Automatic Road Traffic Analyzer using Background Subtraction, Blob Analysis, and Tracking Algorithms

Due to the prodigious road traffic, many cities experience road congestions. Hence, there is a requirement of efficient and automatic road traffic analysis. The computer vision (CV) technique is one of the rapidly utilized fields for traffic analysis. This paper proposes a system that efficiently tracks and counts the number of vehicles for automatic traffic monitoring. The system is composed of several stages viz. demosaicing on Bayer filter pattern video frames, moving object algorithm for vehicle detection, and Blob analysis is for tracking and counting of vehicles on a road. The prepossessing is performed with the help of system generator development tool of Xilinx, and the tracking & counting of vehicles on a road are observed in the MATLAB Simulink environment. The algorithms are applied to all the frames in every video with the labeled vehicle is shown. An average accuracy of 98.78% is obtained on real time videos captured through different traffic conditions. The comparison with the similar reported algorithms validates the superiority of the proposed work for traffic analysis system.