A Study On Algorithm For Moving Vehicle Detection And Tracking Based On Vedio

Traffic information detection technique based on video is widely concerned in recent years, because of its wide detection range, rich collection information and easy installation and maintenance. Moving vehicle detection and tracking are the key techniques in traffic information collection and detection system.In the thesis, traffic scene images are research objects, and the key techniques of background extraction and updating, shadow removal, vehicle detection and vehicle tracking are studied and analyzed.After studying on conventional background extraction algorithms, improved algorithms of background extraction and updating are put forward, and the new background is used to get the binary image. Shadow of vehicle is mistaken for target for the influence of illumination and lighting equipment, an improved algorithm is put forward to remove shadow. After using connected component labeling to process binary image, moving target is detected and tracked by vehicle tracking algorithm based on Kalman filter. And it provides a foundation for traffic incident detection and analysis.The algorithms proposed above are experimented in the situation of heavy vehicle flow, mutual occlusion and adhesion of vehicle. The experimental results show that it can deal with vehicle occlusion problem better and realize the vehicle detection and tracking in real time.