Research And Application Of Human Motion Detecting And Tracking Techniques

Video surveillance system is a new science research field which has more development potential, and it is applied more and more in subway industry. Based on the requirements analysis of the video surveillance system of the rail transportation environment, we give out the video surveillance system architecture. As detecting and tracking of human motion are two key techniques of intelligent video surveillance systems, this dissertation put more focus on it.In this thesis, a method of human motion detecting and tracking based on video is proposed. Background model, threshold segmentation, morphologic filter, connected region analysis, Kalman filter and HSV color histogram are included, and the solution for the problem and difficulties in the implementation are proposed.With regard to human motion detecting, the basic principles of frame-difference method is analyzed, an improved frame-difference method was proposed which separate three frames from the video image for human motion detecting. Moreover, the background model is analyzed, a statistical background model based on selected background by the color of a point change in the time frame is established, and the experimental results are presented. To solve the problem that shadows usually affect the human motion detecting, we discuss the shadow detection method based on the HSV color space model.With regard to human motion tracking, this paper presents a combining method of Kalman filter with the HSV color histogram template matching. A moving human status prediction model based on Kalman filter is building, and according to the result of Kalman filter tracking, human outline are determined; HSV color histogram template acquire the color feature of the image by the color histogram of the image on the three color components and template matching.Finally, a prototype system of human motion detecting and tracking is implemented based on matlab tool.