Research On The Technology Of Vehicles Counting And Tracking At Intersection

Counting and tracking with moving vehicles at intersection is a important function of Intelligent Traffic System(ITS).This paper made detailed researches on some pivotal technology of applying digital image processing and pattern recognition to counting and tracking moving vehicles at intersections in city.First, the concept of random edge feature which has been applied in tracking and recognition of moving targets was proposed, and the result of experiment shows that the proposed feature has a good validity in tracking with moving vehicles. Second, the problem of keeping correct tracking with temporal disappeared targets was solved by an applied simple logic of target tracking state judgement which was devised in manner of feature matching value matrix. Finally, a multiple targets tracking algorithm based on these work was constructed.This algorithm was implemented by Visual C++ 6.0, and the result of the experiment that based on the actual image sequence which was taken out in the city intersection was given . The result of the experiment shows that the proposed algorithm can track and recognize multiple moving targets in image sequence, furthermore , it can adapt well to some disadvantageous factors such as the changing of scale, angle of view point .illuminant ,and moving state of target.