Zhu Han, Jin Yang, Lianru Gao, Zhiqiang Zeng, Bing Zhang, Jocelyn Chanussot
This is a PyTorch implementation of the "Dual-Branch Subpixel-Guided Network for Hyperspectral Image Classification" in IEEE Transactions on Geoscience and Remote Sensing. More specifically, it is detailed as follow.
DSNet includes deep AE unmixing network, CNN-based classifier network and subpixel fusion module. The deep AE unmixing network is designed by considering a general mixing decoder with physically nonlinear constraints, and further extract useful subpixel-level abundance information from the HSI in an unsupervised manner. The CNN-based classifier network extracts the spectral-spatial information within the HSI to obtain pixel-level class features. The subpixel fusion module aims at integrating the abundance information and class features to ensure high-quality information fusion and enhance model representation capability.
Deep learning (DL) has been widely applied into hyperspectral image (HSI) classification owing to its promising feature learning and representation capabilities. However, limited by the spatial resolution of sensors, existing DL-based classification approaches mainly focus on pixel-level spectral and spatial information extraction through complex network architecture design, while ignoring the existence of mixed pixels in actual scenarios. To tackle this difficulty, we propose a novel dual-branch subpixel-guided network for HSI classification, called DSNet, which automatically integrates subpixel information and convolutional class features by introducing a deep autoencoder unmixing architecture to enhance classification performance. DSNet is capable of fully considering physically nonlinear properties within subpixels and adaptively generating diagnostic abundances in an unsupervised manner to achieve more reliable decision boundaries for class label distributions. The subpixel fusion module is designed to ensure high-quality information fusion across pixel and subpixel features, further promoting stable joint classification. Experimental results on three benchmark datasets demonstrate the effectiveness and superiority of DSNet compared with state-of-the-art DL-based HSI classification approaches.
The adopted Berlin and Augsburg datasets can be downloaded in Google Drive Link.
./demo.pyis the script for training DSNet on different hyperspectral datasets. The patch size can be changed according to input dataset.
python demo.py --dataset='Indian' --patches=7 --flag_test='train'- After training DSNet, the saved model is loaded to obtain final classification results.
python demo.py --dataset='Indian' --patches=7 --flag_test='test'Please kindly cite the papers if this code is useful and helpful for your research.
Zhu Han, Jin Yang, Lianru Gao, Zhiqiang Zeng, Bing Zhang, Jocelyn Chanussot. Dual-Branch Subpixel-Guided Network for Hyperspectral Image Classification, IEEE Transactions on Geoscience and Remote Sensing, vol. 62, pp. 1-13, 2024, Art no. 5521813, doi: 10.1109/TGRS.2024.3418583.
@ARTICLE{10570241,
author={Han, Zhu and Yang, Jin and Gao, Lianru and Zeng, Zhiqiang and Zhang, Bing and Chanussot, Jocelyn},
journal={IEEE Transactions on Geoscience and Remote Sensing},
title={Dual-Branch Subpixel-Guided Network for Hyperspectral Image Classification},
year={2024},
volume={62},
pages={1-13},
doi={10.1109/TGRS.2024.3418583}
}
Copyright (C) 2024 Zhu Han
This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, version 3 of the License.
This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
You should have received a copy of the GNU General Public License along with this program.
Zhu Han: [email protected]