| With the high-tech development of the Internet, multimedia and mobile terminals, the demand for high-resolution and ultra-high resolution video is more and more urgent. In order to make high-resolution video be widely used, the Video Coding Expert Group (VCEG) of the International Telecommunication Union and the Moving Picture Experts Group (MPEG) of the International Organization for Standardization jointly established the JCT-VC, studied and developed the latest generation of video codec standard: H.265/HEVC standard reference protocol. This new established codec standard can more efficiently achieve high quality storage and high-resolution video transmission in limited bandwidth transmission conditions. Compared to the previous codec protocol, H.265/HEVC has a corresponding improvement in aspects of image segmentation, intra prediction, inter prediction, space transform and quantization, etc. Theory and practice show that the new codec protocol H.265/HEVC has better coding performance.During video encoding session, one of the key factors which decides the merits of coding performance is motion estimation technology. Motion estimation techniques achieve the goal of effective compression by removing the redundant information between video images, therefore, the study on motion estimation in video codec has important significance.In order to be more convenient and flexible to analyze and research the motion estimation of a video sequence, this paper is based on the new technologies of the H.265/HEVC codec protocol, analyzes and discusses the techniques and algorithms related to motion estimation of the H.265/HEVC codec protocol, designes a motion estimation system which can realize the analysis of motion estimate to the original video sequences.This system can read corresponding original video or image sequences according to the path settings, then carry on real-time analysis and computation, finally obtain the block division results of motion estimation, the motion vectors, the predicted motion vectors, the residual value between images and the associated motion parameters. Each parameter information will be displayed in the software interface in real-time, and recorded in a local log file.Based on the basic process of motion estimation, this paper designs the whole process of H.265/HEVC motion estimation system, especially focuses on conducting intensive studies and designs about the main modules of motion estimation which is to be achieved, such as: encoding units classification module, motion vector prediction module and motion search module. Also introduces the realization of H.265/HEVC motion estimation system which is carried on the Windows operating system, using VS2010 application development platform, discusses the methods of multi-threading parallel optimization used in the implementation process.In order to verify the functionality and performance of the system, the paper carries on a multi-level testing with the help of H.265/HEVC standard test software. Firstly, use the H.265/HEVC motion estimation system realized by this article and the subsequent coding module of the standard test software to encode the test sequences, verify the correctness of the encoded and compressed bit stream. The result showes that, H.265/HEVC motion estimation system can achieve the real-time analysis about video sequences, the correct division on coding units, the effective extraction of motion vectors, also predicted motion vectors and the flexible display of other relative motion parameters. Secondly, use the same test sequence to do encoding test for H.265/HEVC motion estimation system achieved in this paper and the standard test software, evaluate the system performance through counting and comparing the parameters such as time consumption and bit rate for encoding. Experimental results show that, H.265/HEVC motion estimation system has better performance in both encoding complexity and real-time performance. |
PDF Full Text Request