The Intelligent Vehicle Track System Based On The Random Walk Theory

The detection and track of vehicle based on intelligent video analysis are crucial technologies of Intelligent Traffic System.In order to realize the detection and track of multiple vehicles in the traffic surveillance video,this paper develops a system of intelligent vehicle track based on random walk theory.On the basis of achieving GMM and PBAS,Background Subtraction Method is used to get moving region and remove shadow according to the edge information.Then pixels are marked by detecting the skeleton structures of the moving region and the marked pixels are taken as object seeds to obtain accurate moving region by random walk.After eliminating interferential region by judging proportion of region,contour of vehicle is marked to realize automatic detection.Finally,blob track is utilized to track vehicle and draw a path on the video.Experiment results demonstrate the feasibility and effectiveness of this system.