Traffic Surveillance using Computer Vision

Present systems of traffic surveillance in India have not caught up with the technological developments and heavily rely on manual surveillance by traffic police, which is not only inefficient but also a cause for corruption. We are trying to develop a comprehensive autonomous traffic surveillance system that would eliminate the need for manual intervention and improve efficiency. In our approach, we intend to use video footage from traffic cameras and adopt computer vision techniques.

The features we plan on implementing are:

Automatic License Plate Recognition(ALPR): to automatically link a vehicle to its owner through the vehicle registry database. If the system detects traffic violations by the vehicle, challans can be directly issued to the owner.

Speed Detection: to calculate the speed of passing vehicles and penalise those crossing the speed limit.

Traffic Adaptive Signal Timing: to adjust signals based on road traffic and help smooth the flow of city traffic

Helmet and Seatbelt detection: to detect those not following safety rules.

This would be the initial phase of our project. After its successful implementation, we hope to integrate detection of traffic violations, accident detection and crime detection.

Traffic Management