基于信息复用的多路x265视频编码系统设计

发布时间：2018-03-26 13:26

本文选题：视频编码　切入点：HEVC　出处：《电视技术》2017年01期

【摘要】：为了满足不同网络环境下的用户需求,视频节目往往需要经过多次编码以产生不同码率的多个码流。这种多次独立编码方式存在计算资源消耗大和效率低等问题,利用视频在不同码率编码时存在的编码信息相关性,设计了基于编码信息复用的多路x265视频编码系统,使多个从编码器复用主编码器产生的编码信息,如编码单元(CU)的层次划分信息、预测单元(PU)的模式判别信息以及运动估计参考帧信息等,从而大幅提高多路编码效率。该系统支持单路视频输入,进行基于信息复用的多路编码后,同时输出不同码流的多路码流。实验结果表明,与传统的多次独立x265编码相比,该系统在保证视频质量的前提下,对高清1080p和准高清720p视频分别节省了43.32%和52.79%的编码时间。
[Abstract]:In order to meet the needs of users in different network environments, video programs often need to be encoded several times to produce multiple code streams at different bit rates. This multiple independent coding method has many problems, such as high computational resource consumption and low efficiency. A multi-channel x265 video coding system based on the multiplexing of coding information is designed to reuse the encoding information generated from the main encoder by using the correlation of coding information in different coding rates. For example, the hierarchical partition information of the coding unit (CU), the pattern discrimination information of the prediction unit (PUU) and the information of the motion estimation reference frame are used to greatly improve the efficiency of multi-channel coding. The system supports single-channel video input. After multi-channel coding based on information multiplexing, the multi-channel code streams of different bitstreams are output at the same time. The experimental results show that compared with the traditional multi-independent x265 coding, the system can guarantee the video quality. The coding time of 1080p and 720p video is reduced by 43.32% and 52.79% respectively.
【作者单位】：上海交通大学电子工程系图像通信与网络工程研究所;
【基金】：国家自然科学基金项目(61301116)
【分类号】：TN919.81

【相似文献】