength = sprintf('%s%s_%s_%s_zero_%d.isp', vro_dir_name, dynamic_dir_name, compensation_option{1}, feature_pose_name, 118);
%isp_file_name_plus1_minlength2 = sprintf('%s%s_%s_%s_zero_%d.isp', vro_dir_name, dynamic_dir_name, compensation_option{1}, feature_pose_name, 108);
gtsam_isp_file_name_icp = sprintf('%s%s_%s_%s_%d_%s_gtsam.isp', vro_dir_name, dynamic_dir_name, compensation_option{1}, feature_pose_name, 2873, vro_name_list{2}); %5763
gtsam_isp_file_name_icp_ch = sprintf('%s%s_%s_%s_%d_%s_gtsam.isp', vro_dir_name, dynamic_dir_name, compensation_option{1}, feature_pose_name, vro_size, vro_name_list{3});
icp_ct_file_name = sprintf('%s%s_%s_%s_%d_%d_%s_icp.ct', vro_dir_name, dynamic_dir_name, compensation_option{1}, feature_pose_name, 2873, pose_size, vro_name_list{2}); %5763
icp_ct_file_name_ch = sprintf('%s%s_%s_%s_%d_%d_%s_icp.ct', vro_dir_name, dynamic_dir_name, compensation_option{1}, feature_pose_name, vro_size, pose_size, vro_name_list{3});
%gtsam_result_file_name = sprintf('%s%s_%s_%s_zero_%d_gtsam.opt', isam_result_dir_name, dynamic_dir_name, compensation_option{2}, feature_pose_name, vro_size);
%isp_file_name2 = sprintf('%s%s_%s_%s_%d.isp', vro_dir_name, dynamic_dir_name, compensation_option{2}, feature_pose_name, vro_size);
%isam_result2_file_name = sprintf('%s%s_%s_%s_%d_isam.opt', isam_result_dir_name, dynamic_dir_name, compensation_option{2}, feature_pose_name, vro_size);
%isp2_file_name = sprintf('%s%s_%s_%s_%d.isp', vro_dir_name, dynamic_dir_name, compensation_option{2}, feature_pose_name, vro_size);
%toro_result_file_name = sprintf('%s%s_%s_%s_%d-treeopt-final.graph', toro_dir_name, dynamic_dir_name, compensation_option{1}, feature_pose_name, vro_size);
%toro_vro_file_name = sprintf('%s%s_%s_%s_%d.graph', toro_dir_name, dynamic_dir_name, compensation_option{1}, feature_pose_name, vro_size);
%toro_result2_file_name = sprintf('%s%s_%s_%s_%d-treeopt-final.graph', toro_dir_name, dynamic_dir_name, compensation_option{2}, feature_pose_name, vro_size);
%isp_file_name3 = sprintf('%s%s_%s_%s_zero_%d.isp', vro_dir_name, dynamic_dir_name, compensation_option{2}, feature_pose_name, vro_size);
%isam_result3_file_name = sprintf('%s%s_%s_%s_zero_%d_isam.opt', isam_result_dir_name, dynamic_dir_name, compensation_option{2}, feature_pose_name, vro_size);
%[vro_poses vro_edges] = load_graph_g2o(vro_file_name);
%[g2o_poses g2o_edges] = load_graph_g2o(g2o_result_file_name);
[vro_poses] = load_graph_isp(gtsam_isp_file_name_icp);
[vro_poses2] = load_graph_isp(gtsam_isp_file_name_icp_ch);
[vro_ct] = load(icp_ct_file_name);
[vro_ct_ch] = load(icp_ct_file_name_ch);
[ct_mean, ct_median, ct_std] = compute_ct_statistics(vro_ct);
[ct_mean_ch, ct_median_ch, ct_std_ch] = compute_ct_statistics(vro_ct_ch);
figure;
%errorbar(ct_mean, ct_std,'k','LineWidth',2); %,'k*-'
plot(ct_median,'gd-','LineWidth',2);
hold on;
%errorbar(ct_mean_ch, ct_std_ch,'g','LineWidth',2);
plot(ct_median_ch,'k*-','LineWidth',2);
legend('ICP','Fast ICP');
x_interval=1:5;
set(gca,'XTick',x_interval,'FontSize',12,'FontWeight','bold');
h_xlabel = get(gca,'XLabel');
set(h_xlabel,'FontSize',12,'FontWeight','bold');
h_ylabel = get(gca,'YLabel');
set(h_ylabel,'FontSize',12,'FontWeight','bold');
xlabel('Step');
ylabel('Computational Time [sec]');
ylim([2.5 6]);
grid;
hold off;
%[vro_poses] = load_graph_isp(isp_file_name);
%[vro_poses2] = load_graph_isp(isp_file_name_plus1);
%[vro_poses] = load_graph_isp(isp_file_name_plus2);
%[vro_poses3] = load_graph_isp(isp_file_name_plus3);
%[vro_poses3] = load_graph_isp(isp_file_name_plus1_minlength);
%[vro_poses4] = load_graph_isp(isp_file_name_plus1_minlength2);
%[vro_poses2] = load_graph_isp(isp_file_name2);
%[gtsam_poses] = load_graph_isp(gtsam_result_file_name);
%[isam_poses2] = load_graph_isam(isam_result2_file_name);
%[toro_poses toro_edges toro_fpts_poses toro_fpts_edges] = load_graph_toro(toro_result_file_name);
%[toro_poses2 toro_edges2 toro_fpts_poses2 toro_fpts_edges2] = load_graph_toro(toro_result2_file_name);
%[vro_poses3] = load_graph_isp(isp_file_name3);
%[isam_poses3] = load_graph_isam(isam_result3_file_name);
%isam_poses(133:size(isam_poses,1),:) = isam_poses(133:size(isam_poses,1),:) + repmat(isam_poses(132,:), size(isam_poses,1)-132, 1);
%gt_name = sprintf('../data/dynamic/%s/d1_gt.dat',dynamic_dir_name);
%gt = {load(gt_name)};
%gt_x = cumsum(diff(gt{1,1}(6:20,1)))/1000;
start_index = 1;
%end_index = min([size(vro_poses,1) size(toro_poses,1) size(isam_poses,1)]); %80;
end_index = min([size(vro_poses,1) size(vro_poses2,1)]); %80;
if size(vro_poses,2) >= 3 % SE3
figure;
%plot3(vro_poses(start_index:end_index,1), vro_poses(start_index:end_index,2), vro_poses(start_index:end_index,3),'b-','LineWidth',2);
%plot3(vro_poses(:,1), vro_poses(:,2), vro_poses(:,3),'b-','LineWidth',2);
plot(vro_poses(:,1), vro_poses(:,2), 'g-','LineWidth',2);
hold on;
%plot3(vro_poses2(start_index:end_index,1), vro_poses2(start_index:end_index,2), vro_poses2(start_index:end_index,3), 'm-','LineWidth',2);
%plot3(vro_poses3(:,1), vro_poses3(:,2), vro_poses3(:,3), 'g-','LineWidth',2);
plot(vro_poses2(:,1), vro_poses2(:,2), 'k-','LineWidth',2);
%plot3(vro_poses2(:,1), vro_poses2(:,2), vro_poses2(:,3), 'm-','LineWidth',2);
%plot3(vro_poses3(1:98,1), vro_poses3(1:98,2), vro_poses3(1:98,3), 'r-','LineWidth',2);
%plot3(vro_poses4(1:98,1), vro_poses4(1:98,2), vro_poses4(1:98,3), 'g-','LineWidth',2);
%hold on;
%plot(vro_poses3(:,1), vro_poses3(:,2),'k-','LineWidth',2);
%plot(vro_poses(start_index:end_index,1), 'b*-','LineWidth',2);
%xlabel('X [m]');
%ylabel('Y [m]');
%grid;[e_t_pose e_o_pose] = convert_o2p(f_index, t_pose, o_pose)
%figure;
%hold on;
%plot3(toro_poses(:,1), toro_poses(:,2), toro_poses(:,3),'g-','LineWidth',2);
%plot(toro_poses2(:,1), toro_poses2(:,2),'r-','LineWidth',2);
%plot3(isam_poses(:,1), isam_poses(:,2), isam_poses(:,3),'m-','LineWidth',2);
%plot3(gtsam_poses(:,1), gtsam_poses(:,2), gtsam_poses(:,3),'b-','LineWidth',2);
%plot(isam_poses2(:,1), isam_poses2(:,2),'c-','LineWidth',2);
%plot(isam_poses3(:,1), isam_poses3(:,2),'m-','LineWidth',2);
%plot(g2o_poses(start_index:end_index,1), 'ro-','LineWidth',2);
%plot(isam_poses(start_index:end_index,1), 'md-','LineWidth',2);
%plot(gt_x, 'g+-','LineWidth',2);
%plot_groundtruth();
%plot_gt_etas();
xlabel('X [m]');
ylabel('Y [m]');
zlabel('Z [m]');
grid;
legend('ICP','Fast ICP'); %,'iSAM_R','GT_e');
%legend('VRO_R','VRO_L','GT_e');
%legend('vro','Toro_R','Toro_L','isam_R','isam_L','gt_e');
%legend('VRO Dense','VRO Sparse','VRO Adaptive'); %,'iSAM_R','GT_e');
%legend('VRO','TORO_R','TORO_L','GT_e');
%legend('VRO','iSAM_R','iSAM_L','GT_e');
%legend('VRO_L','TORO_L','iSAM_L','GT_e');
%legend('VRO_L','VRO_Zeor','TORO_L','iSAM_L','iSAM_zero','GT_e');
%xlim([-0.1 0.6]);
%ylim([-0.1 0.6]);
set(gca,'FontSize',12,'FontWeight','bold');
h_xlabel = get(gca,'XLabel');
set(h_xlabel,'FontSize',12,'FontWeight','bold');
h_ylabel = get(gca,'YLabel');
set(h_ylabel,'FontSize',12,'FontWeight','bold');
h_zlabel = get(gca,'ZLabel');
set(h_zlabel,'FontSize',12,'FontWeight','bold');
hold off;
axis equal;
end
end
function plot_groundtruth()
gt_x = [0 0 2135 2135 0];
gt_y = [0 1220 1220 0 0];
gt_x = gt_x / 1000; % [mm] -> [m]
gt_y = gt_y / 1000; % [mm] -> [m]
plot(gt_x,gt_y,'g-','LineWidth',2);
end
|
github
|
rising-turtle/slam_matlab-master
|
get_global_transformation_single.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/get_global_transformation_single.m
| 517 |
utf_8
|
f3c4ebbcb15dd51122b659ed76b7d39e
|
% Get global transformation for each data set with the first frame
%
% Author : Soonhac Hong ([email protected])
% Date : 3/27/14
function [h_global] = get_global_transformation_single(data_name)
addpath('D:\Soonhac\SW\slamtoolbox\slamToolbox_11_09_08\FrameTransforms\Rotations');
% rx, ry, rz : [degree]
% tx, ty, tz : [mm]
rx=0; ry=0; rz=0; tx=0; ty=0; tz=0;
euler=plane_fit_to_data_single(data_name);
rx=euler(1);
ry=euler(2);
rz=euler(3);
h_global = [euler_to_rot(rz, rx, ry) [tx ty tz]'; 0 0 0 1];
end
|
github
|
rising-turtle/slam_matlab-master
|
plot_toro.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/plot_toro.m
| 5,122 |
utf_8
|
b208119b20ac256877227aa33a31c7eb
|
% Plot the result of graph SLAM, g2o
%
% Author : Soonhac Hong ([email protected])
% Date : 2/22/12
function plot_toro(org_file_name, opt_file_name, feature_flag)
[ org_poses org_edges org_fpts_poses org_fpts_edges] = load_graph_toro(org_file_name);
[ opt_poses opt_edges opt_fpts_poses opt_fpts_edges] = load_graph_toro(opt_file_name);
% Show the pose
start_index = 1;
end_index = min(size(org_poses,1), size(opt_poses,1)); %80;
if size(org_poses,2) == 3 % SE2
figure;
%plot(org_poses(start_index:end_index,1), org_poses(start_index:end_index,2),'b.-','LineWidth',2);
plot(org_poses(:,1), org_poses(:,2),'b.-','LineWidth',2);
if ~isempty(org_fpts_poses) && feature_flag == 1
hold on;
plot(org_fpts_poses(:,1), org_fpts_poses(:,2), 'bd');
end
%xlabel('X [m]');
%ylabel('Y [m]');
%grid;
%figure;
hold on;
%plot(opt_poses(start_index:end_index,1), opt_poses(start_index:end_index,2),'r.-','LineWidth',2);
plot(opt_poses(:,1), opt_poses(:,2),'r.-','LineWidth',2);
if ~isempty(opt_fpts_poses) && feature_flag == 1
plot(opt_fpts_poses(:,1), opt_fpts_poses(:,2), 'rd');
end
%plot_groundtruth();
xlabel('X [m]');
ylabel('Y [m]');
grid;
%legend('vro','vro fpts', 'g2o', 'g2o fpts','eGT');
legend('vro','toro');
%legend('vro','g2o');
%xlim([-0.1 0.2]);
%ylim([-0.1 0.4]);
hold off;
elseif size(org_poses,2) == 6 % 3D
figure; plot3(org_poses(:,1), org_poses(:,2), org_poses(:,3), 'b:', 'LineWidth',2);
%xlabel('X [m]');
%ylabel('Y [m]');
%zlabel('Z [m]');
%grid;
%figure;
hold on;
plot3(opt_poses(:,1), opt_poses(:,2), opt_poses(:,3), 'r-.', 'LineWidth',2);
%plot_groundtruth_3D();
%plot_gt_etas();
xlabel('X [m]');
ylabel('Y [m]');
zlabel('Z [m]');
%legend('vro','TORO','GT_e');
legend('vro','TORO','GT_e');
grid;
axis equal;
hold off;
show_errors(org_poses, opt_poses);
end
end
function show_errors(org_poses, opt_poses)
% Show last position
last_org = org_poses(end,:)
last_opt = opt_poses(end,:)
distance_org = sqrt(sum(org_poses(end,1:3).^2))
distance_opt = sqrt(sum(opt_poses(end,1:3).^2))
z_rmse_org = sqrt(mean(org_poses(:,3).^2))
z_rmse_opt = sqrt(mean(opt_poses(:,3).^2))
z_me_org = mean(abs(org_poses(:,3)))
z_me_opt = mean(abs(opt_poses(:,3)))
end
function plot_groundtruth()
gt_x = [0 0 2135 2135 0];
gt_y = [0 1220 1220 0 0];
gt_x = gt_x / 1000; % [mm] -> [m]
gt_y = gt_y / 1000; % [mm] -> [m]
plot(gt_x,gt_y,'g-','LineWidth',2);
end
function plot_groundtruth_3D()
% inch2mm = 304.8; % 1 cube = 12 inch = 304.8 mm
% gt_x = [0 0 13*inch2mm 13*inch2mm 0];
% gt_y = [-1*inch2mm 6*inch2mm 6*inch2mm -1*inch2mm -1*inch2mm];
% %gt_x = [0 0 2135 2135 0];
% %gt_y = [0 1220 1220 0 0];
% gt_z = [0 0 0 0 0];
% gt_x = gt_x / 1000; % [mm] -> [m]
% gt_y = gt_y / 1000; % [mm] -> [m]
% gt_z = gt_z / 1000; % [mm] -> [m]
inch2m = 0.0254; % 1 inch = 0.0254 m
gt_x = [0 0 150 910 965 965 910 50 0 0];
gt_y = [0 24 172.5 172.5 122.5 -122.5 -162.5 -162.5 -24 0];
gt_x = [gt_x 0 0 60 60+138 60+138+40 60+138+40 60+138 60 0 0];
gt_y = [gt_y 0 24 38.5+40 38.5+40 38.5 -38.5 -38.5-40 -38.5-40 -24 0];
gt_x = gt_x * inch2m;
gt_y = gt_y * inch2m;
gt_z = zeros(length(gt_x),1);
plot3(gt_x,gt_y,gt_z,'g-','LineWidth',2);
end
% function [poses edges] = load_graph(file_name)
% fid = fopen(file_name);
% data = textscan(fid, '%s %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f'); % 2D format
% fclose(fid);
%
% % Convert data
% % Pose
% data_name = data{1};
% data_name_list = {'VERTEX_SE2','EDGE_SE2','VERTEX_SE3:QUAT','EDGE_SE3:QUAT'};
% vertex_index = 1;
% edge_index = 1;
% for i = 1 : size(data_name,1)
% if strcmp(data_name{i}, data_name_list{1}) % VERTEX_SE2
% unit_data =[];
% for j=3:5
% unit_data = [unit_data data{j}(i)];
% end
% poses(vertex_index,:) = unit_data;
