The Algorithm Of Traffic Flow Statistics Based On DSP

Intelligent Transportation Systems (ITS) is a hot in the field of transportation research and development of every country,and traffic flow statistics is the main research field of Intelligent Transportation Systems and Intelligent Monitoring System. Traffic flow statistics is based on vehicle detection technology. Now vehicle detection technology is mainly proposed in particular constraint or in a application background,there are many issues to be resolved.The main content of this paper is research background updating,moving object segmentation, moving target detection,block detection, shadow suppression and other aspects in traffic flow statistics. The main work and innovation are as follows.(1)Background update phase. This paper using the background subtraction to abstract moving vehicles, in order to improve the traffic flow statistics accuracy. Using the Background subtraction method and Frame-difference method to abstract background mask, use mask to update the background, obtain accurate background image.(2)Images segment phase. This paper purpose a method of Two-Scanning based on connected domain this method can effectively separate connected domain accurately in complex environments, extract the characteristic information of the moving vehicles (size, location, moment features, contours, etc.).(3)Block detection phase. Traditional video detection traffic flow statistics use vehicle tracking to solve block problem, the algorithm is complex and easily interfered by outside. This paper set up some block model of vehicle, Hierarchical matching contours of the vehicles and models on block vehicle to identify block vehicles.(4)Shadow detect phase. Take into account the real-time and the actual situation, utilize the connection of U,V color component of YUV and color and saturation component,propose utilize YUV color space detect and suppress the shade pixel.(5)This paper is based on background updating, image segmentation, block detection, shadow detection module so on. Built a traffic flow statistics system based on DSP,optimize the program in relation with DSP features.we Real-time processing the image from camera to obtain traffic flow information.