基于NSCT和卷积稀疏表示的红外与可见光图像融合

针对红外与可见光图像融合时出现的细节模糊、对比度降低等问题,论文提出了一种基于非下采样轮廓波变换(Nonsubsampled Contourlet Transform,NSCT)和卷积稀疏表示(Convolutional Sparse Representation,CSR)的图像融合方法.首先,分别对红外图像和可见光图...

Infrared and Visible Image Fusion Based on NSCT and Convolutional Sparse Representation

Aiming at the problems of blurred details and low contrast in infrared and visible image fusion,an image fusion method based on nonsubsampled contourlet transform(NSCT)and convolutional sparse representation(CSR)is proposed in this paper. Firstly,the infrared image and visible image are decomposed by NSCT to obtain their high-frequency subbands and low-frequency subband. Then,the high-frequency subbands are enhanced by guided filtering,and the enhanced high-frequency subband coefficients are fused by selecting the maximum strategy. At the same time,the fused sparse coefficient map of low-frequency subband is obtained by using CSR model and maximum selection strategy,and the low-frequency subband coefficients are reconstructed by convolution with learning dictionary. Finally,the fused subband coefficients are transformed by inverse NSCT to obtain the final fused image. The fusion experiments are carried out for typical infrared and visible images. The experimental results show that,compared with other latest methods,the subjective fusion images obtained by this approach retain more detail information from the source images and have less artifacts,and the better fusion performances are also achieved according to the objective evaluation index.