Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -1,16 +1,16 @@
|
|
1 |
-
import cv2
|
2 |
-
import torch
|
3 |
-
import gradio as gr
|
4 |
-
import numpy as np
|
5 |
import os
|
|
|
6 |
import json
|
7 |
import logging
|
|
|
|
|
|
|
8 |
import matplotlib.pyplot as plt
|
9 |
-
import csv
|
10 |
from datetime import datetime
|
11 |
from collections import Counter
|
12 |
-
import
|
13 |
-
|
|
|
14 |
|
15 |
# Set YOLO config directory
|
16 |
os.environ["YOLO_CONFIG_DIR"] = "/tmp/Ultralytics"
|
@@ -29,35 +29,19 @@ FLIGHT_LOG_DIR = "flight_logs"
|
|
29 |
os.makedirs(CAPTURED_FRAMES_DIR, exist_ok=True)
|
30 |
os.makedirs(OUTPUT_DIR, exist_ok=True)
|
31 |
os.makedirs(FLIGHT_LOG_DIR, exist_ok=True)
|
32 |
-
os.chmod(CAPTURED_FRAMES_DIR, 0o777)
|
33 |
-
os.chmod(OUTPUT_DIR, 0o777)
|
34 |
-
os.chmod(FLIGHT_LOG_DIR, 0o777)
|
35 |
|
36 |
# Global variables
|
37 |
-
log_entries = []
|
38 |
detected_counts = []
|
39 |
detected_issues = []
|
40 |
gps_coordinates = []
|
41 |
-
|
42 |
frame_count = 0
|
43 |
-
SAVE_IMAGE_INTERVAL = 1
|
44 |
-
|
45 |
-
# Detection classes
|
46 |
DETECTION_CLASSES = ["Longitudinal", "Pothole", "Transverse"]
|
47 |
|
48 |
-
# Debug: Check environment
|
49 |
-
print(f"Torch version: {torch.__version__}")
|
50 |
-
print(f"Gradio version: {gr.__version__}")
|
51 |
-
|
52 |
# Load custom YOLO model
|
53 |
device = "cuda" if torch.cuda.is_available() else "cpu"
|
54 |
-
|
55 |
-
model = torch.hub.load("ultralytics/yolov5", "custom", path='./data/best.pt').to(device)
|
56 |
-
if device == "cuda":
|
57 |
-
model.half()
|
58 |
-
print(f"Model classes: {model.names}")
|
59 |
|
60 |
-
# Helper functions for video processing, geotagging, flight logs, and quality checks
|
61 |
def zip_directory(folder_path: str, zip_path: str) -> str:
|
62 |
"""Zip all files in a directory."""
|
63 |
try:
|
@@ -72,7 +56,8 @@ def zip_directory(folder_path: str, zip_path: str) -> str:
|
|
72 |
logging.error(f"Failed to zip {folder_path}: {str(e)}")
|
73 |
return ""
|
74 |
|
75 |
-
def generate_map(gps_coords:
|
|
|
76 |
map_path = os.path.join(OUTPUT_DIR, "map_temp.png")
|
77 |
plt.figure(figsize=(4, 4))
|
78 |
plt.scatter([x[1] for x in gps_coords], [x[0] for x in gps_coords], c='blue', label='GPS Points')
|
@@ -84,50 +69,25 @@ def generate_map(gps_coords: list, items: list) -> str:
|
|
84 |
plt.close()
|
85 |
return map_path
|
86 |
|
87 |
-
def write_geotag(image_path: str, gps_coord:
|
|
|
88 |
try:
|
89 |
-
|
90 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
91 |
return True
|
92 |
except Exception as e:
|
93 |
logging.error(f"Failed to geotag {image_path}: {str(e)}")
|
94 |
return False
|
95 |
|
96 |
-
def
|
97 |
-
log_path = os.path.join(FLIGHT_LOG_DIR, f"flight_log_{frame_count:06d}.csv")
|
98 |
-
try:
|
99 |
-
with open(log_path, 'w', newline='') as csvfile:
|
100 |
-
writer = csv.writer(csvfile)
|
101 |
-
writer.writerow(["Frame", "Timestamp", "Latitude", "Longitude", "Speed_ms", "Satellites", "Altitude_m"])
|
102 |
-
writer.writerow([frame_count, timestamp, gps_coord[0], gps_coord[1], 5.0, 12, 60])
|
103 |
-
return log_path
|
104 |
-
except Exception as e:
|
105 |
-
logging.error(f"Failed to write flight log {log_path}: {str(e)}")
|
106 |
-
return ""
|
107 |
-
|
108 |
-
# Generate HTML report using Jinja2 template
|
109 |
-
def generate_report(detections, video_path, issue_images, flight_logs, chart_path, map_path, submission_json):
|
110 |
-
with open("report_template.html", "r") as file:
|
111 |
-
template = Template(file.read())
|
112 |
-
|
113 |
-
report_content = template.render(
|
114 |
-
detections=detections,
|
115 |
-
video_path=video_path,
|
116 |
-
issue_images=issue_images,
|
117 |
-
flight_logs=flight_logs,
|
118 |
-
