Model acquisition and invariant tracking of unknown materials in hyperspectral images

We consider the problem of acquiring models for unknown materials in airborne 0.4-2.5 microm hyperspectral imagery and using these models to identify the unknown materials in image data obtained under significantly different conditions. The material models are generated with use of an airborne sensor spectrum measured under unknown conditions and a physical model for spectral variability. For computational efficiency, the material models are represented by using low-dimensional spectral subspaces. We demonstrate the effectiveness of the material models by using a set of material tracking experiments in HYDICE images acquired in forest and desert environments over widely varying conditions. We show that techniques based on the new representation significantly outperform methods based on direct spectral matching.