File size: 15,393 Bytes
82ae22b |
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 |
"""
Vehicle Detection, Tracking, Counting, and Speed Estimation System
===================================================================
A comprehensive computer vision pipeline for analyzing traffic videos,
detecting vehicles, tracking their movement, counting them, and estimating
their speeds using YOLO object detection and perspective transformation.
Authors:
- Abhay Gupta (0205CC221005)
- Aditi Lakhera (0205CC221011)
- Balraj Patel (0205CC221049)
- Bhumika Patel (0205CC221050)
Technical Approach:
- YOLO for real-time object detection
- ByteTrack for multi-object tracking
- Perspective transformation for speed calculation
- Line zones for vehicle counting
"""
import sys
import logging
from pathlib import Path
from typing import Dict, Optional, Callable
from time import time
import cv2
import numpy as np
import supervision as sv
from ultralytics import YOLO
from src import FrameAnnotator, VehicleSpeedEstimator, PerspectiveTransformer
from src.exceptions import (
VideoProcessingError,
ModelLoadError,
ConfigurationError
)
from config import VehicleDetectionConfig
# Configure logging
logging.basicConfig(
level=logging.INFO,
format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
)
logger = logging.getLogger(__name__)
class VehicleDetectionPipeline:
"""
Main pipeline for vehicle detection, tracking, counting, and speed estimation.
This class orchestrates the entire processing workflow, from loading the model
to processing each frame and generating the output video.
"""
def __init__(self, config: VehicleDetectionConfig):
"""
Initialize the detection pipeline.
Args:
config: Configuration object with all parameters
Raises:
ModelLoadError: If model cannot be loaded
ConfigurationError: If configuration is invalid
"""
self.config = config
self.model = None
self.tracker = None
self.line_zone = None
self.speed_estimator = None
self.annotator = None
self.video_info = None
logger.info(f"Initializing pipeline with config: {config}")
self._initialize_components()
def _initialize_components(self) -> None:
"""Initialize all pipeline components."""
try:
# Load YOLO model
logger.info(f"Loading YOLO model: {self.config.model_path}")
self.model = YOLO(self.config.model_path)
self.model.conf = self.config.confidence_threshold
self.model.iou = self.config.iou_threshold
logger.info("Model loaded successfully")
except Exception as e:
logger.error(f"Failed to load model: {e}")
raise ModelLoadError(f"Could not load YOLO model from {self.config.model_path}: {e}")
def _setup_video_components(self, video_path: str) -> None:
"""
Set up video-specific components.
Args:
video_path: Path to input video
Raises:
VideoProcessingError: If video cannot be opened
"""
try:
# Get video information
self.video_info = sv.VideoInfo.from_video_path(video_path)
logger.info(f"Video info: {self.video_info.width}x{self.video_info.height} @ {self.video_info.fps}fps")
# Initialize ByteTrack tracker
self.tracker = sv.ByteTrack(
frame_rate=self.video_info.fps,
track_activation_threshold=self.config.confidence_threshold
)
logger.info("Tracker initialized")
# Set up counting line zone
line_start = sv.Point(
x=self.config.line_offset,
y=self.config.line_y
)
line_end = sv.Point(
x=self.video_info.width - self.config.line_offset,
y=self.config.line_y
)
self.line_zone = sv.LineZone(
start=line_start,
end=line_end,
triggering_anchors=(sv.Position.BOTTOM_CENTER,)
)
logger.info(f"Line zone created at y={self.config.line_y}")
# Initialize perspective transformer
source_pts = np.array(self.config.source_points, dtype=np.float32)
target_pts = np.array(self.config.target_points, dtype=np.float32)
transformer = PerspectiveTransformer(
source_points=source_pts,
target_points=target_pts
)
logger.info("Perspective transformer initialized")
# Initialize speed estimator
self.speed_estimator = VehicleSpeedEstimator(
fps=self.video_info.fps,
transformer=transformer,
history_duration=self.config.speed_history_seconds,
speed_unit=self.config.speed_unit
)
logger.info("Speed estimator initialized")
# Initialize frame annotator
self.annotator = FrameAnnotator(
video_resolution=(self.video_info.width, self.video_info.height),
show_boxes=self.config.enable_boxes,
show_labels=self.config.enable_labels,
show_traces=self.config.enable_traces,
show_line_zones=self.config.enable_line_zones,
trace_length=self.config.trace_length,
zone_polygon=source_pts
)
logger.info("Frame annotator initialized")
except Exception as e:
logger.error(f"Failed to setup video components: {e}")
raise VideoProcessingError(f"Error setting up video processing: {e}")
def _process_single_frame(self, frame: np.ndarray) -> tuple:
"""
Process a single video frame.
Args:
frame: Input video frame
Returns:
Tuple of (annotated_frame, detections)
"""
# Run YOLO detection
results = self.model(frame, verbose=False)[0]
detections = sv.Detections.from_ultralytics(results)
# Update tracker
detections = self.tracker.update_with_detections(detections)
# Trigger line zone counting
self.line_zone.trigger(detections)
# Estimate speeds
detections = self.speed_estimator.estimate(detections)
# Generate labels
labels = self._create_labels(detections)
# Annotate frame
annotated_frame = self.annotator.draw_annotations(
frame=frame,
detections=detections,
labels=labels,
line_zones=[self.line_zone]
)
return annotated_frame, detections
def _create_labels(self, detections: sv.Detections) -> list:
"""
Create display labels for detected vehicles.
