Research On The Key Technology Of Server For Video Monitoring System Based On ARM And Linux

With particular applications in ARM and Linux being focused on, the present video monitoring system and its technique paths are researched in compare. A novel embedded video monitoring system is proposed with monitor point and video server dividing the work effectively. And then, the key technology of video server is analyzed, focusing on video data transferring principle between monitoring point and video server, streaming media transferring principle between video server and public network, flow and congestion control, video storage, moving image detection, warning principle, etc.Video data transferring principle between monitoring point and video server: The video data gathered by monitoring point is compressed and encapsulation by Ethernet, in order to alleviate the burden of server. The message between them is transferred by Ethernet frame, so it can address by MAC directly and strip the redundancy of IP header. In order to offer transferring service by positive end-to-end QoS, priority marker and resource reservation are used.Video data transferring principle between video server and public network: RTP/RTCP, which is based on multicast, is used instead of multicast directly. Round trip time algorithm based on forecast congestion window is proposed to make sure the network respond in time correctly when congestion happens, and adjust the send rate, so as to reduce packet dropout, get higher send rate and ensure the smoothness of send rate.Video data storage: Two methods are proposed, one is that the data is stored by the IDE hard disk afforded by the server hardware platform; the other is remote storage. The former adopts the way of pre-allocation space, continuous files, frame index and file index.Abnormal image detection and warning principle: Frame-difference Method is used to detect moving image, the acquisition and updating of background image is implemented by statistical method. Triggering warning principle is adopted, which means it will warn and store the video when abnormal image is detected.The technical scheme proposed in this graduate paper is valuable to application, and the performance of test system is stable. So, the system achieves the expected performance, and affords valuable reference for the research in the correlated field.