Graduate Thesis Or Dissertation
 

Foreground signature extraction for a non-linear hyperspectral mixing model

Public Deposited

Downloadable Content

Download PDF
https://ir.library.oregonstate.edu/concern/graduate_thesis_or_dissertations/0k225k949

Descriptions

Attribute NameValues
Creator
Abstract
  • Hyperspectral images consist of intensity measurements associated with a large number of EM wavelengths per pixel. When a pixel contains a single material, the measurements of the pixel match the spectral signature of the material. When multiple materials are present in a single pixel, spectral mixing occurs, and signatures of each of the materials are combined to form the measured signature. Various physical models for spectral mixing have been considered, such as the linear and bilinear mixing models. The process of extracting the original signatures from their mixture is referred to as hyperspectral unmixing. Specific problems in hyperspectral unmixing include material abundance estimation, classification, and material signature (endmember) extraction. The tight coupling of model parameters, and the inherent ambiguity of models with matrix product terms, present significant challenges to performing unmixing. Finding conditions under which reasonable estimates of parameters can be obtained for a given model is a key goal. This thesis considers the problem of foreground material signature extraction in an intimate (nonlinear) mixing setting, in which a foreground material signature can appear in combination with multiple background material signatures. We explore a framework for foreground material signature extraction based on a patch model that accounts for such background variation. We identify data conditions under which a foreground material signature can be extracted up to scaling and elementwise-inverse variations. We present algorithms based on volume minimization and endpoint member identification to recover foreground material signatures under these conditions. Numerical experiments on real and synthetic data illustrate the efficacy of the proposed algorithms.
Contributor
License
Resource Type
Date Issued
Degree Level
Degree Name
Degree Field
Degree Grantor
Commencement Year
Advisor
Committee Member
Academic Affiliation
Rights Statement
Publisher
Peer Reviewed
Language
Embargo reason
  • Pending Publication
Embargo date range
  • 2023-04-04 to 2024-01-11

Relationships

Parents:

This work has no parents.

In Collection:

Items

Thumbnail Title Date Uploaded Visibility Actions
file details HollisJarrodJ2023.pdf 2023-04-27 Public Download