Multispectral Images and Panchromatic Remote Sensing Images Fusion Algorithm Based on Wavelet Transform

Abstract

Multispectral images and panchromatic remote sensing images carry incomplete spectral information, which reduces the application value of spectral images. The fusion of these two types of images can increase the amount of spectral information, highlight the features, and to a certain extent, make up for the shortcomings of the two when they exist separately. However, the existing fusion algorithms cannot solve the problem of spectral distortion well, so this paper proposes a fusion algorithm of multispectral images and panchromatic remote sensing images, through the low-frequency component of multispectral images and the low-frequency component of panchromatic remote sensing images to get the new low-frequency component by the weighted average operation, the high-frequency component of multispectral images and panchromatic remote sensing images high-frequency component to get the new high-frequency component by adopting the rule of great value fusion. The new low-frequency and high-frequency components are finally inverted by wavelet to get the final fused images. Experiments show that the fusion effect of this algorithm is better than other algorithms; especially it can make up for the defects such as spectral distortion or image chunking that exist in some fusion algorithms. Moreover, when the number of wavelet decomposition layers is 2, the fusion images obtained by this algorithm is of better quality.