chart_path=chart_path,
|
119 |
-
map_path=map_path,
|
120 |
-
submission_json=submission_json
|
121 |
-
)
|
122 |
-
|
123 |
-
report_path = "output_report.html"
|
124 |
-
with open(report_path, "w") as report_file:
|
125 |
-
report_file.write(report_content)
|
126 |
-
|
127 |
-
return report_path
|
128 |
-
|
129 |
-
# Function to process video and generate outputs
|
130 |
-
def process_video(video, resize_width=4000, resize_height=3000, frame_skip=5):
|
131 |
global frame_count, detected_counts, detected_issues, gps_coordinates, log_entries
|
132 |
frame_count = 0
|
133 |
detected_counts.clear()
|
@@ -135,91 +95,88 @@ def process_video(video, resize_width=4000, resize_height=3000, frame_skip=5):
|
|
135 |
gps_coordinates.clear()
|
136 |
log_entries.clear()
|
137 |
|
138 |
-
if video is None:
|
139 |
-
log_entries.append("Error: No video uploaded")
|
140 |
-
return None, json.dumps({"error": "No video uploaded"}, indent=2)
|
141 |
-
|
142 |
cap = cv2.VideoCapture(video)
|
143 |
if not cap.isOpened():
|
144 |
-
|
145 |
return None, json.dumps({"error": "Could not open video file"}, indent=2)
|
146 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
147 |
while True:
|
148 |
ret, frame = cap.read()
|
149 |
if not ret:
|
150 |
break
|
151 |
frame_count += 1
|
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 |
-
video_path="processed_video.mp4",
|
188 |
-
issue_images=detected_issues,
|
189 |
-
flight_logs=logs_zip,
|
190 |
-
chart_path=chart_path,
|
191 |
-
map_path=map_path,
|
192 |
-
submission_json=submission_json_path
|
193 |
-
)
|
194 |
-
|
195 |
-
# Create the final zip file containing all report components
|
196 |
-
final_zip = zipfile.ZipFile("final_report.zip", 'w', zipfile.ZIP_DEFLATED)
|
197 |
-
final_zip.write(report_path)
|
198 |
-
final_zip.write(images_zip)
|
199 |
-
final_zip.write(logs_zip)
|
200 |
-
final_zip.write("processed_video.mp4")
|
201 |
-
final_zip.close()
|
202 |
-
|
203 |
-
return final_zip
|
204 |
-
|
205 |
-
# Gradio Interface
|
206 |
with gr.Blocks() as iface:
|
207 |
-
gr.Markdown("#
|
208 |
-
|
209 |
-
|
210 |
-
|
211 |
-
|
212 |
-
|
213 |
-
|
214 |
-
|
215 |
-
with gr.Column(scale=1):
|
216 |
-
metrics_output = gr.Textbox(label="Detection Metrics", lines=5, interactive=False)
|
217 |
-
|
218 |
-
process_btn.click(
|
219 |
-
fn=process_video,
|
220 |
-
inputs=[video_input, width_slider, height_slider, skip_slider],
|
221 |
-
outputs=[metrics_output]
|
222 |
-
)
|
223 |
-
|
224 |
-
if __name__ == "__main__":
|
225 |
-
iface.launch()
|
|
|
|
|
|
|
|
|
|
|
1 |
import os
|
2 |
+
import zipfile
|
3 |
import json
|
4 |
import logging
|
5 |
+
import cv2
|
6 |
+
import torch
|
7 |
+
import numpy as np
|
8 |
import matplotlib.pyplot as plt
|
|
|
9 |
from datetime import datetime
|
10 |
from collections import Counter
|
11 |
+
from ultralytics import YOLO
|
12 |
+
import piexif
|
13 |
+
import time
|
14 |
|
15 |
# Set YOLO config directory
|
16 |
os.environ["YOLO_CONFIG_DIR"] = "/tmp/Ultralytics"
|
|
|
29 |
os.makedirs(CAPTURED_FRAMES_DIR, exist_ok=True)
|
30 |
os.makedirs(OUTPUT_DIR, exist_ok=True)
|
31 |
os.makedirs(FLIGHT_LOG_DIR, exist_ok=True)
|
|
|
|
|
|
|
32 |
|
33 |
# Global variables
|
|
|
34 |
detected_counts = []
|
35 |
detected_issues = []
|
36 |
gps_coordinates = []
|
37 |
+
log_entries = []
|
38 |
frame_count = 0
|
|
|
|
|
|
|
39 |
DETECTION_CLASSES = ["Longitudinal", "Pothole", "Transverse"]
|
40 |
|
|
|
|
|
|
|
|
|
41 |
# Load custom YOLO model
|
42 |
device = "cuda" if torch.cuda.is_available() else "cpu"
|
43 |
+
model = YOLO('./data/best.pt').to(device)
|
|
|
|
|
|
|
|
|
44 |
|
|
|
45 |
def zip_directory(folder_path: str, zip_path: str) -> str:
|
46 |
"""Zip all files in a directory."""
