基于马尔科夫随机场的图像超分辨技术研究综述  

Image Super Resolution Technology based on Markov Random Field

作　　者：黎海雪 林海涛[1] 姜栋瀚 LI Hai-xue;LIN Hai-tao;JIANG Dong-han(School of Electronic Engineering,Naval University of Engineering,Wuhan Hubei 430033,China;Naval Military Representative Office in Shanghai Jiangnan Shipbuiding Group,Shanghai 201913,China)

机构地区：[1]海军工程大学电子工程学院,湖北武汉430033 [2]海军驻上海江南造船(集团)有限责任公司军事代表室,上海201913

出　　处：《通信技术》2018年第10期2356-2364,共9页Communications Technology

摘　　要：介绍了图像超分辨课题研究背景和研究意义,并对超分辨技术国内外研究现状做了梳理,根据图像输入信息的多少,将图像超分辨分为单帧图像超分辨和多帧图像超分辨两大类,同时分析了单帧图像超分辨与多帧图像超分辨的成像过程,建立了二者的图像退化模型,并基于图像退化模型,从贝叶斯角度推理得到了具有一般性的图像超分辨模型,同时还对多种图像质量评价准则进行了对比,结合图像超分辨的特点,分析了不同评价准则的特点。The research background and significance of image super-resolution topics are introduced, and the research status of super-resolution technology both at home and abroad sorted out.According to the image input information, the image super-resolution is divided into two categories： single-frame image super-resolution and multi-frame image super-resolution.At the same time, the imaging process of single-frame image super-resolution and multl-frame image super-resolution is analyzed, and the image degradation models of these two also established.And based on the image degradation model, a general image super- resolution model is reasoned from the Bayesian angle.Finally, various image quality evaluation criteria are compared.Combined with the characteristics of image super-resolution, the features of different evaluation criteria are analyzed.

关 键 词：单帧图像 多帧图像 超分辨技术 马尔科夫随机场 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]