Study On Detection And Tracking Techniques For Moving Vehicles In Video Sequences

Effective and accurate detection and tracking of moving vehicles in video sequences is the key to modern intelligent traffic monitoring systems. Through summing up and analyzing the characteristics of existing methods in vehicle detection and tracking, this paper pays great attention to investigate the practical techniques for moving objects detection and tracking under the condition of a fixed CCD camera, including such techniques as automatic detection of moving objects, elimination of extra shadows of vehicles, tracking of moving vehicles, and so on. The following research has been accomplished:(1) Detection of moving vehicles in video sequencesIn order to deal with the existing problems in automatic vehicle detection for intelligent highway traffic systems, this paper presents a moving vehicle detection and extraction method based on color image difference model, adaptive thresholding and image morphology filtering theory. In addition, according to the difference of the consecutive video frames, a real-time dynamic updating scheme of background image or reference image is described. We also analyze some basic image techniques including image smoothing, enhancement, edge detection and noise elimination, and so on. Finally, extensive experiments were performed to evaluate the performances of the above-mentioned methods.(2) Shadow elimination of moving vehiclesThe shadow of moving object will affect the correctness of object recognition and object tracking. To overcome this problem, this paper develops an effective shadow distinguishing and eliminating method based on the edge and color information of moving vehicles. Experiments were conducted to prove the effectiveness of the method.(3) Tracking of moving vehiclesFocusing on the problem caused by time-consuming computation and multi-vehicles occlusion in object tracking, this paper investigates the basic tracking method of moving vehicles based on Kalman filter theory. The moving model of vehicle is established from Kalman filter technique, and the position and color of the centroid of the vehicle are used for estimation and match of the tracking object in the consecutive frames. According to the tracking results, the important traffic parameters...