Comparison of various techniques for calibration of AIS data

The Airborne Imaging Spectrometer (AIS) samples a region which is strongly influenced by decreasing solar irradiance at longer wavelengths and strong atmospheric absorptions. Four techniques, the Log Residual, the Least Upper Bound Residual, the Flat Field Correction and calibration using field reflectance measurements were investigated as a means for removing these two features. Of the four techniques field reflectance calibration proved to be superior in terms of noise and normalization. Of the other three techniques, the Log Residual was superior when applied to areas which did not contain one dominant cover type. In heavily vegetated areas, the Log Residual proved to be ineffective. After removing anomalously bright data values, the Least Upper Bound Residual proved to be almost as effective as the Log Residual in sparsely vegetated areas and much more effective in heavily vegetated areas. Of all the techniques, the Flat Field Correction was the noisest.