Research Of The Algorithm Of Moving Vehicle Detection &tracking Based On Surveillance Video And Its Implemen Tation

In recent years, with the rapid development of cities, traffic problems became a big problem. Existing urban road is very difficult to expand and the city traffic is increasing, the city traffic bears unprecedented pressure. A variety of solutions have emerged, among which intelligent transportation systems has received extensive attention. Intelligent transportation system is a kind of vehicle detection and tracking. Due to the advantages of easy installation and maintenance, collecting rich information, detection and alarm in a timely manner, ITS has been received widespread concern. Among them, the key technology of extracting traffic information--the accurate detection and tracking of traffic. It has become a research hotspot in recent years, and has broad application prospects.In this thesis, we research on the road traffic surveillance video and intensively study several key issues around the video detection algorithm. Finally, we format a relatively complete traffic video detection algorithm. First of all we calibrate the cameras on the road, and then do background extraction, with updating background. Secondly we do the moving object extraction, choosing the combinative methods of background subtraction and frame difference to reduce the ghost and won the complete moving targets. Thirdly, we make corresponding analysis and research on the problem of the vehicle shadow. We use the shadow elimination algorithm of the fast normalized cross-correlation, connecting with the morphological processing and line scan, to reduce the Phenomenon that mistaken windows, roof parts as the shadows. Then, we use Kalman to predict the location of each moving target in the next frame, and use the matching algorithm to match. Finally we are able to carry out traffic information extraction, and according to the vehicle speed to detect speeding violation behavior and other illegal acts. In the simulation system, we have confirmed the effectiveness of the algorithm presented in this paper.