基于非下采样Contourlet变换和稀疏表示的红外与可见光图像融合方法
首发时间:2013-01-22
摘要:针对非下采样Contourlet变换(NSCT)中低频子带系数稀疏度较低不利于融合的问题,提出基于NSCT和稀疏表示的图像融合方法。首先对红外与可见光图进行NSCT变换;然后对稀疏度较低的低频子带系数提取共有和特有系数,并按照特有系数的活动水平自适应调整权重融合;其次,对稀疏度较高的高频方向子带系数采用同一尺度下方向子带绝对值和取最大的方法融合;最后经NSCT逆变换得到融合图像。实验结果表明,与传统基于变换的DWT、NSCT融合方法以及基于稀疏表示的SOMP、JSR算法比较,本文方法可以获得更好的融合效果。
关键词: 信息处理技术 图像融合 非下采样Contourlet变换(NSCT) 稀疏表示 红外图像 可见光图像
For information in English, please click here
FUSION METHOD FOR VISIBLE AND INFRARED IMAGES BASED ON NON-SUBSAMPLED CONTOURLET TRANSFORM AND SPARSE REPRESENTATION
Abstract:Aiming at the problem of sparseness of low-frequency sub band goes against its fusion in Non-Subsampled Contourlet Transform,a image fusion method based on NSCT and sparse representation was presented. Firstly, infrared and visible images are transformed by NSCT. And then, the common and innovation coefficient of sparseness coefficient of low-frequency sub band was extracted, and the sparse coefficients are consequently weighted by the innovation coefficients adaptively. Secondly, fusion sparse high-frequency sub band by choosing the max add absolute values of high-frequency sub band at the same scale. Finally, the fusion image is reconstructed by the inverse NSCT. The method has better fusion performance both than the traditional fusion method based on DWT, NSCT, and the SOMP, JSR fusion method, which based on sparse representation.
Keywords: Information processing Image fusion Non-Subsampled Contourlet Transform(NSCT) Sparse representation Infrared image Visible image
论文图表:
引用
No.****
同行评议
勘误表
基于非下采样Contourlet变换和稀疏表示的红外与可见光图像融合方法
评论
全部评论0/1000