Mapping of forest fire damages using imaging spectroscopy

An airborne spectroscopic imager (casi) was flown over a region which was damaged by a forest fire while the radiance from the ground was measured from the blue to the near-IR spectral region. A correlation matrix analysis and the normalized difference vegetation index reveals different degrees (classes) of destruction. More details however can be extracted from a neural analysis in course of autoassociative recalls in a space of reference vectors based on a self-organizing feature map.