% vertex_index = vertex_index + 1;
% elseif strcmp(data_name{i}, data_name_list{2}) % EDGE_SE2
% unit_data =[];
% for j=2:12
% unit_data = [unit_data data{j}(i)];
% end
% edges(edge_index,:) = unit_data;
% edge_index = edge_index + 1;
% elseif strcmp(data_name{i}, data_name_list{3}) % VERTEX_SE3:QUAT
% unit_data =[];
% for j=3:9
% unit_data = [unit_data data{j}(i)];
% end
% poses(vertex_index,:) = unit_data;
% vertex_index = vertex_index + 1;
% elseif strcmp(data_name{i}, data_name_list{4}) % EDGE_SE3:QUAT
% unit_data =[];
% for j=2:31
% unit_data = [unit_data data{j}(i)];
% end
% edges(edge_index,:) = unit_data;
% edge_index = edge_index + 1;
% end
%
% end
% end
|
github
|
rising-turtle/slam_matlab-master
|
check_duplication.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/check_duplication.m
| 618 |
utf_8
|
a80eb6b224d62975425c043d78830e0d
|
% Check single data duplication in data set
% Author : Soonhac Hong ([email protected])
% Date : 3/21/12
function [duplication_index, duplication_flag] = check_duplication(data_set, data)
duplication_index = 0;
duplication_flag = 0;
distance_threshold = 41; % [mm]; typical absolute accuracy + 3 * typical repeatibility of SR4000 = 20 + 3 * 7 = 41
for i=1:size(data_set,1)
distance = sqrt(sum((data_set(i,2:4)-data(2:4)).^2));
if distance <= distance_threshold
duplication_flag = 1;
duplication_index = data_set(i,1);
break;
end
end
end
|
github
|
rising-turtle/slam_matlab-master
|
load_graph_isam.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/load_graph_isam.m
| 1,876 |
utf_8
|
4136a74c81fcb67ebfd3066b75051e13
|
% Load the graph of isam
function [poses] = load_graph_isam(file_name)
fid = fopen(file_name);
% Convert data
% Pose
data_name_list = {'Pose2d_Node','Pose2d_Pose2d_Factor','Pose3d_Node'};
vertex_index = 1;
edge_index = 1;
while ~feof(fid) %for i = 1 : size(data_name,1)
header = textscan(fid, '%s',1); % 2D format
data_name = header{1};
if strcmp(data_name, data_name_list{1}) % VERTEX_SE2
data = textscan(fid, '%f (%f,%f,%f)');
unit_data =[];
for j=1:4
unit_data = [unit_data data{j}];
end
f_index(vertex_index, :) = unit_data(1);
t_pose(vertex_index,:) = unit_data(2:3);
o_pose(vertex_index,:) = unit_data(4);
vertex_index = vertex_index + 1;
% elseif strcmp(data_name{i}, data_name_list{2}) % EDGE_SE2
% unit_data =[];
% for j=4:12
% unit_data= [unit_data data{j}(i)];
% end
% edges(edge_index,:) = unit_data;
% edge_index = edge_index + 1;
elseif strcmp(data_name, data_name_list{3}) % VERTEX_SE3:QUAT
data = textscan(fid, '%f (%f,%f,%f;%f,%f,%f)');
unit_data =[];
for j=2:7
unit_data = [unit_data data{j}];
end
t_pose(vertex_index,:) = unit_data(1:3);
o_pose(vertex_index,:) = unit_data(4:6);
vertex_index = vertex_index + 1;
% elseif strcmp(data_name{i}, data_name_list{4}) % EDGE_SE3:QUAT
% unit_data =[];
% for j=2:31
% unit_data = [unit_data data{j}(i)];
% end
% edges(edge_index,:) = unit_data;
% edge_index = edge_index + 1;
end
end
fclose(fid);
poses = [t_pose o_pose];
end
|
github
|
rising-turtle/slam_matlab-master
|
plot_icp_ct.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/plot_icp_ct.m
| 560 |
utf_8
|
0caa9c7fd9f9245a8cd86987ab0e4426
|
% Plot the result of computational time of icp
%
% Author : Soonhac Hong ([email protected])
% Date : 1/22/13
function plot_icp_ct(file_name)
ct_data = load(file_name);
ct = ct_data(:,3);
figure;
plot(ct,'b.');
ylabel('Computatoinal Time [sec]')
set(gca,'FontSize',12,'FontWeight','bold');
h_ylabel = get(gca,'YLabel');
set(h_ylabel,'FontSize',12,'FontWeight','bold');
[ct_mean, ct_std] = compute_ct_statistics(ct_data);
figure;
errorbar(ct_mean, ct_std,'r');
icp_ct_max = max(ct)
icp_ct_mean = mean(ct)
icp_ct_median = median(ct)
icp_ct_min = min(ct)
end
|
github
|
rising-turtle/slam_matlab-master
|
show_camera_pose.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/show_camera_pose.m
| 2,064 |
utf_8
|
7cb0989642cf704f0bc47753fbb09e02
|
% Show camera pose with coordinates
%
% Author : Soonhac Hong ([email protected])
% Date : 2/11/13
function show_camera_pose(poses, is_new_figure, isVRO, isROS_SBA, isCoordinateOn, symbol_color )
% poses : n x 6D (n : number of poses)
if is_new_figure
figure;
end
if strcmp(isROS_SBA, 'ros_sba')
new_poses=[];
for i=1:size(poses,1)
T=[e2R([poses(i,4),poses(i,5),poses(i,6)]) [poses(i,1),poses(i,2),poses(i,3)]'; 0 0 0 1];
sba_T= inv(T);
new_poses(i,:)=[sba_T(1:3,4)' R2e(sba_T(1:3,1:3))'];
end
poses = new_poses;
end
plot3(poses(:,1), poses(:,2), poses(:,3),symbol_color, 'LineWidth', 2);
if strcmp(isCoordinateOn, 'on')
hold on;
%draw coordinates at each pose
scale_unit = 0.05 * max(max(poses(:,1:3))); %[m]
o_unit = [0 0 0 1]';
x_unit = [scale_unit 0 0 1]';
y_unit = [0 scale_unit 0 1]';
z_unit = [0 0 scale_unit 1]';
axis_unit = [o_unit, x_unit, o_unit, y_unit, o_unit, z_unit];
for i=1:size(poses,1)
if strcmp(isVRO, 'vro')
T = p2T(poses(i,:));
else
rot = e2R(poses(i,4:6));
T = [rot poses(i,1:3)'; 0 0 0 1];
end
hat_axis_unit = T*axis_unit;
plot3(hat_axis_unit(1,1:2), hat_axis_unit(2,1:2), hat_axis_unit(3,1:2),'r-','LineWidth',2); %x axis
plot3(hat_axis_unit(1,3:4), hat_axis_unit(2,3:4), hat_axis_unit(3,3:4),'g-','LineWidth',2); %y axis
plot3(hat_axis_unit(1,5:6), hat_axis_unit(2,5:6), hat_axis_unit(3,5:6),'b-','LineWidth',2); %z axis
end
hold off;
end
axis equal;
grid;
xlabel('X');
ylabel('Y');
zlabel('Z');
end
function T = p2T(x)
Rx = @(a)[1 0 0;
0 cos(a) -sin(a);
0 sin(a) cos(a)];
Ry = @(b)[cos(b) 0 sin(b);
0 1 0;
-sin(b) 0 cos(b)];
Rz = @(c)[cos(c) -sin(c) 0;
sin(c) cos(c) 0;
0 0 1];
Rot = @(x)Rz(x(3))*Rx(x(1))*Ry(x(2)); % SR4000 project; see euler_to_rot.m
T = [Rot(x(4:6)) [x(1), x(2), x(3)]'; 0 0 0 1];
end
|
github
|
rising-turtle/slam_matlab-master
|
run_gslam.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/run_gslam.m
| 7,387 |
utf_8
|
78718a0ddde1865e57f3e6723649741f
|
% Conver the results of VRO to the vertex and edges for g2o
%
% Author : Soonhac Hong ([email protected])
% Date : 2/22/12
function [vro_size] = run_gslam(data_index, dynamic_index, nFrame, g2o_file_name, g2o_dir_name, feature_flag, index_interval, dis)
if nargin < 8
dis = 0;
end
% Load the result of VRO
disp('Load the result from VRO.');
[f_index, t_pose, o_pose, feature_points] = load_vro(data_index,dynamic_index, nFrame, feature_flag); % t_pose[mm], o_pose[degree]
if isempty(feature_points)
feature_flag = 0;
end
vro_size = size(t_pose,1)
% Convert the odometry and feature points to the global poses
disp('Convert the VRO and feauture points w.r.t the glabal frame.');
[pose_index, e_t_pose e_o_pose e_ftps] = convert_o2p(f_index, t_pose, o_pose, feature_points);
% graph slam
% Set the dimension of the filter
N = 20;
num_landmarks = 0;
%dim = 2 * (N + num_landmarks); % x1,y1,x2,y2.....
dim = 3 * (N + num_landmarks); % x1,y1,z1,x2,y2,z2.....
motion_noise = 34; % [mm]
o_noise = 0.24 * 2;
% make the constraint information matrix and vector
Omega = zeros(dim, dim);
Omega(1,1) = 1.0;
Omega(2,2) = 1.0;
Omega(3,3) = 1.0;
Xi = zeros(dim, 1);
Xi(1:3) = 0; % first frame locate at (0,0)
data_index = min(find(f_index(:,2) == N));
%data_index = min(find(f_index(:,1) == (N-1)));
% process the data
for k = 0 : data_index-1
% n is the index of the robot pose in the matrix/vector
%n = k * 2;
%n1 = (f_index(k+1,1)-1)*2;
%n2 = (f_index(k+1,2)-1)*2;
%if abs(f_index(k+1,1) - f_index(k+1,2)) >= 2 && abs(f_index(k+1,1) - f_index(k+1,2)) <= 3
if (max(abs((t_pose(k+1,1) - t_pose(k+1,2)))) >= motion_noise || max(abs((o_pose(k+1,1) - o_pose(k+1,2)))) >= o_noise) ...
&& (max(abs((t_pose(k+1,1) - t_pose(k+1,2)))) <= motion_noise*3 || max(abs((o_pose(k+1,1) - o_pose(k+1,2)))) <= o_noise*3)
n = k * 3;
n1 = (f_index(k+1,1)-1)*3
n2 = (f_index(k+1,2)-1)*3
%measurement = feature_points(k+1,:); %data[k][0]
%motion = t_pose(k+1,1:2); %data[k][1]
motion = t_pose(k+1,:); %data[k][1]
R = euler_to_rot(o_pose(k+1,1), o_pose(k+1,2), o_pose(k+1,3));
% integrate the measurements
% for i = 1: size(feature_points)
% m is the index of the landmark coordinate in the matrix/vector
%m = 2 * (N + measurement[i][0])
%update the information maxtrix/vector based on the measurement
% for b in range(2):
% Omega.value[n+b][n+b] += 1.0 / measurement_noise
% Omega.value[m+b][m+b] += 1.0 / measurement_noise
% Omega.value[n+b][m+b] += -1.0 / measurement_noise
% Omega.value[m+b][n+b] += -1.0 / measurement_noise
% Xi.value[n+b][0] += -measurement[i][1+b] / measurement_noise
% Xi.value[m+b][0] += measurement[i][1+b] / measurement_noise
%update the information maxtrix/vector based on the robot motion
% for b = 1:2
% %Omega(n+b, n+b) = Omega(n+b, n+b) + 1.0 / motion_noise; % (x1,x1), (y1,y1), (x2,x2), (y2,y2)
% Omega(n1+b, n1+b) = Omega(n1+b, n1+b) + 1.0 / motion_noise; %(x1,x1), (y1,y1), (x2,x2), (y2,y2)
% Omega(n2+b, n2+b) = Omega(n2+b, n2+b) + 1.0 / motion_noise; %(x1,x1), (y1,y1), (x2,x2), (y2,y2)
% end
%
% for b = 1:2
% Omega(n1+b,n2+b) = Omega(n1+b, n2+b) + (-1.0) / motion_noise;
% Omega(n2+b, n1+b) = Omega(n2+b, n1+b) + (-1.0) / motion_noise;
% Xi(n1+b)= Xi(n1+b) - motion(b) / motion_noise;
% Xi(n2+b)= Xi(n2+b) + motion(b) / motion_noise;
% end
for b = 1:3
%Omega(n+b, n+b) = Omega(n+b, n+b) + 1.0 / motion_noise; % (x1,x1), (y1,y1), (x2,x2), (y2,y2)
Omega(n1+b, n1+b) = Omega(n1+b, n1+b) + 1.0 / motion_noise; %(x1,x1), (y1,y1), (z1,z1)
% Omega(n2+b, n2+b) = Omega(n2+b, n2+b) + (1.0) / motion_noise; %(x2,x2), (y2,y2), (z2, z2)
Omega(n2+b, n2+b) = Omega(n2+b, n2+b) + (1.0)*R(b,b) / motion_noise; %(x2,x2), (y2,y2), (z2, z2)
d_index = [1 2 3];
b_index = find(d_index == b);
d_index(b_index)=[];
for d = 1:size(d_index,2)
Omega(n2+b, n2+d_index(d)) = Omega(n2+b, n2+d_index(d)) + (1.0)*R(b,d_index(d)) / motion_noise;
end
end
for b = 1:3
% Omega(n1+b,n2+b) = Omega(n1+b, n2+b) + (-1.0) / motion_noise;
Omega(n1+b,n2+b) = Omega(n1+b, n2+b) + (-1.0)*R(b,b) / motion_noise;
d_index = [1 2 3];
b_index = find(d_index == b);
d_index(b_index)=[];
for d = 1:size(d_index,2)
Omega(n1+b, n2+d_index(d)) = Omega(n1+b, n2+d_index(d)) + (-1.0)*R(b,d_index(d)) / motion_noise;
end
Omega(n2+b,n1+b) = Omega(n2+b, n1+b) + (-1.0) / motion_noise;
Xi(n1+b)= Xi(n1+b) - motion(b) / motion_noise;
Xi(n2+b)= Xi(n2+b) + motion(b) / motion_noise;
end
end
end
%compute best estimate
mu = (Omega^-1) * Xi;
g_t_pose=[];
temp_index = 1;
for i=1:3:size(mu,1)
g_t_pose(temp_index,:) = [mu(i), mu(i+1), mu(i+2)];
temp_index = temp_index + 1;
end
%Show the result
figure;
plot(g_t_pose(:,1), g_t_pose(:,2),'r*-');
hold on;
plot(e_t_pose(1:N,1), e_t_pose(1:N,2),'bo-');
hold off
end
function [isExist previous_index] = getPreviousIndex(data_set,pts)
isExist = 0;
previous_index = 0;
distance_threshold = 41; % [mm]; typical absolute accuracy + 3 * typical repeatibility of SR4000 = 20 + 3 * 7 = 41
for i=1:size(data_set,1)
if data_set(i,1) > 0 && data_set(i,1) == pts(1) % Skip non-valid data
distance = sqrt(sum((data_set(i,3:5)-pts(4:6)).^2));
if distance <= distance_threshold
isExist = 1;
previous_index = data_set(i,2);
break;
end
end
end
end
function plot_trajectory(e_pose, fpts)
figure;
gt_x = [0 0 2135 2135 0];
gt_y = [0 1220 1220 0 0];
plot(e_pose(:,1),e_pose(:,2),'bo-');
hold on;
plot(gt_x,gt_y,'r-','LineWidth',2);
if ~isempty(fpts)
plot(fpts(:,2),fpts(:,3),'gd');
legend('Estimated Pose','Estimated Truth','feature points');
else
legend('Estimated Pose','Estimated Truth');
end
xlabel('X [mm]');
ylabel('Y [mm]');
grid;
h_xlabel = get(gca,'XLabel');
set(h_xlabel,'FontSize',12,'FontWeight','bold');
h_ylabel = get(gca,'YLabel');
set(h_ylabel,'FontSize',12,'FontWeight','bold');
hold off;
figure;
gt_x = [0 0 2135 2135 0];
gt_y = [0 1220 1220 0 0];
gt_z = [0 0 0 0 0];
plot3(e_pose(:,1),e_pose(:,2),e_pose(:,3),'bo-');
hold on;
plot3(gt_x,gt_y,gt_z,'r-','LineWidth',2);
if ~isempty(fpts)
plot3(fpts(:,2),fpts(:,3),fpts(:,4),'gd');
legend('Estimated Pose','Estimated Truth','feature points');
else
legend('Estimated Pose','Estimated Truth');
end
