This study evaluates the impact of different panchromatic (PAN) images with varying spatial resolutions on hyperspectral (HS) pansharpening performance. While PRISMA (Precursore Iperspettrale della Missione Applicativa) HS imagery provide valuable spectral information at 30-m resolution, many applications require higher spatial detail. Using PRISMA satellite imagery from Agadir, Morocco, the research systematically compares three different PAN sources: PRISMA's native 5-m PAN band, downsampled 10-m PRISMA PAN image, and Sentinel-2 (S2) derived 10-m PAN image. Six pansharpening algorithms were tested across two distinct sites (urban and mixed rural-urban): three conventional methods and three advanced methods. Pansharpening performance was evaluated through both qualitative assessment and quantitative metrics. The methodology considered how each algorithm handles the trade-off between spectral fidelity and spatial enhancement. Results demonstrated that PRISMA's native 5-m PAN consistently delivered superior results across metrics, particularly when pansharpened with advanced algorithms like MTF-GLP and MTF-GLP-HPM, which preserved spectral characteristics while enhancing spatial details. The 10-m PRISMA PAN showed moderate effectiveness, while S2 10-m PAN produced the least favourable outcomes due to limited spatial resolution and spectral compatibility. These findings emphasize the importance of selecting high-resolution PAN imagery from the same sensor platform when conducting HS pansharpening for environmental monitoring applications.
V. Yilmaz (Thu,) studied this question.