A Computer Vision-Based Vehicle Speed Monitoring and Reporting System

The detection and enforcement of over-speeding regulations remain a significant challenge due to limitations in existing vehicle speed estimation techniques. This study addresses this issue by developing a high-accuracy vehicle speed estimation system that incorporates a novel approach for license plate region extraction and over-speeding detection. The methodology employed involves the use of background subtraction methods to estimate vehicle speed, combined with a proprietary algorithm for vehicle plate region extraction. This information is then processed by a reporting system that identifies over-speeding vehicles. Results from extensive testing reveal a mean absolute error of 0.93 and a root mean square error of 1.40 in speed estimation, demonstrating high accuracy and precision of the developed system. Additionally, the mean absolute percentage error of 3.38% further substantiates the effectiveness of the system, leading to an overall accuracy of 96.62% in speed estimation. This advancement in traffic management technology has the potential to improve road safety, reduce traffic violations, and contribute to more efficient and streamlined urban planning.