Design And Implementation Of Video Transmission System Based On 3G Networks

The improvement of wireless communication technology makes it possible to transmit video stream over mobile Internet, while relatively low-speed and more complex network environment is vulnerable enough to bring along many challenges, mainly focuses on the bit rate, compression rate and so on. After researching traditional video compression algorithm, we present a control algorithm based on motion detection and, finally implement the whole embedded-based video transmission system used for long-distance, large-scale video surveillance under 3G networks.Firstly, we investigate the development of video compression and network transmission, sum up the fine features and specially problems of designs facing 3G networks.Secondly, the system model of video transmission is introduced in accordance with video capture, video compression and video transmitted, focusing on the part of video compression, especially motion estimation and quantization that determine rate of compression and encode; when involve special technology about 3G networks, we analyze its advantages or disadvantages.Then, analysis of statistical properties of images under the video surveillance environment is shown, including great relevance in time and space realm. After that we divide monitoring scene into different types at a higher level, by which we present the control algorithm for video compression based on motion detection to improve the compression efficiency of surveillance video.Then we implement the system using an embedded platform; put forward a scheme about scalable software and hardware designation, during which some details especially multiple access of software are discussed. After that, we test the system mainly from video compression and network transmission, address characteristics and problems of the system based on analysis. The results shows that, the transmission system behaves good performance with the compression control algorithm proposed in this paper; speed improves at 67% while bit rate drops at 26%, and it works even better when the scene changes little. The average time delay is about 223.14ms over 3G networks, while 0.18ms over LAN, which shows excellent performance facing real-time video transmission.