A computer vision system to detect traffic violations like red light jumping and wrong lane driving using YOLOv8 and DeepSORT.
- Real-time vehicle detection using YOLOv8
- Vehicle tracking using DeepSORT
- Red light violation detection
- Wrong lane driving detection
- Web dashboard using Flask
- Core Technologies
Python - Primary programming language (v3.6+ recommended)
OpenCV (cv2) - Real-time video processing and computer vision operations
PyTorch - Deep learning framework that powers YOLOv8
Computer Vision & AI Components YOLOv8 (Ultralytics) - For object detection (vehicles, pedestrians, etc.)
Using yolov8n.pt (nano version) for optimal performance
DeepSORT - For object tracking across video frames
Using OSNet (osnet_x0_25) as the ReID model
Backend & Web Interface Flask - Lightweight web framework for creating the dashboard
Jinja2 - Templating engine for Flask (used in index.html)
Supporting Libraries NumPy - Numerical operations and array handling
Threading - For running Flask server alongside video processing
Development & Deployment Git/GitHub - Version control and code hosting
PIP - Package management (requirements.txt)
Optional/Recommended Extensions Git LFS - For managing large model files (like yolov8n.pt)
Docker - For containerized deployment (can be added later)
SQLite/PostgreSQL - For violation logging (future enhancement)