AI-Powered Road Safety: A Unified YOLO Network for Real-Time Helmet Detection, Speed Estimation and Number Plate Recognition

The Indian Road traffic scenario presents significant challenges, as documented by the Ministry of Road Transport and Highways. In 2022, the Indian Road Traffic scenario presented significant challenges. A critical point of concern in these reports is the alarmingly high fatality count in excess of $\mathbf{5 2, 0 0 0}$ for the year linked to not wearing helmets, underlining the urgent need for a more effective traffic regulation enforcement system, particularly in the enforcement of helmet usage. Current traffic monitoring systems confront limitations such as low detection accuracy in real-world circumstances, trouble detecting non-helmet users, and problems recognizing small or non-standardized number plates. Real-time processing limits and the integration of several detection elements, such as helmet detection, speed monitoring, and number plate identification, are yet unsolved. The proposed system intends to create and implement an AI-powered traffic monitoring system that can detect helmet usage, vehicle speed, and number plates with high accuracy, ultimately improving road safety and enforcing greater adherence to traffic regulations in India. The proposed system provides an innovative technique that takes advantage of the powerful capabilities of YOLO algorithms combined with the Deep Sort algorithm. We have successfully identified helmet, number plate and speed with a 90.4 % accuracy rate. This integration holds the potential to significantly reduce the incidence of road accidents and fatalities, marking a vital stride towards safer roads in India. This enhancement in road safety could be instrumental in saving lives and mitigating the financial and emotional strain that road accidents exert on society and individuals alike.