Real-Time Helmet Detection System with Vehicle Number Extraction for Two-Wheeler Vehicles using YOLOv8

In the face of increasing two-wheeler usage and the persistent issue of riders neglecting helmet compliance, this paper proposes a groundbreaking Real-Time Helmet Detection System with Vehicle Number Extraction. The existing literature on helmet detection systems is reviewed, revealing a range of methodologies, accuracies, and limitations in current approaches. Building upon this foundation, the three-stage detector combines the efficiency of YOLO (You Only Look Once) for object detection with a specialized number plate extraction system. The system aims to provide real-time detection of two-wheeler riders, classification of helmet usage, and extraction of vehicle numbers. The proposed methodology, utilizing a pre-trained YOLO model for object detection and EasyOCR library, demonstrates an accuracy rate of 64% for vehicle detection, 78% for helmet detection, and 92% for number plate extraction. The practical implementation of this system showcases its potential for enhancing road safety, automating traffic management, and promoting regulatory enforcement. While showing potential, the ongoing research and development face challenges such as limitation of computational resources and privacy considerations. The Real-Time Helmet Detection System stands as a promising solution for improving road safety, with the potential for broader use in traffic management and safety enforcement. Continuous improvement and adaptation will guarantee its effectiveness in contributing to a safer and more controlled traffic environment.