Nonlocal tensor sparse representation and low-rank regularization for hyperspectral image compressive sensing reconstruction

Jize Xue, Yongqiang Zhao, Wenzhi Liao, Jonathan Cheung-Wai Chan

Research output: Contribution to journalArticle

8 Citations (Scopus)
2 Downloads (Pure)

Abstract

Hyperspectral image compressive sensing reconstruction (HSI-CSR) is an important issue in remote sensing, and has recently been investigated increasingly by the sparsity prior based approaches. However, most of the available HSI-CSR methods consider the sparsity prior in spatial and spectral vector domains via vectorizing hyperspectral cubes along a certain dimension. Besides, in most previous works, little attention has been paid to exploiting the underlying nonlocal structure in spatial domain of the HSI. In this paper, we propose a nonlocal tensor sparse and low-rank regularization (NTSRLR) approach, which can encode essential structured sparsity of an HSI and explore its advantages for HSI-CSR task. Specifically, we study how to utilize reasonably the l1 -based sparsity of core tensor and tensor nuclear norm function as tensor sparse and low-rank regularization, respectively, to describe the nonlocal spatial-spectral correlation hidden in an HSI. To study the minimization problem of the proposed algorithm, we design a fast implementation strategy based on the alternative direction multiplier method (ADMM) technique. Experimental results on various HSI datasets verify that the proposed HSI-CSR algorithm can significantly outperform existing state-of-the-art CSR techniques for HSI recovery.
Original languageEnglish
Article number193
Number of pages24
JournalRemote Sensing
Volume11
Issue number2
DOIs
Publication statusPublished - 19 Jan 2019

Fingerprint

remote sensing
method
state of the art
norm

Keywords

  • hyperspectral image
  • compressive sensing
  • structured sparsity
  • tensor sparse decomposition
  • tensor low-rank approximation

Cite this

@article{1e2bc111cae54b8b8cd03340c7efe4fb,
title = "Nonlocal tensor sparse representation and low-rank regularization for hyperspectral image compressive sensing reconstruction",
abstract = "Hyperspectral image compressive sensing reconstruction (HSI-CSR) is an important issue in remote sensing, and has recently been investigated increasingly by the sparsity prior based approaches. However, most of the available HSI-CSR methods consider the sparsity prior in spatial and spectral vector domains via vectorizing hyperspectral cubes along a certain dimension. Besides, in most previous works, little attention has been paid to exploiting the underlying nonlocal structure in spatial domain of the HSI. In this paper, we propose a nonlocal tensor sparse and low-rank regularization (NTSRLR) approach, which can encode essential structured sparsity of an HSI and explore its advantages for HSI-CSR task. Specifically, we study how to utilize reasonably the l1 -based sparsity of core tensor and tensor nuclear norm function as tensor sparse and low-rank regularization, respectively, to describe the nonlocal spatial-spectral correlation hidden in an HSI. To study the minimization problem of the proposed algorithm, we design a fast implementation strategy based on the alternative direction multiplier method (ADMM) technique. Experimental results on various HSI datasets verify that the proposed HSI-CSR algorithm can significantly outperform existing state-of-the-art CSR techniques for HSI recovery.",
keywords = "hyperspectral image, compressive sensing, structured sparsity, tensor sparse decomposition, tensor low-rank approximation",
author = "Jize Xue and Yongqiang Zhao and Wenzhi Liao and Chan, {Jonathan Cheung-Wai}",
year = "2019",
month = "1",
day = "19",
doi = "10.3390/rs11020193",
language = "English",
volume = "11",
journal = "Remote Sensing",
issn = "2072-4292",
number = "2",

}

Nonlocal tensor sparse representation and low-rank regularization for hyperspectral image compressive sensing reconstruction. / Xue, Jize; Zhao, Yongqiang; Liao, Wenzhi; Chan, Jonathan Cheung-Wai.

In: Remote Sensing, Vol. 11, No. 2, 193, 19.01.2019.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Nonlocal tensor sparse representation and low-rank regularization for hyperspectral image compressive sensing reconstruction

AU - Xue, Jize

AU - Zhao, Yongqiang

AU - Liao, Wenzhi

AU - Chan, Jonathan Cheung-Wai

PY - 2019/1/19

Y1 - 2019/1/19

N2 - Hyperspectral image compressive sensing reconstruction (HSI-CSR) is an important issue in remote sensing, and has recently been investigated increasingly by the sparsity prior based approaches. However, most of the available HSI-CSR methods consider the sparsity prior in spatial and spectral vector domains via vectorizing hyperspectral cubes along a certain dimension. Besides, in most previous works, little attention has been paid to exploiting the underlying nonlocal structure in spatial domain of the HSI. In this paper, we propose a nonlocal tensor sparse and low-rank regularization (NTSRLR) approach, which can encode essential structured sparsity of an HSI and explore its advantages for HSI-CSR task. Specifically, we study how to utilize reasonably the l1 -based sparsity of core tensor and tensor nuclear norm function as tensor sparse and low-rank regularization, respectively, to describe the nonlocal spatial-spectral correlation hidden in an HSI. To study the minimization problem of the proposed algorithm, we design a fast implementation strategy based on the alternative direction multiplier method (ADMM) technique. Experimental results on various HSI datasets verify that the proposed HSI-CSR algorithm can significantly outperform existing state-of-the-art CSR techniques for HSI recovery.

AB - Hyperspectral image compressive sensing reconstruction (HSI-CSR) is an important issue in remote sensing, and has recently been investigated increasingly by the sparsity prior based approaches. However, most of the available HSI-CSR methods consider the sparsity prior in spatial and spectral vector domains via vectorizing hyperspectral cubes along a certain dimension. Besides, in most previous works, little attention has been paid to exploiting the underlying nonlocal structure in spatial domain of the HSI. In this paper, we propose a nonlocal tensor sparse and low-rank regularization (NTSRLR) approach, which can encode essential structured sparsity of an HSI and explore its advantages for HSI-CSR task. Specifically, we study how to utilize reasonably the l1 -based sparsity of core tensor and tensor nuclear norm function as tensor sparse and low-rank regularization, respectively, to describe the nonlocal spatial-spectral correlation hidden in an HSI. To study the minimization problem of the proposed algorithm, we design a fast implementation strategy based on the alternative direction multiplier method (ADMM) technique. Experimental results on various HSI datasets verify that the proposed HSI-CSR algorithm can significantly outperform existing state-of-the-art CSR techniques for HSI recovery.

KW - hyperspectral image

KW - compressive sensing

KW - structured sparsity

KW - tensor sparse decomposition

KW - tensor low-rank approximation

U2 - 10.3390/rs11020193

DO - 10.3390/rs11020193

M3 - Article

VL - 11

JO - Remote Sensing

JF - Remote Sensing

SN - 2072-4292

IS - 2

M1 - 193

ER -