本文已被：浏览 842次 下载 312次
投稿时间：2020-02-28
投稿时间：2020-02-28
中文摘要: 针对高效视频编码（HEVC）实时编码高清视频和超高清视频而带来视频编码速度慢、计算复杂度高的问题，提出了基于异构多核CPU+GPU处理平台上的并行实时编码算法以及在GPU中基于率失真优化快速搜索算法，以提高视频编码的速度和降低计算复杂度。经实验验证，所提算法简单且有效，在不牺牲率失真性能的前提下，使编码速度得到明显提升，并进一步接近了实时编码的要求。
Abstract:Aiming at the problems of slow video encoding speed and high computational complexity caused by HEVC real-time encoding of high-definition video and ultra-high-definition video, this paper proposes the parallel real-time encoding algorithm on multi-core CPU+GPU processing platform and fast search algorithm based on rate distortion optimization in GPU to improve the speed of video encoding and reduce the computational complexity.After verification, the proposed algorithm is simple and effective, which significantly improves the encoding speed and further approaches the requirements of real-time encoding, without sacrifying the rate-distortion performance.
keywords: real-time encoding heterogeneous multiprocessing platform parallel processing complexity rate distortion optimization
文章编号：20213013 中图分类号：TN949.6 文献标志码：
基金项目:
| 作者 | 单位 | |
| 袁三男 | 上海电力大学 电子与信息工程学院 | samuel.yuan@shiep.edu.cn |
| 王孟彬 | 上海电力大学 电子与信息工程学院 | |
| 张艳秋 | 上海电力大学 电子与信息工程学院 | |
| 陶倩昀 | 上海电力大学 电子与信息工程学院 |
引用文本：
袁三男,王孟彬,张艳秋,等.基于异构多处理平台视频编码并行化复杂度与率失真研究[J].上海电力大学学报,2021,37(3):271-276.
YUAN Sannan,WANG Mengbin,ZHANG Yanqiu,et al.Research on Parallel Complexity and Rate Distortion of Video Encoding Based on Heterogeneous Multiprocessing Platform[J].Journal of Shanghai University of Electric Power,2021,37(3):271-276.
袁三男,王孟彬,张艳秋,等.基于异构多处理平台视频编码并行化复杂度与率失真研究[J].上海电力大学学报,2021,37(3):271-276.
YUAN Sannan,WANG Mengbin,ZHANG Yanqiu,et al.Research on Parallel Complexity and Rate Distortion of Video Encoding Based on Heterogeneous Multiprocessing Platform[J].Journal of Shanghai University of Electric Power,2021,37(3):271-276.