File size: 15,000 Bytes
033cb94 686c6f1 033cb94 48d3fca 686c6f1 48d3fca 033cb94 b70c559 033cb94 b70c559 eeb4edc 033cb94 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 |
import numpy as np
import cv2
import onnxruntime as ort
from typing import List, Tuple, Union, Literal, Dict
from pydantic import BaseModel
# Configuration for YOLOX model, set path to model / class - name mappings here!
class ObjectDetectionConfig(BaseModel):
"""Configuration for trained YOLOX object detection model."""
# Model path & hyperparameters
object_detection_model_path: str = "./models/yolox_custom-plates-2cls-0.1.onnx"
confidence_threshold: float = 0.50
nms_threshold: float = 0.65
input_shape: Tuple[int] = (640, 640)
# Class specific inputs
class_map: Dict = {0: 'license-plates', 1: 'License_Plate'}
display_map: Dict = {0: 'license-plate', 1: 'license-plate'}
color_map: Dict = {0: (186, 223, 255), 1: (100, 255, 255)}
class Detection:
def __init__(
self,
points: np.ndarray,
class_id: Union[int, None] = None,
score: Union[float, None] = 0.0,
color: Tuple[int, int, int] = (100, 255, 255),
display_name: str = "Box",
centroid_radius: int = 5,
centroid_thickness: int = -1
):
"""
Represents an object detection in the scene.
Stores bounding box, class_id, and other attributes for tracking and visualization.
"""
self.points_xyxy = points
self.class_id = class_id
self.score = score
self.color_bbox = color
self.color_centroid = color
self.radius_centroid = centroid_radius
self.thickness_centroid = centroid_thickness
self.centroid_location: str = "center"
self.display_name: str = display_name
self.track_id: int = None
self.id: int = None
self.active: bool = False
self.status: str = ""
def __repr__(self) -> str:
return f"Detection({str(self.display_name)})"
@property
def bbox_xyxy(self) -> np.ndarray:
return self.points_xyxy
@property
def size(self) -> float:
"""Return the bounding box area in pixels."""
x1, y1, x2, y2 = self.points_xyxy
return (x2 - x1) * (y2 - y1)
def bbox_image(self, image: np.ndarray, buffer: int = 0) -> np.ndarray:
"""Extract the image patch corresponding to this detection"s bounding box."""
x1, y1, x2, y2 = self.points_xyxy
height, width = image.shape[:2]
x1 = max(0, int(x1 - buffer))
y1 = max(0, int(y1 - buffer))
x2 = min(width, int(x2 + buffer))
y2 = min(height, int(y2 + buffer))
return image[y1:y2, x1:x2]
def centroid(self, location: str = None) -> np.ndarray:
"""Get the centroid of the bounding box based on the chosen centroid location."""
if location is None:
location = self.centroid_location
x1, y1, x2, y2 = self.points_xyxy
if location == "center":
centroid_loc = [(x1 + x2) / 2, (y1 + y2) / 2]
elif location == "top":
centroid_loc = [(x1 + x2) / 2, y1]
elif location == "bottom":
centroid_loc = [(x1 + x2) / 2, y2]
elif location == "left":
centroid_loc = [x1, (y1 + y2) / 2]
elif location == "right":
centroid_loc = [x2, (y1 + y2) / 2]
elif location == "upper-left":
centroid_loc = [x1, y1]
elif location == "upper-right":
centroid_loc = [x2, y1]
elif location == "bottom-left":
centroid_loc = [x1, y2]
elif location == "bottom-right":
centroid_loc = [x2, y2]
else:
raise ValueError("Unsupported location type.")
return np.array([centroid_loc], dtype=np.float32)
def draw(
self,
image: np.ndarray,
draw_boxes: bool = True,
draw_centroids: bool = True,
draw_text: bool = True,
draw_projections: bool = False,
fill_text_background: bool = False,
box_display_type: Literal["minimal", "standard"] = "standard",
box_line_thickness: int = 2,
box_corner_length: int = 20,
obfuscate_classes: List[int] = [],
centroid_color: Union[Tuple[int, int, int], None] = None,
centroid_radius: Union[int, None] = None,
centroid_thickness: Union[int, None] = None,
text_position_xy: Tuple[int] = (25, 25),
text_scale: float = 0.8,
text_thickness: int = 2,
) -> np.ndarray:
"""Draw bounding boxes and centroids for the detection.
