Final YearAdvancedAI/ML 4-5 months
Smart Traffic Management with Computer Vision
A system that detects and counts vehicles from traffic-camera feeds and dynamically suggests signal timings to reduce congestion.
PythonYOLOv8OpenCVFastAPIReact
Find GitHub repos & source code →How to build it — step by step
- 1Detection: Detect and count vehicles per lane using a YOLO model on video frames.
- 2Density estimation: Estimate per-direction density and queue length over time.
- 3Signal logic: Compute adaptive green-time allocation based on relative demand.
- 4Dashboard: Visualise live counts, suggested timings, and simulated throughput gains.
Key features to implement
- ✓Real-time vehicle detection
- ✓Per-lane density
- ✓Adaptive signal timing
- ✓Congestion dashboard
- ✓Throughput simulation
💡 Unique twist to stand out
Add emergency-vehicle detection (siren/visual) that pre-empts the signal to clear a path automatically.
🎓 What you'll learn
Object detection, video processing, control/optimisation logic, and simulation.