Multi-node Intelligent Home Video Surveillance System Based On Embedded Linux

Along with the fast development of the society and economy, people's life was improved a lot, but the burden of work is growing at the same time. However, the complexity of the society makes us pay so much attention to the home security, and therefore the monitoring system for home is becoming more and more widely used. The traditional home video surveillance system needs a constant or frequent checking for the abnormal events, which costs people a lot of energy, and may lead negligence. Because almost each home has more than one room, and it usually has only one IP to the public network, so a special networking technology is needed for muti-node video surveillance system's remote video transmission. Aside, the development of information technology makes people frequently use computers and net whenever they are working or playing, it would be a universal phenomenon that many of them are surfing the net or enjoying a movie. Due to the situation above, designing an real-time video surveillance system, making its nodes form a local network, and send the data to internet with the only IP of home, making it has the function of anomaly detection, and send a short message to the family when abnormal things occurs, meanwhile, making it support the popular web browser and video player, is of great significance.This paper developed a multi-node intelligent home video surveillance system, which uses the internet for remote video transmission, user can watch the monitoring video of home with a PC at the place of work or leisure. By forming the multi-node local network, user can select any monitoring node to watch the home situation, and the monitoring nodes can also use the only access point to public network at home for remote video transmission. By supporting HTTP and RTSP, remote clients can choose web browser or video player to watch the real-time monitoring video of home. By designing an anomaly detection algorithm based on background differencing and the objects mass center displacement, system can automatically detect the abnormal situation and notify the user, at the same time, system will save the exception video.The paper begins with the development situation of monitoring system, analyzes the meaning and feasibility of developing multi-node intelligent home video surveillance system based on embedded Linux. According to the needs and features of home, the paper proposed the basic developing method of using embedded Linux, and then proposed the network organizing method, video capture, compression and transmission scheme. Meanwhile, the paper designed an anomaly detection algorithm based on the classical ones, and once the abnormal situation occurs, system will save the video and send a short message to the user. With the steps of platform building, system migration, software development, as well as repeated debugging, the multi-node intelligent home surveillance system was developed.At the end of this paper, it concludes the research results, and gives the future outlook.