If fill_text_background is True, the text placed near the centroid is drawn over a blurred
background extracted from the image. Extra padding is added so the background box is taller.
"""
image_processed = image.copy()
if draw_boxes:
object_bbox: np.ndarray = self.bbox_xyxy
bbox_color: Tuple[int, int, int] = self.color_bbox if self.color_bbox is not None else (100, 255, 255)
if object_bbox is not None:
x0 = int(object_bbox[0])
y0 = int(object_bbox[1])
x1 = int(object_bbox[2])
y1 = int(object_bbox[3])
if self.class_id in obfuscate_classes:
roi = image_processed[y0:y1, x0:x1]
if roi.size > 0:
image_processed[y0:y1, x0:x1] = cv2.GaussianBlur(roi, (61, 61), 0)
if box_display_type.strip().lower() == "minimal":
box_corner_length = int(
min(box_corner_length, (x1 - x0) / 2, (y1 - y0) / 2)
)
cv2.line(image_processed, (x0, y0), (x0 + box_corner_length, y0), color=bbox_color, thickness=box_line_thickness)
cv2.line(image_processed, (x0, y0), (x0, y0 + box_corner_length), color=bbox_color, thickness=box_line_thickness)
cv2.line(image_processed, (x1, y0), (x1 - box_corner_length, y0), color=bbox_color, thickness=box_line_thickness)
cv2.line(image_processed, (x1, y0), (x1, y0 + box_corner_length), color=bbox_color, thickness=box_line_thickness)
cv2.line(image_processed, (x0, y1), (x0 + box_corner_length, y1), color=bbox_color, thickness=box_line_thickness)
cv2.line(image_processed, (x0, y1), (x0, y1 - box_corner_length), color=bbox_color, thickness=box_line_thickness)
cv2.line(image_processed, (x1, y1), (x1 - box_corner_length, y1), color=bbox_color, thickness=box_line_thickness)
cv2.line(image_processed, (x1, y1), (x1, y1 - box_corner_length), color=bbox_color, thickness=box_line_thickness)
elif box_display_type.strip().lower() == "standard":
cv2.rectangle(
image_processed,
(x0, y0),
(x1, y1),
color=bbox_color,
thickness=box_line_thickness
)
if draw_projections:
projection_start_centroid: np.ndarray = self.centroid(location="bottom")[0]
if self.velocity is not None:
projection_end_centroid: np.array = np.array([self.centroid(location="bottom")[0] + self.velocity])[0]
else:
projection_end_centroid = projection_start_centroid
projection_start_coords: Tuple[int, int] = (int(projection_start_centroid[0]), int(projection_start_centroid[1]))
projection_end_coords: Tuple[int, int] = (int(projection_end_centroid[0]), int(projection_end_centroid[1]))
cv2.arrowedLine(
image_processed,
projection_start_coords,
projection_end_coords,
color=(100, 255, 255),
thickness=3,
tipLength=0.2
)
centroid: np.ndarray = self.centroid()[0]
centroid_coords: Tuple[int, int] = (int(centroid[0]), int(centroid[1]))
if centroid_color is None:
centroid_color = self.color_centroid
if centroid_radius is None:
centroid_radius = self.radius_centroid
if centroid_thickness is None:
centroid_thickness = self.thickness_centroid
if draw_centroids:
cv2.circle(
image_processed,
centroid_coords,
centroid_radius,
centroid_color,
centroid_thickness,
lineType=cv2.LINE_AA
)
if draw_text:
display_text: str = str(self.display_name)
text_position: Tuple[int, int] = (
centroid_coords[0] + text_position_xy[0],
centroid_coords[1] + text_position_xy[1]
)
if hasattr(self, "score") and self.score:
display_text += f" ({round(self.score, 2)})"
if hasattr(self, "status") and self.status:
display_text += f" ({self.status})"
if self.status == "Waiting":
