Research Of Video Compression Rate-control Based On Markov Decision Processes

In recent years, with the development of digital electronic technology, the digital video is having become the most important information carrier of multimedia communication system.Compared with figure,audio and other information form, digital video information has a lot of data, with the storage capacity and bandwidth condition are limited, high efficient compression of the data video increasingly become more urgent requirement. This requirement needs rate control algorithm, it can prevent the encoded video bit rate exceeding the practical network bandwidth and reducing loss of video information and network congestion. Consider the importance of rate control algorithm, this paper studying rate control technology deeply.Now H.265 is the newest standard after H.264, but it has a high demand for computer hardware, the environment is just beginning, and the problem of high power when it is used in mobile devices is unsolved. So we studied the rate control proposed by mainstream coding standard-H.264. To research rate control algorithm, we must build a more accurate rate-distortion model and control buffer state. This paper makes the systematic analysis about the classical rate control algorithm and rate distortion optimization technology, meanwhile summaries and references them, and improves and amends the classical rate distortion model,then this paper combines Markov decision processes with rate control algorithm, finally proposes a new improved frame layer rate control algorithm based on Markov decision processes.In this paper, JM 18.0 is used as a test platform,this paper tests through many different complexity CIF sequences. The result shows that the rate control algorithm based on Markov decision processes has superior performance in quality of coding and utilization ratio of buffer.Meanwhile, the proposed rate distortion model can make more accurate estimates about coding bits, distortion and quantization parameter.