期刊
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
卷 15, 期 -, 页码 184-200出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/JSTARS.2021.3133009
关键词
Feature extraction; Logic gates; Hyperspectral imaging; Task analysis; Convolutional neural networks; Training; Earth; Band grouping; deep learning (DL); features fusion; global-local; hyperspectral image (HSI)
类别
资金
- National Natural Science Foundation of China [61977022]
- Science Foundation for Distinguished Young Scholars of Hunan Province [2020JJ2017]
- Key Research and Development Program of Hunan Province [2019SK2012]
- Foundation of Department of Water Resources of Hunan Province [XSKJ2021000-12, XSKJ2021000-13]
- Natural Science Foundation of Hunan Province [2020JJ4340, 2021JJ40226]
- Foundation of Education Bureau of Hunan Province [20B257, 20B266]
- Scientific Research Fund of Education Department of Hunan Province [19A200]
- Open Fund of Education Department of Hunan Province [20K062]
- Postgraduate Scientific Research Innovation Project of Hunan Province [QL20210254]
This article proposes a global-local hierarchical weighted fusion end-to-end classification architecture that enhances the discrimination of spectral-spatial features. Experimental results demonstrate its competitiveness in terms of accuracy and generalization.
The fusion of spectral-spatial features based on deep learning has become the focus of research in hyperspectral image (HSI) classification. However, previous deep frameworks based on spectral-spatial fusion usually performed feature aggregation only at the branch ends. Furthermore, only first-order statistical features are considered in the fusion process, which is not conducive to improving the discrimination of spectral-spatial features. This article proposes a global-local hierarchical weighted fusion end-to-end classification architecture. The architecture includes two subnetworks for spectral classification and spatial classification. For the spectral subnetwork, two band-grouping strategies are designed, and bidirectional long short-term memory is used to capture spectral context information from global to local perspectives. For the spatial subnetwork, a pooling strategy based on local attention is combined to construct a global-local pooling fusion module to enhance the discriminability of spatial features learned by a convolutional neural network. For the fusion stage, a hierarchical weighting fusion mechanism is developed to obtain the nonlinear relationship between both spectral and spatial features. The experimental results on four real HSI datasets and a GF-5 satellite dataset demonstrate that the method proposed is more competitive in terms of accuracy and generalization.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据