ESTIMATION OF VEHICULAR TRAFFIC PARAMETERS USING VIDEO SURVEILLANCE

Abstract

The advanced applications of video based intelligent transportation systems (ITS) have enabled new strategies for better traffic management by improving road safety, relieving traffic congestion and enhancing productivity. This report provides an overview of various techniques used for estimating the vehicular traffic parameters, like vehicle count, speed, congestion estimation etc., in the video based surveillance systems. The virtual detection line (VDL) based method has been implemented and its performance has been compared with the traditional background subtraction based method. The performance evaluation of VDL based method has been carried out in different environmental and traffic conditions. For these test conditions, even though the VDL based method gives higher detection rates as well as better real time performance as compared to other techniques, it has high false accept rate and false reject rate. To minimize this problem, different error corrections have been performed, which showed considerable reduction in the false detections. Then the performance of VDL based method in congestion estimation was evaluated