基于NSCT的多光谱和全色图像的融合

Fusion Algorithm of Multispectral and Panchromatic Images Based on Unsubsampled Contourlet Transform

提出了一种基于无下采样Contourlet变换的多光谱和全色图像的融合方法。该方法在对多光谱影像进行IHS变换的基础上,对多光谱的I分量和高分辨率的全色影像分别进行无下采样Contourlet变换(NSCT),然后对分解得到的近似分量以及各层金字塔各方向的细节分量利用本文提出的一定的融合准则分别对近似分量和细节分量进行影像融合,最后通过无下采样Conlourlet逆变换得到新的I分量,与H,S分量一起还原到RGB空间,最终得到融合后的高分辨率多光谱彩色图像。本文采用一组多光谱图像和全色图像数据进行融合实验,其实验融合图像的目视效果和统计指标均优于传统的IHS融合方法、小波融合方法以及Contourlet变换方法。

A new algorithm is proposed to merge a high-resolution panchromatic image and a low-resolution multispectral image based on unsubsampled contourlet transform （NSCT）. Firstly, the multispectral image is transformed into IIIS space, and I, H, S components are obtained. Then, the original I component and panchromatic image were decomposed into multiresolution representation by using NSCT. The approximated images and detail images resulting from unsubsampled transform were fused in each level of pyramid and different direction of input resources based on certain rtdes. Fusion algorithms aiming at the approximated images and detail images are proposed respectively. Then the fusion image can be obtained by performing inverse unsubsampled transform and inverse IHS transform. A set of multispectral and panchromatic images were used in our test ,the experimental results show that the effect of our method is more satisfacto- ry than the usual traditional IHS method,wavelet fusion method and contourlet transform based method.