Research On Front-end Image Processing Of Intelligent Surveillance Systems

Intelligent surveillance system and intelligent video analytics have been substantially growing from practical needs in the past decade, being driven by a wide range of applications in transportation and healthcare. Through the analysis of the video image, the intelligent surveillance system can detect the occurrence of abnormal events, and automatically send alarm information. Compared with traditional video surveillance, it can reduce the working pressure of watchstander, improve work efficiency. In the aspect of video analysis, compared with the platform mode, the front mode improves the real-time performance of the target detection, and greatly saves the bandwidth. This thesis is concerned with the application and Realization of image processing technology in the intelligent front end.In this thesis, a series of common image processing methods and object detection framework are studied, including the method principle, the method steps, the performance comparison, the result optimization and so on. In particular, this thesis implements a fast moving object detection framework based on background difference, the framework in the test environment can real-time accurately detect moving targets, not influenced by light intensity and small objects. With raspberry pi2 mode B+ as video analysis server, We designed and implemented an intelligent front-end video analysis software, which integrates panoramic video surveillance, object detection and tracking, video synthesis, image cloud storage, sensor monitoring and other functions, can meet the needs of the school, home and other scenes. The main work of this thesis includes the following aspects:First, through the analysis of the overall demand for the system, it is clear that the system requirements including three important parts: real-time panoramic monitoring, data cloud storage, intelligent front-end video analysis. Based on system requirements, this thesis analyzes the front-end video analysis software, back-end business management software, the client application.Second, this thesis analyzes the advantages and implementation mode of the intelligent front-end video analysis. A design scheme of intelligent surveillance system based on panoramic video sensing and cloud video distribution is proposed. According to the design scheme, this thesis provides system's overall architecture, including the logical architecture and physical architecture, framework and business flow of intelligent front-end monitoring software, and back-end services software hierarchical structure model and hierarchical communication model.Third, according to the system requirements and software architecture, we completed the intelligent front end, business back-end, client program design. The implementation steps of main modules are described in this thesis.Last, we test the function of each module by experiments, analyze and summarize the results of the experiment. The experimental results show that the proposed algorithm is effective and robust.