SAR and Multispectral Image Fusion Using Generalized IHS Transform Based on à Trous Wavelet and EMD Decompositions

Abstract

In the intensity hue saturation (IHS) based image fusion, the enhancement of the spatial information often leads to the distortion of the information in the spectral domain. In this paper, a spectral preserve fusion method is developed by introducing a trous wavelet transform (AWT) and empirical mode decomposition (EMD) into one generalized IHS transform, which extends panchromatic sharpening of low-resolution multispectral images (LRMIs) through detail injection to the fusion of LRMIs and synthetic aperture radar imagery (SARI). The original LRMIs are first transformed into one low-resolution intensity component (LRIC) and the details from the SARI are abstracted using AWT. The LRIC and the approximation image of the SARI are fused through an intrinsic mode functions based model. Subsequently, high RIC produced through an inverse AWT is then substituted for the old one. High RMIs are obtained through an inverse IHS transform. Experiments reveal that this approach performs better than the AWT and IHS-AWT-based methods in preserving the spectral information of the LRMIs. The performance of the image fusion is examined both visually and quantitatively.