Automated Real-Time Detection of Uninsured Vehicles and Unlicensed Drivers Using Advanced Computer Vision and Face Recognition Algorithms in Indian Traffic Management

Abstract

Traffic control is the key to road safety and legal regulations, but the problem of detecting uninsured vehicles and the licensing status of drivers is not an easy task for the Indian authorities. Old-fashioned manual systems are not always efficient and effective, and may easily introduce errors, and can hardly be implemented in high-density traffic. This study is a comparative study on an automated real time detection system that is aimed at correcting these gaps with the help of the use of an advanced computer vision. The paper is an assessment of a multi-layered pipeline, which consists of the use of YOLOv7 to detect license plates, Bilateral Filters to enhance image quality, and a mix of easyOCR and CRNN to identify characters. Additionally, ArcFace, DeepFace, and DeepID are integrated into the system to test the driver identification system. The results prove that the suggested integrated framework is much more effective than the traditional ones. Comparative test shows that YOLOv7 is the best tool to detect the intricacies of the Indian license plates, with a detection rate of 99.8%. Also, the facial recognition module that used ArcFace was better in a dynamic environment, with 97 % success in identifying individuals. In the comparison of these contemporary algorithms to the traditional standards, this study will prove a very dependable and efficient approach to automated policing. The findings indicate that the implementation of such a system can increase the capacity of traffic agencies to implement insurance and licensing laws in real time without having to break traffic.