A fast algorithm to find all paths for hyperspectral unmixing

Yang Liu; Yi Guo; Feng Li; Lei Xin; Puming Huang

doi:10.1109/IGARSS.2018.8519090

A fast algorithm to find all paths for hyperspectral unmixing

Yang Liu, Yi Guo, Feng Li, Lei Xin, Puming Huang

Western Sydney University

Research output: Chapter in Book / Conference Paper › Conference Paper › peer-review

Abstract

![CDATA[Abundance estimation is one of the most important procedures in spectral unmixing. When the spectral library is fixed, the abundance estimation is to find the optimal subset of the library. This is solved by linear regression with sparsity constraint with nonnegativity i.e. the socalled nonnegative L-1 regression (NNL1). However, it is not clear how to choose the regularisation parameter for a given spectrum to be unmixed. In this paper, a fast algorithm is proposed to find all regularisation paths of NNL1, named as FastNNL1, which selects an optimal result from all paths as the final active set of fractional abundances. The simulation results show that the proposed method performs much better than conventional sparse unmixing algorithms in abundance estimation.]]

Original language	English
Title of host publication	Proceedings of the 2018 IEEE International Geoscience and Remote Sensing Symposium (IGARSS 2018): Observing, Understanding and Forecasting the Dynamics of our Planet, 22-27 July, 2018, Valencia, Spain
Publisher	IEEE
Pages	4269-4272
Number of pages	4
ISBN (Print)	9781538671504
DOIs	https://doi.org/10.1109/IGARSS.2018.8519090
Publication status	Published - 2018
Event	International Geoscience and Remote Sensing Symposium - Duration: 28 Jul 2019 → …

Conference

Conference	International Geoscience and Remote Sensing Symposium
Period	28/07/19 → …

Keywords

algorithms
coding theory
hyperspectral imaging
sparse matrices

Access to Document

10.1109/IGARSS.2018.8519090

Cite this

Liu, Y., Guo, Y., Li, F., Xin, L., & Huang, P. (2018). A fast algorithm to find all paths for hyperspectral unmixing. In Proceedings of the 2018 IEEE International Geoscience and Remote Sensing Symposium (IGARSS 2018): Observing, Understanding and Forecasting the Dynamics of our Planet, 22-27 July, 2018, Valencia, Spain (pp. 4269-4272). IEEE. https://doi.org/10.1109/IGARSS.2018.8519090

@inproceedings{c890092436ee4fb3b248a868a537afde,

title = "A fast algorithm to find all paths for hyperspectral unmixing",

abstract = "![CDATA[Abundance estimation is one of the most important procedures in spectral unmixing. When the spectral library is fixed, the abundance estimation is to find the optimal subset of the library. This is solved by linear regression with sparsity constraint with nonnegativity i.e. the socalled nonnegative L-1 regression (NNL1). However, it is not clear how to choose the regularisation parameter for a given spectrum to be unmixed. In this paper, a fast algorithm is proposed to find all regularisation paths of NNL1, named as FastNNL1, which selects an optimal result from all paths as the final active set of fractional abundances. The simulation results show that the proposed method performs much better than conventional sparse unmixing algorithms in abundance estimation.]]",

keywords = "algorithms, coding theory, hyperspectral imaging, sparse matrices",

author = "Yang Liu and Yi Guo and Feng Li and Lei Xin and Puming Huang",

year = "2018",

doi = "10.1109/IGARSS.2018.8519090",

language = "English",

isbn = "9781538671504",

pages = "4269--4272",

booktitle = "Proceedings of the 2018 IEEE International Geoscience and Remote Sensing Symposium (IGARSS 2018): Observing, Understanding and Forecasting the Dynamics of our Planet, 22-27 July, 2018, Valencia, Spain",

publisher = "IEEE",

note = "International Geoscience and Remote Sensing Symposium ; Conference date: 28-07-2019",

}

Liu, Y, Guo, Y , Li, F, Xin, L & Huang, P 2018, A fast algorithm to find all paths for hyperspectral unmixing. in Proceedings of the 2018 IEEE International Geoscience and Remote Sensing Symposium (IGARSS 2018): Observing, Understanding and Forecasting the Dynamics of our Planet, 22-27 July, 2018, Valencia, Spain. IEEE, pp. 4269-4272, International Geoscience and Remote Sensing Symposium, 28/07/19. https://doi.org/10.1109/IGARSS.2018.8519090

A fast algorithm to find all paths for hyperspectral unmixing. / Liu, Yang; Guo, Yi ; Li, Feng et al.
Proceedings of the 2018 IEEE International Geoscience and Remote Sensing Symposium (IGARSS 2018): Observing, Understanding and Forecasting the Dynamics of our Planet, 22-27 July, 2018, Valencia, Spain. IEEE, 2018. p. 4269-4272.

Research output: Chapter in Book / Conference Paper › Conference Paper › peer-review

TY - GEN

T1 - A fast algorithm to find all paths for hyperspectral unmixing

AU - Liu, Yang

AU - Guo, Yi

AU - Li, Feng

AU - Xin, Lei

AU - Huang, Puming

PY - 2018

Y1 - 2018

N2 - ![CDATA[Abundance estimation is one of the most important procedures in spectral unmixing. When the spectral library is fixed, the abundance estimation is to find the optimal subset of the library. This is solved by linear regression with sparsity constraint with nonnegativity i.e. the socalled nonnegative L-1 regression (NNL1). However, it is not clear how to choose the regularisation parameter for a given spectrum to be unmixed. In this paper, a fast algorithm is proposed to find all regularisation paths of NNL1, named as FastNNL1, which selects an optimal result from all paths as the final active set of fractional abundances. The simulation results show that the proposed method performs much better than conventional sparse unmixing algorithms in abundance estimation.]]

AB - ![CDATA[Abundance estimation is one of the most important procedures in spectral unmixing. When the spectral library is fixed, the abundance estimation is to find the optimal subset of the library. This is solved by linear regression with sparsity constraint with nonnegativity i.e. the socalled nonnegative L-1 regression (NNL1). However, it is not clear how to choose the regularisation parameter for a given spectrum to be unmixed. In this paper, a fast algorithm is proposed to find all regularisation paths of NNL1, named as FastNNL1, which selects an optimal result from all paths as the final active set of fractional abundances. The simulation results show that the proposed method performs much better than conventional sparse unmixing algorithms in abundance estimation.]]

KW - algorithms

KW - coding theory

KW - hyperspectral imaging

KW - sparse matrices

UR - http://handle.westernsydney.edu.au:8081/1959.7/uws:51301

U2 - 10.1109/IGARSS.2018.8519090

DO - 10.1109/IGARSS.2018.8519090

M3 - Conference Paper

SN - 9781538671504

SP - 4269

EP - 4272

BT - Proceedings of the 2018 IEEE International Geoscience and Remote Sensing Symposium (IGARSS 2018): Observing, Understanding and Forecasting the Dynamics of our Planet, 22-27 July, 2018, Valencia, Spain

PB - IEEE

T2 - International Geoscience and Remote Sensing Symposium

Y2 - 28 July 2019

ER -

A fast algorithm to find all paths for hyperspectral unmixing

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this