Research Of Video-based Vehicle Detection Algorithms

Along with the booming of the world economy and the amount of vehicle is growing day after day. Every city faces the problem as traffic jam and frequent traffic accidents. This means it is necessary and emergence to built an effective ITS to monitor, dispatch and control the traffic.Right now we have several methods to detection the traffic, such as by using electromagnetic induction, micro waves, ultrasound, and mark of ultra rays. Recently, the method of detecting the video frequency which based on image processing has enhanced with some obvious advantages. It is replacing the old methods and taking domination.There are mainly two methods of video vehicle detection, by using characteristics of the images or model matching. This paper is about to deeply comparing these two methods by comparing the basic materials and analyzing the two basic problems which are characteristics detection and following, and comparing the methods of connecting and virtual coil, then by using the method of model matching we can process the video vehicle detection. It makes the method of following different targets in the same time possible.The methods of flow detection of virtual coil and model following have been made though on computer. Experiments proved that they have high precision.