The Research And Design Of Embedded Network Video Surveillance Based On ARM

Though the traditional network video monitor solve the problem of remote monitoring, there are still a lot of improvement in terms of human consumption and intelligent processing. Domestic and foreign scholars’research in the field of video surveillance has gradually transferred from PC control to embedded microprocessor control, using embedded chip to achieve the smart video surveillance.This paper design and implement an Ethernet digital home monitoring system that based on ARM-9embedded platform, this system use Samsung S3C2440microprocessor as the control core and use Linux operating system as the running environment of application software, with the improved motion detection algorithm and adaptive threshold selection used in compressed video streams, the detected target image can be send to users’mobile phone by GPRS MMS alarm module, through the design and implementation of the embedded Web server based on B/S structure, users can access the static HTML and view the real-time video through a browser, and users who has access can also set or change the device configuration. The design content of this paper mainly has three aspects: hardware design of the system, software design of the system, research of motion detection algorithm. Hardware design includes hardware selection and circuit design of the minimum system and external extension device; The software design of the system is the key part of the system, first of all is the design of underlying driver software, which include the production and transplantation of Boot-loader, Linux kernel, root file system and USB camera driver, video capture accomplished through the V4L2interface function in kernel, video compression through the MJPEG, video transmission accomplished the TCP/IP protocol, and through the design of Boa server and CGI component to achieve the network view and control in B/S model; At last, according to the fixity of background under family video detection environment and the system processing performance, we choose an improved algorithm that combined with the frame difference method and background subtraction method, with the Otsu adaptive threshold method to segment image, to extract the moving target.By the practical operation shows that system achieves good results in remote video collection and motion detection effect is accurate, and the modular design is convenient for system maintenance and upgrades, which means this system has a high application value.