SIMULATION OF PAN-SHARPENING USING HYPERSPECTRAL DATA TO EVALUATE THE METHOD AND BAND COMBINATIONS

摘要

This study proposes a framework for simulating pan-sharpening using hyperspectral data in order to evaluate pan-sharpening methods and band combinations of panchromatic and multispectral data. This framework was applied to two pan-sharpening methods (block-SVR, Gram-Schmidt spectral sharpening) to characterize the spectral and spatial qualities of pan-sharpened images. Panchromatic images were generated from Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) hyperspectral data by changing the position and width of the spectral waveband, and these were subjected to pan-sharpening with multispectral bands, also generated from hyperspectral data. In this study the multispectral bands had fixed spectral positions and bandwidths for simplicity. The pan-sharpened multispectral data were evaluated from the viewpoints of both spectral and spatial qualities using three indices: ERGAS, Q_(AB)~f, and QILV.