Research On Video Analysis Technology In Security Surveillance

With the development of high technology on computer science, machine vision, image project, pattern recognition and so on, video surveillance technology has gone through three stages, they are early analog video surveillance, data video surveillance in the medium-term and intelligent video surveillance in the moment. video content analysis technology was brought into the video surveillance, then intelligent video surveillance technology emerged. Intelligent video surveillance is a substitute for the monitors and overcome the lack of traditional video surveillance. Security surveillance which the paper disign and implement is one of the widest fields where the intelligent video surveillance is applied.The paper introduces the design and implement method of abandoned objects detection and intrusion detection in security suveillance using video content analysis technology. Video content analysis technology is the key technology of intelligent video surveillance. Firstly, the paper analyses the architecture of video analysis system. Secondly, the paper discusses and researches object detection algorithm currently used and analyzes and compares its strengths and weaknesses. Lastly, the paper lays emphasis on the study of Hu invariant moment based on the deep study and analysis of feature extraction method.In abandoned objects detection algorithm design, the paper detects abandoned objects using double background modeling technology based on gaussian modeling method. At the same time, the centroid distance judge method is needed to judge the characters of abandoned objects and therefore the system sends alarm messages automatically. In intrusion detection algorithm design, the object detection algorithm based on time average method and frame difference algorithm is presented for intrusion detection. Support vector machines method is used to classify the intrus objects. The system sends alarm message automatically depending on classification judgement.