Research On Adaptive Acquisition And Dynamic Stream Extraction Technology Of SIP Video Conference

Which makes peoples in different places communicate as same as face to face, in this way our communication become more efficiency, that video conferencing applications gradually occupied an increasingly non substituted position. In particular,which has become possible that video conferencing system is more simple, convenient and inexpensive, by developing of the Internet and network infrastructure. But by the large of bandwidth fluctuation and uncontrollable of the network transfer quality of the Internet, quality of video communication become very bad. That’s all, it become a important subject to guarantee the transport Qo S(Quality of Service) of the video conferencing.This research tries to improve the transmission Qo S in the unstable network from the point of rate control of video stream. The core content is the SIP(Initiation Protocol Session) video conferencing based on the adaptive acquisition(video capture) and the extraction technology. The main research content includes the analysis of the SIP video conferencing principle and the related stream technology, improvement the AIMD video transmission rate control model based on the RTP/RTCP feedback mechanism, design and implementation of SIP video conferencing system.Research and analyze SIP video conferencing protocol structure and SIP video conferencing video stream technology. In theory, analyze the interactive process and architecture model of video conferencing based on SIP. Research and analyze the structure of H.264 video encoding and the corresponding rate control, video transcoding scheme. According to the video conferencing scene selection rate control scheme and proposed a GOP based total solution for video transcoding scheme.Research and analyze the network congestion control scheme based on RTP/RTCP video transmission, and apply the improved traditional congestion control algorithm AIMD to the rate of application in adaptive video capture and dynamic stream extraction rate decision part. Use RTP/RTCP feedback control mechanism for packet loss rate, loop time and jitter parameters. A double sliding window mechanism is designed to predict the packet loss rate. Smooth loop time using low pass filter. Design video rate decision according to the packet loss rate, loop time and jitter parameters. On the basis of this, design adaptive video capture and dynamic video stream extraction model.Based on Open MCU, Open Sips and other open source projects, do the secondary development, design and implement SIP video conferencing system which have adaptive video capture and dynamic video stream extraction. It is mainly to add the adaptive video capture module to the UA SIP and add the dynamic video streamextraction module to the forwarding server side. This system is characterized by a higher network sensitivity, which can improve the video frequency in the system according to the difference of the network, and can effectively improve the video communication quality.In this research, we consider the application scenarios of video conferencing and the prediction of the transmission rate. Based on the AIMD algorithm, the whole set of network sensitive video conferencing system is designed, which can realize adaptive video capture and dynamic stream extraction. The system can effectively improve the video conferencing system’s Qo S.