| In recent years, with the rapid development of information technology and network, intelligent city construction is in full swing all over the country, it makes video security surveillance system have been increasingly widely applied in all industries. The media stream playback, storage, transmission become the important basic applications of video surveillance system, what produce a variety of mass media streaming data is growing. How to solve the reliability and security of the video data storage, It’s facing a realistic evaluation. Aimed at the parameters configuration of the media stream in video monitoring system, many standard protocol had been development and promotion. But based on the position of protecting their interests, many enterprises manufacturers and business organizations set the private coding standard which meet the their own needs, you need to specify the SDK suite that can realize video transcoding data. It’s difficult to meet the demand of playing everywhere for terminal users. It have suffered a great deal of challenges to achieve the goal of video surveillance resources integration and resource sharing.To solve the problems of video surveillance resources integration and resource sharing in video security surveillance system, improve video transcoding efficiency. Aimed at media streaming data encapsulated by PS Stream(Program Steam), this paper adopt Map Reduce programming model to achieve distributed video transcoding. Because the media streaming data is very different from text data in data type, analytic way, etc. This paper research the parallel processing features of Map Reduce programming model, dependencies within video frames about video streaming data segmentation and development process about FFmpeg for video transcoding, the parallel processing, distributed computing are integrated into transcoding processing for video data encapsulated by PS Stream, it also modifies the create method about key-value pair which is processed by the map function, and put forward the GOP-based video segmentation algorithm and FFmpeg-based distributed video transcoding algorithm to improve the efficiency of parallel video transcoding, and design solution and implementation process are introduced in detail in paper.Finally, the paper set up experiment for performance analysis about distributed video transcoding system solutions, the experiment tests show that it promotes efficiency of transcoding process for the media streaming data encapsulated by PS through distributed cluster, and the solution compared with the existing traditional video transcoding system, by increasing the number of machine nodes, it significantly reduces video transcoding time overhead. In the follow-up research work, the improvement process about the optimization of PS Stream data analysis and the algorithmic reliability of video transcoding, shall be as the future research focus in this paper. |
