HEVC视频编码标准的快速编码算法研究

发布时间：2018-06-26 18:12

本文选题：HEVC + 帧间预测　；参考：《吉林大学》2017年硕士论文

【摘要】：随着数字信息化技术迅速发展,现代人对视觉需求日益增加,传统标清视频已逐渐不能满足人们的诉求,而高清、超高清视频数据量大并且传输速度慢,这对视频压缩编码技术提出新挑战。高效视频编码标准(High Efficiency Video Coding,HEVC)在对原有编码技术改进的基础上,引入部分新技术,使其压缩效率较上一代视频编码标准H.264/AVC提高一倍,是解决高清、超高清视频存储和传输方面问题的有效方法。虽然HEVC对原有编码技术进行提高和增加,但其编码复杂度亦大幅提升,这严重阻碍了高清、超高清视频在实时传输和海量存储等重要领域的应用。如何降低编码复杂度,提高编码效率,使HEVC全面替代H.264/AVC,是当今国内外研究学者共同研究的热门课题。本文首先对视频编码标准的发展历程以及HEVC的制定过程进行了简单介绍,并详细分析其编码过程和关键技术,在进一步分析现有算法的基础上,提出一种基于平均深度信息的帧间编码单元快速选择算法。本文提出的改进算法分为两步:首先提出平均预测深度的概念,即对已编码单元求取平均深度,利用编码帧间的时间相关性以及编码单元间的空间相关性,对当前待编码单元求取平均预测深度,由平均预测深度对当前编码单元进行初步分类。对于纹理简单区域,剪裁掉后面不必要的遍历;对于纹理复杂区域,跳过前面不必要的遍历。而对于纹理中等区域,本文引入第二个概念,利用参考帧已编码单元的率失真代价与总深度的比值作为对中等区域的划分指导,进一步剪裁掉不必要的遍历。本文算法将深度信息和率失真代价信息结合并有效利用,可以准确地反映已编码单元的纹理特征,这也是本文算法可以合理跳过冗余遍历的必要条件。通过对HEVC官方测试序列的测试,该算法可平均减少编码时间55%左右,而在高清视频测试下,对编码时间的减少可达到60%左右。对实验结果的率失真曲线分析可见,本文算法可以保证编码后的视频质量,从而验证了算法的可行性。
[Abstract]:With the rapid development of digital information technology and the increasing demand for vision by modern people, the traditional standard and clear video has been unable to meet the demands of people, and the high-definition, ultra-high-definition video has a large amount of data and a slow transmission speed. This poses a new challenge to video compression and coding technology. High efficiency Video coding Standard (HEVC), based on the improvement of the original coding technology, introduces some new technologies to double the compression efficiency of H.264 / AVC, which is the solution to high definition. High-definition video storage and transmission of the problem of effective methods. Although HEVC improves and increases the original coding technology, its coding complexity is also greatly increased, which seriously hinders the application of high-definition, ultra-high-definition video in real-time transmission and mass storage and other important fields. How to reduce coding complexity and improve coding efficiency so that HEVC can completely replace H.264 / AVC is a hot topic for researchers at home and abroad. In this paper, the development of video coding standards and the establishment of HEVC are introduced briefly, and the coding process and key technologies are analyzed in detail. On the basis of further analysis of existing algorithms, A fast selection algorithm for interframe coding units based on average depth information is proposed. The improved algorithm proposed in this paper is divided into two steps: firstly, the concept of average predictive depth is proposed, that is, the average depth is obtained for the coded unit, the temporal correlation between the coded frames and the spatial correlation between the coding units are used. The average prediction depth is obtained for the current coding unit, and the current coding unit is preliminarily classified by the average prediction depth. For the simple texture region, the traversal is cut off; for the texture complex region, the unnecessary traversal is skipped. For the middle texture region, the second concept is introduced in this paper. The ratio of rate-distortion cost to total depth of the reference frame coding unit is used as the guide to divide the middle region, and the unnecessary traversal is further cut out. By combining depth information with rate-distortion cost information, the proposed algorithm can accurately reflect the texture features of the coded units, which is also a necessary condition for the algorithm to avoid redundant traversal reasonably. By testing the official test sequence of HEVC, the algorithm can reduce the coding time by about 55% on average, and under high-definition video test, the coding time can be reduced by about 60%. By analyzing the rate-distortion curve of the experimental results, it can be seen that the proposed algorithm can guarantee the video quality after coding, thus validating the feasibility of the algorithm.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN919.81

【相似文献】