We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Hyperspectral Reconnaissance: Joint Characterization of the Spectral Mixture Residual Delineates Geologic Unit Boundaries in the White Mountains, CA.
- Authors
Sousa, Francis J.; Sousa, Daniel J.
- Abstract
We use a classic locale for geology education in the White Mountains, CA, to demonstrate a novel approach for using imaging spectroscopy (hyperspectral imaging) to generate base maps for the purpose of geologic mapping. The base maps produced in this fashion are complementary to, but distinct from, maps of mineral abundance. The approach synthesizes two concepts in imaging spectroscopy data analysis: the spectral mixture residual and joint characterization. First, the mixture residual uses a linear, generalizable, and physically based continuum removal model to mitigate the confounding effects of terrain and vegetation. Then, joint characterization distinguishes spectrally distinct geologic units by isolating residual, absorption-driven spectral features as nonlinear manifolds. Compared to most traditional classifiers, important strengths of this approach include physical basis, transparency, and near-uniqueness of result. Field validation confirms that this approach can identify regions of interest that contribute significant complementary information to PCA alone when attempting to accurately map spatial boundaries between lithologic units. For a geologist, this new type of base map can complement existing algorithms in exploiting the coming availability of global hyperspectral data for pre-field reconnaissance and geologic unit delineation.
- Subjects
WHITE Mountains (N.H. &; Me.); GEOLOGY education; GEOLOGICAL mapping; RECONNAISSANCE operations; MIXTURES; GEOLOGICAL maps; DATA analysis; MODULATIONAL instability
- Publication
Remote Sensing, 2022, Vol 14, Issue 19, p4914
- ISSN
2072-4292
- Publication type
Article
- DOI
10.3390/rs14194914