Research On Detection Method Of Human Abnormal Action Behavior In Video Surveillance

With the development of technology, video surveillance has been widely applied to various traffic, communities, factories and other places, the traditional monitoring system is just a simple video recording and storage, once the incident occurred, you need to manually from a large number of video files in the search for clues, low efficiency and can not do a real-time warning, and realize automatic detection of abnormal the behavior can effectively reduce the labor cost in the monitoring system, to remind staff to respond in a timely manner when an exception occurs, so as to reduce human body risk and economic loss, is of great significance to social stability and people's life safety, so more and more researchers are devoted to the research of abnormal behavior recognition the human body.In this paper, the research focuses on the detection methods of two kinds of abnormal behavior of wandering and fall, and the effectiveness of the proposed method is verified by the program. The main work of this paper is as follows:1 the overall framework of human behavior recognition is analyzed, and the classical methods and characteristics of the current moving human detection are summarized, and the background modeling algorithm is proposed to detect the human body.2 in the aspect of the movement of human body identification, the external matrix is proposed to represent the movement of the body, and the movement of the center of mass is used to represent the movement of the human body.3 through characteristic analysis of wandering behavior and falling action defined length to width ratio, centroid change ratio, centroid displacement characteristics of operator, to quantify the human abnormal behavior using the characteristic operator, operator characteristic changes of abnormal behavior were evaluated synthetically according to, thus the to wander and fall two abnormal behavior detection method.At the end of this paper, under the support of visual development environment and opencv library realized wandering behavior and behavior detection algorithm falls, also recorded a video for the experiment, and the experimental results are analyzed and summarized. The experimental results show that the detection method of human abnormal behavior based on the proposed method can detect the two kinds of abnormal behavior of hovering and falling.