The Research Of Vehicle Flow Statistics System Based On Target Tracking

Vehicle flow statistics system based on video image was one of the important research topics of intelligent transportation system,which took advantage of image processing and artificial intelligence technology to process the recording traffic,got the number of vehicles on the road in a certain period,provided the basic data for the subsequent processing of the intelligent transportation system,realized intelligent dispatching of the road,improved the utilization rate of road resources.Due to the impact of illumination,based on target tracking method for vehicle flow statistics in practical application often produces shadow problems.The shadow area is larger and the distance between vehicles is closer will cause the vehicle adhesion in the image,which affects the accuracy of vehicle flow statistics system.Therefore,the vehicle shadow elimination algorithm and the adherent vehicle segmentation algorithm are the key technologies of vehicle flow statistics system.The thesis presents a vehicle statistics system based on target tracking on the basis of the research of shadow elimination and adhesion segmentation algorithm.The main research works are as follows:1.In order to solve the problem of low accuracy in the occlusion vehicle curve segmentation algorithm of the vehicle edge,an improved segmentation algorithm based on concave analysis was presented.Simulation results show that the algorithm in the thesis has a higher accuracy rate compared with the occlusion curve segmentation algorithm of the vehicle edge in segmenting the vehicle,which effectively improves the accuracy of vehicle flow statistics system.2.In this thesis,the shadow elimination algorithms(HSV color feature and gradient feature algorithm)were studied.In order to solve the problem of low removal rate of the two algorithms mentioned above,a shadow elimination algorithm combined the HSV color feature with gradient feature was proposed.The simulation results show that the proposed algorithm has a higher shadow elimination rate,which improves the robustness of vehicle detection to a certain extent.3.In the Visual Studio 2010 integrated development environment,we achieve the vehicle flow statistics system based on target tracking through using the MFC application framework and Open CV computer vision library.The experimental results show that the system designed in this thesis has better real-time performance and overcomes the influence of the illumination to a certain extent.