display_text += f" ({int(self.queue_time_duration)}s)"
if fill_text_background:
font = cv2.FONT_HERSHEY_SIMPLEX
(text_width, text_height), baseline = cv2.getTextSize(display_text, font, text_scale, text_thickness)
pad_x = 0
pad_y = 10
# Calculate rectangle coordinates
rect_x1 = text_position[0] - pad_x
rect_y1 = text_position[1] - text_height - pad_y
rect_x2 = text_position[0] + text_width + pad_x
rect_y2 = text_position[1] + baseline + pad_y
# Ensure coordinates are within image boundaries
rect_x1 = max(0, rect_x1)
rect_y1 = max(0, rect_y1)
rect_x2 = min(image_processed.shape[1], rect_x2)
rect_y2 = min(image_processed.shape[0], rect_y2)
# Extract the region of interest and apply a Gaussian blur
roi = image_processed[rect_y1:rect_y2, rect_x1:rect_x2]
if roi.size > 0:
image_processed[rect_y1:rect_y2, rect_x1:rect_x2] = cv2.GaussianBlur(roi, (31, 31), 0)
cv2.putText(
image_processed,
display_text,
text_position,
fontFace=cv2.FONT_HERSHEY_SIMPLEX,
fontScale=text_scale,
color=centroid_color,
thickness=text_thickness,
lineType=cv2.LINE_AA
)
return image_processed
class YOLOXDetector:
def __init__(
self,
model_path: str,
input_shape: Tuple[int] = (640, 640),
confidence_threshold: float = 0.6,
nms_threshold: float = 0.65,
providers: List[str] = ["CoreMLExecutionProvider", "CUDAExecutionProvider", "CPUExecutionProvider"],
sess_options=ort.SessionOptions(),
):
self.model_path: str = model_path
self.dims: Tuple[int] = input_shape
self.ratio: float = 1.0
self.confidence_threshold: float = confidence_threshold
self.nms_threshold: float = nms_threshold
self.classes: List[str] = ["license-plates", "License_Plate"]
self.categories: List[str] = ["DEFAULT" for _ in range(len(self.classes))]
self.providers: List[str] = providers
self.session = ort.InferenceSession(
self.model_path,
providers=self.providers,
sess_options=sess_options,
)
def nms(self, boxes, scores, nms_thr):
"""Single class NMS implemented in Numpy."""
x1 = boxes[:, 0]
y1 = boxes[:, 1]
x2 = boxes[:, 2]
y2 = boxes[:, 3]
areas = (x2 - x1 + 1) * (y2 - y1 + 1)
order = scores.argsort()[::-1]
keep = []
while order.size > 0:
i = order[0]
keep.append(i)
xx1 = np.maximum(x1[i], x1[order[1:]])
yy1 = np.maximum(y1[i], y1[order[1:]])
xx2 = np.minimum(x2[i], x2[order[1:]])
yy2 = np.minimum(y2[i], y2[order[1:]])
w = np.maximum(0.0, xx2 - xx1 + 1)
h = np.maximum(0.0, yy2 - yy1 + 1)
inter = w * h
ovr = inter / (areas[i] + areas[order[1:]] - inter)
inds = np.where(ovr <= nms_thr)[0]
order = order[inds + 1]
return keep
def multiclass_nms_class_aware(self, boxes, scores, nms_thr, score_thr):
"""Multiclass NMS implemented in Numpy. Class-aware version."""
final_dets = []
num_classes = scores.shape[1]
for cls_ind in range(num_classes):
cls_scores = scores[:, cls_ind]
valid_score_mask = cls_scores > score_thr
if valid_score_mask.sum() == 0:
continue
else:
valid_scores = cls_scores[valid_score_mask]
valid_boxes = boxes[valid_score_mask]
keep = self.nms(valid_boxes, valid_scores, nms_thr)
if len(keep) > 0:
cls_inds = np.ones((len(keep), 1)) * cls_ind
dets = np.concatenate(
[valid_boxes[keep], valid_scores[keep, None], cls_inds], 1
)
final_dets.append(dets)
if len(final_dets) == 0:
return None
return np.concatenate(final_dets, 0)
def multiclass_nms_class_agnostic(self, boxes, scores, nms_thr, score_thr):
"""Multiclass NMS implemented in Numpy. Class-agnostic version."""
