ORASIS framework - benefits to working within the linear mixing model

The Optical Real-time Adaptive Spectral Identification System (ORASIS) continues to be developed by the Naval Research Laboratory. Recently, a new method to perform target identification based on Spectral Angle Mapper (SAM) has been demonstrated. When projected into a reduced dimension space, angles between spectra are not conserved. Proper determination of the dimensions and their coordinates of the subspace can result in maximizing the angles between a spectrum of interest and the general background. This approach can be used for target identification where the "target" can be a rare spectrum or something found in the general background. In addition, new methods to compress hyperspectral data using the ORASIS in combination with wavelets have been demonstrated. Variations on this approach continue to be worked on and will be discussed. Recent improvements to these methods as well as new directions being pursued will be discussed.