Probabilistic iterative inversion of combined surface/atmosphere models for atmospheric correction of imaging spectrometer data(NPO-50704-1)

We have developed a software package that permits inversion of imaging spectrometer data using a Bayesian probabilistic formalism and least squares optimization with a "forward model" of photon transport to match an estimate of atmosphere and surface properties to the measured spectrum. It is instantiated as a python package and can take advantage of externally-available radiative transfer packages such as MODTRAN. It runs on data formats such as ASCII text files and binary files similar to those used by NASA instruments such as AVIRIS-NG.