基于降码率设计的图像编码处理技术研究  

Research on Image Coding Processing Technology Based on Reduced Bit Rate Design

作　　者：李耿 刘怡[1] 占力 杨家斌 张翔 马蒙蒙 LI Geng;LIU Yi;ZHAN Li;YANG Jia-bin;ZHANG Xiang;MA Meng-meng(Southwest Institute of Technical Physics,Chengdu 610046,China)

机构地区：[1]西南技术物理研究所,四川成都610046

出　　处：《红外》2023年第2期41-48,共8页Infrared

基　　金：深圳市科技计划项目(KQTD20190929173954826);国家自然科学基金项目(U1930116;U206720039;12004331)。

摘　　要：为了降低老化的图像采集设备引入的加性噪声所带来的图像数据冗余,研究了加性噪声对图像和视频的影响,提出用三维块匹配(Block Matching 3D,BM3D)算法处理受噪的静态图像,验证了通过降噪来降低图像信息熵和数据量的方法,进而提出降噪--编码方案。首先采用四维视频块匹配(Video Block Matching 4D,VBM4D)算法对受噪视频进行降噪处理,随后进行H.264编码。经过此方案处理的图像信息熵降低16%,数据量减少38%;在视频质量未显著降低的前提下,编码码流减少59.5%。数据表明该方法能在保证视频质量的前提下显著提升H.264的编码压缩率,促进视频在小带宽通信系统中的传输。与H.264以及神经网络等编码方案相比,该方案复杂度低,能相对实时地处理视频。In order to reduce the redundancy of image data caused by additive noise introduced by aging image acquisition equipment,the effect of additive noise on image and video is studied,and the BM3D(block matching 3D)algorithm to process noisy static images is proposed in this paper.It is verified that the image information entropy and data volume can be reduced by reducing noise.Then the noise reduction-coding scheme is proposed.In this scheme,the noisy video is firstly denoised by VBM4D algorithm,and then encoded by H.264.After the scheme processing,the image information entropy is reduced by 16%,the data volume is reduced by 38%,and the encoded code stream is reduced by 59.5%on the premise that the video quality is not significantly reduced.The data show that the encoding compression rate of H.264 can be significantly improved and the transmission of video in small bandwidth communication systems is promoted by this method under the premise of guaranteeing video quality.Compared with H.264 and neural networks,this scheme has low complexity and can process video relatively in real time.

关 键 词：信息熵 加性噪声 视频 VBM4D H.264编码 

分 类 号：TN919.81[电子电信—通信与信息系统]