Vehicle Tracking Based On The Video

Today, the conflicts about the population, resources and environment have become increasingly prominent, but it can't solve the increasingly serious traffic problems by building more roads, expanding the size of the network or other methods. In this case, the Intelligent Transportation Systems develop rapidly, and become the future direction of development of transport. Due to the obvious advantages compared with the traditional method about the road traffic information, the video-based intelligent transportation system is concerned and become the hotspot of research in recent years. In it, the vehicle tracking technology can provide direction movement of vehicles, road tracks, and status and other important information, it lay the foundation for the next step analysis of the vehicle behavior.This paper briefly describes the background of the development of intelligent transportation, domestic and international status quo and future trends of development and difficulties, on the basis of summing up and analysis of the inadequacies of existing technologies, we focus on the movement of vehicles shadow suppression and vehicle tracking.In this paper, about the moving shadow suppression of vehicles, at first we describe the formation mechanism of the shadow, then presents a vehicle detection method based on shadow suppression. In this method, first we calculate the foreground regional which is obtained by the background difference method using denoising treatment, and margin or the ratio of each component of the corresponding background regions in the HSV color space. Then the optimal threshold is divided by the OTSU algorithm based on the information of average and square error of image pixels method, and the process is optimized through the particle swarm optimization algorithm based on immune clone principle, then bring the threshold into the shadow discrimination formula to judge the shadow and then remove the real shadow. In this way we achieve the adaptive process of the shadow discrimination and suppression.And about the aspect of the vehicles tracking, we solve the occlusion problem of the vehicles by a combination method of mean-shift and adaptive kalman filtering. The color features is used as the vehicle's features, and the target is matched by using the former template, which is searched by combining the results of mean-shift algorithm and adaptive kalman filter algorithm through a certain weight decided by the matching degreeρ. This method can solve tracking problem about the vehicle being obscured, and it can improve the applicability and accuracy of the tracking algorithm, too.