Models and methods for automated material identification in hyperspectral imagery acquired under unknown illumination and atmospheric conditions

Abstract

Abstract-The spectral radiance measured by an airborne imaging spectrometer for a material on the Earth's surface depends strongly on the illumination incident of the material and the atmospheric conditions. This dependence has limited the success of material-identification algorithms that rely on hyperspectral image data without associated ground-truth information. In this paper, we use a comprehensive physical model to show that the set of observed 0.4-2.5 m spectral-radiance vectors for a material lies in a low-dimensional subspace of the hyperspectral-measurement space. The physical model captures the dependence of the reflected sunlight, reflected skylight, and path-radiance terms on the scene geometry and on the distribution of atmospheric gases and aerosols over a wide range of conditions. Using the subspace model, we develop a local maximum-likelihood algorithm for automated material identification that is invariant to illumination, atmospheric conditions, and the scene geometry. The algorithm requires only the spectral reflectance of the target material as input. We show that the low dimensionality of material subspaces allows for the robust discrimination of a large number of materials over a wide range of conditions. We demonstrate the invariant algorithm for the automated identification of material samples in HYDICE imagery acquired under different illumination and atmospheric conditions.

Anomaly detection from hyperspectral imagery

“…We compare the predicted performance of two methods of implementing anomaly detection algorithms. For each possible target signature of interest, one might develop a subspace that encapsulates the uncertainty as to illumination and atmospheric transmission [10] and utilize a bank of subspace detectors. Alternatively, one might develop a single subspace that encapsulates all uncertainty as to the target model.…”

“…Furthermore, if the parameters are known within a range of values, then subspace target models may be developed such that for a large set of conditions the target in radiance at the output of the sensor is well represented by the sub-space, i.e., the angle between the target and the sub-space is small [10]. Algorithms have been developed to detect targets contained within a subspace [10]- [12]. If the target subspaces are designed to be invariant to likely variation in significant parameters, then atmospheric compensation is not required for target detection purposes.…”

Does Independent Component Analysis Play a~Role in Unmixing Hyperspectral Data?

“…Hyperspectral observations are described by a generative model. This model takes into account the degradation mechanisms normally found in hyperspectral applications, namely signature variability [38]- [40], abundance constraints, topography modulation, and system noise. The computation of mutual information is based on fitting mixtures of Gaussians (MOG) to data.…”

Face recognition in hyperspectral images

“…Spectral measurements from human tissue, for example, have been used for many years for characterization and monitoring applications in biomedicine. In remote sensing, researchers have shown that hyperspectral data are effective for material identification in scenes where other sensing modalities are ineffective [1]. The introduction of hyperspectral cameras has led to the development of techniques that combine spectral and spatial information.…”