Effects of empirical versus model-based reflectance calibration on automated analysis of imaging spectrometer data: a case study from the Drum Mountains, Utah

Data collected by the Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) have been calibrated to surface reflectance using an empirical method and an atmospheric model-based method. Single spectra extracted from both calibrated data sets for locations with known mineralogy compared favorably with laboratory and field spectral measurements of samples from the same locations. Generally, spectral features were somewhat subdued in data calibrated using the model-based method when compared with those calibrated using the empirical method. Automated feature extraction and expert system analysis techniques have been successfully applied to both data sets to produce similar endmember probability images and spectral endmember libraries. Linear spectral unmixing procedures applied to both calibrated data sets produced similar image maps. These comparisons demonstrated the utility of the model-based approach for atmospherically correcting imaging spectrometer data prior to extraction of scientific information. The results indicated that imaging spectrometer data can be calibrated and analyzed without a priori knowledge of the remote target.