Comparative study of pansharpening remote sensing techniques

Pansharpening involves merging multispectral (MS) imagery and panchromatic (PAN) data to generate an output that matches the spatial resolution of the PAN data and retains the spectral resolution of the MS image. Over the past three decades, numerous approaches have emerged to address this challenge. This study aims to provide a comprehensive assessment by benchmarking various MS pansharpening techniques. Specifically, it compares optimized classical methods such as multiresolution analysis (MRA) and component substitution (CS) with third-generation pansharpening methods like variational optimization (VO) and machine learning (ML) techniques. The evaluation is conducted using urban IKONOS (IK) images. Both reduced and full-resolution quantitative assessments are performed.