Video Surveillance System Based Tmsdm643 Design And Implementation

Currently, Video Surveillance Systems used widely in various fields, major companies, banks, public transport needs a stable, reliable and real-time video surveillance system. The needing of reliability and real-time increases the difficulty of the video surveillance system development which involves the design of hardware modules, software and the algorithm. This paper mainly develops the video-driven model and realizes the moving target detection algorithm on the existing hardware platform.In this paper, it firstly describes the TMS320DM643 hardware platform, peripherals on the chip, ITU-R BT.656 standards and the working style of video processing system in detail. Then it describes the capture and display mode of the video port, and after the compartion of the existing model of video-driven, this paper describes the realization and design characteristic of the video-driven model on the DSP. Also it realizes the communications between the DSP and the camera to get better image.Then it compares the complexity of two methods realized the Gaussian Smoothing Filter, also it utilizes the corresponding algorithm to implement the moving target detection under the static and dynamic background respectively, and then realizes the Video Surveillance System based on the TMS320DM643. Simultaneously it optimizes the image processing algorithms based on the C6000 hardware platform, improves the system performance.At the end of this paper, it summaries the methods which are used to realize the system, and then makes a forecast.