|
47 |
try:
|
|
|
56 |
logging.error(f"Failed to zip {folder_path}: {str(e)}")
|
57 |
return ""
|
58 |
|
59 |
+
def generate_map(gps_coords: List[List[float]], items: List[Dict[str, Any]]) -> str:
|
60 |
+
"""Generate and save map of detected issue locations."""
|
61 |
map_path = os.path.join(OUTPUT_DIR, "map_temp.png")
|
62 |
plt.figure(figsize=(4, 4))
|
63 |
plt.scatter([x[1] for x in gps_coords], [x[0] for x in gps_coords], c='blue', label='GPS Points')
|
|
|
69 |
plt.close()
|
70 |
return map_path
|
71 |
|
72 |
+
def write_geotag(image_path: str, gps_coord: List[float]) -> bool:
|
73 |
+
"""Add GPS coordinates as EXIF data to an image."""
|
74 |
try:
|
75 |
+
lat, lon = abs(gps_coord[0]), abs(gps_coord[1])
|
76 |
+
lat_ref, lon_ref = ("N" if gps_coord[0] >= 0 else "S"), ("E" if gps_coord[1] >= 0 else "W")
|
77 |
+
exif_dict = piexif.load(image_path) if os.path.exists(image_path) else {"GPS": {}}
|
78 |
+
exif_dict["GPS"] = {
|
79 |
+
piexif.GPSIFD.GPSLatitudeRef: lat_ref,
|
80 |
+
piexif.GPSIFD.GPSLatitude: ((int(lat), 1), (0, 1), (0, 1)),
|
81 |
+
piexif.GPSIFD.GPSLongitudeRef: lon_ref,
|
82 |
+
piexif.GPSIFD.GPSLongitude: ((int(lon), 1), (0, 1), (0, 1))
|
83 |
+
}
|
84 |
+
piexif.insert(piexif.dump(exif_dict), image_path)
|
85 |
return True
|
86 |
except Exception as e:
|
87 |
logging.error(f"Failed to geotag {image_path}: {str(e)}")
|
88 |
return False
|
89 |
|
90 |
+
def process_video(video):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
91 |
global frame_count, detected_counts, detected_issues, gps_coordinates, log_entries
|
92 |
frame_count = 0
|
93 |
detected_counts.clear()
|
|
|
95 |
gps_coordinates.clear()
|
96 |
log_entries.clear()
|
97 |
|
|
|
|
|
|
|
|
|
98 |
cap = cv2.VideoCapture(video)
|
99 |
if not cap.isOpened():
|
100 |
+
logging.error("Could not open video file")
|
101 |
return None, json.dumps({"error": "Could not open video file"}, indent=2)
|
102 |
|
103 |
+
fps = cap.get(cv2.CAP_PROP_FPS)
|
104 |
+
total_frames = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
|
105 |
+
|
106 |
+
out_path = os.path.join(OUTPUT_DIR, "processed_output.mp4")
|
107 |
+
out = cv2.VideoWriter(out_path, cv2.VideoWriter_fourcc(*'mp4v'), fps, (4000, 3000))
|
108 |
+
|
109 |
+
all_detections = []
|
110 |
+
data_lake_submission = {"images": [], "flight_logs": [], "metrics": {}}
|
111 |
+
|
112 |
while True:
|
113 |
ret, frame = cap.read()
|
114 |
if not ret:
|
115 |
break
|
116 |
frame_count += 1
|
117 |
+
results = model(frame)
|
118 |
+
annotated_frame = results[0].plot()
|
119 |
+
|
120 |
+
# Simulate GPS coordinates for each frame
|
121 |
+
gps_coord = [17.385044 + (frame_count * 0.0001), 78.486671 + (frame_count * 0.0001)]
|
122 |
+
gps_coordinates.append(gps_coord)
|
123 |
+
|
124 |
+
frame_detections = []
|
125 |
+
for detection in results[0].boxes:
|
126 |
+
label = model.names[int(detection.cls)]
|
127 |
