Joint kernelized sparse representation classification for hyperspectral imagery

Research output: Contribution to conferencePaper

Abstract

In recent years, the hyperspectral image (HSI) classification has received much attention due to its importance on the military applications, food quality assessment [1], and land cover analysis [2-5], etc. Multiple classifiers have been adopted to label pixels of HSI images, including support vector machine (SVM), random forest (RF), and recently, the deep learning methods. Considering that HSI pixels belonging to the same class are usually lying in a low-dimensional space, those pixels can be represented by training samples from the same class. Based on that, Sparse Representation Classification (SRC) methods have also introduced in the HSI imagery. For an unlabeled pixel, a few atoms from the constructed training dictionary can sparsely represent it. With the recovered sparse coefficients, the class label can be determined by the residual between the test pixel and its approximation.

With the development of SRC in HIS [5, 6], there is one severe problem during the process of classification. Due to the high dimensions of the HSI data, it may result the Hughes phenomenon. Sufficient training samples are required to overcome the curse of dimensionality. However, sufficient training data are not always available in real application. For example, the ground truth labelling work for remote sensing data is rather inconvenient. Therefore, to solve the above problem, we decide to combine multiple types of features extracted from HSI data, and a joint kernelized SRC will be operated on those extracted features. The aim of our work is to improve the performance of SRC with less training samples.
LanguageEnglish
Number of pages2
Publication statusPublished - 10 Oct 2018
EventHyperspectral Imaging Applications (HSI) 2018 -
Duration: 10 Oct 201811 Oct 2018
https://www.hsi2018.com

Conference

ConferenceHyperspectral Imaging Applications (HSI) 2018
Period10/10/1811/10/18
Internet address

Fingerprint

Pixels
Labels
Military applications
Image classification
Glossaries
Labeling
Support vector machines
Remote sensing
Classifiers
Atoms

Keywords

  • hyperspectral image (HSI) classification
  • HSI images
  • HSI pixels
  • Sparse Representation Classification (SRC)

Cite this

@conference{ddcb568fb455450cbd79fd54eaeb9ac6,
title = "Joint kernelized sparse representation classification for hyperspectral imagery",
abstract = "In recent years, the hyperspectral image (HSI) classification has received much attention due to its importance on the military applications, food quality assessment [1], and land cover analysis [2-5], etc. Multiple classifiers have been adopted to label pixels of HSI images, including support vector machine (SVM), random forest (RF), and recently, the deep learning methods. Considering that HSI pixels belonging to the same class are usually lying in a low-dimensional space, those pixels can be represented by training samples from the same class. Based on that, Sparse Representation Classification (SRC) methods have also introduced in the HSI imagery. For an unlabeled pixel, a few atoms from the constructed training dictionary can sparsely represent it. With the recovered sparse coefficients, the class label can be determined by the residual between the test pixel and its approximation.With the development of SRC in HIS [5, 6], there is one severe problem during the process of classification. Due to the high dimensions of the HSI data, it may result the Hughes phenomenon. Sufficient training samples are required to overcome the curse of dimensionality. However, sufficient training data are not always available in real application. For example, the ground truth labelling work for remote sensing data is rather inconvenient. Therefore, to solve the above problem, we decide to combine multiple types of features extracted from HSI data, and a joint kernelized SRC will be operated on those extracted features. The aim of our work is to improve the performance of SRC with less training samples.",
keywords = "hyperspectral image (HSI) classification, HSI images, HSI pixels, Sparse Representation Classification (SRC)",
author = "He Sun and Jinchang Ren and Yijun Yan and Jaime Zabalza and Stephen Marshall",
year = "2018",
month = "10",
day = "10",
language = "English",
note = "Hyperspectral Imaging Applications (HSI) 2018 ; Conference date: 10-10-2018 Through 11-10-2018",
url = "https://www.hsi2018.com",

}

Sun, H, Ren, J, Yan, Y, Zabalza, J & Marshall, S 2018, 'Joint kernelized sparse representation classification for hyperspectral imagery' Paper presented at Hyperspectral Imaging Applications (HSI) 2018, 10/10/18 - 11/10/18, .

Joint kernelized sparse representation classification for hyperspectral imagery. / Sun, He; Ren, Jinchang; Yan, Yijun; Zabalza, Jaime; Marshall, Stephen.

2018. Paper presented at Hyperspectral Imaging Applications (HSI) 2018, .

Research output: Contribution to conferencePaper

TY - CONF

T1 - Joint kernelized sparse representation classification for hyperspectral imagery

AU - Sun, He

AU - Ren, Jinchang

AU - Yan, Yijun

AU - Zabalza, Jaime

AU - Marshall, Stephen

PY - 2018/10/10

Y1 - 2018/10/10

N2 - In recent years, the hyperspectral image (HSI) classification has received much attention due to its importance on the military applications, food quality assessment [1], and land cover analysis [2-5], etc. Multiple classifiers have been adopted to label pixels of HSI images, including support vector machine (SVM), random forest (RF), and recently, the deep learning methods. Considering that HSI pixels belonging to the same class are usually lying in a low-dimensional space, those pixels can be represented by training samples from the same class. Based on that, Sparse Representation Classification (SRC) methods have also introduced in the HSI imagery. For an unlabeled pixel, a few atoms from the constructed training dictionary can sparsely represent it. With the recovered sparse coefficients, the class label can be determined by the residual between the test pixel and its approximation.With the development of SRC in HIS [5, 6], there is one severe problem during the process of classification. Due to the high dimensions of the HSI data, it may result the Hughes phenomenon. Sufficient training samples are required to overcome the curse of dimensionality. However, sufficient training data are not always available in real application. For example, the ground truth labelling work for remote sensing data is rather inconvenient. Therefore, to solve the above problem, we decide to combine multiple types of features extracted from HSI data, and a joint kernelized SRC will be operated on those extracted features. The aim of our work is to improve the performance of SRC with less training samples.

AB - In recent years, the hyperspectral image (HSI) classification has received much attention due to its importance on the military applications, food quality assessment [1], and land cover analysis [2-5], etc. Multiple classifiers have been adopted to label pixels of HSI images, including support vector machine (SVM), random forest (RF), and recently, the deep learning methods. Considering that HSI pixels belonging to the same class are usually lying in a low-dimensional space, those pixels can be represented by training samples from the same class. Based on that, Sparse Representation Classification (SRC) methods have also introduced in the HSI imagery. For an unlabeled pixel, a few atoms from the constructed training dictionary can sparsely represent it. With the recovered sparse coefficients, the class label can be determined by the residual between the test pixel and its approximation.With the development of SRC in HIS [5, 6], there is one severe problem during the process of classification. Due to the high dimensions of the HSI data, it may result the Hughes phenomenon. Sufficient training samples are required to overcome the curse of dimensionality. However, sufficient training data are not always available in real application. For example, the ground truth labelling work for remote sensing data is rather inconvenient. Therefore, to solve the above problem, we decide to combine multiple types of features extracted from HSI data, and a joint kernelized SRC will be operated on those extracted features. The aim of our work is to improve the performance of SRC with less training samples.

KW - hyperspectral image (HSI) classification

KW - HSI images

KW - HSI pixels

KW - Sparse Representation Classification (SRC)

M3 - Paper

ER -