四元数小波变换联合稀疏表示的图像融合

发布时间：2018-05-04 22:20

本文选题：图像融合 + 四元数小波　；参考：《系统工程与电子技术》2017年07期

【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。
[Abstract]:Based on quaternion wavelet transform and sparse representation theory, an image fusion method is proposed, which makes up for the shortcomings of traditional multi-scale theory analysis and sparse representation theory in the fusion process. The proposed method is divided into three steps: firstly, the high pass subband and the low pass subband are obtained by using the source image decomposed by the quaternion wavelet transform. The maximum absolute value of coefficients and the sparse representation rule are used to fuse the high pass subbands and the fusion coefficients are obtained. Finally, the fusion image is obtained by inverse quaternion wavelet transform for the fusion coefficients. In addition, the proposed fusion method is theoretically analyzed. In the numerical experiment, six groups of images are used to test the performance of the proposed method, and the fusion results are combined with sparse representation, discrete wavelet transform, even complex wavelet transform. The results obtained by quaternion wavelet transform and other fusion methods are compared subjectively and objectively. The experimental results show that this method is very effective.
【作者单位】：西安电子科技大学数学与统计学院;宝鸡文理学院数学与信息科学学院;
【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
【分类号】：TP391.41

【相似文献】