Intelligent Video Surveillance System Based On The Dm642 And Overlapping Target Detection Algorithm

The development of economy and society improves the living situations of people. At the same time, people pay more attentions to the society security. And with the development of electrical techniques, computer techniques and signal processing techniques, it's possible to apply high-tech, information-based and intelligent techniques to the society security. Among other things, video surveillance systems are used widely because it can provide intuitive and reliable information. At present, video surveillance systems have been developed to its 4th generation which can provide digital, network-based and intelligent video surveillance service.The task of this paper is to develop and design an intelligent video surveillance platform based on DM642, including hardware and software environment, and research the overlapped objects detecting algorithm in the tracking of vehicles based on this platform. This paper involves three aspects of the system which is hardware, software and algorithms. And this paper will pay more attentions to the software part. Thus, after firstly introduces general design and the hardware structure of the system, then the software design will be introduced. Finally, this paper will discuss the overlapped objects detecting algorithm.The task of software development, include the system software framework design based on the DSP/BIOS, the development of the software modules, and the video driver design based on the TI IOM model. For the task of algorithm research, include the summarizing of the overlapped objects detecting algorithm at home and abroad, the research and improvement of the overlapped objects detecting algorithm based on rectangle model detection, then propose an overlapped objects detecting algorithm based on convex hull. At last, compare two of the algorithms.