四元数小波变换联合稀疏表示的图像融合  

Image fusion scheme based on quaternion wavelet transform and sparse representation

作　　者：常莉红[1,2] 冯象初[1] 张瑞[1,2] CHANG Lihong FENG Xiangchu ZHANG Rui(School of McUhemaUcs and SuuisUcs,Xidian University, Xi,an 710171, China Maihemaiics Department , Baoji University o f Arts and Sciences,Baoji 721013, China)

机构地区：[1]西安电子科技大学数学与统计学院,陕西西安710171 [2]宝鸡文理学院数学与信息科学学院,陕西宝鸡721013

出　　处：《系统工程与电子技术》2017年第7期1633-1639,共7页Systems Engineering and Electronics

基　　金：国家自然科学基金(61271294;61362029;61379030;61472303);陕西省自然科学基金(2013JM1001)资助课题

摘　　要：基于四元数小波变换和稀疏表示理论,提出了一种图像融合方法,该方法弥补了传统的多尺度理论分析和稀疏表示理论在融合过程中的不足。所提方法分为3步:首先,利用四元数小波变换分解所给的源图像,得到各个尺度下的高通子带和低通子带;其次,对高通子带选用系数绝对值最大和低通子带采用稀疏表示的规则进行融合,获得融合系数;最后,对融合系数进行四元数小波逆变换得到融合图像。此外,对所提融合方法进行了理论分析。在数值实验中用6组测试图像测试所提方法性能,并将融合结果与稀疏表示、离散小波变换、对偶数复小波变换、四元数小波变换等融合方法所得结果进行了主观与客观的比较。实验结果表明,该方法是十分有效的。A novel image fusion scheme based on quaternion wavelet transform and sparse representation（QWTS） is proposed. This method complements the defects existing in traditional multi-scale analysis based image fusion and sparse representation based image fusion. The proposed QWTS based image fusion scheme performs the three stages. Firstly, the quaternion wavelet multiscale decomposition of each source image is per-formed to obtain their low-pass and high-pass coefficients. Secondly, the fused quaternion wavelet coefficients are obtained in the low-pass bands with a sparse representation based fusion approach while in the high-pass bands by fusing the largest absolution values. Thirdly, the fused image is reconstructed by taking inverse qua-ternion wavelet transform. Furthermore,the proposed image fusion scheme is analyzed theoretically. The per-formance of the proposed image fusion scheme is evaluated on six groups testing images , and the fused results is compared with those results obtained by the sparse representation, the discrete wavelet transform, the dual-tree complex wavelet transform and the quaternion wavelet transform image fusion schemes subjectively and objec-tively. Experimental results show that the proposed fusion scheme is very effective.

关 键 词：图像融合 四元数小波 稀疏表示 多尺度分解 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]