Data Management for Imaging Spectrometers

Imaging spectrometers are a powerful new tool for performing hyperspectral measurements for remote sensing, environmental monitoring, and target detection and characterization applications. The major problem associated with hyperspectral measurements now, however, is not the technology for performing the measurements but rather the technology for analyzing and recording the high data rates generated by imaging spectrometers.In this Phase I program, we will develop a comprehensive approach for analyzing imaging spectrometer data in real time. A major focus of this program will be the evaluation of existing algorithms and the development of new algorithms for the real time "screening" analysis that must be performed with imaging spectrometer data due to the current impossibility of storing an entire stream of data for any substantial time. A particularly promising algorithm is identified and its implementation in the time domain of Fourier transform spectrometer data is shown to be effective and efficient. Algorithms for detailed analysis of the "screened" data, including a powerful approach recently developed at Aerodyne Research, will also be evaluated and improved. The impact of imaging spectrometer hardware and data processing hardware on the performance of the data analysis routines will be characterized. The Phase I program will develop an optimized comprehensive approach for the real time analysis of hyperspectral data from imaging spectrometers for a remote sensing/monitoring problem of high interest to the Air Force.