Research And Development Of Video Surveillance System Based On Streaming Media

With the development of streaming-media and network technology, video surveillance plays an increasingly important role in the national economy. How to utilize the hosts'resource and network resource more efficiently is the core technology of video surveillance.In this dissertation, the R&D of a streaming-based digital video surveillance system is discussed detailedly surrounding with Monitoring and Control System of Robot Welding Product Line, which is a subtask of the Tenth Five Tackle Key Project of Jiangsu Province'R&D of Open Network Industrial Robot and the Application Project of Automobile Welding Product Line of Yuejin Motor Company'. The system implements a lot of functions include video capture, video record, real-time multicast, video-on-demand (VOD) and the opening interface for third-party systems such as parameter monitoring system. The video data transmission part of the system employs the RTP/RTCP protocol; the video playback part is developed based on DirectShow architecture. A dynamical record algorithm is proposed in the dissertation. This algorithm can save a lot of disk space while obtaining well smooth record files. The system provides a control interface for the third systems with which they can control the system in video record. That improves the system in flexibility and opening.To distribute video data on the Intranet better, a layered coding congestion control scheme is proposed. Using the scheme, the RGB video stream will be splitted into several substreams in temporal direction equably, and then each susstream will be encoded and send to the Intranet. The dissertation shows the process and the result of the experiment and the way of implementation. Also, the properties of this scheme are discussed. In this scheme, all layers will be encoded separately, the traditional constain, which separates all the layers as base layer and enhanced layers, and all layers have different important level, is wiped off, so that the scheme can be adapted to a larger scale of network bandwidth. And the scheme provides more flexibility and robustness; also the scheme can be implemented easily.