Real-Time Helmet and License Plate Recognition for Enhanced Traffic Monitoring

This paper presents a machine learning solution to enhance road safety and enforce traffic rules by detecting helmets and recognizing vehicle number plates in motorcycles and cars. It integrates YOLO (You Only Look Once) for helmet detection and EasyOCR for Optical Character Recognition (OCR) of number plates. The system preprocesses input images using techniques like bilateral filtering for noise reduction and Canny edge detection for contour identification, isolating number plates for accurate OCR under varying conditions. Concurrently, a YOLO model trained on a custom dataset identifies helmeted and non-helmeted riders. The automated approach streamlines the identification of violators by recognizing helmet use and capturing vehicle registration details. The system achieves high accuracy, even in challenging scenarios, and is scalable for real-time applications such as traffic surveillance. Future enhancements could include robustness to weather variations, support for diverse vehicle types, and live video processing. This model advances automated traffic enforcement, contributing to improved safety and compliance on roads.