xlabel('X [mm]');
ylabel('Y [mm]');
zlabel('Z [mm]');
grid;
h_xlabel = get(gca,'XLabel');
set(h_xlabel,'FontSize',12,'FontWeight','bold');
h_ylabel = get(gca,'YLabel');
set(h_ylabel,'FontSize',12,'FontWeight','bold');
legend('Estimated Pose','Estimated Truth');
end
|
github
|
rising-turtle/slam_matlab-master
|
generate_video.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/generate_video.m
| 5,239 |
utf_8
|
1e4f4012d3f049bf499b74bcb98967b4
|
% Conver isp to ply for meshlab
%
% Author : Soonhac Hong ([email protected])
% Date : 4/1/13
function generate_video(isp_file_name, opt_file_name, file_index, dynamic_index, isgframe, vro_name)
close all;
[poses] = load_graph_isp(isp_file_name);
[opt_poses] = load_graph_isp(opt_file_name);
% convert_feature2ply(ply_headers, ply_file_name, poses, file_index, dynamic_index, isgframe);
data_name_list={'pitch', 'pan', 'roll','x2', 'y2', 'c1', 'c2','c3','c4','m','etas','loops2','kinect_tum','sparse_feature','swing'};
% Write headers
pose_interval = 4;
sample_interval = 3;
image_width = 176;
image_height = 144;
show_image_width = floor(image_width/sample_interval);
show_image_height = floor(image_height/sample_interval);
% nfeature = size(poses,1)*show_image_width*show_image_height;
% element_vertex_n = sprintf('element vertex %d',nfeature);
% ply_headers{4} = element_vertex_n;
% for i=1:size(ply_headers,2)
% fprintf(fd,'%s\n',ply_headers{i});
% end
% Write data
for i = 1:size(opt_poses,1)
% if vro_cpp == 1
% h{i} = [euler_to_rot(vro_o_pose(i,2), vro_o_pose(i,1), vro_o_pose(i,3)) vro_t_pose(i,:)'; 0 0 0 1];
% else
if strcmp(isgframe, 'gframe')
h{i} = [e2R([opt_poses(i,4), opt_poses(i,5), opt_poses(i,6)]) opt_poses(i,1:3)'; 0 0 0 1]; % e2R([rx,ry,rz]) [radian]
else
h{i} = [euler_to_rot(opt_poses(i,5)*180/pi, opt_poses(i,4)*180/pi, opt_poses(i,6)*180/pi) opt_poses(i,1:3)'; 0 0 0 1]; % euler_to_rot(ry, rx, rz) [degree]
end
end
distance_threshold_max = 5;
distance_threshold_min = 0.8;
x_min = min(opt_poses(:,1)) - 1;
x_max = max(opt_poses(:,1)) + 1;
y_min = min(opt_poses(:,2)) - 1;
y_max = max(opt_poses(:,2)) + 1;
z_min = min(opt_poses(:,3)) - 1;
z_max = max(opt_poses(:,3)) + 1;
scrsz = get(0,'ScreenSize');
% figure(video_figure);
video_figure=figure('Position',[scrsz(3)/4 scrsz(4)/4 scrsz(3)/2 scrsz(4)/2]);
[prefix, confidence_read] = get_sr4k_dataset_prefix(data_name_list{file_index+3}, dynamic_index);
vidObj = VideoWriter(sprintf('%s_gslam_vf.avi', prefix));
vidObj.Quality=100;
vidObj.FrameRate=5;
%vidObj.Height=scrsz(4)/2;
%vidObj.Width=scrsz(3)/2;
open(vidObj);
for i=1:size(opt_poses,1)
i
% show images
if check_stored_visual_feature(data_name_list{file_index+3}, dynamic_index, i, true, 'intensity') == 0
[img, x, y, z, c, elapsed_pre] = LoadSR(data_name_list{file_index+3}, 'gaussian', 0, dynamic_index, i, 1, 'int');
else
[frm, des, elapsed_sift, img, x, y, z, c, elapsed_pre] = load_visual_features(data_name_list{file_index+3}, dynamic_index, i, true, 'intensity');
if i>=2
[match_num, ransac_iteration, op_pset1_image_index, op_pset2_image_index, op_pset_cnt, elapsed_match, elapsed_ransac, op_pset1, op_pset2] = load_matched_points(data_name_list{file_index+3}, dynamic_index, i-1, i, 'none', true);
end
end
subplot(1,2,1);imshow(img);colormap(gray);
if i>=2
hold on;
for j=1:size(op_pset2_image_index,1)
subplot(1,2,1);plot(round(op_pset2_image_index(j,1))+1, round(op_pset2_image_index(j,2))+1,'r+','Markersize',10);
end
end
axis image;
% set(gca,'XTick',[]);
% set(gca,'YTick',[]);
% show poses
subplot(1,2,2);plot3(poses(i,1),poses(i,2),poses(i,3),'g.','LineWidth', 2);
hold on;
subplot(1,2,2);plot3(opt_poses(i,1),opt_poses(i,2),opt_poses(i,3),'r.', 'LineWidth',2);
grid on;
axis equal;
axis([x_min x_max y_min y_max z_min z_max]);
xlabel('X');ylabel('Y');zlabel('Z');
%axis equal;
%hold off;
% if mod(i,pose_interval) == 1
%
% confidence_threshold = floor(max(max(c))/2);
%
% for j=1:show_image_width
% for k=1:show_image_height
% col_idx = (j-1)*sample_interval + 1;
% row_idx = (k-1)*sample_interval + 1;
% %unit_pose = [x(row_idx,col_idx), y(row_idx, col_idx), z(row_idx, col_idx)];
% unit_pose = [-x(row_idx,col_idx), z(row_idx, col_idx), y(row_idx, col_idx)];
% unit_pose_distance = sqrt(sum(unit_pose.^2));
% %if img(row_idx, col_idx) > 50
% if unit_pose(3) <= -0.1
% if c(row_idx,col_idx) >= confidence_threshold && unit_pose_distance < distance_threshold_max && unit_pose_distance > distance_threshold_min
% unit_pose_global = h{i}*[unit_pose, 1]';
% unit_color = [img(row_idx, col_idx),img(row_idx, col_idx),img(row_idx, col_idx)]./255;
% %fprintf(fd,'%f %f %f %d %d %d\n',unit_pose_global(1:3,1)', unit_color);
% %ply_data(ply_data_index,:) = [unit_pose_global(1:3,1)', double(unit_color)];
% %ply_data_index = ply_data_index + 1;
% subplot(1,2,2);plot3(unit_pose_global(1,1),unit_pose_global(2,1),unit_pose_global(3,1),'Color',unit_color);
% grid;
% end
% end
% end
% end
% end
currFrame = getframe(video_figure);
writeVideo(vidObj,currFrame);
drawnow;
end
hold off;
close(vidObj);
end
|
github
|
rising-turtle/slam_matlab-master
|
plot_kinect_tum.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/plot_kinect_tum.m
| 3,022 |
utf_8
|
e0bff3482681f917c9d1b66f877a96f7
|
% Plot the result of graph SLAM for kinect_tum data
%
% Author : Soonhac Hong ([email protected])
% Date : 2/22/12
function plot_kinect_tum(org_file_name, opt_file_name, optimizer_name, dir_index, nFrame)
switch optimizer_name
case 'isam'
[org_poses] = load_graph_isp(org_file_name);
if ~strcmp(opt_file_name, 'none')
[opt_poses] = load_graph_isam(opt_file_name);
end
case 'toro'
[org_poses org_edges org_fpts_poses org_fpts_edges] = load_graph_toro(org_file_name);
[opt_poses opt_edges opt_fpts_poses opt_fpts_edges] = load_graph_toro(opt_file_name);
case 'g2o'
[org_poses org_edges org_fpts_poses org_fpts_edges] = load_graph_g2o(org_file_name);
[opt_poses opt_edges opt_fpts_poses opt_fpts_edges] = load_graph_g2o(opt_file_name);
end
%load ground truth
[gt rgbdslam rtime] = Load_kinect_gt(dir_index);
first_timestamp = get_timestamp_kinect_tum(dir_index,1);
gt_start_index = find(gt(:,1) > first_timestamp, 1);
if gt(gt_start_index,1) - first_timestamp > first_timestamp - gt(gt_start_index-1,1)
gt_start_index = gt_start_index - 1;
end
%gt(gt_start_index,:)
last_timestamp = get_timestamp_kinect_tum(dir_index,nFrame);
gt_last_index = find(gt(:,1) > last_timestamp, 1);
if gt(gt_last_index,1) - last_timestamp > last_timestamp - gt(gt_last_index-1,1)
gt_last_index = gt_last_index - 1;
end
% Show the pose
start_index = 1;
if ~strcmp(opt_file_name, 'none')
end_index = min(size(org_poses,1), size(opt_poses)); %80;
else
end_index = size(org_poses,1);
end
if size(org_poses,2) == 2 % SE2
figure;
plot(org_poses(start_index:end_index,1), org_poses(start_index:end_index,2),'g:-');
hold on;
if ~strcmp(opt_file_name, 'none')
plot(opt_poses(start_index:end_index,1), opt_poses(start_index:end_index,2),'b.-','LineWidth',2);
end
plot(gt(:,2), gt(:,3),'r.-','LineWidth',2);
xlabel('X [m]');
ylabel('Y [m]');
grid;
if ~strcmp(opt_file_name, 'none')
legend('vro',optimizer_name,'GT');
else
legend('vro','GT');
end
hold off;
elseif size(org_poses,2) == 3 % SE3:QUAT
figure;
plot3(org_poses(start_index:end_index,1), org_poses(start_index:end_index,2), org_poses(start_index:end_index,3), 'g-', 'LineWidth',1);
hold on;
if ~strcmp(opt_file_name, 'none')
plot3(opt_poses(start_index:end_index,1), opt_poses(start_index:end_index,2), opt_poses(start_index:end_index,3), 'b.-', 'LineWidth',2);
end
plot3(gt(gt_start_index:gt_last_index,2), gt(gt_start_index:gt_last_index,3), gt(gt_start_index:gt_last_index,4), 'r-','LineWidth',1);
%plot3(rgbdslam(start_index:end_index,2), rgbdslam(start_index:end_index,3), rgbdslam(start_index:end_index,4), 'b:', 'LineWidth',1);
xlabel('X [m]');
ylabel('Y [m]');
zlabel('Z [m]');
if ~strcmp(opt_file_name, 'none')
legend('vro',optimizer_name,'GT_e');
else
legend('vro','GT');
end
grid;
axis equal;
hold off;
end
end
|
github
|
rising-turtle/slam_matlab-master
|
generate_map.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/generate_map.m
| 1,184 |
utf_8
|
3157a39500f461ed019f656d17af83e5
|
% Generate *.ply files for visualization
%
% Author : Soonhac Hong ([email protected])
% Date : 6/5/14
function generate_map()
graphslam_addpath;
file_index = 12; %16; % 5 = whitecane 6 = etas 7 = loops 8 = kinect_tum 9 = loops2 10= amir_vro 11 = sparse_feature 12 = swing 13 = swing2 14 = motive 15 = object_recognition 16=map 17=it
dynamic_index = 25; %5;
isgframe='none';
vro_name='vro';
%gtsam_isp_file_name='data\3d\whitecane\revisiting2_10m_Replace_pose_zero_411_vro_gtsam.isp';
%gtsam_opt_file_name='results\isam\3d\map5_Replace_pose_zero_3842_vro_gtsam.opt';
%gtsam_opt_file_name='results/isam/3d/revisiting2_10m_Replace_pose_zero_411_vro_gtsam.opt'; %18
%gtsam_opt_file_name='results/isam/3d/revisiting6_10m_Replace_pose_zero_910_vro_gtsam.opt'; %22
gtsam_opt_file_name='results/isam/3d/revisiting9_10m_Replace_pose_zero_827_vro_gtsam.opt'; %25
%gtsam_opt_file_name='results/isam/3d/revisiting8_10m_Replace_pose_zero_1084_vro_gtsam.opt'; %24
% VRO trajectory
%convert_isp2ply(gtsam_isp_file_name, file_index, dynamic_index, isgframe, vro_name);
% PGO trajectory and its map
convert_isp2ply(gtsam_opt_file_name, file_index, dynamic_index, isgframe, vro_name);
end
|
github
|
rising-turtle/slam_matlab-master
|
run_pose_graph_optimization_v0.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/run_pose_graph_optimization_v0.m
| 989 |
utf_8
|
7b2b6f882f54303b50a4255b1264f1df
|
% Run pose graph optimization
%
% Author : Soonhac Hong ([email protected])
% History :
% 3/26/14 : Created
function [result, graph, initial, h_global]=run_pose_graph_optimization_v0(data_name, dynamic_index, vro_result, vro_pose_std, graph, initial, h_global, dis)
import gtsam.*
t = gtsam.Point3(0, 0, 0);
if isempty(h_global)
h_global = get_global_transformation_dataname(data_name, dynamic_index, 'none');
rot = h_global(1:3,1:3);
R = gtsam.Rot3(rot);
origin= gtsam.Pose3(R,t);
initial.insert(0,origin);
end
pgc_t=tic;
[graph,initial] = construct_pose_graph(vro_result, vro_pose_std, graph, initial);
first = initial.at(0);
pgc_ct =toc(pgc_t)
graph.add(NonlinearEqualityPose3(0, first));
gtsam_t=tic;
optimizer = LevenbergMarquardtOptimizer(graph, initial);
result = optimizer.optimizeSafely();
gtsam_ct =toc(gtsam_t)
% Show the results in the plot and generate location information
if dis==true
plot_graph_initial_result_v0(initial, result);
end
end
|
github
|
rising-turtle/slam_matlab-master
|
get_timestamp_kinect_tum_color.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/get_timestamp_kinect_tum_color.m
| 572 |
utf_8
|
cd7c6a57abe82b21dac2db599552499f
|
% Get time stamp of color image in kinect_tum dataset
%
% Author : Soonhac Hong ([email protected])
% Date : 12/20/12
function [time_stamp] = get_timestamp_kinect_tum_color(dm,j)
dir_name = get_kinect_tum_dir_name();
[depth_data_dir, err] = sprintf('E:/data/kinect_tum/%s/rgb',dir_name{dm});
dirData = dir(depth_data_dir); %# Get the data for the current directory
dirIndex = [dirData.isdir]; %# Find the index for directories
file_list = {dirData(~dirIndex).name}';
[file_name, err]=sprintf('%s',file_list{j});
time_stamp = str2num(strrep(file_name, '.png',''));
end
|
github
|
rising-turtle/slam_matlab-master
|
pose_optimization_example_1d.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/pose_optimization_example_1d.m
| 1,981 |
utf_8
|
692c2cb99f72f86abae593d592dbdb8f
|
% 1D Example of Pose Graph Optimization
% Data : 4/12/12
% Author : Soonhac Hong ([email protected])
function pose_optimization_example_1d()
% 1D case
pose_data=[1 1 0; 1 2 9 ; 2 3 5; 3 4 8; 1 3 11; 1 4 19; 2 4 12]; % [first pose, second pose, constraint]
xinit = [0 9 14 22]';
motion_noise = 0.3;
pose_num = size(unique(pose_data(:,1:2)),1); % first position is not optimized.
