Effective denoising and classification of hyperspectral images using curvelet transform and singular spectrum analysis

Tong Qiao, Jinchang Ren, Zheng Wang, Jaime Zabalza, Meijun Sun, Huimin Zhao, Shutao Li, Jon Atli Benediktsson, Qingyun Dai, Stephen Marshall

Research output: Contribution to journalArticle

46 Citations (Scopus)

Abstract

Hyperspectral imaging (HSI) classification has become a popular research topic in recent years, and effective feature extraction is an important step before the classification task. Traditionally, spectral feature extraction techniques are applied to the HSI data cube directly. This paper presents a novel algorithm for HSI feature extraction by exploiting the curvelet transformed domain via a relatively new spectral feature processing technique – singular spectrum analysis (SSA). Although the wavelet transform has been widely applied for HSI data analysis, the curvelet transform is employed in this paper since it is able to separate image geometric details and background noise effectively. Using the support vector machine (SVM) classifier, experimental results have shown that features extracted by SSA on curvelet coefficients have better performance in terms of classification accuracies over features extracted on wavelet coefficients. Since the proposed approach mainly relies on SSA for feature extraction on the spectral dimension, it actually belongs to the spectral feature extraction category. Therefore, the proposed method has also been compared with some state-of-the-art spectral feature extraction techniques to show its efficacy. In addition, it has been proven that the proposed method is able to remove the undesirable artefacts introduced during the data acquisition process as well. By adding an extra spatial post-processing step to the classified map achieved using the proposed approach, we have shown that the classification performance is comparable with several recent spectral-spatial classification methods.
LanguageEnglish
Pages119-133
Number of pages15
JournalIEEE Transactions on Geoscience and Remote Sensing
Volume55
Issue number1
Early online date9 Nov 2016
DOIs
Publication statusPublished - 4 Jan 2017

Fingerprint

Spectrum analysis
Feature extraction
transform
wavelet
Processing
data acquisition
Wavelet transforms
Support vector machines
artifact
analysis
Data acquisition
Classifiers
Hyperspectral imaging
method

Keywords

  • hyperspectral imaging
  • the curvelet transform
  • singular spectrum analysis
  • classification
  • support vector machine
  • spatial post-processing

Cite this

Qiao, Tong ; Ren, Jinchang ; Wang, Zheng ; Zabalza, Jaime ; Sun, Meijun ; Zhao, Huimin ; Li, Shutao ; Benediktsson, Jon Atli ; Dai, Qingyun ; Marshall, Stephen. / Effective denoising and classification of hyperspectral images using curvelet transform and singular spectrum analysis. In: IEEE Transactions on Geoscience and Remote Sensing. 2017 ; Vol. 55, No. 1. pp. 119-133.
@article{3c2d8bcfa6514483a782217cbb539ba6,
title = "Effective denoising and classification of hyperspectral images using curvelet transform and singular spectrum analysis",
abstract = "Hyperspectral imaging (HSI) classification has become a popular research topic in recent years, and effective feature extraction is an important step before the classification task. Traditionally, spectral feature extraction techniques are applied to the HSI data cube directly. This paper presents a novel algorithm for HSI feature extraction by exploiting the curvelet transformed domain via a relatively new spectral feature processing technique – singular spectrum analysis (SSA). Although the wavelet transform has been widely applied for HSI data analysis, the curvelet transform is employed in this paper since it is able to separate image geometric details and background noise effectively. Using the support vector machine (SVM) classifier, experimental results have shown that features extracted by SSA on curvelet coefficients have better performance in terms of classification accuracies over features extracted on wavelet coefficients. Since the proposed approach mainly relies on SSA for feature extraction on the spectral dimension, it actually belongs to the spectral feature extraction category. Therefore, the proposed method has also been compared with some state-of-the-art spectral feature extraction techniques to show its efficacy. In addition, it has been proven that the proposed method is able to remove the undesirable artefacts introduced during the data acquisition process as well. By adding an extra spatial post-processing step to the classified map achieved using the proposed approach, we have shown that the classification performance is comparable with several recent spectral-spatial classification methods.",
keywords = "hyperspectral imaging , the curvelet transform, singular spectrum analysis , classification, support vector machine , spatial post-processing",
author = "Tong Qiao and Jinchang Ren and Zheng Wang and Jaime Zabalza and Meijun Sun and Huimin Zhao and Shutao Li and Benediktsson, {Jon Atli} and Qingyun Dai and Stephen Marshall",
note = "(c) 2016 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other users, including reprinting/ republishing this material for advertising or promotional purposes, creating new collective works for resale or redistribution to servers or lists, or reuse of any copyrighted components of this work in other works.",
year = "2017",
month = "1",
day = "4",
doi = "10.1109/TGRS.2016.2598065",
language = "English",
volume = "55",
pages = "119--133",
journal = "IEEE Transactions on Geoscience and Remote Sensing",
issn = "0196-2892",
publisher = "IEEE",
number = "1",

