Multiscale superpixel-level subspace-based support vector machines for hyperspectral image classification

Haoyang Yu, Lianru Gao, Wenzhi Liao, Bing Zhang, Aleksandra Pizurica, Wilfried Philips

Research output: Contribution to journalArticle

12 Citations (Scopus)

Abstract

This letter introduces a new spectral-spatial classification method for hyperspectral images. A multiscale superpixel segmentation is first used to model the distribution of classes based on spatial information. In this context, the original hyperspectral image is integrated with segmentation maps via a feature fusion process in different scales such that the pixel-level data can be represented by multiscale superpixel-level (MSP) data sets. Then, a subspace-based support vector machine (SVMsub) is adopted to obtain the classification maps with multiscale inputs. Finally, the classification result is achieved via a decision fusion process. The resulting method, called MSP-SVMsub, makes use of the spatial and spectral coherences, and contributes to better feature characterization. Experimental results based on two real hyperspectral data sets indicate that the MSP-SVMsub exhibits good performance compared with other related methods.
LanguageEnglish
Pages2142-2146
Number of pages5
JournalIEEE Geoscience and Remote Sensing Letters
Volume14
Issue number11
DOIs
Publication statusPublished - 4 Oct 2017

Fingerprint

image classification
Image classification
Support vector machines
segmentation
Fusion reactions
fusion
pixel
Pixels
pixels
support vector machine
method

Keywords

  • hyperspectral image classification
  • multiscale superpixel segmentation
  • subspace projection
  • support vector machines (SVM)
  • principal component analysis
  • geophysical image processing

Cite this

Yu, Haoyang ; Gao, Lianru ; Liao, Wenzhi ; Zhang, Bing ; Pizurica, Aleksandra ; Philips, Wilfried. / Multiscale superpixel-level subspace-based support vector machines for hyperspectral image classification. In: IEEE Geoscience and Remote Sensing Letters. 2017 ; Vol. 14, No. 11. pp. 2142-2146.
@article{be8472c9d34a4464968b15251b8ec6d3,
title = "Multiscale superpixel-level subspace-based support vector machines for hyperspectral image classification",
abstract = "This letter introduces a new spectral-spatial classification method for hyperspectral images. A multiscale superpixel segmentation is first used to model the distribution of classes based on spatial information. In this context, the original hyperspectral image is integrated with segmentation maps via a feature fusion process in different scales such that the pixel-level data can be represented by multiscale superpixel-level (MSP) data sets. Then, a subspace-based support vector machine (SVMsub) is adopted to obtain the classification maps with multiscale inputs. Finally, the classification result is achieved via a decision fusion process. The resulting method, called MSP-SVMsub, makes use of the spatial and spectral coherences, and contributes to better feature characterization. Experimental results based on two real hyperspectral data sets indicate that the MSP-SVMsub exhibits good performance compared with other related methods.",
keywords = "hyperspectral image classification, multiscale superpixel segmentation, subspace projection, support vector machines (SVM), principal component analysis, geophysical image processing",
author = "Haoyang Yu and Lianru Gao and Wenzhi Liao and Bing Zhang and Aleksandra Pizurica and Wilfried Philips",
year = "2017",
month = "10",
day = "4",
doi = "10.1109/lgrs.2017.2755061",
language = "English",
volume = "14",
pages = "2142--2146",
journal = "IEEE Geoscience and Remote Sensing Letters",
issn = "1545-598X",
publisher = "IEEE",
number = "11",

}

Multiscale superpixel-level subspace-based support vector machines for hyperspectral image classification. / Yu, Haoyang; Gao, Lianru; Liao, Wenzhi; Zhang, Bing; Pizurica, Aleksandra; Philips, Wilfried.

In: IEEE Geoscience and Remote Sensing Letters, Vol. 14, No. 11, 04.10.2017, p. 2142-2146.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Multiscale superpixel-level subspace-based support vector machines for hyperspectral image classification

AU - Yu, Haoyang

AU - Gao, Lianru

AU - Liao, Wenzhi

AU - Zhang, Bing

AU - Pizurica, Aleksandra

AU - Philips, Wilfried

PY - 2017/10/4

Y1 - 2017/10/4

N2 - This letter introduces a new spectral-spatial classification method for hyperspectral images. A multiscale superpixel segmentation is first used to model the distribution of classes based on spatial information. In this context, the original hyperspectral image is integrated with segmentation maps via a feature fusion process in different scales such that the pixel-level data can be represented by multiscale superpixel-level (MSP) data sets. Then, a subspace-based support vector machine (SVMsub) is adopted to obtain the classification maps with multiscale inputs. Finally, the classification result is achieved via a decision fusion process. The resulting method, called MSP-SVMsub, makes use of the spatial and spectral coherences, and contributes to better feature characterization. Experimental results based on two real hyperspectral data sets indicate that the MSP-SVMsub exhibits good performance compared with other related methods.

AB - This letter introduces a new spectral-spatial classification method for hyperspectral images. A multiscale superpixel segmentation is first used to model the distribution of classes based on spatial information. In this context, the original hyperspectral image is integrated with segmentation maps via a feature fusion process in different scales such that the pixel-level data can be represented by multiscale superpixel-level (MSP) data sets. Then, a subspace-based support vector machine (SVMsub) is adopted to obtain the classification maps with multiscale inputs. Finally, the classification result is achieved via a decision fusion process. The resulting method, called MSP-SVMsub, makes use of the spatial and spectral coherences, and contributes to better feature characterization. Experimental results based on two real hyperspectral data sets indicate that the MSP-SVMsub exhibits good performance compared with other related methods.

KW - hyperspectral image classification

KW - multiscale superpixel segmentation

KW - subspace projection

KW - support vector machines (SVM)

KW - principal component analysis

KW - geophysical image processing

UR - http://hdl.handle.net/1854/LU-8542174

U2 - 10.1109/lgrs.2017.2755061

DO - 10.1109/lgrs.2017.2755061

M3 - Article

VL - 14

SP - 2142

EP - 2146

JO - IEEE Geoscience and Remote Sensing Letters

T2 - IEEE Geoscience and Remote Sensing Letters

JF - IEEE Geoscience and Remote Sensing Letters

SN - 1545-598X

IS - 11

ER -