Spectral Dark Subtraction: A MODTRAN-Based Algorithm for Estimating Ground Reflectance without Atmospheric Information
Spectral Dark Subtraction (SDS) provides good ground reflectance estimates across a variety of atmospheric conditions with no knowledge of those conditions. The algorithm may be sensitive to errors from stray light, calibration, and excessive haze/water vapor. SDS seems to provide better estimates than traditional algorithms using on-site atmospheric measurements much of the time.
Freedman, Ellis↗