Improvements in Aerosol Retrieval for Atmospheric Correction

摘要

First-principles atmospheric correction that converts Visible- NIRSWIR spectral imagery to surface reflectance requires an estimate of the scene visibility / aerosol optical depth. This paper describes aerosol upgrades to FLAASH, a first-principles atmospheric correction algorithm developed by Spectral Sciences, Inc. and the US Air Force Research Laboratory. FLAASH utilizes an automated band ratio method for retrieving an average scene visibility from dark pixels. The visibility estimate is combined with a MODTRAN- tm aerosol representation to describe the atmosphere. Recent FLAASH upgrades improve both visibility retrieval and atmospheric correction accuracy. These result in better handling of highly off-nadir viewing geometries, high aerosol optical depths and sensors lacking infrared spectral channels.