Research On Estimation Algorithm Based On H.264

Video Technology, which is the most important branch of information processing, is constantly upgrading and its applications are continuously expanding, with the rapid development of Multimedia Technology, Computer Network Technology and Communications Technology. However, the increasing amount of data brings the challenges to storage device and network bandwidth. H.264, as a new generation of video codec standard proposed by VCEG (Video Coding Expert Group) of ITU-T and MPEG(Moving Pictures Experts Group) of ISO/IEC, have the advantage of higher compression ratio, lower codes rating, better ability of adapting to network bandwidth and other compression. Then Complex algorithms of H.264 also caused that more time is needed for coding. Encoding time of motion estimation accounted for more than 60% of the total encoding time, thus it can be seen coding efficiency of the module have direct influence to the efficiency of H.264 video coding algorithm. So optimizing the algorithms not only can reduce the coding time but also benefit the real-time application of H.264.This paper began with an introduction of the basic theory and core technology in the H.264 encoding standard, then expounds the principle of block matching motion estimation, cards the advantages and disadvantages of five classic motion estimation search algorithm, mainly focus on the Unsymmetrical-cross Multi-Hexagon-grid Search (UMHexagonS). It proposes optimization of three aspects for the original algorithm:Firstly, the partition of the type of motion shorten the searching process; Secondly, the reduction of prediction step of the initial search point for the under moderate movement saves coding time; Finally, the improvement of 5×5 template and asymmetric large hexagonal grid search template reduces the number of search points.In JM16.1 test model, the improved algorithm was implemented in C language, and the simulation results were obtained using VC6.0 compiler. The comparisons of the experimental data and subjective quality showed that PSNR of the improved UMHexagonS algorithms raised an average of 0.004dB, the average rate increase of 0.204%. However, the motion estimation time saves of 13.498% on average, maximum saves 20.03%. The improved UMHexagonS algorithm described in this paper can assure the quality and bit rate of the video while reducing the motion estimation time, and it verifies the availability of the improved algorithm.