|
Title
|
|
|
|
Robust supervised method for nonlinear spectral unmixing accounting for endmember variability
|
|
|
Author
|
|
|
|
|
|
|
Abstract
|
|
|
|
| Due to the complex interaction of light with mixed materials, reflectance spectra are highly nonlinear mixtures of the reflectances of the pure materials that are contained within a mixed pixel, making it hard to estimate the fractional abundances of the materials. Changing illumination conditions and crosssensor situations cause endmember variability, further complicating the unmixing. In this work, we propose a supervised approach to unmix mineral powder mixtures, containing endmember variability. The method combines a supervised regression method to learn the nonlinearity with a geodesic distance-based approach, that is invariant to endmember variability. Experiments are conducted on simulated and real mineral mixtures. In particular, we developed datasets of homogeneously mixed mineral powder mixtures, acquired by 2 different sensors, an Agrispec spectrometer and a snapscan shortwave infrared hyperspectral camera, under strictly controlled experimental settings. The proposed approach is compared to other supervised approaches and nonlinear mixture models. |
|
|
|
Language
|
|
|
|
English
|
|
|
Source (journal)
|
|
|
|
IEEE transactions on geoscience and remote sensing / Institute of Electrical and Electronics Engineers. - New York, N.Y., 1980, currens
|
|
|
Publication
|
|
|
|
New York, N.Y.
:
2021
|
|
|
ISSN
|
|
|
|
0196-2892
[print]
1558-0644
[online]
|
|
|
DOI
|
|
|
|
10.1109/TGRS.2020.3031012
|
|
|
Volume/pages
|
|
|
|
59
:9
(2021)
, p. 7434-7448
|
|
|
ISI
|
|
|
|
000690968800026
|
|
|
Full text (Publisher's DOI)
|
|
|
|
|
|
|
Full text (open access)
|
|
|
|
|
|
|
Full text (publisher's version - intranet only)
|
|
|
|
|
|