Omega = zeros(pose_num,pose_num);
Xi = zeros(pose_num,1);
Odometry=zeros(pose_num,1);
%Odometry = zeros(pose_num,1);
% Fill the elements of Omega and Xi
% Initial point
Omega(1,1) = 1;
Xi(1) = pose_data(1,3);
Odometry(1) = pose_data(1,3);
for i=2:size(pose_data,1)
unit_data = pose_data(i,:);
current_index = unit_data(1);
next_index = unit_data(2);
movement = unit_data(3);
% Fill diagonal elements of Omega
Omega(current_index,current_index) = Omega(current_index,current_index) + 1/motion_noise;
Omega(next_index,next_index) = Omega(next_index,next_index) + 1/motion_noise;
% Fill Off-diagonal elements of Omega
Omega(current_index,next_index) = Omega(current_index,next_index) + (-1)/motion_noise;
Omega(next_index,current_index) = Omega(next_index,current_index) + (-1)/motion_noise;
% Fill Xi
Xi(current_index) = Xi(current_index) + (-1)*movement/motion_noise;
Xi(next_index) = Xi(next_index) + movement/motion_noise;
% Update Odometry
if abs(current_index - next_index) == 1
Odometry(next_index) = Odometry(next_index-1) + movement;
end
end
%Omega
%Xi
%mu = Omega^-1 * Xi
% Using LM
%myfun = @(x,xdata)Rot(x(1:3))*xdata+repmat(x(4:6),1,length(xdata));
myfun = @(x,xdata)xdata*x(1:size(xdata,1));
xdata = Omega;
ydata = Xi;
options = optimset('Algorithm', 'levenberg-marquardt');
%x = lsqcurvefit(myfun, zeros(6,1), p, q, [], [], options);
x = lsqcurvefit(myfun, xinit, xdata, ydata, [], [], options);
plot(Odometry,'bd-');
hold on;
plot(x,'ro-');
hold off;
legend('Odometry','Optimized');
end
|
github
|
rising-turtle/slam_matlab-master
|
convert_isam2rossba.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/convert_isam2rossba.m
| 15,493 |
utf_8
|
039d105831c4081ddbc63fca4bd798e9
|
% Convert *.isp and *.sam to files for ROS sba(http://www.ros.org/wiki/sba/Tutorials/IntroductionToSBA)
%
% Author : Soonhac Hong ([email protected])
% Date : 2/11/13
function convert_isam2rossba()
graphslam_addpath;
addpath('D:\soonhac\Project\PNBD\SW\ASEE\slamtoolbox\slamToolbox_11_09_08\FrameTransforms\Rotations');
addpath('D:\soonhac\Project\PNBD\SW\ASEE\slamtoolbox\slamToolbox_11_09_08\DataManagement');
%% Load isam data
%sam_filename = 'D:\soonhac\Project\PNBD\SW\ASEE\GraphSLAM\data\3d\whitecane\exp8__Replace_pose_feature_76_18_vro_cov.sam';
%isp_filename = 'D:\soonhac\Project\PNBD\SW\ASEE\GraphSLAM\data\3d\whitecane\exp8__Replace_pose_feature_76_18_vro.isp';
%sam_filename = 'D:\soonhac\Project\PNBD\SW\ASEE\GraphSLAM\data\3d\whitecane\exp8__Replace_pose_feature_10138_580_vro_cov.sam';
%isp_filename = 'D:\soonhac\Project\PNBD\SW\ASEE\GraphSLAM\data\3d\whitecane\exp8__Replace_pose_feature_10138_580_vro.isp';
%isp_filename = 'D:\soonhac\Project\PNBD\SW\ASEE\GraphSLAM\data\3d\whitecane\exp3_bus_straight_150_Replace_pose_feature_1385_84_vro_icp_ch.isp';
%sam_filename = 'D:\soonhac\Project\PNBD\SW\ASEE\GraphSLAM\data\3d\whitecane\exp3_bus_straight_150_Replace_pose_feature_1385_84_vro_icp_ch_cov.sam';
file_index = 9; % 5 = whitecane 6 = etas 7 = loops 8 = kinect_tum 9 = loops2 10= amir_vro 11 = sparse_feature
dynamic_index =1; % 15:square_500, 16:square_700, 17:square_swing
vro_name_list={'vro','vro_icp','vro_icp_ch','icp'};
vro_name_index = 1;
vro_name=vro_name_list{vro_name_index};
compensation_option ={'Replace','Linear'};
compensation_index = 1;
feature_pose_name = 'pose_feature';
vro_size = 426; %1614; %1632;
pose_size = 48; %101; %104;
isgframe = 'gframe';
[g2o_result_dir_name, isam_result_dir_name, vro_dir_name, dynamic_dir_name] = get_file_names(file_index, dynamic_index);
sam_filename = sprintf('%s%s_%s_%s_%d_%d_%s_cov.sam', vro_dir_name, dynamic_dir_name, compensation_option{compensation_index}, feature_pose_name, vro_size, pose_size, vro_name);
isp_filename = sprintf('%s%s_%s_%s_%d_%d_%s.isp', vro_dir_name, dynamic_dir_name, compensation_option{compensation_index}, feature_pose_name, vro_size, pose_size, vro_name);
sba_filename =sprintf('data/ba/ros_sba/%s_%s_%s_%d_%d_ros_sba.out',dynamic_dir_name, compensation_option{compensation_index}, feature_pose_name, vro_size, pose_size);
%sam_filename = 'D:\soonhac\Project\PNBD\SW\ASEE\GraphSLAM\data\3d\whitecane\exp1_bus_door_straight_150_Replace_pose_zero_1632_104_vro_icp_ch_cov.sam';
%isp_filename = 'D:\soonhac\Project\PNBD\SW\ASEE\GraphSLAM\data\3d\whitecane\exp1_bus_door_straight_150_Replace_pose_zero_1632_104_vro_icp_ch.isp';
%sba_filename = 'D:\soonhac\Project\PNBD\SW\ASEE\GraphSLAM\data\ba\ros_sba\exp8__Replace_pose_feature_10138_580_ros_sba.out';
[landmarks, initial_landmarks] = load_landmark_vro(sam_filename);
[initial_camera_pose] = load_camera_pose(isp_filename); % x, y, z, rx, ry, rz
sr4k_parameter = initialize_cam();
%check_camera_projection(initial_camera_pose, landmarks, isgframe);
%show_camera_pose(initial_camera_pose, true, 'vro', 'k.-');
%debug
% show_camera_pose(initial_camera_pose,true, 'k.-');
% hold on;
% plot3(initial_landmarks(:,2),initial_landmarks(:,3),initial_landmarks(:,4),'m*')
% axis equal;
%% Generate cameara pose
% Data format (http://phototour.cs.washington.edu/bundler/bundler-v0.3-manual.html#S6)
% # Bundle file v0.3
% <num_cameras> <num_points> [two integers]
% <camera1>
% <camera2>
% ...
% <cameraN>
% <point1>
% <point2>
% ...
% <pointM>
%
% <cameraI>
% <f> <k1> <k2> [the focal length, followed by two radial distortion coeffs]
% <R> [a 3x3 matrix representing the camera rotation]
% <t> [a 3-vector describing the camera translation]
%
% Global coordindate
% Z : depth
% -------------> X
% |
% |
% V
% Y
%
fu = sr4k_parameter.f;
u0 = sr4k_parameter.Cx;
v0 = sr4k_parameter.Cy;
kd = [sr4k_parameter.k1, sr4k_parameter.k2];
%ros_sba_T = inv(p2T([0,0,0,-90*pi/180,0,0]));
ros_sba_T = sr4k_p2T([0,0,0,pi/2,0,0]);
%ros_sba_R2 = e2R([-pi/2, 0, -pi/2]);
%ros_sba_T = [ros_sba_R2 [0, 0, 0]'; 0 0 0 1];
%file_name_pose = sprintf('%s%s_%s_%s_%d_%d_%s.isp',file_name, dir_name, cmp_option, feature_pose_name, vro_size, e_t_pose_size, vro_name);
%filename = 'D:\soonhac\Project\PNBD\SW\ASEE\GraphSLAM\data\ba\ros_sba\exp8__Replace_pose_feature_76_18_ros_sba.out';
%sba_filename = 'D:\soonhac\Project\PNBD\SW\ASEE\GraphSLAM\data\ba\ros_sba\exp8__Replace_pose_feature_10138_580_ros_sba.out';
%filename = 'D:\soonhac\Project\PNBD\SW\ASEE\GraphSLAM\data\ba\ros_sba\exp3_bus_straight_150_Replace_pose_feature_1385_84_ros_sba.out';
converted_camera_pose=[];
cams_fd = fopen(sba_filename, 'w');
fprintf(cams_fd,'# Bundle file v0.3\n');
fprintf(cams_fd,'%d %d\n', size(initial_camera_pose,1), size(initial_landmarks,1));
%r2d = 180/pi;
%first camera pose as the reference frame
% rot = e2R([pi, 0, 0]);
% trans = [0, 0, 0]';
% fprintf(cams_fd,'%f %f %f\n', fu, kd);
% for r=1:3
% fprintf(cams_fd,'%f %f %f\n', rot(r,:));
% end
% fprintf(cams_fd,'%f %f %f\n', trans(1), trans(2), trans(3));
% converted_camera_pose=[converted_camera_pose; trans(1:3)' [0, 0, 0]];
% ros_sba_T_camera = [rot trans; 0 0 0 1];
trans = [0 0 0];
rot = e2R([pi(), 0, 0]);%e2R([initial_camera_pose(1,4), initial_camera_pose(1,5), initial_camera_pose(1,6)]);
bundler_T = [rot trans'; 0 0 0 1];
for i=1:size(initial_camera_pose,1)
trans = initial_camera_pose(i,1:3);
if strcmp(isgframe, 'gframe')
rot = e2R([initial_camera_pose(i,4), initial_camera_pose(i,5), initial_camera_pose(i,6)]);
else
rot = euler_to_rot(initial_camera_pose(i,5)*180/pi,initial_camera_pose(i,4)*180/pi,initial_camera_pose(i,6)*180/pi);
end
T = [rot trans'; 0 0 0 1];
%T = ros_sba_T * inv(first_camera_T) * T;
%T = ros_sba_T * T;
%T = bundler_T*ros_sba_T * T;
sba_T = inv(T); % ROS SBA convention !!!! camera N = inv(T)* camera N+1
if strcmp(isgframe, 'gframe')
T = bundler_T*sba_T; % rotate 180 around x for Bundler frame
else
T = bundler_T*ros_sba_T*sba_T*inv(ros_sba_T); % rotate 180 around x to Bundler frame
end
trans = T(1:3,4)';
rot = T(1:3,1:3);
%trans = ros_sba_T*[trans'; 1]; % convert coordinate
%debug_rot = euler_to_rot(initial_camera_pose(i,5)*r2d,initial_camera_pose(i,4)*r2d,initial_camera_pose(i,6)*r2d); % [ry, rx, rz]
%rot = e2R([initial_camera_pose(i,4), -initial_camera_pose(i,6), initial_camera_pose(i,5)]);
%rot = e2R([0, 0, 0]);
%rot = ros_sba_T_camera(1:3,1:3)*ros_sba_T(1:3,1:3)*e2R([initial_camera_pose(i,4), initial_camera_pose(i,5), initial_camera_pose(i,6)]);
fprintf(cams_fd,'%f %f %f\n', fu, kd);
for r=1:3
fprintf(cams_fd,'%f %f %f\n', rot(r,:));
end
fprintf(cams_fd,'%f %f %f\n', trans(1), trans(2), trans(3));
%fprintf(cams_fd,'%f %f %f\n', trans);
%converted_camera_pose=[converted_camera_pose; T(1:3,4)' R2e(T(1:3,1:3))'];
converted_camera_pose=[converted_camera_pose; trans R2e(rot)'];
end
show_camera_pose(converted_camera_pose,true, 'typical', 'ros_sba', 'on', 'k.-');
%% Generate landmarks points
%
% <pointM>
% <position> [a 3-vector describing the 3D position of the point]
% <color> [a 3-vector describing the RGB color of the point]
% <view list> [a list of views the point is visible in]
% <view list>
% <nlist> <camera> <key> <x> <y> .... [The pixel positions are floating point numbers in a coordinate system where the origin is the center of the image, the x-axis increases to the right, and the y-axis increases towards the top of the image.]
%
% Note : The reference coordinate of landmark point is the first camera coordinate !!!!
%ros_sba_T_landmark = sr4k_p2T([0,0,0,-pi/2,0,0]);
%TODO : initial_landmarks and landmarks position is different in position
%and measured pixel
converted_landmark=[];
color = [255, 255, 255];
key_offset = landmarks(1,2);
for i=1:size(initial_landmarks,1)
xyz = initial_landmarks(i,2:4);
if strcmp(isgframe, 'gframe')
xyz = [xyz'; 1];
else
xyz = ros_sba_T*[xyz'; 1]; % convert coordinate because z-axis should represent the depth.
end
%xyz = bundler_T * ros_sba_T*[xyz'; 1]; % convert coordinate. See Note !!!!
%xyz = ros_sba_T*inv(first_camera_T)*[xyz'; 1]; % convert coordinate. See Note !!!!
landmark_idx = initial_landmarks(i,1);
nframes_idx=find(landmarks(:,2)==landmark_idx);
fprintf(cams_fd,'%f %f %f\n', xyz(1), xyz(2), xyz(3));
%fprintf(cams_fd,'%f %f %f\n', xyz);
fprintf(cams_fd,'%d %d %d\n', color);
%View list
fprintf(cams_fd,'%d ', size(nframes_idx,1));
for c=1:size(nframes_idx,1)
frame_idx = landmarks(nframes_idx(c,1),1);
key_idx = landmarks(nframes_idx(c,1),2);
img_idx = landmarks(nframes_idx(c,1),6:7);
img_idx(1) = img_idx(1) - u0; % Adjust image coordinate for ROS SBA
img_idx(2) = (-1) * (img_idx(2) - v0); % Adjust image coordinate for ROS SBA
fprintf(cams_fd,'%d %d %f %f ', frame_idx, key_idx-key_offset, img_idx);
converted_landmark=[converted_landmark; frame_idx, key_idx-key_offset, xyz(1), xyz(2), xyz(3),img_idx];
end
fprintf(cams_fd,'\n');
%coverted_landmark=[coverted_landmark; xyz(1), xyz(2), xyz(3)];
end
fclose(cams_fd);
%Debug - show landmarks
hold on;
plot3(converted_landmark(:,3), converted_landmark(:,4), converted_landmark(:,5), 'm.');
hold off;
axis equal;
%check_camera_projection_with_converted_data(converted_camera_pose, converted_landmark, initial_camera_pose, landmarks, isgframe);
end
function check_camera_projection_with_converted_data(camera_poses, feature_points, initial_camera_pose, initial_landmarks, isgframe)
cam = initialize_cam();
features_info=[];
ros_sba_T = sr4k_p2T([0,0,0,pi/2,0,0]);
u0 = cam.Cx;
v0 = cam.Cy;
trans = [0 0 0];
rot = e2R([pi(), 0, 0]);%e2R([initial_camera_pose(1,4), initial_camera_pose(1,5), initial_camera_pose(1,6)]);
bundler_T = [rot trans'; 0 0 0 1];
camera_index_list=unique(feature_points(:,1));
for i=1:size(camera_index_list,1)
temp_t = camera_poses(i,1:3)';
initial_temp_t = initial_camera_pose(i,1:3)';
%temp_R = e2R(camera_poses(i,4:6)');
if strcmp(isgframe, 'gframe')
temp_R = e2R(camera_poses(i,4:6)');
initial_temp_R = e2R(initial_camera_pose(i,4:6)');
else
temp_R = euler_to_rot(camera_poses(i,5)*180/pi,camera_poses(i,4)*180/pi,camera_poses(i,6)*180/pi);
end
T = [temp_R, temp_t; 0 0 0 1];
T = bundler_T*T; % Convert VRO frame to SBA frame
T = inv(T);
temp_t = T(1:3,4);
temp_R = T(1:3,1:3);
abs(temp_t - initial_temp_t)
abs(temp_R - initial_temp_R)
initial_camera_index = camera_index_list(i);
initial_data_index_list = find(initial_landmarks(:,1) == initial_camera_index);
initial_unit_data = initial_landmarks(initial_data_index_list,:);
initial_estimated_uv = [];
initial_measured_uv = initial_unit_data(:,6:7);
camera_index = camera_index_list(i);
data_index_list = find(feature_points(:,1) == camera_index);
unit_data = feature_points(data_index_list,:);
estimated_uv = [];
measured_uv = [unit_data(:,6)+u0, unit_data(:,7).*(-1) + v0];
for j=1:size(unit_data,1)
%temp_p = ros_sba_T * [unit_data(j,3:5) 1]';
temp_p = [unit_data(j,3:5) 1]';
initial_temp_p = [initial_unit_data(j,3:5) 1]';
%estimated_uv(j,:) = hi_cartesian_test(temp_p(1:3), [0;0;0], eye(3), cam, features_info )';
if strcmp(isgframe, 'gframe')
estimated_uv(j,:) = hi_cartesian_test(temp_p(1:3), temp_t, temp_R, cam, features_info )';
%estimated_uv(j,:) = hi_cartesian_test(initial_temp_p(1:3), temp_t, temp_R, cam, features_info )';
initial_estimated_uv(j,:) = hi_cartesian_test(initial_temp_p(1:3), initial_temp_t, initial_temp_R, cam, features_info )';
else
estimated_uv(j,:) = vro_camera_projection(temp_p(1:3), temp_t, temp_R, cam, features_info, ros_sba_T)';
end
end
figure;
plot(estimated_uv(:,1), estimated_uv(:,2),'b+');
hold on;
%plot(unit_data(:,6)+u0,unit_data(:,7).*(-1) + v0,'ro');
plot(measured_uv(:,1), measured_uv(:,2),'ro');
plot(initial_estimated_uv(:,1), initial_estimated_uv(:,2),'gd');
legend('Estimated','Measured','Init Estimated');
hold off;
projection_error = sqrt(sum((estimated_uv - measured_uv).^2,2));
projection_error_mean_std(i,:) = [mean(projection_error), std(projection_error)];
initial_projection_error = sqrt(sum((initial_estimated_uv - initial_measured_uv).^2,2));
initial_projection_error_mean_std(i,:) = [mean(initial_projection_error), std(initial_projection_error)];
%compare with initial values
%abs(projection_error - initial_projection_error)
%abs(projection_error_mean_std - initial_projection_error_mean_std)
end
figure;
errorbar(projection_error_mean_std(:,1), projection_error_mean_std(:,2),'b');
figure;
errorbar(initial_projection_error_mean_std(:,1), initial_projection_error_mean_std(:,2),'r');
end
function check_camera_projection(camera_poses, feature_points, isgframe)
cam = initialize_cam();
features_info=[];
ros_sba_T = sr4k_p2T([0,0,0,pi/2,0,0]);
camera_index_list=unique(feature_points(:,1));
for i=1:size(camera_index_list,1)
temp_t = camera_poses(i,1:3)';
%temp_R = e2R(camera_poses(i,4:6)');
if strcmp(isgframe, 'gframe')
temp_R = e2R(camera_poses(i,4:6)');
else
temp_R = euler_to_rot(camera_poses(i,5)*180/pi,camera_poses(i,4)*180/pi,camera_poses(i,6)*180/pi);
end
% T = [temp_R, temp_t; 0 0 0 1];
% T = ros_sba_T*T; % Convert VRO frame to SBA frame
% temp_t = T(1:3,4);
% temp_R = T(1:3,1:3);
camera_index = camera_index_list(i);
data_index_list = find(feature_points(:,1) == camera_index);
unit_data = feature_points(data_index_list,:);
estimated_uv = [];
for j=1:size(unit_data,1)
%temp_p = ros_sba_T * [unit_data(j,3:5) 1]';
temp_p = [unit_data(j,3:5) 1]';
%estimated_uv(j,:) = hi_cartesian_test(temp_p(1:3), [0;0;0], eye(3), cam, features_info )';
if strcmp(isgframe, 'gframe')
estimated_uv(j,:) = hi_cartesian_test(temp_p(1:3), temp_t, temp_R, cam, features_info )';
else
estimated_uv(j,:) = vro_camera_projection(temp_p(1:3), temp_t, temp_R, cam, features_info, ros_sba_T)';
end
end
figure;
plot(estimated_uv(:,1), estimated_uv(:,2),'b+');
hold on;
plot(unit_data(:,6),unit_data(:,7),'ro');
legend('Estimated','Measured');
hold off;
projection_error = sqrt(sum((estimated_uv - unit_data(:,6:7)).^2,2));
projection_error_mean_std(i,:) = [mean(projection_error), std(projection_error)];
end
figure;
errorbar(projection_error_mean_std(:,1), projection_error_mean_std(:,2),'b');
end
% function T = p2T(x)
%
% Rx = @(a)[1 0 0;
% 0 cos(a) -sin(a);
% 0 sin(a) cos(a)];
%
% Ry = @(b)[cos(b) 0 sin(b);
% 0 1 0;
% -sin(b) 0 cos(b)];
%
% Rz = @(c)[cos(c) -sin(c) 0;
% sin(c) cos(c) 0;
% 0 0 1];
%
% Rot = @(x)Rz(x(3))*Rx(x(1))*Ry(x(2)); % SR4000 project; see euler_to_rot.m
% T = [Rot(x(4:6)) [x(1), x(2), x(3)]'; 0 0 0 1];
%
% end
|
github
|
rising-turtle/slam_matlab-master
|
sampling_feature_points.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/sampling_feature_points.m
| 2,203 |
utf_8
|
50716c6d229373f5a112bf5f9b1292e5
|
% Conver the results of VRO to the vertex and edges for isam
%
% Author : Soonhac Hong ([email protected])
% Date : 3/14/13
function new_feature_points = sampling_feature_points(feature_points, min_pixel_distance)
camera_index_list=unique(feature_points(:,1));
temp_feature_points=[];
%min_pixel_distance = 2;
for i=1:size(camera_index_list,1)
camera_index = camera_index_list(i);
data_index_list = find((feature_points(:,1) == camera_index & feature_points(:,3) == 1) | (feature_points(:,2) == camera_index & feature_points(:,3) == 2));
unit_data = feature_points(data_index_list,:);
sampled_unit_data = sampling_unit_data(unit_data, min_pixel_distance);
temp_feature_points(data_index_list,:) = sampled_unit_data;
end
sampled_index = find(temp_feature_points(:,end) == 1);
new_feature_points = temp_feature_points(sampled_index,1:end-1);
% Delete near feature points which distance is less than 0.8m
for i=1:size(new_feature_points,1)
distance = sqrt(sum(new_feature_points(:,4:6).^2,2));
end
near_feature_idx = find(distance < 800); % 800 [mm]
new_feature_points(near_feature_idx,:)=[];
end
function sampled_unit_data_final = sampling_unit_data(unit_data, min_pixel_distance)
sampled_unit_data = [];
%sampled_data_index_list = [];
center_data_index = find_nearest_center(unit_data(:,7:8));
%sampled_data_index_list = [sampled_data_index_list; center_data_index];
sampled_unit_data = [center_data_index, unit_data(center_data_index,7:8)];
for i=1:size(unit_data,1)
[duplication_index, duplication_flag] = check_duplication_imgidx(sampled_unit_data, unit_data(i,7:8), min_pixel_distance);
if duplication_flag == 0 % check duplicated points
sampled_unit_data = [sampled_unit_data; [i, unit_data(i,7:8)]];
end
end
sampled_unit_data_final = unit_data;
sampled_unit_data_final(sampled_unit_data(:,1), 9) = 1;
end
function data_index = find_nearest_center(unit_pixel_data)
image_center_x = 176/2;
image_center_y = 144/2;
distance = sqrt(sum((unit_pixel_data- repmat([image_center_x, image_center_y], size(unit_pixel_data,1), 1)).^2, 2));
[~, data_index] = min(distance);
end
|
github
|
rising-turtle/slam_matlab-master
|
modify_graph_isam.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/modify_graph_isam.m
| 3,121 |
utf_8
|
c0aacf62c3a5af46b5574b70ed447559
|
% Load the graph of isam
function modify_graph_isam(input_file_name)
temp_file_name = textscan(input_file_name,'%s','Delimiter','.');
output_file_name = sprintf('%s_modified.sam',temp_file_name{1}{1,1});
fid = fopen(input_file_name);
% Convert data
% Pose
data_name_list = {'EDGE3','POINT3'};
vertex_index = 1;
edge_index = 1;
while ~feof(fid) %for i = 1 : size(data_name,1)
%header = textscan(fid, '%s',1); % 2D format
%data_name = header{1};
%if strcmp(data_name, data_name_list{1}) % VERTEX_SE2
data = textscan(fid, '%s %d %d %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f\n');
% unit_data =[];
% for j=1:4
% unit_data = [unit_data data{j}];
% end
% f_index(vertex_index, :) = unit_data(1);
% t_pose(vertex_index,:) = unit_data(2:3);
% o_pose(vertex_index,:) = unit_data(4);
% vertex_index = vertex_index + 1;
% elseif strcmp(data_name{i}, data_name_list{2}) % EDGE_SE2
% unit_data =[];
% for j=4:12
% unit_data= [unit_data data{j}(i)];
% end
% edges(edge_index,:) = unit_data;
% edge_index = edge_index + 1;
%elseif strcmp(data_name, data_name_list{2}) % VERTEX_SE3:QUAT
% data = textscan(fid, '%s %d %d %f %f %f %f %f %f %f %f %f\n');
% unit_data =[];
% for j=2:7
% unit_data = [unit_data data{j}];
% end
% t_pose(vertex_index,:) = unit_data(1:3);
% o_pose(vertex_index,:) = unit_data(4:6);
% vertex_index = vertex_index + 1;
% elseif strcmp(data_name{i}, data_name_list{4}) % EDGE_SE3:QUAT
% unit_data =[];
% for j=2:31
% unit_data = [unit_data data{j}(i)];
% end
% edges(edge_index,:) = unit_data;
% edge_index = edge_index + 1;
%end
end
fclose(fid);
fod = fopen(output_file_name,'w');
pose_index = 0 ;
for j=1:size(data{1},1)
j
for i=1:size(data{1},1)
if strcmp(data{1}(i), data_name_list{1}) && data{1,2}(i) == pose_index
unit_data=[];
for k=1:size(data,2)
unit_data = [unit_data data{k}(i)];
end
fprintf(fod,'%s %d %d %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f\n',unit_data{1:30});
elseif strcmp(data{1}(i), data_name_list{2}) && data{1,2}(i) == (pose_index+1)
unit_data=[];
for k=1:12
unit_data = [unit_data data{k}(i)];
end
fprintf(fod,'%s %d %d %f %f %f %f %f %f %f %f %f\n',unit_data{1}, unit_data{2}-1, unit_data{3}-1, unit_data{4:12});
end
end
pose_index = pose_index + 1;
end
fclose(fod);
%poses = [t_pose o_pose];
end
|
github
|
rising-turtle/slam_matlab-master
|
get_file_names.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/get_file_names.m
| 4,480 |
utf_8
|
a1283f12893ede7f6b3241f3fc754853
|
% Get file name, directory name and number of frame
function [g2o_result_dir_name, isam_result_dir_name, vro_dir_name, dynamic_dir_name, toro_dir_name] = get_file_names(dir_index, dynamic_index)
% Select Data Set
vro_file_name_list ={'data/2d/manhattan3500/manhattanOlson3500.g2o', 'data/2d/intel/intel.g2o','data/3d/sphere/sphere_bignoise_vertex3.g2o','data/3d/garage/parking-garage.g2o','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/'};
isam_result_file_name_list ={'results/isam/2d/manhattanOlson3500.opt','results/isam/2d/intel.opt','results/isam/3d/sphere_bignoise_vertex3.opt','results/isam/3d/parking-garage.opt','results/isam/3d/','results/isam/3d/','results/isam/3d/','results/isam/3d/','results/isam/3d/','results/isam/3d/','results/isam/3d/','results/isam/3d/','results/isam/3d/','results/isam/3d/','results/isam/3d/','results/isam/3d/','results/isam/3d/'};
g2o_result_file_name_list ={'results/g2o/2d/manhattanOlson3500.opt','results/g2o/2d/intel.opt','results/g2o/3d/sphere_bignoise_vertex3.opt','results/g2o/3d/parking-garage.opt','results/g2o/3d/','results/g2o/3d/','results/g2o/3d/','results/g2o/3d/','results/g2o/3d/','results/g2o/3d/','results/g2o/3d/','results/g2o/3d/','results/g2o/3d/','results/g2o/3d/','results/g2o/3d/','results/g2o/3d/','results/g2o/3d/'};
toro_file_name_list={'2D/w10000-odom','3D/sphere_smallnoise','3D/sphere_mednoise','3D/sphere_bignoise','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/','data/3d/whitecane/'};
dynamic_name_list={'pitch_3degree','pitch_9degree','pan_3degree','pan_9degree','pan_30degree','pan_60degree','roll_3degree','roll_9degree','x_30mm','x_150mm','x_300mm','y_30mm','y_150mm','y_300mm','square_500','square_700','square_swing','square_1000','test'};
etas_name_list={'3th_straight','3th_swing','4th_straight','4th_swing','5th_straight','5th_swing'};
loops_name_list = {'bus','bus_3','bus_door_straight_150','bus_straight_150','data_square_small_100','eit_data_150','exp1','exp2','exp3','exp4','exp5','lab_80_dynamic_1','lab_80_swing_1','lab_it_80','lab_lookforward_4','s2','second_floor_150_not_square','second_floor_150_square','second_floor_small_80_swing','third_floor_it_150'};
kinect_tum_name_list = get_kinect_tum_dir_name();
loops2_name_list = get_loops2_filename();
sparse_feature_name_list = get_sparse_feature_filename();
