The Research And Realization On Image Surveillance System Based On Motion Detection

Image as one of mass media won increasingly acceptance in virtue of intuitionistic, convenient and content-abundance form. With these advantages, detection system based on video was widely taken by home surveillance, industry surveillance, traffic scout, iatrical image and etc. Embedded video detection system which integrated the embedded technology, video compression technology and net transmission technology, could be more robust, real-time, low-complexity, low-cost and portable. The advancement of embedded video detection technology directs the trend of video detection system. By the development of computer technology, digital image technology and computer networks, video detection system based on embedded could not be confined in mono-functions, such as image capture, image transfer and image storage. The demanding requirement on intelligence, automation and easy-operation is absolutely inevitable.We construct the new video detection system which was based on motion detection. The motion detection algorithm based on luminance difference algorithm, which was proved to be low complexity but accurate and efficient, became the core of the whole system. According to the key motion detection, this paper proposed the auto exposure algorithm and image compression both based on motion detection. The applications of the two algorithms enhance the quality of the image effectively and compress the size of the image to a contented level by optimizing the storage and net transfers. The whole software system designed and realized on ARM9 platform were divided into four modules, named as capture module, detection and control module, compression module and storage module. The software system was based on Linux and its reliability was satisfied.