Journal of Computer Applications ›› 2020, Vol. 40 ›› Issue (2): 381-385.DOI: 10.11772/j.issn.1001-9081.2019081385

• DPCS 2019 • Previous Articles     Next Articles

Hyperspectral band selection based on deep adversarial subspace clustering

Meng ZENG1,2, Bin NING1(), Zhihua CAI2, Qiong GU1   

  1. 1.Computer School,Hubei University of Arts and Science,Xiangyang Hubei 441053,China
    2.School of Computer Science,China University of Geosciences (Wuhan),Wuhan Hubei 430074,China
  • Received:2019-07-31 Revised:2019-09-21 Accepted:2019-09-23 Online:2019-09-29 Published:2020-02-10
  • Contact: Bin NING
  • About author:ZENG Meng,born in 1995, M. S. candidate. Her research interests include deep learning, data mining.
    CAI Zhihua, born in 1963, Ph. D., professor. His research interests include data mining, evolutionary calculation, parallel computing.
    GU Qiong, born in 1973, Ph. D., professor. Her research interests include intelligent computing, machine learning, network data mining.
  • Supported by:
    the Colleges and Universities’ Industry-University-Research Innovation Fund of Science and Technology Development Center of the Ministry of Education — New Generation Information Technology Innovation Project(2018A02028);the National Natural Science Foundation of China(61773355);the Fundamental Research Founds for National University,China University of Geosciences (Wuhan)(G1323541717);the Natural Science Foundation of Hubei Province(2018CFB528);the Open Research Project of Hubei Key Laboratory of Intelligent Geo-Information Processing(KLIGIP-2017B01)

使用深度对抗子空间聚类实现高光谱波段选择

曾梦1,2, 宁彬1(), 蔡之华2, 谷琼1   

  1. 1.湖北文理学院 计算机工程学院,湖北 襄阳 441053
    2.中国地质大学(武汉) 计算机学院,武汉 430074
  • 通讯作者: 宁彬
  • 作者简介:曾梦(1995—),女,湖北襄阳人,硕士研究生,主要研究方向:深度学习、数据挖掘
    蔡之华(1963—),男,湖北麻城人,教授,博士,CCF高级会员,主要研究方向:数据挖掘、演化计算、并行计算
    谷琼(1973—),女,湖北荆门人,教授,博士,CCF高级会员,主要研究方向:智能计算、机器学习、网络数据挖掘。
  • 基金资助:
    教育部科技发展中心高校产学研创新基金——新一代信息技术创新项目(2018A02028);国家自然科学基金资助项目(61773355);中国地质大学(武汉)中央高校基本科研业务费专项资金资助项目(G1323541717);湖北省自然科学基金资助项目(2018CFB528);智能地学信息处理湖北省重点实验室开放研究项目(KLIGIP-2017B01)

Abstract:

HyperSpectral Image (HSI) consists of hundreds of bands with strong intra-band correlations between bands and high redundancy, resulting in dimensional disaster and increased classification complexity. Therefore, a Deep Adversarial Subspace Clustering (DASC) method was used for hyperspectral band selection, and Laplacian regularization was introduced to make the network performance more robust, which reduces the classification complexity under the premise of ensuring classification accuracy. A self-expressive layer was introduced between the encoder and the decoder to imitate the “self-expression” attribute of traditional subspace clustering, making full use of the spectral information and nonlinear feature transformation to obtain the relationships between the bands, and solving the problem that traditional band selection methods cannot consider spectral-spatial information simultaneously. At the same time, adversarial learning was introduced to supervise the sample representation of the auto-encoder and subspace clustering, so that the subspace clustering has better self-expression performance. In order to make the network performance more robust, Laplacian regularization was added to consider the manifold structure reflecting geometric information. Experimental results on two public hyperspectral datasets show that compared with several mainstream band selection methods, DASC method has higher accuracy, and the selected band subset of the method can satisfy application requirements.

Key words: HyperSpectral Image (HSI), band selection, Deep Adversarial Subspace Clustering (DASC), Laplacian regularization, deep learning

摘要:

高光谱图像(HSI)由数百个波段组成,波段之间的相关性强且具有较高的冗余度,导致出现维度灾难并且分类的复杂性很高。为此,使用深度对抗子空间聚类(DASC)网络进行高光谱的波段选择,并引入拉普拉斯正则化使网络更优,在保证分类精度的前提下降低分类的复杂度。该网络通过在编码器和解码器中引入自表达层来模仿传统子空间聚类的“自表达”属性,充分运用光谱信息和非线性特征转换得到波段之间的相互关系,解决传统波段选择方法无法同时考虑光谱和空间信息的问题。同时,引入对抗学习来监督自编码器的样本表示和子空间聚类,使得子空间聚类具有更好的自表达性能。为了使网络性能更优,加入拉普拉斯正则化来考虑反映图像几何信息的局部流形结构。实验在两个公开的高光谱数据集上进行,所提出的方法和几种主流的波段选择方法进行对比的结果表明,DASC方法在分类精度上优于对比方法,其选出的波段子集可以满足应用需求。

关键词: 高光谱图像, 波段选择, 深度对抗子空间聚类, 拉普拉斯正则化, 深度学习

CLC Number: