Surface emissivity retrieval from Digital Airborne Imaging Spectrometer data

Abstract

A study has been carried out on the most recent algorithms for the estimation of land surface emissivity (ε) using high‐resolution data (Digital Airborne Imaging Spectrometer, DAIS) over the Rhine Valley (France) and Castilla La Mancha (Spain). Three published methods have been applied for extracting absolute spectral emissivity information from images recorded during the DAISEX experiment in 1999. They are NDVI Thresholds Method (NDVITHM), Normalized Emissivity Method (NEM) and Temperature/Emissivity Separation (TES). These lather two methods were originally designed to work over geological surfaces. Five methods have been used for extracting relative spectral emissivity. They are temperature‐independent spectral indices method (TISI), reference channel method (REF), emissivity normalization method (NOR), emissivity re‐normalization method (RE), and alpha emissivity method (ALPHA), respectively. NDVITHM and NEM give the same absolute emissivity values with differences between 1% and 0.2% depending on the thermal channel considered, while NDVITHM and TES give the greatest differences, around 2%. The comparison with in situ values shows that NDVITHM gives the best results for vegetation plots, while NEM gives the best ones for bare soil and water plots. The TISI and NOR relative methods give the same relative emissivity values within less than 0.4%, in accordance with Becker and Li's [1990] conclusion as regards the superiority of these methods compared to the others.