H.264 Motion Estimation And Rate Control Algorithm Research

Accompany with the fast development of network and mobile communication, disposition and transmission technology in the field of internet and mobile network of video information and multimedia information have become a hotspot in Chinese informatization, nowadays. The new generation video encoding standard H.264 adopts large amount of advanced encoding decoding technology, and its compressive performance and decoding quality have a large improvement. H.264 is much better than other video encoding standard in signal to noise ratio, compression efficiency and vision effect aspects.The main content of this paper are: motion estimation and rate control, which are two important parts of H.264 standard.H.264 codec enhances the compressive performance, but the operative complication of H.264 is three times of MPEG-4. Motion estimation search takes account of about 80% of the whole encoding complexity. Scholars at home and aboard do a large amount of job to explore various inter frame,intra frame fast research algorithm to solve the difficulty. This paper proposes a kind of fast search algorithm using different search strategy based on different video image, which realizes the aim of fast search, ensures the quality of image and avoids problems that the existing fast search algorithm is too complex when search optimal point.Rate control algorithm affects video codec efficiency and video quality directly. The main target of rate control is minimizing encoding distortion as best as possible under the condition that bit rate is limited. Better rate control algorithm is necessary for acquiring optimization compression efficiency, insuring that encoding data transmits in channel successfully, and obtaining optimal video quality. Some previous scholars have proposed rate control algorithm of using mean absolution distortion to reflect complexity of image. This paper adopts mean absolute change, and express complexity of image, which open out the real reason of image distortion and reflect complexity of image. It uses mean absolute transform instead of mean absolute error, which could quantize error correctly to output rate better.The experiment result shows that: the rate control algorithm adopted by this paper could obtain smoother rate, the motion estimation search algorithm proposed by this paper, cut shot the encoding time largely. And the peak value of signal to noise rate of the image satisfies the requirement.