The Research And Realization Of Key Techniques Of Video Coding Based On H.264/AVC Standard

New generation video compression standard H.264/AVC, which had a profound effect on the video storage and transmission of video mobile communication, stream media and high definition television by its high efficient compression performance, was proposed formally at the end of 2004. Some key technologies of H.264 coding system were researched in this paper to develop its advantage as far as possible. Major achievements of research were listed as follows:According to fundamental principle of intra prediction coding, this paper analyzes some technique details of intra prediction mode selection algorithms in H.264/AVC standard.Then, it introduces Sobel edge detection algorithm and does a histogram simulation with this algorithm, summarizes the advantage and disadvantage of the Sobel edge detection algorithm in terms of simulation result. In the light of huge computational complexity of intra prediction mode selection algorithm in H.264/AVC reference code, this paper proposes a fast intra prediction mode selection algorithm which based on edge information detection and subsample technique. This fast algorithm emphasis on DC prediction mode, and it combines with SATD threshold method to pick out from some candidates in advance to speed up the process of 4x4 luma block mode selection which needs a lot of time in intra prediction mode.Experiment results indicate that the proposed algorithm improves about 45% and 10% on encoding rate with little change on PSNR and bit rate compared with FS and a mode selection algorithm which adopted Sobel edge detection technique.This paper also does a systematic research on motion estimation algorithm of inter prediction coding in H.264/AVC video coding.It proposes an adaptive search algorithm based on block classification for motion estimation on the basis of motion estimation theorem and classic block matching algorithm.The originality of the new algorithm is : it exploits the information gathered from the previous frame based on the relationship in space and time region of motion vectors to derive a parameter, called CP (Correlation Parameter), which indicates the relationship of motion vectors,and employs CP to classify the blocks in the current frame into potentially dependent blocks and potentially independent blocks to reduce search range.The result of algorithm simulation shows that the new algorithm has lowest SP(Search Point) among all the algorithms,also,it can be seen that the proposed new algorithm has highest PSNR and search precision among all the algorithms except the full search.The advantage of high PSNR and low SAD implies that the prediction of the search region in new alorithm is very accurate so that the possibility of finding the best matching reference block within the predictive search area is very high.In a word, the proposed new algorithm has both higher PNSR and lower computational complexity than many classic motion estimation algorithms.