+
if label in DETECTION_CLASSES:
|
128 |
+
frame_detections.append({"label": label, "box": detection.xyxy[0].cpu().numpy().tolist()})
|
129 |
+
log_entries.append(f"Detected {label} in frame {frame_count}")
|
130 |
+
|
131 |
+
if frame_detections:
|
132 |
+
captured_frame_path = os.path.join(CAPTURED_FRAMES_DIR, f"detected_{frame_count:06d}.jpg")
|
133 |
+
cv2.imwrite(captured_frame_path, annotated_frame)
|
134 |
+
write_geotag(captured_frame_path, gps_coord)
|
135 |
+
|
136 |
+
detected_issues.append(captured_frame_path)
|
137 |
+
data_lake_submission["images"].append({"path": captured_frame_path, "frame": frame_count, "gps": gps_coord})
|
138 |
|
139 |
+
log_path = os.path.join(FLIGHT_LOG_DIR, f"flight_log_{frame_count:06d}.csv")
|
140 |
+
with open(log_path, 'w', newline='') as csvfile:
|
141 |
+
writer = csv.writer(csvfile)
|
142 |
+
writer.writerow(["Frame", "Latitude", "Longitude", "Timestamp"])
|
143 |
+
writer.writerow([frame_count, gps_coord[0], gps_coord[1], datetime.now().strftime("%Y-%m-%d %H:%M:%S")])
|
144 |
|
145 |
+
data_lake_submission["flight_logs"].append({"path": log_path, "frame": frame_count})
|
146 |
+
|
147 |
+
out.write(annotated_frame)
|
148 |
+
|
149 |
+
cap.release()
|
150 |
+
out.release()
|
151 |
+
|
152 |
+
# Generate the map and trend chart
|
153 |
+
map_path = generate_map(gps_coordinates, all_detections)
|
154 |
+
trend_chart_path = os.path.join(OUTPUT_DIR, "detection_trend.png")
|
155 |
+
plt.plot(detected_counts)
|
156 |
+
plt.savefig(trend_chart_path)
|
157 |
+
plt.close()
|
158 |
+
|
159 |
+
# Compile everything into a single ZIP file
|
160 |
+
zip_path = os.path.join(OUTPUT_DIR, "final_report.zip")
|
161 |
+
with zipfile.ZipFile(zip_path, 'w', zipfile.ZIP_DEFLATED) as zipf:
|
162 |
+
zipf.write(out_path, os.path.basename(out_path)) # Add processed video
|
163 |
+
zipf.write(map_path, os.path.basename(map_path)) # Add map
|
164 |
+
zipf.write(trend_chart_path, os.path.basename(trend_chart_path)) # Add trend chart
|
165 |
+
zipf.write("data_lake_submission.json", "data_lake_submission.json") # Add submission JSON
|
166 |
+
zipf = zip_directory(CAPTURED_FRAMES_DIR, zip_path) # Add captured frames
|
167 |
+
zipf = zip_directory(FLIGHT_LOG_DIR, zip_path) # Add flight logs
|
168 |
+
|
169 |
+
return zip_path
|
170 |
+
|
171 |
+
|
172 |
+
# Gradio interface (keep unchanged)
|
173 |
+
import gradio as gr
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
174 |
with gr.Blocks() as iface:
|
175 |
+
gr.Markdown("# Drone Analysis Report")
|
176 |
+
video_input = gr.Video(label="Upload Video")
|
177 |
+
process_btn = gr.Button("Generate Report")
|
178 |
+
zip_output = gr.File(label="Download Final Report (ZIP)")
|
179 |
+
|
180 |
+
process_btn.click(fn=process_video, inputs=[video_input], outputs=[zip_output])
|
181 |
+
|
182 |
+
iface.launch()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|