Matching Items (2)
Filtering by

Clear all filters

Differentiable Programming for Physics-based 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. In this thesis, physics-based approaches…

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. In this thesis, physics-based approaches are incorporated into an end-to-end spectral unmixing algorithm via differentiable programming. First, sparse regularization and constraints are implemented by adding differentiable penalty terms to a cost function to avoid unrealistic predictions. Secondly, a physics-based dispersion model is introduced to simulate realistic spectral variation, and an efficient method to fit the parameters is presented. Then, this dispersion model is utilized as a generative model within an analysis-by-synthesis spectral unmixing algorithm. Further, a technique for inverse rendering using a convolutional neural network to predict parameters of the generative model is introduced to enhance performance and speed when training data are available. Results achieve state-of-the-art 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.
ContributorsJaniczek, John (Author) / Jayasuriya, Suren (Thesis advisor) / Dasarathy, Gautam (Thesis advisor) / Christensen, Phil (Committee member) / Arizona State University (Publisher)
Created2020
132143-Thumbnail Image.png

Brine Stability at Recurring Slope Lineae in Valles Marineris, Mars

Description
Recurring Slope Lineae (RSL) are dark, narrow features which form on steep Martian slopes during warm seasons, lengthening, fade in cold seasons and recurring annually. There are many hypotheses on the formation mechanism of RSL. A number of these hypotheses suggest that RSL are liquid brines flowing on the surface.…

Recurring Slope Lineae (RSL) are dark, narrow features which form on steep Martian slopes during warm seasons, lengthening, fade in cold seasons and recurring annually. There are many hypotheses on the formation mechanism of RSL. A number of these hypotheses suggest that RSL are liquid brines flowing on the surface. Brine based hypotheses often state that sub-surface aquifers are necessary to supply the water needed to recharge RSL. One problem with this is that RSL are observed forming on isolated peaks and ridgelines where a sub-surface aquifer is unlikely. This study uses a thermal model called KRC to examine the correlation between RSL activity and surface temperature at several RSL sites in Valles Marineris. This correlation is compared to the freezing temperature of several brines. Results show an interesting relationship between RSL activity and the surface temperature of very steep (> 60º) slopes. This could indicate that RSL are caused by thermal stresses loosening material on the face of bedrock outcroppings instead of briny flows.
ContributorsPatterson, Bradley (Author) / Christensen, Phil (Thesis director) / Piqueux, Slyvain (Committee member) / Ruff, Steven (Committee member) / Mechanical and Aerospace Engineering Program (Contributor) / Barrett, The Honors College (Contributor)
Created2019-05