Defining shaded spectra by model inversion for spectral unmixing of hyperspectral datasets - theory and preliminary application

The potential of using hyperspectral imagery of canopies to retrieve vegetation and soil information using spectral mixture analysis (SMA) techniques has been the focus of several recent studies. The SMA method estimates the proportion of pixel area that can be attributed to a cover type with a unique spectral profile. Shaded leaf, shaded residue, and shaded soil areas are generally ignored, or treated as equivalent. This paper presents a method of determining shaded spectral reflectance profiles for component cover types by determining the mean multi-scattering ratio (the ratio of shaded-to-sunlit reflectance) and applying that mean to measured sunlit component spectral reflectance. In this method, the multi-scattering ratio is determined by FLAIR model inversion. The resulting component shaded spectral reflectance can then be used as part of the SMA.