Pan sharpening aims to create a multispectral, high spatial resolution image by combining the multispectral image (MSI) with a high spatial resolution panchromatic image (PAN). Pan sharpening methods are performed between the MS image, which is the MSI image brought to PAN dimensions with the help of interpolation, and the PAN image. In this study, PAN sharpening is approached as an optimization problem. It is assumed that the optimal solution consists of multiplying the pixels of the MS image by optimized coefficients. It would be costly to optimize all the coefficients in this coefficient matrix one by one. For this reason, these coefficients were tried to be found with 5 different optimizationbased methods. It was also compared with 19 different methods commonly used in the literature. 6 different evaluation criteria were used for this comparison. These comparisons were made on 3 different datasets. It has been observed that the proposed methods are superior to other methods.
Primary Language | English |
---|---|
Subjects | Image Processing |
Journal Section | Computer Engineering |
Authors | |
Early Pub Date | January 30, 2024 |
Publication Date | March 28, 2024 |
Submission Date | December 21, 2023 |
Acceptance Date | January 17, 2024 |
Published in Issue | Year 2024 |