<title>Spectroradiometric remote sensing for military applications</title>

Abstract

The U.S. Government is sponsoring a number ofresearch initiatives to explore the potential of spectroradiometric remote sensing in satisfying the time-sensitive information needs associated with military operations and national security applications. These programs address spectroradiometric phenomenology from the ultraviolet through longwave infrared domains, using imaging and nonimaging sensor technology. In contrast with conventional broadband multispectral imaging surveillance systems; e.g., Landsat, spectrally agile spectror&iiometers, offering high spectral resolution and excellent signal-to-noise performance, have introduced a significant paradigm shift i.e., remote sensing based on a priori knowledge of spectral contrast opportunities, as opposed to serendipity. Hyperspectral and ultraspectral sensors, respectively defined by hundreds or thousands of contiguous spectral bands, may be employed as calibrated instruments or adapted to serve as tunable multispectral imaging systems for spectral reconnaissance and surveillance missions.The information needs associated with military operations may be synthesized into a suite of "levels of information (LOIs)", which may be used for designing experiments, conducting analyses, and quantifying the utility of spectral collection and nonliteral exploitation techniques. Experiments are underway to explore and quantify the satisfaction of key LOIs: characterization, detection, material identification and quantification. Qassification and discrimination are also explored, as part of a material identification taxonomy that is based on inherent spectral properties of material, as opposed to human perception. In contrast with the "literal" spatial and geometric features that have been historically exploited by imagery analysts using conventional techniques, "nonliteral" spectraL textural, temporal, or polarization features are being extracted and exploited.The utility of imaging spectroradiometers and nonliteral exploitation algorithms, models, software, and techniques are being evaluated using analytical, empirical and relative measures of effectiveness. In conjunction with ongoing experiments and demonstrations, radiosity modeling and simulation tools are being developed to study the effects of shadow and adjacency. Leveraging from the link margin calculations employed in telecommunications, spectroradiometric system error budget and performance forecasting tools are being developed. Spatial versus spectral trades are also being studied through simulation. The utility assessment results, simulation tools, and advanced exploitation techniques will be transferred to DoD system acquisition program offices, to facilitate the development of advanced spectral reconnaissance and surveillance systems, as well as their supporting exploitation and dissemination infrastructures.