Support Human Abnormal Behavior Recognition Of Video Monitoring System Design And Implementation

The system of this thesis belongs to the project required by Chengdu RailwayBureau and UESTC laboratory. The client needs to extend the existing videosurveillance system based on C/S structure to a system taking good advantage ofnetwork based on B/S structure, and make the system to some extent intelligentializedas well.Video Surveillance System of this thesis is designed by distributed databasemethod. It constructs a three-level tree structure database to store information andrealize this framework by MS SQL Server database. This system distribute the videostream needed by users through network, no longer relying on the traditional C/Sstructure but converting to B/S structure to achieve this function. Distributeddatabase/Link server/Web Server/.NET Remoting are used in specific implementation.Each node of the three-level tree structure is called server node. The node containsDatabase server/Video distribution server/Web Server. In physical structure, WebServer and Database server may not be corresponding to the physical location of thenode. In logical structure, a Server node contains Database server/Video distributionserver/Web Server. Database maintenance program uses C/S structure based onthe.NET Remoting technology. Database maintenance program contains a serverprogram and client program. In database, we use Role and Stored Procedure to accesssecurity needs. At the same time, unusual into detection System based on Gaussianmixture modeling theory and human abnormal behavior detection system based on opticflow field energy statistics are designed in this thesis. The image processing in thesystem are located in the third grade of nodes deployed in embedded equipment.The video surveillance system introduced by this thesis has been tested by somerailway stations of Chengdu Railway Bureau; it can meet the main requirements ofusers, and the subsequent works have been carried out.