Video Vehicle Detection And Tracking Method

Recently, Society economy has been developed to make the existing traffic control system more efficient. The number of vehicles has been increased, in order to solve the problems which cased by the development of ground traffic, first, it need to integrate image processing and pattern recognition to get some traffic parameters, then solve the problems by behavior understanding and intelligence analysis. So it can be used to solve the problems of traffic block,road macro planning and automatic charge system of the highways. Video detection and tracking system detect, recognition and track the vehicles from the video sequence then get the exact traffic parameters, the exactness of detection and track effects the exactness of intelligent traffic system decision-making.This paper detects and tracks vehicles from video sequence and extracts the traffic parameters to help to solve the traffic problems. This paper includes three parts: vehicle objects detection, shadow removing and vehicle tracking. We will give the details as follows:1. Vehicle detection. This paper gains the moving objects by background difference. First, we extract the background and then get the vehicles by background different. The method of background extraction and background updating integrate gay-level quantification and two attenuation weights which introduced to reduce the impact of environment lighting condition, two discriminant functions are employed to distinguish false moving objects and true moving objects for solving the deadlock problem of background updating. After background difference, we get the moving objects by the adaptive thresholds segmentation method. The experimental results show that the proposed method is more robust, accurate and powerful.2. Shadow removing. Moving shadow removal is critical for robust moving object tracking system. A shadow removal algorithm based on shadow attributes is presented in this paper. The direction of shadow is estimated first, and then shadow points are sampled based on the direction of shadow. The shadow and vehicles attributes is calculated using the sampled shadow points. Calculate the shadow and vehicles attributes. At last remove the shadow by maximizing the posteriori probability. The method is effective own to using the true shadow property.3. Vehicle tracking. This paper proposes the vehicle tracking method which integrated discrete wavelet transform and extended kalman filter. Using index table to estimate the time when occlusion occur and relieve, and solve the occlusion problem by feature matching, test image data indicate the system which possesses good robust and accurateness, more easy and has good real peculiarity.We simulate the algorithm in the computer and test the performance of traffic detection, shadow removal, and vehicle tracking. Experimental results show that the algorithm is feasible and robust.