Classification of Hyperspectral Images Using Kernel Fully Constrained Least Squares
As a widely used classifier, sparse representation classification (SRC) has shown its good performance for hyperspectral image classification. Recent works have highlighted that it is the collaborative representation mechanism under SRC that makes SRC a highly effective technique for classification...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2017-11-01
|
Series: | ISPRS International Journal of Geo-Information |
Subjects: | |
Online Access: | https://www.mdpi.com/2220-9964/6/11/344 |
_version_ | 1811345073022959616 |
---|---|
author | Jianjun Liu Zebin Wu Zhiyong Xiao Jinlong Yang |
author_facet | Jianjun Liu Zebin Wu Zhiyong Xiao Jinlong Yang |
author_sort | Jianjun Liu |
collection | DOAJ |
description | As a widely used classifier, sparse representation classification (SRC) has shown its good performance for hyperspectral image classification. Recent works have highlighted that it is the collaborative representation mechanism under SRC that makes SRC a highly effective technique for classification purposes. If the dimensionality and the discrimination capacity of a test pixel is high, other norms (e.g., ℓ 2 -norm) can be used to regularize the coding coefficients, except for the sparsity ℓ 1 -norm. In this paper, we show that in the kernel space the nonnegative constraint can also play the same role, and thus suggest the investigation of kernel fully constrained least squares (KFCLS) for hyperspectral image classification. Furthermore, in order to improve the classification performance of KFCLS by incorporating spatial-spectral information, we investigate two kinds of spatial-spectral methods using two regularization strategies: (1) the coefficient-level regularization strategy, and (2) the class-level regularization strategy. Experimental results conducted on four real hyperspectral images demonstrate the effectiveness of the proposed KFCLS, and show which way to incorporate spatial-spectral information efficiently in the regularization framework. |
first_indexed | 2024-04-13T19:57:37Z |
format | Article |
id | doaj.art-9d1f2f30cf134d3abc8b339d977a6d2f |
institution | Directory Open Access Journal |
issn | 2220-9964 |
language | English |
last_indexed | 2024-04-13T19:57:37Z |
publishDate | 2017-11-01 |
publisher | MDPI AG |
record_format | Article |
series | ISPRS International Journal of Geo-Information |
spelling | doaj.art-9d1f2f30cf134d3abc8b339d977a6d2f2022-12-22T02:32:17ZengMDPI AGISPRS International Journal of Geo-Information2220-99642017-11-0161134410.3390/ijgi6110344ijgi6110344Classification of Hyperspectral Images Using Kernel Fully Constrained Least SquaresJianjun Liu0Zebin Wu1Zhiyong Xiao2Jinlong Yang3The Key Laboratory of Advanced Process Control for Light Industry (Ministry of Education), School of Internet of Things Engineering, Jiangnan University, Wuxi 214122, ChinaSchool of Computer Science, Nanjing University of Science and Technology, Nanjing 210094, ChinaThe Key Laboratory of Advanced Process Control for Light Industry (Ministry of Education), School of Internet of Things Engineering, Jiangnan University, Wuxi 214122, ChinaThe Key Laboratory of Advanced Process Control for Light Industry (Ministry of Education), School of Internet of Things Engineering, Jiangnan University, Wuxi 214122, ChinaAs a widely used classifier, sparse representation classification (SRC) has shown its good performance for hyperspectral image classification. Recent works have highlighted that it is the collaborative representation mechanism under SRC that makes SRC a highly effective technique for classification purposes. If the dimensionality and the discrimination capacity of a test pixel is high, other norms (e.g., ℓ 2 -norm) can be used to regularize the coding coefficients, except for the sparsity ℓ 1 -norm. In this paper, we show that in the kernel space the nonnegative constraint can also play the same role, and thus suggest the investigation of kernel fully constrained least squares (KFCLS) for hyperspectral image classification. Furthermore, in order to improve the classification performance of KFCLS by incorporating spatial-spectral information, we investigate two kinds of spatial-spectral methods using two regularization strategies: (1) the coefficient-level regularization strategy, and (2) the class-level regularization strategy. Experimental results conducted on four real hyperspectral images demonstrate the effectiveness of the proposed KFCLS, and show which way to incorporate spatial-spectral information efficiently in the regularization framework.https://www.mdpi.com/2220-9964/6/11/344hyperspectralimage classificationleast squarescollaborative representationsparse representationposterior probabilityregularization |
spellingShingle | Jianjun Liu Zebin Wu Zhiyong Xiao Jinlong Yang Classification of Hyperspectral Images Using Kernel Fully Constrained Least Squares ISPRS International Journal of Geo-Information hyperspectral image classification least squares collaborative representation sparse representation posterior probability regularization |
title | Classification of Hyperspectral Images Using Kernel Fully Constrained Least Squares |
title_full | Classification of Hyperspectral Images Using Kernel Fully Constrained Least Squares |
title_fullStr | Classification of Hyperspectral Images Using Kernel Fully Constrained Least Squares |
title_full_unstemmed | Classification of Hyperspectral Images Using Kernel Fully Constrained Least Squares |
title_short | Classification of Hyperspectral Images Using Kernel Fully Constrained Least Squares |
title_sort | classification of hyperspectral images using kernel fully constrained least squares |
topic | hyperspectral image classification least squares collaborative representation sparse representation posterior probability regularization |
url | https://www.mdpi.com/2220-9964/6/11/344 |
work_keys_str_mv | AT jianjunliu classificationofhyperspectralimagesusingkernelfullyconstrainedleastsquares AT zebinwu classificationofhyperspectralimagesusingkernelfullyconstrainedleastsquares AT zhiyongxiao classificationofhyperspectralimagesusingkernelfullyconstrainedleastsquares AT jinlongyang classificationofhyperspectralimagesusingkernelfullyconstrainedleastsquares |