The Design Of Intelligent Recognition System For City Appearance Violation Behavior

This paper focuses on the application of intelligent video surveillance analysis technology in monitoring city appearance violation behavior. Firstly, the research and development of intelligent video surveillance system in the computer vision field is introduced in detail. Then four types of video dataset used in this paper are presented and the method of the de-noising and enhancement in image pre-processing is also discussed, which lays the foundation for the follow-up work for the image processing.Through the analysis of the characteristics of abnormal targets, this paper proposes a multi-layer background modeling approach with the feedback correction mechanism, which forms the basis for extracting the static foreground by utilizing confidence map. Then to improve the accuracy, illumination/shadow filter and abandoned/removed object classifier are also designed. The test for the different video sequences shows that our scheme can accurately extract and recognize the abnormal objects in indoor and outdoor environment.Lastly, in order to apply the aforementioned theoretical research to the practical monitoring scene, this paper presents the design and implementation of intelligent recognition system for city appearance violation behavior. Visual Studio2008IDE is used as the development tool for the software platform. C++programming language and object-oriented programming idea is utilized. A framework for designing software platform is provided by MFC and OpenCV. Then the input module of the image information, the module of the intelligent recognition algorithms, and the interface module are designed and implemented.The system designed by this paper achieves the basic function required by the intelligent recognition system. It can not only timely analyze and detect the monitoring scene, but also automatically snap the anomaly into the alarm image library.