Vehicle Detection And Tracking Classification Technology Research Based On The Video

To manage vehicles effectively and collect information of traffic quickly and accurately, an intelligent system of vehicles monitoring for vision-based detection and tracking and classification of vehicles in image sequences of traffic scenes is presented. Through summing up and analyzing the characteristics of existing methods based the technology of computer vision,this paper pays great attention to investigate the practical techniques for moving vehicls detection tracking and classification under the condition of a fixed camera,including such techniques as automatic detection of moving vehicles, elimination of extra shadows of vehicles,tracking and classification of moving vehicles,and so on. The following research has been accomplished:(1) Moving vehicle detection: Through analayzing and studying some methods in background extraction and updating backgrougnd,this paper presents a modified adaptive background model to update the background and detect moving vehicles effectively.(2) Moving vehicle shadows elimination:This paper makes a study of gray-level distribution features and V component of HSV color space of shadow and develops an effective shadow elimination method.(3) Moving vehicle tracking and classification:This paper analyses the characteristics of vehicle edge and implements a vehicle tracking and classification algorithm based on improved Hausdorff distance. One key step is Euclidean distance transform,this paper presents a center probe method to reduce computation cost.This system is implemented using C# in windows environment. Experimental results show that the system can detect and track the moving vehicle effectively, and it is robust and practical.