Systems and Methods for Differentiable Programming for Hyperspectral Unmixing
Description
Hyperspectral unmixing is an important remote sensing task with applications including material identification and analysis. Characteristic spectral features make many pure materials identifiable from their visible-to-infrared spectra, but quantifying their presence within a mixture is a challenging task due to nonlinearities and factors of variation. This method provides an end-to-end spectral unmixing algorithm via differentiable programming. Excellent results are achieved on both infrared and visible-to-near-infrared (VNIR) datasets as compared to baselines and show promise for the synergy between physics-based models and deep learning in hyperspectral unmixing in the future.
Additional information
Patent number and inventor
17/353,077
Christopher Edwards, John Janiczek, Suren Jayasuriya, Gautam Dasarathy, and Philip Christensen.
Potential applications
Applications include remote sensing with capabilities in agriculture, minerology, landscape surveying, and more using hyperspectral imaging.
Benefits and advantages
These methods provide improved spectral unmixing performance including material identification and classification.
Case number and licensing status
2020-037
This invention is available for licensing.