AI-Powered Traffic Violation Detection and Electronic Challan Generation Using Computer Vision

Traffic rule violations, such as riding a motorcycle without a helmet and carrying more passengers than permitted, pose significant road safety risks. Manual monitoring of traffic surveillance footage is time-consuming and difficult to scale. This paper presents an AI-powered traffic violation detection system that integrates computer vision, object detection, vehicle tracking, and optical character recognition (OCR) to support automated traffic monitoring. The system uses YOLOv8-based object detection to identify relevant road users and safety equipment, analyzes helmet-related violations and triple riding, and applies vehicle tracking to associate detections across video frames. OCR is used to extract vehicle registration numbers from detected number plates. The system records violation details, including timestamps, recognized registration numbers, violation categories, and supporting image evidence, for review through an electronic challan management workflow. Experimental evaluation achieved an accuracy of 92.5%, precision of 91.8%, recall of 90.6%, and F1-score of 91.2%. The results indicate the potential of the integrated approach to support automated traffic monitoring. The system is intended to assist traffic enforcement personnel, with human verification retained before further challan processing