Method of selecting an algorithm for use in processing hyperspectral data

1. A method of selecting an algorithm for use in processing hyperspectral data comprising:

providing a set of algorithms, each having qualities for processing certain characteristics of hyperspectral data, and including at least Spectral Information Divergence (SID), Spectral Angle Mapping (SAM), Zero Mean Differential Area (ZMDA), Mahalanobis Distance, and Bhattacharyya Distance;

accessing frame characteristics of the hyperspectral data;

selecting at least one characteristic of the hyperspectral data;

establishing a tolerance for variations in the at least one characteristic from a reference sample of the at least one characteristic;

comparing the at least one characteristic in the hyperspectral data to the tolerance; and

if the at least one characteristic exceeds the tolerance, selecting an algorithm from the set best associated with the at least one characteristic to process the hyperspectral data.

2. The method of claim 1 where the frame characteristics of hyperspectral data include the variability of the illumination of hyperspectral data, the variability of pixels with similar signatures of hyperspectral data.

3. The method of claim 1 including the step of selecting at least two characteristics and if the at least one characteristic does not exceed the tolerance, selecting an algorithm from the set best associated the second characteristic to process the hyperspectral data.