A Research Of Network Video Surveillance System Based On B/S Mode

With the development of information technology, network video surveillance system isapplied to various kinds of occasions, thus people are surrounded by different monitoring andcontrol systems where they are living, studying and working. Meanwhile, the software ofmonitoring and control system is more and more intelligent with the monitoring and controlsystem’s wider application. With the popularity of the Internet, there are a variety of network videomonitoring systems based on B/S mode, and people gradually realize the superiority of Webtechnologies in use. Web-based systems can realize the function of specialized software by genericbrowser. It not only simplify the operation, but also more convenience and higher efficiency, andwhat’s more, it saves the cost of development and post-maintenance of the system. Users can videomonitor by browsing the web with any computer and any browser the system owns, withoutinstalling any client software. With the development of the network video monitoring system, alarge number of researches on the intelligent algorithm of video monitoring emerge. Theresearchers realized the recognition algorithm of all kinds of moving object, which can be appliedin different situations. However, there has not been any good development in the teachingmonitoring scenario research, and there is no method can effectively identify the students’behavior of standing up referring to various literature.In this paper, studies on the current popular B/S mode of video monitoring systems andintelligent recognition algorithm of video monitoring is made as following aspects:(1) Under the analysis of the current situation and tendency of the development of networkvideo monitoring system home and abroad, and the comparison and summary of the relatedtechnologies on network video monitoring system, a implementation scheme of video monitoringsystem has made. Based on RTP network transmission protocol, ActiveX control technology,H.264video compression technology and other basis technical, a kind of network videomonitoring system based on Web browser has developed, which implements the access authentication, the function such as real-time video monitoring, snap shooting of the network videomonitoring system.(2) Experimental verification and comparison on moving target recognition algorithm suchas inter-frame difference method and background difference method in simple monitoringscenarios, which finds out that the inter-frame difference method’s identification results may havehollow and double faults, and several kinds of background difference method’s adaptability fordynamic environment is relatively poor, which lay the foundation for further research onidentifying the monitoring of the students’ standing up behavior.(3) A method that can effectively identify the students’ standing up behavior is put forwardaiming at teaching video monitoring scenario which has not been mature so far. Determine theapproximate movement area by using the combination of the inter-frame difference method andbackground subtraction method. Then use the optical flow method for further more accurateidentification, using the movement direction features of the optical flow vector to identify thestudents’ stood up behavior. The experimental contrast has found several other classicalrecognition algorithms for moving target, such as frame differential method and several kinds ofbackground difference methods cannot effectively and accurately identify the standing up behaviorin teaching scenarios, but identify all the moving objects, such as students’ swinging, whichmeanwhile demonstrates effectiveness of the identifying method in this paper.