Joint bilateral filtering and spectral similarity-based sparse representation: a generic framework for effective feature extraction and data classification in hyperspectral imaging

Tong Qiao, Zhijing Yang, Jinchang Ren, Peter Yuen, Huimin Zhao, Genyun Sun, Stephen Marshall, Jon Atli Benediktsson

Research output: Contribution to journalArticle

6 Citations (Scopus)

Abstract

Classification of hyperspectral images (HSI) has been a challenging problem under active investigation for years especially due to the extremely high data dimensionality and limited number of samples available for training. It is found that hyperspectral image classification can be generally improved only if the feature extraction technique and the classifier are both addressed. In this paper, a novel classification framework for hyperspectral images based on the joint bilateral filter and sparse representation classification (SRC) is proposed. By employing the first principal component as the guidance image for the joint bilateral filter, spatial features can be extracted with minimum edge blurring thus improving the quality of the band-to-band images. For this reason, the performance of the joint bilateral filter has shown better than that of the conventional bilateral filter in this work. In addition, the spectral similarity-based joint SRC (SS-JSRC) is proposed to overcome the weakness of the traditional JSRC method. By combining the joint bilateral filtering and SS-JSRC together, the superiority of the proposed classification framework is demonstrated with respect to several state-of-the-art spectral-spatial classification approaches commonly employed in the HSI community, with better classification accuracy and Kappa coefficient achieved.
LanguageEnglish
Pages316-328
Number of pages13
JournalPattern Recognition
Volume77
Early online date10 Oct 2017
DOIs
Publication statusPublished - 31 May 2018

Fingerprint

Feature extraction
Image classification
Hyperspectral imaging
Classifiers

Keywords

  • hyperspectral imaging
  • joint bilateral filtering
  • sparse representation
  • feature extraction
  • data classification

Cite this

@article{72b0fe75efa34ed6b904b64657fc2f79,
title = "Joint bilateral filtering and spectral similarity-based sparse representation: a generic framework for effective feature extraction and data classification in hyperspectral imaging",
abstract = "Classification of hyperspectral images (HSI) has been a challenging problem under active investigation for years especially due to the extremely high data dimensionality and limited number of samples available for training. It is found that hyperspectral image classification can be generally improved only if the feature extraction technique and the classifier are both addressed. In this paper, a novel classification framework for hyperspectral images based on the joint bilateral filter and sparse representation classification (SRC) is proposed. By employing the first principal component as the guidance image for the joint bilateral filter, spatial features can be extracted with minimum edge blurring thus improving the quality of the band-to-band images. For this reason, the performance of the joint bilateral filter has shown better than that of the conventional bilateral filter in this work. In addition, the spectral similarity-based joint SRC (SS-JSRC) is proposed to overcome the weakness of the traditional JSRC method. By combining the joint bilateral filtering and SS-JSRC together, the superiority of the proposed classification framework is demonstrated with respect to several state-of-the-art spectral-spatial classification approaches commonly employed in the HSI community, with better classification accuracy and Kappa coefficient achieved.",
keywords = "hyperspectral imaging, joint bilateral filtering, sparse representation, feature extraction, data classification",
author = "Tong Qiao and Zhijing Yang and Jinchang Ren and Peter Yuen and Huimin Zhao and Genyun Sun and Stephen Marshall and Benediktsson, {Jon Atli}",
year = "2018",
month = "5",
day = "31",
doi = "10.1016/j.patcog.2017.10.008",
language = "English",
volume = "77",
pages = "316--328",
journal = "Pattern Recognition",
issn = "0031-3203",

}

Joint bilateral filtering and spectral similarity-based sparse representation : a generic framework for effective feature extraction and data classification in hyperspectral imaging. / Qiao, Tong; Yang, Zhijing; Ren, Jinchang; Yuen, Peter; Zhao, Huimin; Sun, Genyun; Marshall, Stephen; Benediktsson, Jon Atli.

In: Pattern Recognition, Vol. 77, 31.05.2018, p. 316-328.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Joint bilateral filtering and spectral similarity-based sparse representation

T2 - Pattern Recognition

AU - Qiao, Tong

AU - Yang, Zhijing

AU - Ren, Jinchang

AU - Yuen, Peter

AU - Zhao, Huimin

AU - Sun, Genyun

AU - Marshall, Stephen

AU - Benediktsson, Jon Atli

PY - 2018/5/31

Y1 - 2018/5/31

N2 - Classification of hyperspectral images (HSI) has been a challenging problem under active investigation for years especially due to the extremely high data dimensionality and limited number of samples available for training. It is found that hyperspectral image classification can be generally improved only if the feature extraction technique and the classifier are both addressed. In this paper, a novel classification framework for hyperspectral images based on the joint bilateral filter and sparse representation classification (SRC) is proposed. By employing the first principal component as the guidance image for the joint bilateral filter, spatial features can be extracted with minimum edge blurring thus improving the quality of the band-to-band images. For this reason, the performance of the joint bilateral filter has shown better than that of the conventional bilateral filter in this work. In addition, the spectral similarity-based joint SRC (SS-JSRC) is proposed to overcome the weakness of the traditional JSRC method. By combining the joint bilateral filtering and SS-JSRC together, the superiority of the proposed classification framework is demonstrated with respect to several state-of-the-art spectral-spatial classification approaches commonly employed in the HSI community, with better classification accuracy and Kappa coefficient achieved.

AB - Classification of hyperspectral images (HSI) has been a challenging problem under active investigation for years especially due to the extremely high data dimensionality and limited number of samples available for training. It is found that hyperspectral image classification can be generally improved only if the feature extraction technique and the classifier are both addressed. In this paper, a novel classification framework for hyperspectral images based on the joint bilateral filter and sparse representation classification (SRC) is proposed. By employing the first principal component as the guidance image for the joint bilateral filter, spatial features can be extracted with minimum edge blurring thus improving the quality of the band-to-band images. For this reason, the performance of the joint bilateral filter has shown better than that of the conventional bilateral filter in this work. In addition, the spectral similarity-based joint SRC (SS-JSRC) is proposed to overcome the weakness of the traditional JSRC method. By combining the joint bilateral filtering and SS-JSRC together, the superiority of the proposed classification framework is demonstrated with respect to several state-of-the-art spectral-spatial classification approaches commonly employed in the HSI community, with better classification accuracy and Kappa coefficient achieved.

KW - hyperspectral imaging

KW - joint bilateral filtering

KW - sparse representation

KW - feature extraction

KW - data classification

UR - http://www.sciencedirect.com/science/journal/00313203

U2 - 10.1016/j.patcog.2017.10.008

DO - 10.1016/j.patcog.2017.10.008

M3 - Article

VL - 77

SP - 316

EP - 328

JO - Pattern Recognition

JF - Pattern Recognition

SN - 0031-3203

ER -