An efficient fusion algorithm of panchromatic and multi-spectral remote sensing images based on wavelet transform

Abstract

An efficient fusion algorithm of panchromatic and multi-spectral remote sensing images based on wavelet transform is proposed. The multi-spectral remote sensing image is firstly decomposed with IHS transform, the corresponding I component, H component, S component are obtained, the I component and panchromatic remote sensing image are decomposed with wavelet transform. The corresponding low-frequency components of wavelet decomposition are fused with the fusion rule based on the feature matching, and the corresponding high-frequency components of wavelet decomposition are fused with the fusion rule based on the sub-region variance. The fusion components are transformed with the inverse wavelet transform, and the new component Inew is gotten. The Inew component, H component, and S component are transformed with the inverse IHS transformation, and the fusion image is obtained. Finally, the performance of the image fusion is evaluated with some criteria such as entropy and average gradient, correlation coefficient, standard deviatio. The experiment results show that the proposed fusion method improves entropy, average gradient, correlation coefficient, standard deviation, and enriches the detailed information of images besides keeping original images information.