Subset selection for the analysis of hyperspectral imagery software package

This project is a software package implementing a heuristic algorithm that applies Subset Selection Analysis as a mechanism for data dimension reduction in hyperspectral imagery. The heuristic algorithm is known as Band Subset Selection for Hyperspectral. Imagery, and has been developed in the Laboratory for Applied Remote Sensing and Image Processing LARSIP of the University of Puerto Rico at Mayaguez.; The Band Subset Selection for Hyperspectral Imagery heuristic algorithm takes advantage of the Singular Value Decomposition mechanism (SVD) to estimate the effective dimensionality p of the hyperspectral image, and the QR Factorization with Column Pivoting mechanism (QR-PIV) to obtain the most linearly independent p bands from the original image, which can explain a high percentage of the hyperspectral image data variability.