swing_name_list = get_dir_name('swing'); %{'forward1','forward2','forward3','forward4','forward5','forward6'};
swing2_name_list = get_dir_name('swing2');
object_recognition_name_list = get_dir_name('object_recognition');
motive_name_list = get_dir_name('motive');
map_name_list=get_dir_name('map');
start_frame = [11 8 11 8 8 6 11 11 11 8 6 21 8 8 40 40 50 101];
finish_frame = [70 25 70 25 16 11 135 50 150 35 21 150 35 22 316 376 397 130]; %316
%nFrame = finish_frame(dynamic_data_index) - start_frame(dynamic_data_index) - 1;
%nFrame = 3920; %1253; %5468; %2399; %1253; %2440; %829; %270;
isam_result_dir_name = isam_result_file_name_list{dir_index};
g2o_result_dir_name = g2o_result_file_name_list{dir_index};
vro_dir_name = vro_file_name_list{dir_index};
toro_dir_name = toro_file_name_list{dir_index};
if dir_index == 5
dynamic_dir_name = dynamic_name_list{dynamic_index};
elseif dir_index == 6
dynamic_dir_name = etas_name_list{dynamic_index};
elseif dir_index == 7
dynamic_dir_name = loops_name_list{dynamic_index};
elseif dir_index == 8
dynamic_dir_name = kinect_tum_name_list{dynamic_index};
elseif dir_index == 9 || dir_index == 10
dynamic_dir_name = loops2_name_list{dynamic_index};
elseif dir_index == 11
dynamic_dir_name = sparse_feature_name_list{dynamic_index};
elseif dir_index == 12
dynamic_dir_name = swing_name_list{dynamic_index};
elseif dir_index == 13
dynamic_dir_name = swing2_name_list{dynamic_index};
elseif dir_index == 14
dynamic_dir_name = motive_name_list{dynamic_index};
elseif dir_index == 15
dynamic_dir_name = object_recognition_name_list{dynamic_index};
elseif dir_index == 16
dynamic_dir_name = map_name_list{dynamic_index};
else
dynamic_dir_name = 'none';
end
end
|
github
|
rising-turtle/slam_matlab-master
|
graphslam_addpath.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/graphslam_addpath.m
| 486 |
utf_8
|
9c4c53f3ba485b8dc6124acfb412717c
|
% Add the path for graph slam
%
% Author : Soonhac Hong ([email protected])
% Date : 10/16/12
function graphslam_addpath
addpath('D:\Soonhac\SW\gtsam-toolbox-2.3.0-win64\toolbox');
addpath('D:\soonhac\SW\kdtree');
addpath('D:\soonhac\SW\LevenbergMarquardt');
addpath('D:\soonhac\SW\Localization');
addpath('D:\soonhac\SW\SIFT\sift-0.9.19-bin\sift');
addpath('D:\soonhac\SW\slamtoolbox\slamToolbox_11_09_08\FrameTransforms\Rotations');
addpath('D:\Soonhac\SW\plane_fitting_code');
end
|
github
|
rising-turtle/slam_matlab-master
|
Load_kinect_gt.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/Load_kinect_gt.m
| 1,147 |
utf_8
|
504b68074215f361df0680fcfb6e0467
|
% Load data from Kinect data
%
% Parameters
% data_name : the directory name of data
% dm : index of directory of data
% j : index of frame
%
% Author : Soonhac Hong ([email protected])
% Date : 4/20/11
function [gt rgbdslam rtime] = Load_kinect_gt(dm, dis)
if nargin < 2
dis = 0;
end
t_load = tic;
%dir_name={'rgbd_dataset_freiburg1_xyz'};
dir_name_list = get_kinect_tum_dir_name();
[file_name, err] = sprintf('E:/data/kinect_tum/%s/groundtruth.txt',dir_name_list{dm});
[time tx ty tz qx qy qz qw] = textread(file_name,'%f %f %f %f %f %f %f %f','commentstyle','shell');
prefix_name = strrep(dir_name_list{dm}, 'rgbd_dataset_','');
[rgbdslam_file_name, err] = sprintf('E:/data/kinect_tum/%s/%s-rgbdslam.txt',dir_name_list{dm},prefix_name);
[stime stx sty stz sqx sqy sqz sqw] = textread(rgbdslam_file_name,'%f %f %f %f %f %f %f %f','commentstyle','shell');
rtime = toc(t_load);
gt = [time tx ty tz qx qy qz qw];
rgbdslam = [stime stx sty stz sqx sqy sqz sqw];
if dis == 1
plot3(tx, ty, tz, 'r-');
hold on;
plot3(stx, sty, stz, 'b:');
hold off;
grid;
xlabel('X');
ylabel('Y');
zlabel('Z');
end
end
|
github
|
rising-turtle/slam_matlab-master
|
convert_vro_toro.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/convert_vro_toro.m
| 9,855 |
utf_8
|
e24f66fba9a7a135e5d337a5e0c131dc
|
% Conver the results of VRO to the vertex and edges for TORO
%
% Author : Soonhac Hong ([email protected])
% Date : 2/22/12
function [vro_size] = convert_vro_toro(data_index, dynamic_index, nFrame, g2o_file_name, g2o_dir_name, feature_flag, index_interval, cmp_option, dense_index, sparse_interval, dis)
if nargin < 9
dis = 0;
end
% Load the result of VRO
disp('Load the result from VRO.');
[f_index, t_pose, o_pose, feature_points] = load_vro(data_index,dynamic_index, nFrame, feature_flag); % t_pose[mm], o_pose[degree]
if isempty(feature_points)
feature_flag = 0;
end
vro_size = size(t_pose,1)
% Interploation for missing constraints
%[f_index, t_pose, o_pose] = compensate_vro(f_index, t_pose, o_pose, cmp_option);
% Interploation for missing constraints
if feature_flag == 0
[f_index, t_pose, o_pose, feature_points] = compensate_vro(f_index, t_pose, o_pose, feature_points, cmp_option);
else
[f_index, t_pose, o_pose, feature_points(:,1:2)] = compensate_vro(f_index, t_pose, o_pose, feature_points(:,1:2), cmp_option);
end
% Convert the odometry and feature points to the global poses
disp('Convert the VRO and feauture points w.r.t the glabal frame.');
%dense_index = 1;
[pose_index, e_t_pose e_o_pose e_ftps] = convert_o2p(data_index, dynamic_index, f_index, t_pose, o_pose, feature_points, dense_index, sparse_interval);
% Generate an index of feature points on the global poses
disp('Generate an index of feature points on the global poses.');
g_fpts_edges = [];
g_fpts = [];
if feature_flag == 1
global_fpts_index = max(pose_index);
for i=1:size(e_ftps,1)
unit_cell = e_ftps{i,1};
for j = 1:size(unit_cell,1)
if isempty(g_fpts_edges)/home/soonhac
global_fpts_index = global_fpts_index + 1;
new_index = global_fpts_index;
g_fpts = [g_fpts; new_index unit_cell(j,2:4)];
else
[duplication_index, duplication_flag] = check_duplication(g_fpts_edges(:,2:5), unit_cell(j,:));
if duplication_flag == 0
global_fpts_index = global_fpts_index + 1;
new_index = global_fpts_index;
g_fpts = [g_fpts; new_index unit_cell(j,2:4)];
else
new_index = duplication_index;
end
end
g_fpts_edges = [g_fpts_edges; unit_cell(j,1) new_index unit_cell(j,2:4)];
end
end
end
% plot pose
if dis == 1
if data_index == 10 || data_index == 11
gt_dis = 1;
else
gt_dis = 0;
end
plot_trajectory(e_t_pose, g_fpts, gt_dis, dynamic_index);
end
if feature_flag == 1
feature_pose_name = 'pose_feature';
else
feature_pose_name = 'pose';
end
% Write Vertexcies and Edges
data_name_list = {'VERTEX2','EDGE2','VERTEX3','EDGE3'};
fpts_name_list ={'VERTEX_XY','EDGE_SE2_XY'};
g2o_file_name_final = sprintf('%s%s_%s_%s_%d.graph',g2o_file_name, g2o_dir_name, cmp_option, feature_pose_name, vro_size)
g2o_fd = fopen(g2o_file_name_final,'w');
t_cov = 0.034; % [m] in SR4000
o_cov = 0.5 * pi / 180; % [rad] in SR4000
%cov_mat = [t_cov 0 0; 0 t_cov 0; 0 0 o_cov];
cov_mat = zeros(6,6);
for i=1:size(cov_mat,1)
if i > 3
cov_mat(i,i) = o_cov;
else
cov_mat(i,i) = t_cov;
end
end
info_mat = cov_mat^-1;
sqrt_info_mat = info_mat; %sqrt(info_mat);
e_t_pose = e_t_pose/1000; % [mm] -> [m]
t_pose = t_pose / 1000; %[mm] -> [m]
o_pose = o_pose * pi / 180; % [degree] -> [radian]
%Convert the euler angles to quaterion
% disp('Convert the euler angles to quaterion.');
% for i=1:size(o_pose,1)
% temp_rot = euler_to_rot(o_pose(i,1),o_pose(i,2),o_pose(i,3));
% o_pose_quat(i,:) = R2q(temp_rot);
% %o_pose_quat(i,:) = e2q([o_pose(i,2),o_pose(i,1),o_pose(i,3)]);
% end
%
% for i=1:size(e_o_pose,1)
% temp_rot = euler_to_rot(e_o_pose(i,1), e_o_pose(i,2), e_o_pose(i,3));
% e_o_pose_quat(i,:) = R2q(temp_rot);
% %e_o_pose_quat(i,:) = e2q([e_o_pose(i,2),e_o_pose(i,1),e_o_pose(i,3)]);
% end
if ~isempty(g_fpts)
g_fpts(:,2:4) = g_fpts(:,2:4) / 1000; %[mm] -> [m]
end
if ~isempty(g_fpts_edges)
g_fpts_edges(:,3:5) = g_fpts_edges(:,3:5) / 1000; % [mm] -> [m]
end
f_index(:,2) = f_index(:,2); % + index_interval; TODO : Generalize !!!
for i=1:size(e_t_pose,1)
%fprintf(g2o_fd,'%s %d %f %f %f\n', data_name_list{3}, pose_index(i)-1, e_t_pose(i,1:2), e_o_pose(i,3));
%if (max(t_pose(i,:)) >= t_cov *4) || (max(o_pose(i,:)) >= o_cov*4)
fprintf(g2o_fd,'%s %d %f %f %f %f %f %f \n', data_name_list{3}, pose_index(i)-1, e_t_pose(i,1:3), e_o_pose(i,2), e_o_pose(i,1), e_o_pose(i,3));
%end
end
%TODO Change the format for 3D
for i=1:size(f_index,1)-1
%if (max(t_pose(i,:)) >= t_cov *4) || (max(o_pose(i,:)) >= o_cov*4)
%fprintf(g2o_fd,'%s %d %d %f %f %f %f %f %f %f %f %f\n', data_name_list{4}, f_index(i,1)-1, f_index(i,2)-1, t_pose(i,1), t_pose(i,2), o_pose(i,3), sqrt_info_mat(1,1), sqrt_info_mat(1,2), sqrt_info_mat(1,3), sqrt_info_mat(2,2), sqrt_info_mat(2,3), sqrt_info_mat(3,3));
%fprintf(g2o_fd,'%s %d %d %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f \n', data_name_list{4}, f_index(i,1)-1, f_index(i,2)-1, t_pose(i,1:3), o_pose(i,:), sqrt_info_mat(1,:), sqrt_info_mat(2,2:6), sqrt_info_mat(3,3:6), sqrt_info_mat(4,4:6), sqrt_info_mat(5,5:6), sqrt_info_mat(6,6));
fprintf(g2o_fd,'%s %d %d %f %f %f %f %f %f \n', data_name_list{4}, f_index(i,1)-1, f_index(i,2)-1, t_pose(i,1:3), o_pose(i,2), o_pose(i,1), o_pose(i,3));
%end
end
if feature_flag == 1
for i=1:size(g_fpts,1)
fprintf(g2o_fd,'%s %d %f %f \n', fpts_name_list{1}, g_fpts(i,1:3));
end
for i=1:size(g_fpts_edges,1)
fprintf(g2o_fd,'%s %d %d %f %f %f %f %f \n', fpts_name_list{2}, g_fpts_edges(i,1:4), t_covariance, 0, t_covariance);
end
end
fclose(g2o_fd);
end
function [isExist previous_index] = getPreviousIndex(data_set,pts)
isExist = 0;
previous_index = 0;
distance_threshold = 41; % [mm]; typical absolute accuracy + 3 * typical repeatibility of SR4000 = 20 + 3 * 7 = 41
for i=1:size(data_set,1)
if data_set(i,1) > 0 && data_set(i,1) == pts(1) % Skip non-valid data
distance = sqrt(sum((data_set(i,3:5)-pts(4:6)).^2));
if distance <= distance_threshold
isExist = 1;
previous_index = data_set(i,2);
break;
end
end
end
end
function plot_trajectory(e_pose, fpts, gt_dis, dynamic_index)
e_pose = e_pose/1000;
figure;
if dynamic_index == 18
cube2m = 0.3048; % 1cube = 12 inch = 0.3048 m
gt_x=[];
gt_y=[];
[px,py] = plot_arc([1*cube2m;5*cube2m],[0;5*cube2m],[1*cube2m;6*cube2m]);
gt_x =[gt_x px];
gt_y =[gt_y py];
[px,py] = plot_arc([12*cube2m;5*cube2m],[12*cube2m;6*cube2m],[13*cube2m;5*cube2m]);
gt_x =[gt_x px];
gt_y =[gt_y py];
[px,py] = plot_arc([12*cube2m;0],[13*cube2m;0],[12*cube2m;-1*cube2m]);
gt_x =[gt_x px];
gt_y =[gt_y py];
[px,py] =plot_arc([1*cube2m;0],[1*cube2m;-1*cube2m],[0;0]);
gt_x =[gt_x px gt_x(1)];
gt_y =[gt_y py gt_y(1)];
elseif dynamic_index >= 15 && dynamic_index <= 17
gt_x = [0 0 2.135 2.135 0];
gt_y = [0 1.220 1.220 0 0];
else
inch2m = 0.0254; % 1 inch = 0.0254 m
gt_x = [0 0 150 910 965 965 910 50 0 0];
gt_y = [0 24 172.5 172.5 122.5 -122.5 -162.5 -162.5 -24 0];
gt_x = [gt_x 0 0 60 60+138 60+138+40 60+138+40 60+138 60 0 0];
gt_y = [gt_y 0 24 38.5+40 38.5+40 38.5 -38.5 -38.5-40 -38.5-40 -24 0];
gt_x = gt_x * inch2m;
gt_y = gt_y * inch2m;
end
plot(e_pose(:,1),e_pose(:,2),'b.-','LineWidth',2);
hold on;
if gt_dis == 1
plot(gt_x,gt_y,'r-','LineWidth',2);
end
if ~isempty(fpts)
plot(fpts(:,2),fpts(:,3),'gd');
if gt_dis == 1
legend('Estimated Pose','Estimated Truth','feature points');
else
legend('Estimated Pose','feature points');
end
else
if gt_dis == 1
legend('Estimated Pose','Estimated Truth');
else
legend('Estimated Pose');
end
end
xlabel('X [mm]');
ylabel('Y [mm]');
grid;
h_xlabel = get(gca,'XLabel');
set(h_xlabel,'FontSize',12,'FontWeight','bold');
h_ylabel = get(gca,'YLabel');
set(h_ylabel,'FontSize',12,'FontWeight','bold');
ylim([-18 6]);
axis equal;
hold off;
figure;
gt_x = [0 0 2135 2135 0];
gt_y = [0 1220 1220 0 0];
gt_z = [0 0 0 0 0];
plot3(e_pose(:,1),e_pose(:,2),e_pose(:,3),'bo-');
hold on;
plot3(gt_x,gt_y,gt_z,'r-','LineWidth',2);
if ~isempty(fpts)
plot3(fpts(:,2),fpts(:,3),fpts(:,4),'gd');
legend('Estimated Pose','Estimated Truth','feature points');
else
legend('Estimated Pose','Estimated Truth');
end
xlabel('X [mm]');
ylabel('Y [mm]');
zlabel('Z [mm]');
grid;
h_xlabel = get(gca,'XLabel');
set(h_xlabel,'FontSize',12,'FontWeight','bold');
h_ylabel = get(gca,'YLabel');
set(h_ylabel,'FontSize',12,'FontWeight','bold');
legend('Estimated Pose','Estimated Truth');
end
function [px, py] = plot_arc(P0,P1,P2)
n = 50; % The number of points in the arc
v1 = P1-P0;
v2 = P2-P0;
c = det([v1,v2]); % "cross product" of v1 and v2
a = linspace(0,atan2(abs(c),dot(v1,v2)),n); % Angle range
v3 = [0,-c;c,0]*v1; % v3 lies in plane of v1 and v2 and is orthog. to v1
v = v1*cos(a)+((norm(v1)/norm(v3))*v3)*sin(a); % Arc, center at (0,0)
px = v(1,:) + P0(1);
py = v(2,:) + P0(2);
%plot(v(1,:)+P0(1),v(2,:)+P0(2),'r-','LineWidth',2) % Plot arc, centered at P0
%hold on;
%axis equal
end
|
github
|
rising-turtle/slam_matlab-master
|
get_timestamp_kinect_tum.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/get_timestamp_kinect_tum.m
| 568 |
utf_8
|
fd152c23df0229d3e879a91f12ea7ed7
|
% Get time stamp of depth image in kinect_tum dataset
%
% Author : Soonhac Hong ([email protected])
% Date : 12/20/12
function [time_stamp] = get_timestamp_kinect_tum(dm,j)
dir_name = get_kinect_tum_dir_name();
[depth_data_dir, err] = sprintf('E:/data/kinect_tum/%s/depth',dir_name{dm});
dirData = dir(depth_data_dir); %# Get the data for the current directory
dirIndex = [dirData.isdir]; %# Find the index for directories
file_list = {dirData(~dirIndex).name}';
[file_name, err]=sprintf('%s',file_list{j});
time_stamp = str2num(strrep(file_name, '.png',''));
end
|
github
|
rising-turtle/slam_matlab-master
|
analyze_pose_std.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/analyze_pose_std.m
| 2,255 |
utf_8
|
d715dbcac4fa5d655d84870c9c2162d4
|
% Analyze standard deviation of motion estimation by VRO, VRO_ICP.
