Characterizing Clutter in the Context of Detecting Weak Gaseous Plumes in Hyperspectral Imagery

Weak gaseous plume detection in hyperspectral imagery requires thatbackground clutter consisting of a mixture of components such as water, grass, and asphaltbe well characterized. The appropriate characterization depends on analysis goals.Although we almost never see clutter as a single-component multivariate Gaussian(SCMG), alternatives such as various mixture distributions that have been proposed mightnot be necessary for modeling clutter in the context of plume detection when the chemicaltargets that could be present are known at least approximately. Our goal is to show to whatextent the generalized least squares (GLS) approach applied to real data to look for evidenceof known chemical targets leads to chemical concentration estimates and to chemicalprobability estimates (arising from repeated application of the GLS approach) that aresimilar to corresponding estimates arising from simulated SCMG data. In some cases,approximations to decision thresholds or confidence estimates based on assuming the clutterhas a SCMG distribution will not be sufficiently accurate. Therefore, we also describe astrategy that uses a scene-specific reference distribution to estimate decision thresholds forplume detection and associated confidence measures.