Comparison between Multi-Scale Decomposition and Spatial Frequency based Methods for the Integration of Panchromatic and Multispectral Pictures

Abstract

A technique for merging a high-resolution PAN image with a low-resolution multispectral image from LISS-III and World View-2 is suggested in this paper. The spatial-frequency technique, in particular the pseudo-Wigner distribution (PWD), offers pixel-by-pixel analysis, shift invariance, and characterization of local spectral features of nonstationary images. The suggested approach is compared with multi-scale decomposition techniques like NSCT and SWT. Using root mean square error (RMSE), peak signal to noise ratio (PSNR), correlation coefficient (CC), relative dimensionless global error (ERGAS), and RASE, the performance of the suggested technique, the non-subsampled contour-let transform, and the stationary wavelet transform (SWT), is compared. Based on thorough tests using the LISS-III and World View-2 datasets, it is observed that the suggested method is statistically and visually more effective than traditional picture fusion algorithms. The suggested approach enhances spatial resolution while retaining spectral data.