Research And Implementation On Human Behavior Analysis In Intelligent Monitoring System

Movement analysis of human behavior has becoming currently a hot subject of research for its widely application foreground and potential economic value in all kinds of area such as intelligent monitoring, human-computer interaction and video conferencing. This paper has designed and implemented an intelligence monitoring system of human behavior based on the analysis of the research situation and the development trend both at home and abroad.In the aspect of moving target detection, we proposed the Gaussian Mixture Model algorithm to build the background, which can extract the moving target clearly and completely in the complicated background. By using the adaptive threshold to update the background, the influence of the environment has been eliminated. Though detecting and suppressing the shadow in HSV space, the performance of the algorithm has been further improved.In the analysis of human behavior, we propose an algorithm which took the improved weighted Hu moment invariants as the character for template matching. We use sliding window to comprehensive analyze the matching of the successive video frames, thus to recognize the human behavior. With combination of system-specific applications, we proposed the method using targets' contour feature to recognize the human action. In the end, the results showed our method could detect and analysis the human behavior in video sequences effectively and real-timely.