}

Effective denoising and classification of hyperspectral images using curvelet transform and singular spectrum analysis. / Qiao, Tong; Ren, Jinchang; Wang, Zheng; Zabalza, Jaime; Sun, Meijun; Zhao, Huimin; Li, Shutao; Benediktsson, Jon Atli; Dai, Qingyun; Marshall, Stephen.

In: IEEE Transactions on Geoscience and Remote Sensing, Vol. 55, No. 1, 04.01.2017, p. 119-133.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Effective denoising and classification of hyperspectral images using curvelet transform and singular spectrum analysis

AU - Qiao, Tong

AU - Ren, Jinchang

AU - Wang, Zheng

AU - Zabalza, Jaime

AU - Sun, Meijun

AU - Zhao, Huimin

AU - Li, Shutao

AU - Benediktsson, Jon Atli

AU - Dai, Qingyun

AU - Marshall, Stephen

N1 - (c) 2016 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other users, including reprinting/ republishing this material for advertising or promotional purposes, creating new collective works for resale or redistribution to servers or lists, or reuse of any copyrighted components of this work in other works.

PY - 2017/1/4

Y1 - 2017/1/4

N2 - Hyperspectral imaging (HSI) classification has become a popular research topic in recent years, and effective feature extraction is an important step before the classification task. Traditionally, spectral feature extraction techniques are applied to the HSI data cube directly. This paper presents a novel algorithm for HSI feature extraction by exploiting the curvelet transformed domain via a relatively new spectral feature processing technique – singular spectrum analysis (SSA). Although the wavelet transform has been widely applied for HSI data analysis, the curvelet transform is employed in this paper since it is able to separate image geometric details and background noise effectively. Using the support vector machine (SVM) classifier, experimental results have shown that features extracted by SSA on curvelet coefficients have better performance in terms of classification accuracies over features extracted on wavelet coefficients. Since the proposed approach mainly relies on SSA for feature extraction on the spectral dimension, it actually belongs to the spectral feature extraction category. Therefore, the proposed method has also been compared with some state-of-the-art spectral feature extraction techniques to show its efficacy. In addition, it has been proven that the proposed method is able to remove the undesirable artefacts introduced during the data acquisition process as well. By adding an extra spatial post-processing step to the classified map achieved using the proposed approach, we have shown that the classification performance is comparable with several recent spectral-spatial classification methods.

AB - Hyperspectral imaging (HSI) classification has become a popular research topic in recent years, and effective feature extraction is an important step before the classification task. Traditionally, spectral feature extraction techniques are applied to the HSI data cube directly. This paper presents a novel algorithm for HSI feature extraction by exploiting the curvelet transformed domain via a relatively new spectral feature processing technique – singular spectrum analysis (SSA). Although the wavelet transform has been widely applied for HSI data analysis, the curvelet transform is employed in this paper since it is able to separate image geometric details and background noise effectively. Using the support vector machine (SVM) classifier, experimental results have shown that features extracted by SSA on curvelet coefficients have better performance in terms of classification accuracies over features extracted on wavelet coefficients. Since the proposed approach mainly relies on SSA for feature extraction on the spectral dimension, it actually belongs to the spectral feature extraction category. Therefore, the proposed method has also been compared with some state-of-the-art spectral feature extraction techniques to show its efficacy. In addition, it has been proven that the proposed method is able to remove the undesirable artefacts introduced during the data acquisition process as well. By adding an extra spatial post-processing step to the classified map achieved using the proposed approach, we have shown that the classification performance is comparable with several recent spectral-spatial classification methods.

KW - hyperspectral imaging

KW - the curvelet transform

KW - singular spectrum analysis

KW - classification

KW - support vector machine

KW - spatial post-processing

U2 - 10.1109/TGRS.2016.2598065

DO - 10.1109/TGRS.2016.2598065

M3 - Article

VL - 55

SP - 119

EP - 133

JO - IEEE Transactions on Geoscience and Remote Sensing

T2 - IEEE Transactions on Geoscience and Remote Sensing

JF - IEEE Transactions on Geoscience and Remote Sensing

SN - 0196-2892

IS - 1

ER -