Args:
detections: Detection results
Returns:
List of label strings
"""
labels = []
if not hasattr(detections, 'tracker_id') or detections.tracker_id is None:
return labels
for idx, tracker_id in enumerate(detections.tracker_id):
# Get class name
class_name = "Vehicle"
if "class_name" in detections.data:
class_name = detections.data["class_name"][idx]
# Get speed
speed_text = ""
if "speed" in detections.data:
speed = detections.data["speed"][idx]
if speed > 0:
speed_text = f" {speed:.0f}{self.config.speed_unit}"
# Create label
label = f"{class_name} #{tracker_id}{speed_text}"
labels.append(label)
return labels
def process_video(
self,
progress_callback: Optional[Callable[[float], None]] = None
) -> Dict:
"""
Process the entire video.
Args:
progress_callback: Optional callback for progress updates
Returns:
Dictionary with processing statistics
Raises:
VideoProcessingError: If video processing fails
"""
start_time = time()
try:
# Validate input video
if not Path(self.config.input_video).exists():
raise VideoProcessingError(f"Input video not found: {self.config.input_video}")
# Setup components
self._setup_video_components(self.config.input_video)
# Create output directory if needed
output_path = Path(self.config.output_video)
output_path.parent.mkdir(parents=True, exist_ok=True)
# Initialize statistics
frame_count = 0
total_frames = self.video_info.total_frames or 0
all_speeds = []
# Setup display window if enabled (disabled in headless environments like HF Spaces)
if self.config.display_enabled:
try:
cv2.namedWindow(self.config.window_name, cv2.WINDOW_NORMAL)
cv2.resizeWindow(
self.config.window_name,
self.video_info.width,
self.video_info.height
)
except Exception as e:
logger.warning(f"Could not create display window (headless environment?): {e}")
self.config.display_enabled = False
# Process video
logger.info("Starting video processing...")
frame_generator = sv.get_video_frames_generator(self.config.input_video)
with sv.VideoSink(self.config.output_video, self.video_info) as sink:
for frame in frame_generator:
try:
# Process frame
annotated_frame, detections = self._process_single_frame(frame)
# Collect speed statistics
if "speed" in detections.data:
speeds = detections.data["speed"]
all_speeds.extend([s for s in speeds if s > 0])
# Write to output
sink.write_frame(annotated_frame)
# Display if enabled
if self.config.display_enabled:
cv2.imshow(self.config.window_name, annotated_frame)
# Check for quit
if cv2.waitKey(1) & 0xFF == ord('q'):
logger.info("Processing interrupted by user")
break
# Check if window was closed
if cv2.getWindowProperty(
self.config.window_name,
cv2.WND_PROP_VISIBLE
) < 1:
logger.info("Window closed by user")
break
# Update progress
frame_count += 1
if progress_callback and total_frames > 0:
progress = frame_count / total_frames
progress_callback(progress)
except Exception as e:
logger.warning(f"Error processing frame {frame_count}: {e}")
continue
# Cleanup
if self.config.display_enabled:
cv2.destroyAllWindows()
# Calculate statistics
processing_time = time() - start_time
stats = {
'total_count': self.line_zone.in_count + self.line_zone.out_count,
'in_count': self.line_zone.in_count,
'out_count': self.line_zone.out_count,
'avg_speed': np.mean(all_speeds) if all_speeds else 0.0,
'max_speed': np.max(all_speeds) if all_speeds else 0.0,
'min_speed': np.min(all_speeds) if all_speeds else 0.0,
'frames_processed': frame_count,
'processing_time': processing_time,
'fps': frame_count / processing_time if processing_time > 0 else 0
}
logger.info(f"Processing complete: {frame_count} frames in {processing_time:.2f}s")
logger.info(f"Vehicles counted: {stats['total_count']} (In: {stats['in_count']}, Out: {stats['out_count']})")
return stats
except Exception as e:
logger.error(f"Video processing failed: {e}", exc_info=True)
raise VideoProcessingError(f"Failed to process video: {e}")
def process_video(
config: VehicleDetectionConfig,
progress_callback: Optional[Callable[[float], None]] = None
) -> Dict:
"""
Convenience function to process a video with given configuration.
Args:
config: Configuration object
progress_callback: Optional progress callback
Returns:
Processing statistics dictionary
"""
pipeline = VehicleDetectionPipeline(config)
return pipeline.process_video(progress_callback)
def main():
"""Main entry point for CLI usage."""
try:
logger.info("=" * 60)
logger.info("Vehicle Speed Estimation & Counting System")
logger.info("=" * 60)
# Load configuration
config = VehicleDetectionConfig()
logger.info(f"Configuration: {config}")
# Process video
stats = process_video(config)
# Display results
print("\n" + "=" * 60)
print("PROCESSING RESULTS")
print("=" * 60)
print(f"Output saved to: {config.output_video}")
print(f"\nVehicle Count:")
print(f" Total: {stats['total_count']}")
print(f" In: {stats['in_count']}")
print(f" Out: {stats['out_count']}")
print(f"\nSpeed Statistics ({config.speed_unit}):")
print(f" Average: {stats['avg_speed']:.1f}")
print(f" Maximum: {stats['max_speed']:.1f}")
print(f" Minimum: {stats['min_speed']:.1f}")
print(f"\nProcessing Info:")
print(f" Frames: {stats['frames_processed']}")
print(f" Time: {stats['processing_time']:.2f}s")
print(f" FPS: {stats['fps']:.1f}")
print("=" * 60)
return 0
except KeyboardInterrupt:
logger.info("Processing interrupted by user")
return 1
except Exception as e:
logger.error(f"Fatal error: {e}", exc_info=True)
print(f"\n❌ Error: {e}", file=sys.stderr)
return 1
if __name__ == "__main__":
sys.exit(main())
|