Remote Sensing Image Fusion Method Based on Nonsubsampled Shearlet Transform and Sparse Representation

摘要

The remote sensing image fusion is an important preprocessing techniquernin remote sensing image processing. In this paper, a remote sensing imagernfusion method based on the nonsubsampled shearlet transform (NSST) with sparsernrepresentation (SR) is proposed. Firstly, the low resolution multispectral (MS)rnimage is upsampled and color space is transformed from Red–Green–Blue (RGB) tornIntensity–Hue–Saturation (IHS). Then, the high resolution panchromatic (PAN)rnimage and intensity component of MS image are decomposed by NSST to high andrnlow frequency coefficients. The low frequency coefficients of PAN and the intensityrncomponent are fused by the SR with the learned dictionary. The high frequencyrncoefficients of intensity component and PAN image are fused by local energy basedrnfusion rule. Finally, the fused result is obtained by performing inverse NSST andrninverse IHS transform. The experimental results on IKONOS and QuickBirdrnsatellites demonstrate that the proposed method provides better spectral quality andrnsuperior spatial information in the fused image than other remote sensing imagernfusion methods both in visual effect and object evaluation.