cls_inds = scores.argmax(1)
cls_scores = scores[np.arange(len(cls_inds)), cls_inds]
valid_score_mask = cls_scores > score_thr
if valid_score_mask.sum() == 0:
return None
valid_scores = cls_scores[valid_score_mask]
valid_boxes = boxes[valid_score_mask]
valid_cls_inds = cls_inds[valid_score_mask]
keep = self.nms(valid_boxes, valid_scores, nms_thr)
if keep:
dets = np.concatenate(
[valid_boxes[keep], valid_scores[keep, None], valid_cls_inds[keep, None]], 1
)
return dets
def multiclass_nms(self, boxes, scores, nms_thr, score_thr, class_agnostic=False):
"""Multiclass NMS implemented in Numpy"""
if class_agnostic:
return self.multiclass_nms_class_agnostic(boxes, scores, nms_thr, score_thr)
else:
return self.multiclass_nms_class_aware(boxes, scores, nms_thr, score_thr)
def preprocess(self, image: np.ndarray, bgr2rgb: bool = False):
"""Preprocess image for YOLOX model."""
if len(image.shape) == 3:
padded_image = np.ones((self.dims[0], self.dims[1], 3), dtype=np.uint8) * 114
else:
padded_image = np.ones(self.dims, dtype=np.uint8) * 114
if bgr2rgb:
padded_image = cv2.cvtColor(padded_image, cv2.COLOR_BGR2RGB)
self.ratio = min(self.dims[0] / image.shape[0], self.dims[1] / image.shape[1])
resized_image = cv2.resize(
image,
(int(image.shape[1] * self.ratio), int(image.shape[0] * self.ratio)),
interpolation=cv2.INTER_LINEAR,
).astype(np.uint8)
padded_image[: int(image.shape[0] * self.ratio), : int(image.shape[1] * self.ratio)] = resized_image
padded_image = padded_image.transpose((2, 0, 1))
padded_image = np.ascontiguousarray(padded_image, dtype=np.float32)
return padded_image
def postprocess(self, outputs, p64=False):
"""Post-process YOLOX model outputs into usable bounding boxes and scores."""
grids = []
expanded_strides = []
strides = [8, 16, 32] if not p64 else [8, 16, 32, 64]
hsizes = [self.dims[0] // stride for stride in strides]
wsizes = [self.dims[1] // stride for stride in strides]
for hsize, wsize, stride in zip(hsizes, wsizes, strides):
xv, yv = np.meshgrid(np.arange(wsize), np.arange(hsize))
grid = np.stack((xv, yv), 2).reshape(1, -1, 2)
grids.append(grid)
shape = grid.shape[:2]
expanded_strides.append(np.full((*shape, 1), stride))
grids = np.concatenate(grids, 1)
expanded_strides = np.concatenate(expanded_strides, 1)
outputs[..., :2] = (outputs[..., :2] + grids) * expanded_strides
outputs[..., 2:4] = np.exp(outputs[..., 2:4]) * expanded_strides
outputs = outputs[0]
boxes = outputs[:, :4]
scores = outputs[:, 4:5] * outputs[:, 5:]
boxes_xyxy = np.ones_like(boxes)
boxes_xyxy[:, 0] = boxes[:, 0] - boxes[:, 2] / 2.0
boxes_xyxy[:, 1] = boxes[:, 1] - boxes[:, 3] / 2.0
boxes_xyxy[:, 2] = boxes[:, 0] + boxes[:, 2] / 2.0
boxes_xyxy[:, 3] = boxes[:, 1] + boxes[:, 3] / 2.0
boxes_xyxy /= self.ratio
return boxes_xyxy, scores
def predict(self, image: np.ndarray):
"""Run YOLOX detector on an image and return detected bounding boxes and scores."""
image = self.preprocess(image=image)
onnx_pred = self.session.run(None, {self.session.get_inputs()[0].name: np.expand_dims(image, axis=0)})[0]
boxes_xyxy, scores = self.postprocess(onnx_pred)
detections = self.multiclass_nms(
boxes=boxes_xyxy,
scores=scores,
nms_thr=self.nms_threshold,
score_thr=self.confidence_threshold,
class_agnostic=False if len(self.classes) > 1 else True
)
if detections is not None and len(detections) > 0:
final_boxes, final_scores, final_cls_inds = detections[:, :4], detections[:, 4], detections[:, 5]
else:
final_boxes, final_scores, final_cls_inds = np.empty((0, 4)), np.empty((0,)), np.empty((0,))
return final_boxes, final_scores, final_cls_inds |