Nonlinear spectral unmixing of hyperspectral images using Gaussian processes

Yoann Altmann, Nicolas Dobigeon, Steve McLaughlin, Jean-Yves Tourneret

Research output: Contribution to journalArticle

Abstract

This paper presents an unsupervised algorithm for nonlinear unmixing of hyperspectral images. The proposed model assumes that the pixel reflectances result from a nonlinear function of the abundance vectors associated with the pure spectral components. We assume that the spectral signatures of the pure components and the nonlinear function are unknown. The first step of the proposed method estimates the abundance vectors for all the image pixels using a Bayesian approach an a Gaussian process latent variable model for the nonlinear function (relating the abundance vectors to the observations). The endmembers are subsequently estimated using Gaussian process regression. The performance of the unmixing strategy is first evaluated on synthetic data. The proposed method provides accurate abundance and endmember estimations when compared to other linear and nonlinear unmixing strategies. An interesting property is its robustness to the absence of pure pixels in the image. The analysis of a real hyperspectral image shows results that are in good agreement with state of the art unmixing strategies and with a recent classification method.
Original languageEnglish
Pages (from-to)2442-2453
Number of pages12
JournalIEEE Transactions on Signal Processing
Volume61
Issue number10
DOIs
Publication statusPublished - 15 May 2013

Fingerprint

pixel
reflectance
method
state of the art
analysis

Cite this

@article{fe5db8c25b5a4329af92df47727c9c43,
title = "Nonlinear spectral unmixing of hyperspectral images using Gaussian processes",
abstract = "This paper presents an unsupervised algorithm for nonlinear unmixing of hyperspectral images. The proposed model assumes that the pixel reflectances result from a nonlinear function of the abundance vectors associated with the pure spectral components. We assume that the spectral signatures of the pure components and the nonlinear function are unknown. The first step of the proposed method estimates the abundance vectors for all the image pixels using a Bayesian approach an a Gaussian process latent variable model for the nonlinear function (relating the abundance vectors to the observations). The endmembers are subsequently estimated using Gaussian process regression. The performance of the unmixing strategy is first evaluated on synthetic data. The proposed method provides accurate abundance and endmember estimations when compared to other linear and nonlinear unmixing strategies. An interesting property is its robustness to the absence of pure pixels in the image. The analysis of a real hyperspectral image shows results that are in good agreement with state of the art unmixing strategies and with a recent classification method.",
author = "Yoann Altmann and Nicolas Dobigeon and Steve McLaughlin and Jean-Yves Tourneret",
year = "2013",
month = "5",
day = "15",
doi = "10.1109/TSP.2013.2245127",
language = "English",
volume = "61",
pages = "2442--2453",
journal = "IEEE Transactions on Signal Processing",
issn = "1053-587X",
publisher = "IEEE",
number = "10",

}

Nonlinear spectral unmixing of hyperspectral images using Gaussian processes. / Altmann, Yoann; Dobigeon, Nicolas; McLaughlin, Steve; Tourneret, Jean-Yves.

In: IEEE Transactions on Signal Processing, Vol. 61, No. 10, 15.05.2013, p. 2442-2453.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Nonlinear spectral unmixing of hyperspectral images using Gaussian processes

AU - Altmann, Yoann

AU - Dobigeon, Nicolas

AU - McLaughlin, Steve

AU - Tourneret, Jean-Yves

PY - 2013/5/15

Y1 - 2013/5/15

N2 - This paper presents an unsupervised algorithm for nonlinear unmixing of hyperspectral images. The proposed model assumes that the pixel reflectances result from a nonlinear function of the abundance vectors associated with the pure spectral components. We assume that the spectral signatures of the pure components and the nonlinear function are unknown. The first step of the proposed method estimates the abundance vectors for all the image pixels using a Bayesian approach an a Gaussian process latent variable model for the nonlinear function (relating the abundance vectors to the observations). The endmembers are subsequently estimated using Gaussian process regression. The performance of the unmixing strategy is first evaluated on synthetic data. The proposed method provides accurate abundance and endmember estimations when compared to other linear and nonlinear unmixing strategies. An interesting property is its robustness to the absence of pure pixels in the image. The analysis of a real hyperspectral image shows results that are in good agreement with state of the art unmixing strategies and with a recent classification method.

AB - This paper presents an unsupervised algorithm for nonlinear unmixing of hyperspectral images. The proposed model assumes that the pixel reflectances result from a nonlinear function of the abundance vectors associated with the pure spectral components. We assume that the spectral signatures of the pure components and the nonlinear function are unknown. The first step of the proposed method estimates the abundance vectors for all the image pixels using a Bayesian approach an a Gaussian process latent variable model for the nonlinear function (relating the abundance vectors to the observations). The endmembers are subsequently estimated using Gaussian process regression. The performance of the unmixing strategy is first evaluated on synthetic data. The proposed method provides accurate abundance and endmember estimations when compared to other linear and nonlinear unmixing strategies. An interesting property is its robustness to the absence of pure pixels in the image. The analysis of a real hyperspectral image shows results that are in good agreement with state of the art unmixing strategies and with a recent classification method.

U2 - 10.1109/TSP.2013.2245127

DO - 10.1109/TSP.2013.2245127

M3 - Article

VL - 61

SP - 2442

EP - 2453

JO - IEEE Transactions on Signal Processing

JF - IEEE Transactions on Signal Processing

SN - 1053-587X

IS - 10

ER -