%
% Author : Soonhac Hong ([email protected])
% Date : 3/1/13
function [median_pose_std, std_pose_std, unique_step, step1_pose_std_total, step1_pose_std] = analyze_pose_std(f_index, pose_std, verbosity)
step = abs(f_index(:,1)-f_index(:,2));
unique_step = unique(step);
plot_colors={'gd','k*','md','rs','bv','r+','m*','g^','b<','r>','mp','gh'};
title_list = {'ry','rx','rz','x','y','z'};
median_pose_std = zeros(size(unique_step,1),6);
std_pose_std = zeros(size(unique_step,1),6);
% Show all standard deviation
for k=1:6
% if verbosity == 1
% figure;
% end
for i=1:size(unique_step,1)
step_idx = find(step == unique_step(i));
% if verbosity == 1
% if k < 4
% plot(pose_std(step_idx,k).*180./pi,plot_colors{i});
% else
% plot(pose_std(step_idx,k),plot_colors{i});
% end
% hold on;
% end
median_pose_std(i,k) = median(pose_std(step_idx,k));
std_pose_std(i,k) = std(pose_std(step_idx,k));
end
if verbosity == 1
%hold off;
%title(title_list{k});
%legend('s1','s2','s3','s4','s5');
figure;
%plot(median_pose_std(:,k),'o-');
if k < 4
errorbar(median_pose_std(:,k).*180./pi,std_pose_std(:,k).*180./pi,'bo-');
ylabel('[degree]');
else
errorbar(median_pose_std(:,k),std_pose_std(:,k),'bo-');
ylabel('[m]');
end
title(title_list{k});
end
end
%Save step1 pose std
step1_pose_std_total=[];
step1_pose_std=[];
step1_pose_std_temp=[];
for i=1:size(step,1)
if step(i) == 1
step1_pose_std_total(i,:) = pose_std(i,:);
step1_pose_std_temp = pose_std(i,:);
step1_pose_std = [step1_pose_std; pose_std(i,:)];
else
if ~isempty(step1_pose_std_temp)
step1_pose_std_total(i,:) = step1_pose_std_temp;
else
step1_pose_std_total(i,:) = pose_std(i,:);
end
end
end
% % plot step1
% figure(3);
% plot(step1_pose_std_debug(:,1:3));
% legend('x','y','z');
% grid;
% figure(4);
% plot(step1_pose_std_debug(:,4:6).*(180/pi));
% legend('rx','ry','rz');
% grid;
end
|
github
|
rising-turtle/slam_matlab-master
|
convert_vro_g2o.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/convert_vro_g2o.m
| 7,261 |
utf_8
|
c5b3ce08193f250cd2c7f03abd91f261
|
% Conver the results of VRO to the vertex and edges for g2o
%
% Author : Soonhac Hong ([email protected])
% Date : 2/22/12
function [vro_size] = convert_vro_g2o(data_index, dynamic_index, nFrame, g2o_file_name, g2o_dir_name, feature_flag, index_interval, cmp_option, dense_index, sparse_interval, dis)
if nargin < 9
dis = 0;
end
% Load the result of VRO
disp('Load the result from VRO.');
[f_index, t_pose, o_pose, feature_points] = load_vro(data_index,dynamic_index, nFrame, feature_flag); % t_pose[mm], o_pose[degree]
if isempty(feature_points)
feature_flag = 0;
end
vro_size = size(t_pose,1)
% Interploation for missing constraints
if feature_flag == 0
[f_index, t_pose, o_pose, feature_points] = compensate_vro(f_index, t_pose, o_pose, feature_points, cmp_option);
else
[f_index, t_pose, o_pose, feature_points(:,1:2)] = compensate_vro(f_index, t_pose, o_pose, feature_points(:,1:2), cmp_option);
end
% Convert the odometry and feature points to the global poses
disp('Convert the VRO and feauture points w.r.t the glabal frame.');
%dense_index = 1;
[pose_index, e_t_pose, e_o_pose, e_ftps] = convert_o2p(data_index, dynamic_index, f_index, t_pose, o_pose, feature_points, dense_index, sparse_interval); %convert_o2p(f_index, t_pose, o_pose, feature_points, dense_index);
% e_t_pose [mm]
% e_o_pose [radian]
% Generate an index of feature points on the global poses
disp('Generate an index of feature points on the global poses.');
if feature_flag == 1
[g_fpts g_fpts_edges] = convert_fpts2g(pose_index, e_fpts);
else
g_fpts=[];
g_fpts_edges=[];
end
% plot pose
if dis == 1
if data_index == 10
gt_dis = 1;
else
gt_dis = 0;
end
plot_trajectory(e_t_pose, g_fpts, gt_dis);
end
% Write Vertexcies and Edges
if feature_flag == 1
feature_pose_name = 'pose_feature';
else
feature_pose_name = 'pose';
end
data_name_list = {'VERTEX_SE2','EDGE_SE2','VERTEX_SE3:QUAT','EDGE_SE3:QUAT'};
fpts_name_list ={'VERTEX_XY','EDGE_SE2_XY'};
g2o_file_name_final = sprintf('%s%s_%s_%s_%d.g2o',g2o_file_name, g2o_dir_name, cmp_option, feature_pose_name, vro_size)
g2o_fd = fopen(g2o_file_name_final,'w');
t_cov = 0.034; % [m] in SR4000
o_cov = 0.5 * pi / 180; % [rad] in SR4000
%cov_mat = [t_cov 0 0; 0 t_cov 0; 0 0 o_cov];
cov_mat = zeros(6,6);
for i=1:size(cov_mat,1)
if i > 3
cov_mat(i,i) = o_cov;
else
cov_mat(i,i) = t_cov;
end
end
info_mat = cov_mat^-1;
sqrt_info_mat = info_mat; %sqrt(info_mat);
e_t_pose = e_t_pose/1000; % [mm] -> [m]
t_pose = t_pose / 1000; %[mm] -> [m]
%o_pose = o_pose * pi / 180; % [degree] -> [radian]
e_o_pose = e_o_pose * 180 /pi; % [radian] -> [degree]
%Convert the euler angles to quaterion
disp('Convert the euler angles to quaterion.');
for i=1:size(o_pose,1)
temp_rot = euler_to_rot(o_pose(i,1),o_pose(i,2),o_pose(i,3)); % input(ry, rx, rz) [degree]
o_pose_quat(i,:) = R2q(temp_rot);
%o_pose_quat(i,:) = e2q([o_pose(i,2),o_pose(i,1),o_pose(i,3)]);
end
for i=1:size(e_o_pose,1)
temp_rot = euler_to_rot(e_o_pose(i,1), e_o_pose(i,2), e_o_pose(i,3)); % input(ry, rx, rz) [degree]
e_o_pose_quat(i,:) = R2q(temp_rot);
%e_o_pose_quat(i,:) = e2q([e_o_pose(i,2),e_o_pose(i,1),e_o_pose(i,3)]);
end
if ~isempty(g_fpts)
g_fpts(:,2:4) = g_fpts(:,2:4) / 1000; %[mm] -> [m]
end
if ~isempty(g_fpts_edges)
g_fpts_edges(:,3:5) = g_fpts_edges(:,3:5) / 1000; % [mm] -> [m]
end
f_index(:,2) = f_index(:,2) + index_interval;
for i=1:size(e_t_pose,1)
%fprintf(g2o_fd,'%s %d %f %f %f\n', data_name_list{3}, pose_index(i)-1, e_t_pose(i,1:2), e_o_pose(i,3));
%if (max(t_pose(i,:)) >= t_cov *4) || (max(o_pose(i,:)) >= o_cov*4)
%fprintf(g2o_fd,'%s %d %f %f %f %f %f %f %f\n', data_name_list{3}, pose_index(i)-1, e_t_pose(i,1:3), e_o_pose_quat(i,2), e_o_pose_quat(i,1), e_o_pose_quat(i,3));
fprintf(g2o_fd,'%s %d %f %f %f %f %f %f %f\n', data_name_list{3}, pose_index(i)-1, e_t_pose(i,1:3), e_o_pose_quat(i,:));
%end
end
for i=1:size(f_index,1)-1 %vro_size-1
%if (max(t_pose(i,:)) >= t_cov *4) || (max(o_pose(i,:)) >= o_cov*4)
%fprintf(g2o_fd,'%s %d %d %f %f %f %f %f %f %f %f %f\n', data_name_list{4}, f_index(i,1)-1, f_index(i,2)-1, t_pose(i,1), t_pose(i,2), o_pose(i,3), sqrt_info_mat(1,1), sqrt_info_mat(1,2), sqrt_info_mat(1,3), sqrt_info_mat(2,2), sqrt_info_mat(2,3), sqrt_info_mat(3,3));
fprintf(g2o_fd,'%s %d %d %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f %f\n', data_name_list{4}, f_index(i,1)-1, f_index(i,2)-1, t_pose(i,1:3), o_pose_quat(i,:), sqrt_info_mat(1,:), sqrt_info_mat(2,2:6), sqrt_info_mat(3,3:6), sqrt_info_mat(4,4:6), sqrt_info_mat(5,5:6), sqrt_info_mat(6,6));
%end
end
if feature_flag == 1
for i=1:size(g_fpts,1)
fprintf(g2o_fd,'%s %d %f %f \n', fpts_name_list{1}, g_fpts(i,1:3));
end
for i=1:size(g_fpts_edges,1)
fprintf(g2o_fd,'%s %d %d %f %f %f %f %f \n', fpts_name_list{2}, g_fpts_edges(i,1:4), t_covariance, 0, t_covariance);
end
end
fclose(g2o_fd);
end
function [isExist previous_index] = getPreviousIndex(data_set,pts)
isExist = 0;
previous_index = 0;
distance_threshold = 41; % [mm]; typical absolute accuracy + 3 * typical repeatibility of SR4000 = 20 + 3 * 7 = 41
for i=1:size(data_set,1)
if data_set(i,1) > 0 && data_set(i,1) == pts(1) % Skip non-valid data
distance = sqrt(sum((data_set(i,3:5)-pts(4:6)).^2));
if distance <= distance_threshold
isExist = 1;
previous_index = data_set(i,2);
break;
end
end
end
end
function plot_trajectory(e_pose, fpts, gt_dis)
figure;
gt_x = [0 0 2135 2135 0];
gt_y = [0 1220 1220 0 0];
plot(e_pose(:,1),e_pose(:,2),'bo-');
hold on;
if gt_dis == 1
plot(gt_x,gt_y,'r-','LineWidth',2);
end
if ~isempty(fpts)
plot(fpts(:,2),fpts(:,3),'gd');
if gt_dis == 1
legend('Estimated Pose','Estimated Truth','feature points');
else
legend('Estimated Pose','feature points');
end
else
if gt_dis == 1
legend('Estimated Pose','Estimated Truth');
else
legend('Estimated Pose');
end
end
xlabel('X [mm]');
ylabel('Y [mm]');
grid;
h_xlabel = get(gca,'XLabel');
set(h_xlabel,'FontSize',12,'FontWeight','bold');
h_ylabel = get(gca,'YLabel');
set(h_ylabel,'FontSize',12,'FontWeight','bold');
hold off;
figure;
gt_x = [0 0 2135 2135 0];
gt_y = [0 1220 1220 0 0];
gt_z = [0 0 0 0 0];
plot3(e_pose(:,1),e_pose(:,2),e_pose(:,3),'bo-');
hold on;
plot3(gt_x,gt_y,gt_z,'r-','LineWidth',2);
if ~isempty(fpts)
plot3(fpts(:,2),fpts(:,3),fpts(:,4),'gd');
legend('Estimated Pose','Estimated Truth','feature points');
else
legend('Estimated Pose','Estimated Truth');
end
xlabel('X [mm]');
ylabel('Y [mm]');
zlabel('Z [mm]');
grid;
h_xlabel = get(gca,'XLabel');
set(h_xlabel,'FontSize',12,'FontWeight','bold');
h_ylabel = get(gca,'YLabel');
set(h_ylabel,'FontSize',12,'FontWeight','bold');
legend('Estimated Pose','Estimated Truth');
end
|
github
|
rising-turtle/slam_matlab-master
|
load_rossba.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/load_rossba.m
| 5,851 |
utf_8
|
90b136e61c9c391804c988e74286a1b7
|
% Load input and output files of ROS SBA (http://www.ros.org/wiki/sba/Tutorials/IntroductionToSBA)
%
% Author : Soonhac Hong ([email protected])
% Date : 2/28/13
% Note : Camera pose, T, in input files and output files of ROS-SBA is inverse
% transformaton from camera N to camera N+1. In other words,
% camera0=inv(T)*camera1. So to speak, camera1 = T * camera0
function [camera_pose, camera_parameters, landmark_position, landmark_projection] = load_rossba(file_name)
addpath('D:\soonhac\Project\PNBD\SW\ASEE\slamtoolbox\slamToolbox_11_09_08\FrameTransforms\Rotations');
addpath('D:\soonhac\Project\PNBD\SW\ASEE\slamtoolbox\slamToolbox_11_09_08\DataManagement');
fid = fopen(file_name);
if fid < 0
error(['load2D: Cannot open file ' file_name]);
end
% scan all lines into a cell array
columns=textscan(fid,'%s','delimiter','\n');
fclose(fid);
lines=columns{1};
N=size(lines,1);
camera_pose_idx = 1;
landmark_index=1;
landmark_projection=[];
trans = [0 0 0];
rot = e2R([pi(), 0, 0]); % rotate 180 around x-axis from bundler frame to SBA frame;
bundler_T = [rot trans'; 0 0 0 1];
for i=1:N
line_i=lines{i};
%line_data = textscan(line_i,'%s','delimiter',' ');
if i>=2
%line_data = str2num(line_data{1});
line_data = textscan(line_i,'%f %f %f',1);
end
if i==2
camera_index_total = line_data{1};
landmark_index_total = line_data{2};
elseif i>2
if camera_pose_idx <= camera_index_total % camera pose
unit_idx = mod((i - 3),5);
if unit_idx == 0
camera_parameters(camera_pose_idx,:)=[line_data{1}, line_data{2}, line_data{3}];
elseif unit_idx >=1 && unit_idx <=3 % rotation matrix
unit_rot(unit_idx,:)=[line_data{1}, line_data{2}, line_data{3}];
elseif unit_idx == 4 %translation
unit_trans = [line_data{1}, line_data{2}, line_data{3}];
T=[unit_rot unit_trans'; 0 0 0 1];
%T= bundler_T * inv(T); % Convert transform from ROS-SBA convention to VRO convention !!!!!
T = inv(bundler_T * T); % Convert transform from ROS-SBA convention to VRO convention !!!!!
