Karthik Tatineni

Helmet & Triple-Ride Detection

✕

This project uses deep learning (YOLOv5 + CNN) to detect motorcyclists violating traffic rules—specifically those not wearing helmets or riding with more than two passengers.

The system is trained on a custom dataset and deployed on a system capable of processing live video feed from a roadside camera. Once a violation is detected, it highlights the rider and extracts the number plate for logging.

The detection model is optimized for accuracy and real-time performance using OpenCV and PyTorch.

Future enhancements include automatic challan generation and direct integration with traffic databases.