Research And Implementation On Abnormal Human Behavior Detection In Monitoring System Of The Safe City

With the development of science and technology, city security video monitoring is more and more widely applied to various occasions the safe city. The compression, storage, analysis, display and alarm processing automation video information for computer city security video monitoring system, so as to realize the unattended; through the network platform to realize the remote video surveillance, even thousands of kilometers away can also visit the scene to achieve results; not only in a few minutes to complete the analysis of large amounts of data of traditional video surveillance using advanced software system, improve the efficiency of video monitoring, and can obtain more realistic digital image quality, clear and more convenient, practical video surveillance management and maintenance.Through the construction of city security monitoring project, realize the city security monitoring covers the whole city, city alarm linkage, vehicle tracking, intelligent card and other functions, to achieve the whole city safety monitoring, monitoring and command platform provides uniform for the public security system, through information technology advanced city, improve the safety level of the building. City Center is a highly complex, crowded places, effective use of city security monitoring can ensure public law enforcement officers and the safety of the masses. Monitoring of the places the masses daily behavior and special circumstances, safe city management work is indispensable. In the safe city scene, we care about abnormal behavior security monitoring in the city, the identification of anomalies in monitoring image, and can alarm in real time and reliable manner, effective assistance to the video monitoring staff access to prevent hazardous activities occur more accurate information and deal with emergencies. So how will these city securities monitoring in the detection and identification of abnormal behavior accurately and timely notify the treatment are the main contents of the article.This paper studies Ping abnormal behavior recognition of city video surveillance system, mainly for video sequences are foreground extraction, target behavior recognition are studied, and implemented the real-time detection of abnormal behavior of the prospect of a and a plurality of motion of the. Firstly, the method of recognition of some new types of abnormal behavior of algorithm, to solve this kind of strenuous exercise such as fighting and other characteristics of complex technical problems and extraction of image speed slow, and then puts forward a new method of abnormal behavior recognition of target contour feature detection, target center tracking and region detection based on optical flow features combination. Realize the use of VC technology and open-source computer vision library for urban security video monitoring abnormal behavior of algorithm, and through field experiments verify the effectiveness of the algorithm.