Research On Remote Sensing Image Fusion Based On Color Compensation And Wavelet Transform

Satellite multisensor remote sensing image fusion is a new developing edge cross discipline in remote sensing image processing. It can combine multisensor images and produce a more precise, integrated and reliable estimation and description of them than a single one, and thus has attracted extensive attention in remote sensing application and image project field. The image fusion of multisensor and panchromatic images is one of the faster developing technologies in remote sensing image fusion. Due to the superior multiresolution analytical characteristics of wavelet transform in time and frequency domains and the color enhancement effect of IHS transform for multispectral images, image fusion algorithms based on IHS and wavelet transform have become the main stream in multispectral remote sensing image fusion. Be different from orthogonal wavelet transform, biorthogonal wavelet transform has linear phase and can reconstruct signals exactly. Consequently, this paper puts emphasis on the multispectral and panchromatic remote sensing image fusion based on IHS and biorthogonal wavelet transform. However, for most remote sensing image fusion, the requirements of spectral preservation and spatial characteristics enhancement are usually incompatible. It is hard to satisfy these two requirements only by single fusion method. To solve this inconsistency effectively, we don't confine to one single fusion method or wavelet coefficients fusion rule,but combine them together and improve them. We consider several eigenvectors of the input images synthetically, and use the discrete biorthogonal wavelet transform at the lowest decomposition level to fulfill the new CWT-IHS multisensor remote sensing image fusion algorithm based on color compensation rule.First, the panchromatic image and the intensity component of multispectral image are decomposed based on the discrete biorthogonal wavelet transform at first level. Then, fuse the decomposed wavelet coefficients as follow rules: for detail wavelet coefficients, we introduce the color composition (CC) into local deviation (LD) fusion method, and propose a local compensation (LC) fusion rule; and for approximate wavelet coefficients, we combine CC with High Pass Filtering (HPF), to propose a high-pass-filtering compensation (HC) fusion rule.Finally, perform inverse IHS transform to the wavelet reconstructed new intensity component and the hue and saturation components from the original MS image to get the fused image. The New algorithm has been tested on various multispectral and panchromatic satellite remote sensing images on IKONOS, QuickBird, SPOT, Landsat TM / ETM+ for different areas, and has been compared with some other typical image fusion methods. Experimental results demonstrate that it is simple and effective. The new fusion algorithm possesses superior fusion effect either from subjective visual evaluation, or from objective performance statistical analysis. It can not only enhances the spatial detail features of result image greatly, but also preserves more spectral characteristics from original multispectral images, and fulfills the unification of spatial details enhancement and spectral information preservation effectively.