This proposal responds to technical subtopic a91-021: topography, image intelligence, and space exploitation, which requests that unique approaches for a real time knowledge engineering based tactical decision aids system be developed involving parallel computer architectures. The opportunity is to develop a system that acquires and translates the expert's knowledge into a representation utilizable by a moderately-trained analyst under a robust set of scenarios, including probability constraints. The conceptual design for a knowledge-based system that utilizes hyperspectral data for the remote detection and identification of gaseous materials (called hygas) is to be developed. The knowledge base for the system shall be a database of high resolution spectra and the associated tables of diagnostic spectral parameters, as defined by expert analysts. The system is to use hyperspectral processing methods in a parallel-processing computer environment. The objectives of phase i are to develop a detailed conceptual design for the entire knowledge-based system, and to define the requirements, sources, and methods for populating the gas database. Sets technology, inc. Personnel have many years experience developing and using image processing systems that access multiple spectral databases of diverse materials, and process data at variable spectral resolutions.
