Design And Realization Of The Illegal Vehicle Detection System On Single Intersection Based On Video Detection

With increasing number of private cars,traffic jam and illegal vehicle problem have been becoming terrible,especially the single intersection.In view of the situation,this topic designs a set of single intersection detect system which is based on video detection.This system has realized the violation detection and recognition,such as traffic light,vehicle stop line,running the red light,wrong directions and overspeed,with applying Visual Studio 2010,while used C++ language and Open CV development.Consequently,it is more convenient to maintenance of traffic management.This text has researched the vehicle detection and tracking technology which is based on dynamic background updating algorithm of wavelet transform and come up with a illegal vehicle detection approach based on improving background and updating core algorithm.The approach uses weight coefficient to update background effectively and avoids the influence of detection output from background change and external environment change.Next,we do denoising to three-component sub-detail graph from one-level wavelet transform and do wavelet construction with low frequency part to get the new background image.This has effectively solve the vehicle track retention problem.Then we use dynamic tracking method and background difference method to monitor and compute the position of vehicle centroid automatically to track vehicles’ position and recognize illegal vehicles.Through test,we find that the system is able to realize behaviour of moving vehicles detection accurately normally which is up to 85 percent.