基于HEVC的高效视频编码算法设计与实现

发布时间：2018-02-23 04:06

本文关键词： HEVC 帧内编码帧间编码编码复杂度　出处：《西安电子科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：HEVC是继H.264之后出现的新一代视频编码标准,其目的是为了满足现在人类对于高质量视频的需求。HEVC编码器在设计框架上保留了传统的混合编码框架结构,但是在相同视频质量的条件下,HEVC相比H.264节省了大约50%的码率。伴随着编码性能的大幅度提升,编码器的复杂度也大大增加了。因此如何在保证视频质量的前提下,最大限度的减少视频编码的复杂度是现在HEVC研究的主要方向。本文在深入研究了HEVC的编码标准之后,结合已有的优化算法,从帧内预测和帧间预测两个方面出发对编码算法进行了优化,大大降低了视频编码的复杂度。为了达到降低编码器编码复杂度的要求,本文首先介绍了HEVC的编码结构以及涉及到的关键技术,对其中复杂度较高的模块进行了详细说明和分析,并根据模块特点提出了优化的大体方向；然后针对复杂度较高的帧内和帧间模块提出了自己的优化方法。本文的主要工作如下：1.帧内预测算法。本文首先介绍了帧内模式选择上已有的优化算法,并对算法中存在的问题进行了分析,并在此基础之上利用PU像素值的走势特征改进了获得PU边缘主方向的方法,简化了粗选择(RMD)所需的模式数量；其次,对于相邻预测单元之间预测模式的相关性进行了统计分析,提出了一种利用相邻已编码PU预测当前PU帧内模式的算法；最后,利用相邻的CU在深度选择上的相关性,提出了一种利用相邻已编码CU估计当前CU深度的提前终止算法。仿真结果表明,采用上述三种算法可以保证在码率增加1%的前提下,编码时间减少10%以上。2.帧间预测算法。本文首先分析了帧间编码模式选择的特点,统计帧间模式选择与CU划分之间的关系,根据统计的数据提出了一种利用帧间模式选择来提前终止CU划分的算法,测试结果表明该算法能够大幅度的提升编码的速率。最后分析发现系数中非零系数越多编码所需比特数就越多,本文对于测试序列中的量化单元中非零系数所占的比例进行了统计,提出了一种利用量化单元中非零系数的比例来提前终止变换单元继续划分的算法。测试结果表明,变换单元的提前终止算法可以减少编码时间,降低传输码率,大大的提升视频编码效率。
[Abstract]:HEVC is a new generation of video coding standard after H.264. Its purpose is to meet the human needs for high-quality video. The design framework of HEVC encoder retains the traditional hybrid coding framework. But under the same video quality condition, the code rate of H. 264 is saved by about 50%. The encoder complexity increases greatly with the improvement of coding performance. So how to ensure the video quality, To minimize the complexity of video coding is the main research direction of HEVC. After deeply studying the coding standard of HEVC, this paper combines the existing optimization algorithm. The coding algorithm is optimized from the two aspects of intra prediction and inter prediction, which greatly reduces the complexity of video coding. Firstly, this paper introduces the coding structure of HEVC and the key technologies involved, explains and analyzes the modules with high complexity in detail, and puts forward the general direction of optimization according to the characteristics of the modules. Then we propose our own optimization method for intra-frame and inter-frame modules with high complexity. The main work of this paper is as follows: 1. Intra-frame prediction algorithm. Firstly, this paper introduces the existing optimization algorithms for intra-frame mode selection. Based on the analysis of the existing problems in the algorithm, the method of obtaining the main direction of pu edge is improved by using the trend feature of pu pixel value, which simplifies the number of patterns required for rough selection of RMDs. Based on the statistical analysis of the correlation of prediction patterns between adjacent prediction units, an algorithm is proposed to predict the current intra mode of pu using adjacent coded pu, and finally, the correlation between adjacent CU in depth selection is used. An early termination algorithm for estimating the current CU depth using adjacent coded CU is proposed. The simulation results show that the proposed three algorithms can guarantee that the code rate can be increased by 1%. The coding time is reduced by more than 10%. 2. Inter-frame prediction algorithm. Firstly, the characteristics of inter-frame coding mode selection are analyzed, and the relationship between inter-frame mode selection and CU partition is analyzed. Based on the statistical data, an algorithm for preemptively terminating CU partition using inter-frame mode selection is proposed. The test results show that the algorithm can greatly improve the coding rate. Finally, it is found that the more the non-zero coefficients of coefficients, the more bits are required for coding. In this paper, the proportion of non-zero coefficients of quantization units in the test sequence is counted, and an algorithm is proposed to terminate the partition of transformation units ahead of time by using the ratio of non-zero coefficients of quantization units. The test results show that, The early termination algorithm of the transform unit can reduce the coding time, reduce the transmission rate, and greatly improve the efficiency of video coding.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN919.81

【相似文献】