unit_rot=T(1:3,1:3);
unit_trans=T(1:3,4);
[e] = R2e(unit_rot);
camera_pose(camera_pose_idx,:) = [unit_trans', e(1), e(2), e(3) ];
camera_pose_idx = camera_pose_idx + 1;
unit_rot=[];
% elseif unit_idx == 0 && i>=5
end
else % landmark pose
% unit_idx = mod((i-(3+camera_index_total*5)), 3);
% if unit_idx == 0
% landmark_pose(landmark_idx,:) = [line_data{1}, line_data{2}, line_data{3}];
% landmark_idx = landmark_idx + 1;
% end
index_modulus = mod(i-(camera_index_total*5 + 2), 3);
switch index_modulus
case 1 % position
v = textscan(line_i,'%f %f %f',1);
landmark_position(landmark_index,:) = [v{1},v{2},v{3}];
case 2 % color
v = textscan(line_i,'%d %d %d',1);
landmark_color(landmark_index,:) = [v{1},v{2},v{3}];
case 0 % projected positions
v = textscan(line_i,'%f','delimiter',' ');
v = v{1};
for k=1:v(1)
landmark_projection = [landmark_projection; v(2+(k-1)*4),v(3+(k-1)*4),v(4+(k-1)*4), v(5+(k-1)*4)];
end
landmark_index = landmark_index + 1;
end
end
end
end
% if nargin < 4,
% successive=false;
% end
%
% fid = fopen(file_name);
% if fid < 0
% error(['load2D: Cannot open file ' file_name]);
% end
%
% % scan all lines into a cell array
% columns=textscan(fid,'%s','delimiter','\n');
% fclose(fid);
% lines = columns{1};
%
% camera_index = 1;
% landmark_index = 1;
% landmark_projection=[];
% transform = {};
% landmark_position=[];
% landmark_color=[];
% for i=1:size(lines,1)
% line_i=lines{i};
% if i == 2
% v = textscan(line_i,'%d %d',1);
% nCamera=v{1};
% nLandmarks=v{2};
% elseif i > 2 && i <= (nCamera*5 + 2) % camera pose
% v = textscan(line_i,'%f %f %f',1);
% index_modulus = mod(i-2,5);
% switch index_modulus
% case 1
% camera_parameters(camera_index,:)=[v{1},v{2},v{3}];
% case {2,3,4}
% rotation_matrix(index_modulus-1,:) = [v{1},v{2},v{3}];
% case 0
% transform{camera_index} = [rotation_matrix, [v{1},v{2},v{3}]'; 0 0 0 1];
% camera_index = camera_index + 1;
% end
% elseif i >2 %landmark position
% index_modulus = mod(i-(nCamera*5 + 2), 3);
% switch index_modulus
% case 1 % position
% v = textscan(line_i,'%f %f %f',1);
% landmark_position(landmark_index,:) = [v{1},v{2},v{3}];
% case 2 % color
% v = textscan(line_i,'%d %d %d',1);
% landmark_color(landmark_index,:) = [v{1},v{2},v{3}];
% case 0 % projected positions
% v = textscan(line_i,'%f','delimiter',' ');
% v = v{1};
% for k=1:v(1)
% landmark_projection = [landmark_projection; v(2+(k-1)*4),v(3+(k-1)*4),v(4+(k-1)*4), v(5+(k-1)*4)];
% end
% landmark_index = landmark_index + 1;
% end
% end
% end
%
% %% Convert data to the VRO convention
% camera_poses = []; % x, y, z, rx, ry, rz
% for i=1:size(transform,2)
% unit_transform = transform{i};
% camera_poses(i,:) = [unit_transform(1:3,4)' R2e(unit_transform(1:3,1:3))'];
% end
end
|
github
|
rising-turtle/slam_matlab-master
|
Convert_trajectory_threshold.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/Convert_trajectory_threshold.m
| 1,762 |
utf_8
|
72bcdd34eefa4c383e3ba6d58d35b109
|
% Convert the color of a trajectory in the polygon file format (*.ply) for Meshlab
%
% Input parameters :
% input_file_name : Input file name in the polygon file format
% output_file_name : Output file name in the polygon file format
% threshold : Cut-off threshold of intensity values
% Usage :Convert_trajectory_color('pm_trajectory.ply', 'pm_trajectory_new_color.ply', [0,0,255])
%
% Author : Soonhac Hong ([email protected])
% Date : 5/16/14
function Convert_trajectory_threshold(input_file_name, output_file_name, threshold)
%% Load an input file
fid = fopen(input_file_name);
if fid < 0
error(['Cannot open file ' input_file_name]);
end
% scan all lines into a cell array
columns=textscan(fid,'%s','delimiter','\n');
lines=columns{1};
fclose(fid);
%% Convert the color of a trajectory
new_data=[];
for i=1:size(lines,1)
i
line_i=lines{i};
if i > 12 % check intensity data
v = textscan(line_i,'%f %f %f %d %d %d',1);
if v{4} > threshold
new_data=[new_data; [v{1},v{2},v{3},double(v{4}),double(v{5}),double(v{6})]];
%fprintf(fd,'%f %f %f %d %d %d\n',v{1},v{2},v{3}, target_color);
end
end
end
%% Write data
ply_headers={'ply','format ascii 1.0','comment Created with XYZRGB_to_PLY', 'element_vertex_dummy', 'property float x', 'property float y','property float z','property uchar red','property uchar green','property uchar blue','end_header'};
nply_data = size(new_data,1)
element_vertex_n = sprintf('element vertex %d',nply_data);
ply_headers{4} = element_vertex_n;
fd = fopen(output_file_name, 'w');
for i=1:size(ply_headers,2)
fprintf(fd,'%s\n',ply_headers{i});
end
for i=1:nply_data
fprintf(fd,'%f %f %f %d %d %d\n',new_data(i,:));
end
fclose(fd);
end
|
github
|
rising-turtle/slam_matlab-master
|
run_pose_graph_optimization.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/run_pose_graph_optimization.m
| 1,292 |
utf_8
|
1836ed46d2469d0f4edec264b5926f13
|
% Run pose graph optimization
%
% Author : Soonhac Hong ([email protected])
% History :
% 3/26/14 : Created
function [result, graph, initial, h_global, location_file_index, location_info_history]=run_pose_graph_optimization(vro_result, vro_pose_std, graph, initial, h_global, dis, location_flag, location_file_index, location_info_history)
import gtsam.*
t = gtsam.Point3(0, 0, 0);
if isempty(h_global)
h_global = get_global_transformation_single('smart_cane');
rot = h_global(1:3,1:3);
R = gtsam.Rot3(rot);
origin= gtsam.Pose3(R,t);
initial.insert(0,origin);
end
pgc_t=tic;
[graph,initial] = construct_pose_graph(vro_result, vro_pose_std, graph, initial);
first = initial.at(0);
pgc_ct =toc(pgc_t)
graph.add(NonlinearEqualityPose3(0, first));
gtsam_t=tic;
optimizer = LevenbergMarquardtOptimizer(graph, initial);
result = optimizer.optimizeSafely();
gtsam_ct =toc(gtsam_t)
% Show the results in the plot and generate location information
if dis==true
[location_file_index, location_info_history]=plot_graph_initial_result(initial, result, location_flag, location_file_index, location_info_history);
elseif location_flag == true
[location_info, location_file_index] = generate_location_info(result,[], location_file_index); % for finding rooms
end
end
|
github
|
rising-turtle/slam_matlab-master
|
get_global_transformation.m
|
.m
|
slam_matlab-master/ground_truth_zh/GraphSLAM/get_global_transformation.m
| 11,998 |
utf_8
|
57e044f4869271c0316b4c195dadbd53
|
% Get global transformation for each data set
%
% Author : Soonhac Hong ([email protected])
% Date : 10/22/12
function [h_global] = get_global_transformation(data_index, dynamic_index, isgframe)
addpath('..\slamtoolbox\slamToolbox_11_09_08\FrameTransforms\Rotations');
% rx, ry, rz : [degree]
% tx, ty, tz : [mm]
rx=0; ry=0; rz=0; tx=0; ty=0; tz=0;
switch data_index
case 10 % square
switch dynamic_index
case 16
%h_global = [euler_to_rot(0, -15.4, 0) [0 0 0]'; 0 0 0 1]; % square_700
rx = -15.4;
end
case 11 % etas
switch dynamic_index
case 1
%h_global = [euler_to_rot(0, -36.3, 0) [0 0 0]'; 0 0 0 1]; % etas
%rx = -36.3;
rx = -29.5089; ry = 1.1837; rz=3.6372;
case 3
rx = -28.9278; ry = 1.1894; rz=0.8985;
case 5
rx = -29.2181; ry = 1.1830; rz=1.8845;
end
case 12 % loops
switch dynamic_index
case 2
%h_global = [euler_to_rot(1.23, -25.6, 3.51) [0 0 0]'; 0 0 0 1]; % loops_2
rx=-25.6; ry=1.23; rz=3.51;
case 3
%h_global =[euler_to_rot(1.2670, -24.3316, 4.6656) [0 0 0]'; 0 0 0 1]; %loops_3
rx=-24.3316; ry=1.2670; rz=0; %4.6656;
case 13
rx=-25.6109; ry=1.2418; rx=3.8448
end
case 13 % kinect_tum
switch dynamic_index
case 2
% start image : 13050033527.670034
% Groundtruth : 1305033527.6662 1.4906 -1.1681 0.6610 0.8959 0.0713 -0.0460 -0.4361
init_qauterion = [0.8959, 0.0713, -0.046, -0.4361];
% temp_rot = q2R(init_qauterion);
% [r1, r2, r3] = rot_to_euler(temp_rot);
% r2d = 180 / pi;
% temp_rot = euler_to_rot(r2*r2d, r1*r2d, r3*r2d);
temp_r = q2e(init_qauterion) * 180 / pi;
%temp_rot = euler_to_rot(temp_r(2), temp_r(1), temp_r(3)); % degree
%temp_rot = euler_to_rot(45, -45, 90) * temp_rot ;
%[temp_r(2) temp_r(1) temp_r(3)] = rot_to_euler(temp_rot);
temp_trans = [1.4906 -1.1681 0.661]*1000;
%temp_r = temp_r * 180 / pi;
rx = temp_r(1); ry = temp_r(2); rz = temp_r(3);
%rx =0; ry =0; rz=0;
tx = temp_trans(1); ty = temp_trans(2); tz = temp_trans(3);
end
case {14,15} % loops2, amir_vro
switch dynamic_index
case 1
%h_global =[euler_to_rot(1.2670, -24.3316, 4.6656) [0 0 0]'; 0 0 0 1]; %loops_3
rx=-24.4985419798496; ry=1.26190320102629; rz=4.69088809909393; %4.6656;
case 2
rx=-23.3420; ry=1.2739; rz=4.3772;
case 3
rx=-24.5234; ry=1.2565; rz=4.2682;
case 4
rx=-24.2726; ry=1.2606; rz=2.7462;
case 5
rx=-26.6191; ry=1.2173; rz=5.0601; %2.7462;
case 6
rx=-25.7422; ry=1.2396; rz=3.0821;
case 7
rx=-24.6791; ry=1.2562; rz=4.1917; %3.0821;
case 8
%h_global =[euler_to_rot(1.2670, -24.3316, 4.6656) [0 0 0]'; 0 0 0 1]; %loops_3
rx=-23.8464; ry=1.2721; rz= 3.7964; %3.75437022813459;
%init_qauterion = [0.977813222516827,-0.206522692608165,0.00392574917797234,0.0348463456121639];
%temp_r = q2e(init_qauterion) * 180 / pi;
%rx = temp_r(1); ry = temp_r(2); rz = temp_r(3);
case 9
rx=-25.6806; ry=1.2383; rz=3.4203;
case 10
rx=-25.1216960394087; ry=1.25311495911881; rz=-0.909688742543562;
case 11
%h_global = [euler_to_rot(1.23, -25.6, 3.51) [0 0 0]'; 0 0 0 1]; % loops_2
%rx=-25.6; ry=1.23; rz=3.51;
rx=-33.7709; ry=1.0903; rz=3.0738;
case 12 % same as exp 7
rx=-24.6791; ry=1.2562; rz=4.1917; %3.0821;
end
case 16 % sparse_feature
switch dynamic_index
case 1
rx=-29.9267; ry=1.1978; rz=0; %-2.6385; %4.69088809909393; %4.6656;
case 2
rx=-28.7386; ry=1.2247; rz=0.8001; %-2.6385; %4.69088809909393; %4.6656;
case 3
rx=-31.1218; ry=1.1713; rz=-1.7365;
case 4
rx=-29.3799; ry=1.2133; rz=-2.4075;
case {5,6,7,8}
rx = -39.1930; ry = 1.0134; rz = -1.2122;
case {9, 10, 11}
rx = -37.1533; ry = 1.0455; rz = -3.0266;
case {12,16}
rx = -38.3570; ry = 1.0315; rz = -0.3114;
case 13
rx=-28.3021; ry=1.2357; rz=-1.3391;
case 14
rx=-29.1073; ry=1.2133; rz=-0.5520;
end
case {17,18} % swing
switch dynamic_index
case 1
rx=-32.2179; ry=1.1567; rz=-1.1344;
case 2
rx=-28.6379; ry=1.2207; rz=-1.5099;
case 3
rx=-33.3490; ry=1.1279; rz=-3.1225;
case 4
rx=-31.0073; ry=1.1770; rz=-0.6052;
case 5
rx=-31.7611; ry=1.1627; rz=-0.9156;
case 7
rx=-19.2560; ry=0.8450; rz=0.5809;
case 8
rx=-17.5219; ry=0.8546; rz=-4.5522;
case 9
rx=-17.2286; ry=0.8616; rz=-3.4105;
case 10
rx=-18.1945; ry=0.8501; rz=1.4029;
case 11
rx=-18.2734; ry=0.8556; rz=0.2481;
case 12
rx=-24.4356; ry=0.7918; rz=-0.9333;
case 13
rx=-17.2649; ry=0.8549; rz=-2.8848;
case 14
rx=-20.1519; ry=0.8284; rz=-1.3661;
case 15
rx=-20.5691; ry=0.8302; rz=-0.7713;
case 16
rx=-19.8892; ry=0.8389; rz=-1.3603;
case 17
rx=-24.4211; ry=0.7959; rz=-3.0972;
case 18
rx=-21.7207; ry=0.8163; rz=-4.3600;
case 19
rx=-19.3202; ry=0.8454; rz=-2.7863;
case 20
rx=-20.2662; ry=0.8354; rz=-2.5763;
case 21
rx=-19.5310; ry=0.8403; rz=-1.9310;
case 22
rx=-18.3081; ry=0.8536; rz=-4.6978;
case 23
rx=-16.7653; ry=0.8678; rz=-4.2483;
case 24
rx=-16.4797; ry=0.8697; rz=-2.3778;
case 25
rx=-17.3060; ry=0.8587; rz=-3.1753;
case 26
rx=-16.9306; ry=0.8576; rz=-2.6512;
end
case {19} % motive
switch dynamic_index
case 1
rx=-22.2924; ry=0.8198; rz=-4.3367;
case 2
rx=-23.9217; ry=0.7990; rz=-3.1450;
case 11
rx=-17.5025; ry=1.3989; rz=-1.5120;
case 12
rx=-17.1856; ry=1.4018; rz=-0.3795;
case 13
rx=-16.8428; ry=0.8657; rz=-4.1148;
case 15
rx=-20.3861; ry=0.8323; rz=-4.4565;
case 16
rx=-21.2226; ry=0.8261; rz=-6.5253;
case 17
rx=-19.2100; ry=1.3724; rz=-4.0953;
case 18
rx=-18.4614; ry=1.3822; rz=-2.2221;
case 19
rx=-14.2685; ry=0.8810; rz=-2.4514;
case 20
rx=-14.8030; ry=0.8806; rz=-3.6865;
case 21
rx=-18.6913; ry=0.8436; rz=-1.3773;
case 22
rx=-15.3041; ry=0.8815; rz=0.3623;
case 23
%rx=-14.6923; ry=0.8813; rz=0.5994;
rx=-15.8020; ry=0.8704; rz=-1.0940;
case 24
%rx=-13.9262; ry=0.8867; rz=2.9371;
rx=-13.4739; ry=0.8915; rz=-3.8607;
case 25
%rx=-12.5003; ry=0.8973; rz=-1.6229;
rx=-12.5852; ry=0.8966; rz=-1.7487;
case 26
%rx=-13.1931; ry=0.8922; rz=0.3750;
rx=-13.1931; ry=0.8922; rz=0.3750;
case 27
rx=-13.7715; ry=0.8868; rz=0.7914;
case 28
rx=-14.2207; ry=0.8836; rz=-1.8911;
case 29
rx=-13.4205; ry=0.8920; rz=-1.8912;
case 30
rx=-15.2320; ry=0.8793; rz=-0.8757;
case 31
rx=-15.8020; ry=0.8704; rz=-1.0940;
case 32
rx=-18.3082; ry=1.3832; rz=2.6882;
case 33
rx=-18.3023; ry=1.3870; rz=-0.4907;
case 34
rx=-16.3416; ry=1.4141; rz=-1.3016;
case 35
rx=-15.9030; ry=1.4182; rz=-2.7589;
case 36
rx=-18.8863; ry=1.3755; rz=-4.6635;
case 37
rx=-15.9639; ry=1.4180; rz=0.1582;
end
case {20} % object_recognition
switch dynamic_index
case 1
rx=-37.7656; ry=1.0435; rz=-5.2416;
case 2
rx=-9.4696; ry=0.9351; rz=8.0194;
case 3
rx=-49.0312; ry=0.4693; rz=17.9271;
case 4
rx = -102.3278; ry=-0.4342; rz=-4.9235;
case 5
rx=-36.4384; ry=1.0755; rz=22.0853;
case 6
rx=-36.8751; ry=1.0886; rz=24.7719;
case 7
rx = -26.0697; ry=1.2667; rz=3.3011;
case 8
rx = -34.3278; ry=0.6732; rz=3.7436;
case 9
rx = -28.8132; ry=0.7380; rz=10.7045;
case 10
rx = -35.7148; ry=0.6569; rz=-10.9565;
case 11
rx = -27.3853; ry=0.7568; rz=-7.5140;
end
case {21} % map
switch dynamic_index
case 2
rx=-20.9872; ry=0.8237; rz= -1.6569;
case 3
rx=-21.9661; ry=0.8124; rz=1.3211;
case 4
rx=-20.0099; ry=0.8344; rz=3.4412;
case 5
rx=-20.7988; ry=0.8285; rz=-1.4896;
case 6
%rx=-20.8341; ry=0.8281; rz=-1.5208;
rx=-17.5434; ry=1.4023; rz=-0.4495;
case 7
rx=-19.7766; ry=1.3672; rz=-2.9528;
case 8
rx=-20.7362; ry=1.3549; rz=-3.8060;
case 9
rx=-17.1903; ry=1.4010; rz=0.2855;
end
end
%h_global_tf = [euler_to_rot(0, 90, 0) [0 0 0]'; 0 0 0 1];
if strcmp(isgframe, 'gframe')
h_global = [e2R([rx*pi/180, ry*pi/180, rz*pi/180]) [tx ty tz]'; 0 0 0 1];
else
%h_global = [euler_to_rot(ry, rx, rz) [tx ty tz]'; 0 0 0 1];
h_global = [euler_to_rot(rz, rx, ry) [tx ty tz]'; 0 0 0 1];
end
%h_global = h_global * h_global_tf;
%h_global = [euler_to_rot(0, 0, 15) [tx ty tz]'; 0 0 0 1];
%h_global = [euler_to_rot(rz, rx, ry) [tx ty tz]'; 0 0 0 1];
%h_global = [euler_to_rot(0, 0, 0) [0 0 0]'; 0 0 0 1];
%h_global = [euler_to_rot(0, -23.8, 0) [0 0 0]'; 0 0 0 1]; % square_1000
%init_qauterion = [0.8959, 0.0695, -0.0461, -0.4364];
%temp_rot = q2R(init_qauterion);
%[r1, r2, r3] = rot_to_euler(temp_rot);
%r2d = 180 / pi;
%temp_rot = euler_to_rot(r2*r2d, r1*r2d, r3*r2d);
%temp_r = q2v(init_qauterion) * 180 / pi;
%temp_rot = euler_to_rot(temp_r(2), temp_r(1), temp_r(3)); % degree
%temp_trans = [1.4908 -1.1707 0.6603]*1000;
%h_global = [temp_rot temp_trans'; 0 0 0 1]; % kinect_tum
%h_global_temp = [euler_to_rot(90, 0, 90) [0 0 0]'; 0 0 0 1]; % ?????
%h_global_temp = [[1 0 0; 0 0 -1; 0 1 0] [0 0 0]'; 0 0 0 1];
%h_global = h_global * h_global_temp;
%h_global = [euler_to_rot(0, 0, 0) temp_trans'; 0 0 0 1];
end
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.