Band Selection-Based Dimensionality Reduction for Change Detection in Multi-Temporal Hyperspectral Images
This paper proposes to use band selection-based dimensionality reduction (BS-DR) technique in addressing a challenging multi-temporal hyperspectral images change detection (HSI-CD) problem. The aim of this work is to analyze and evaluate in detail the CD performance by selecting the most informative...
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2017-09-01
|
Series: | Remote Sensing |
Subjects: | |
Online Access: | https://www.mdpi.com/2072-4292/9/10/1008 |
id |
doaj-de56669793b94a6a8881bd92cfba6bd8 |
---|---|
record_format |
Article |
spelling |
doaj-de56669793b94a6a8881bd92cfba6bd82020-11-25T00:09:01ZengMDPI AGRemote Sensing2072-42922017-09-01910100810.3390/rs9101008rs9101008Band Selection-Based Dimensionality Reduction for Change Detection in Multi-Temporal Hyperspectral ImagesSicong Liu0Qian Du1Xiaohua Tong2Alim Samat3Haiyan Pan4Xiaolong Ma5College of Surveying and Geoinformatics, Tongji University, Shanghai 200092, ChinaCollege of Surveying and Geoinformatics, Tongji University, Shanghai 200092, ChinaCollege of Surveying and Geoinformatics, Tongji University, Shanghai 200092, ChinaXinjiang Institute of Ecology and Geography, CAS and the CAS Research Center for Ecology and Environment of Central Asia, Urumqi 830011, ChinaCollege of Surveying and Geoinformatics, Tongji University, Shanghai 200092, ChinaCollege of Surveying and Geoinformatics, Tongji University, Shanghai 200092, ChinaThis paper proposes to use band selection-based dimensionality reduction (BS-DR) technique in addressing a challenging multi-temporal hyperspectral images change detection (HSI-CD) problem. The aim of this work is to analyze and evaluate in detail the CD performance by selecting the most informative band subset from the original high-dimensional data space. In particular, for cases where ground reference data are available or unavailable, either supervised or unsupervised CD approaches are designed. The following sub-problems in HSI-CD are investigated, including: (1) the estimated number of multi-class changes; (2) the binary CD; (3) the multiple CD; (4) the estimated optimal number of selected bands; and (5) computational efficiency. The main contribution of this paper is to provide for the first time a thorough analysis of the impacts of band selection on the HSI-CD problem, thus to fix the gap in the state-of-the-art techniques either by simply utilizing the full dimensionality of the data or exploring a complex hierarchical change analysis. It is applicable to CD problems in multispectral or PolSAR images when the feature space is expanded for discriminant feature extraction. Two real multi-temporal hyperspectral Hyperion datasets are used to validate the proposed approaches. Quantitative and qualitative experimental results demonstrated that by selecting a subset of the most informative and distinct spectral bands, the proposed approaches offered better CD performance than the state-of-the-art techniques using original full bands, without losing the change representative and discriminable capabilities of a detector.https://www.mdpi.com/2072-4292/9/10/1008change detection (CD)hyperspectral imagesdimensionality reductionband selectionmulti-temporal imagesremote sensing |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Sicong Liu Qian Du Xiaohua Tong Alim Samat Haiyan Pan Xiaolong Ma |
spellingShingle |
Sicong Liu Qian Du Xiaohua Tong Alim Samat Haiyan Pan Xiaolong Ma Band Selection-Based Dimensionality Reduction for Change Detection in Multi-Temporal Hyperspectral Images Remote Sensing change detection (CD) hyperspectral images dimensionality reduction band selection multi-temporal images remote sensing |
author_facet |
Sicong Liu Qian Du Xiaohua Tong Alim Samat Haiyan Pan Xiaolong Ma |
author_sort |
Sicong Liu |
title |
Band Selection-Based Dimensionality Reduction for Change Detection in Multi-Temporal Hyperspectral Images |
title_short |
Band Selection-Based Dimensionality Reduction for Change Detection in Multi-Temporal Hyperspectral Images |
title_full |
Band Selection-Based Dimensionality Reduction for Change Detection in Multi-Temporal Hyperspectral Images |
title_fullStr |
Band Selection-Based Dimensionality Reduction for Change Detection in Multi-Temporal Hyperspectral Images |
title_full_unstemmed |
Band Selection-Based Dimensionality Reduction for Change Detection in Multi-Temporal Hyperspectral Images |
title_sort |
band selection-based dimensionality reduction for change detection in multi-temporal hyperspectral images |
publisher |
MDPI AG |
series |
Remote Sensing |
issn |
2072-4292 |
publishDate |
2017-09-01 |
description |
This paper proposes to use band selection-based dimensionality reduction (BS-DR) technique in addressing a challenging multi-temporal hyperspectral images change detection (HSI-CD) problem. The aim of this work is to analyze and evaluate in detail the CD performance by selecting the most informative band subset from the original high-dimensional data space. In particular, for cases where ground reference data are available or unavailable, either supervised or unsupervised CD approaches are designed. The following sub-problems in HSI-CD are investigated, including: (1) the estimated number of multi-class changes; (2) the binary CD; (3) the multiple CD; (4) the estimated optimal number of selected bands; and (5) computational efficiency. The main contribution of this paper is to provide for the first time a thorough analysis of the impacts of band selection on the HSI-CD problem, thus to fix the gap in the state-of-the-art techniques either by simply utilizing the full dimensionality of the data or exploring a complex hierarchical change analysis. It is applicable to CD problems in multispectral or PolSAR images when the feature space is expanded for discriminant feature extraction. Two real multi-temporal hyperspectral Hyperion datasets are used to validate the proposed approaches. Quantitative and qualitative experimental results demonstrated that by selecting a subset of the most informative and distinct spectral bands, the proposed approaches offered better CD performance than the state-of-the-art techniques using original full bands, without losing the change representative and discriminable capabilities of a detector. |
topic |
change detection (CD) hyperspectral images dimensionality reduction band selection multi-temporal images remote sensing |
url |
https://www.mdpi.com/2072-4292/9/10/1008 |
work_keys_str_mv |
AT sicongliu bandselectionbaseddimensionalityreductionforchangedetectioninmultitemporalhyperspectralimages AT qiandu bandselectionbaseddimensionalityreductionforchangedetectioninmultitemporalhyperspectralimages AT xiaohuatong bandselectionbaseddimensionalityreductionforchangedetectioninmultitemporalhyperspectralimages AT alimsamat bandselectionbaseddimensionalityreductionforchangedetectioninmultitemporalhyperspectralimages AT haiyanpan bandselectionbaseddimensionalityreductionforchangedetectioninmultitemporalhyperspectralimages AT xiaolongma bandselectionbaseddimensionalityreductionforchangedetectioninmultitemporalhyperspectralimages |
_version_ |
1725413375919259648 |