An in-scene method for atmospheric compensation of thermal hyperspectral data

Interpretation of surface radiation characteristics by overhead hyperspectral imagery often requires a compensation of the measured radiance data for the absorption and emission effects of the intervening atmosphere. We describe a procedure that accounts for these effects by direct use of the hyperspectral radiance data without recourse to ancillary meteorological data or atmospheric modeling. This in-scene atmospheric compensation (ISAC) procedure is applicable to a broad class of problems. The present work concerns terrestrial surface characterization by remote sensing in the 8-13 mum atmospheric window band. A complete ISAC analysis is carried out in two steps. In the first step, unscaled atmospheric compensation spectra (transmittance and upwelling radiance) are extracted from the data. The step is carried out by application of a specially designed, line-fitting procedure to a scatterplot constructed from the hyperspectral data. The compensation spectra are defined in terms of the slope and intercept parameters of the line. In the second step, these unscaled compensation spectra are scaled to quantitative compensation spectra. Here, this step is carried out using the strength of absorption in the 11.7-mum water band. The compensation procedure is demonstrated by application to hyperspectral imagery obtained with the Aerospace Corporation's Spatially Enhanced Broadband Array Spectrograph System (SEBASS) hyperspectral imaging sensor at the Department of Energy's (DOE) Atmospheric Radiation Measurement (ARM) facility. Example applications demonstrate the ability of the method to remove atmospheric spectral structure from the observed data and reveal the spectral structure intrinsic to the underlying surface. This removal is demonstrated both for near-blackbody surfaces composed of grass and for surfaces containing limestone gravel and Red Clay soil.