An efficient implementation of video based traffic analysis system

As the problem of traffic congestion is increasing day by day, there is a pressing need for the introduction of advanced system to monitor and control traffic efficiently. The analysis of traffic data collected from the camera based systems are the main method used for this purpose. The method presented here can detect vehicles, track and count them with high accuracy on videos from the intersection of traffic. A combination of Running Gaussian Average and local binary pattern are used for vehicle detection. For multiple object tracking, Kalman filtering is used. The system will count the number of vehicles passed through a particular area and determine the speed of vehicles passed. The system will help in traffic flow monitoring and management. The experimental analysis have shown that the proposed system show high detection rate and is feasible.