The Algorithm Research Of Traffic Flow Detection Based On Video

With the development of our social economy and the popularity of the car,the city traffic congestion phenomenon and traffic accident occurrence volume increases gradually,serious impact on the urban economic development and people's life.In order to reduce the incidence of traffic accidents,computer science,information processing technology and traffic video image processing based intelligent transportation system came into being.In order to realize the management of intelligent transportation systems must be timely grasp of the traffic road information,and to respond quickly to the road traffic information,the traffic road flow detection is very critical part of the intelligent transportation.This paper mainly studied based on the traffic flow of traffic video detection technology.Full text around the moving vehicle detection,moving vehicle traffic shadow removal and traffic statistics,etc,the main work accomplished as follows:First of all,by analysis and comparison the several common targets detection method of complex environment,puts forward a kind of adaptive Gaussian mixture model methods.The method has very strong ability to adapt to the environment,can adapt to the dynamic impact of the background scene,the better to extract the foreground objects,and solve the problems of scene light mutations and slow moving targets in the scene caused by the target mistake examining.The experiments show that the proposed algorithm can effectively extract moving targets in complex environmen,while overcoming the influence of scene light mutations.Second,we focused on the traffic scene movement characteristics of the shadow,based on the neighborhood light intensity ratio of moving shadow unchanged and color information than become characteristics,proposed an effective algorithm to removal the moving shadow,avoid the influence of the moving shadow on the vehicle detection.The experiments show that the proposed algorithm can overcome the defects existing shadow removal algorithm,fast and effective removel of the moving shadow.Finally,we designed a traffic flow detection methods based on virtual coil,the realization of the traffic statistics.By use of the block of virtual area of traffic flow testing,and the correlation characteristics of adjacent frames,correction of the data flow of traffic detection.The experiments show that the proposed algorithm is effective statistical number of vehicle in the road andnd improve the accuracy of traffic detection.