文献类型：期刊文献

中文题名：基于DWT-IRLS的压缩感知图像融合

英文题名：Image fusion based on CS of DWT-IRLS

作者：李尊[1];苗同军[1]

第一作者：李尊

通讯作者：Li, Z[1]

机构：[1]新乡学院物理与电子工程学院

第一机构：新乡学院物理与电子工程学院

通讯机构：[1]corresponding author), Xinxiang Univ, Colledge Phys & Elect Engn, Xinxiang 453000, Peoples R China.|[11071]新乡学院;

年份：2017

卷号：32

期号：1

起止页码：35-39

中文期刊名：液晶与显示

外文期刊名：Chinese Journal of Liquid Crystals and Displays

收录：CSTPCD;;Scopus(收录号：2-s2.0-85010064605);WOS:【ESCI(收录号:WOS:000415697100007)】；北大核心:【北大核心2014】；CSCD:【CSCD2017_2018】；

语种：中文

中文关键词：图像融合;压缩感知;小波变换;迭代权重最小二乘法

外文关键词：image fusion;compressive sensing;DWT;IRLS

摘要：基于压缩感知的图像融合属于像素级层次的图像融合。传统的DWT压缩感知图像融合研究对象是整个稀疏系数,但小波系数中低频系数并非稀疏,因此影响融合质量。针对此,提出一种基于DWT-IRLS的压缩感知图像融合。首先对图像进行DWT转换,针对高频系数采样测量;然后对高频系数和低频系数进行融合,并且引入迭代权重最小二乘法(IRLS)算法,重构高频系数;最后经DWT逆转换,得到融合图像。实验证明:通过4个客观评价指标和主观评价对比,2组实验融合效果均得到提高。此方法在一定程度上可提高图像融合效果,具有一定的实用价值。
Image fusion based on compressed sensing belongs to the level of image fusion at pixel level.Traditional image fusion based on DWT compressed sensing is for whole sparse coefficient.Because the coefficient of low frequency is not sparse,the quality of image fusion is bad.Aiming at this problem,this paper proposes the image fusion based on CS of DWT-IRLS.First,the image is decomposed by DWT and the coefficient of high frequency is measured;Then,the coefficient of low frequency and the coefficient of high frequency are fused in different domain,the coefficient of high frequency is reconstructed by IRLS algorithm;Finally,the fused image is obtained through inverse DWT.The experiment shows the fusion effect of 2 group is improved by comparison with four objective evaluation and subjective evaluation.This method can improve the image fusion,it has practical value.