Processing OMEGA/Mars Express hyperspectral imagery from radiance-at-sensor to surface reflectance

Abstract

OMEGA/Mars Express hyperspectral imagery is an excellent source of data for exploring the surface composition of the planet Mars. Compared to terrestrial hyperspectral imagery, the data are challenging to work with; scene-specific transmission models are lacking, spectral features are shallow making them difficult to detect, and field data for validation is mostly not available. Software is available for calibrating the raw data to radiance-at sensor, but a processing methodology for processing the radiance data to surface reflectance is not straightforward. In this paper, we present a processing methodology for converting OMEGA data from radiance-at-sensor to surface reflectance. To overcome specific issues with OMEGA data, we developed novel methods for determining a per-scene transmittance spectrum, performing an atmospheric correction, and filtering systematic and random noise. The processing methodology produces surface reflectance images, from which single-pixel spectra can be extracted to identify surface features. Consistent with several previous studies, our results indicate the presence on the Martian surface of iron-rich clays.