🚀 Spectral Dynamic Attention Network for Hyperspectral Image Super-Resolution, IEEE GRSL 2026.

📌 Introduction

This repository contains the official implementation of our paper:
📄 Spectral Dynamic Attention Network for Hyperspectral Image Super-Resolution (IEEE GRSL 2026)

🛠 Requirements

• basicsr 1.3.4
• Python 3.8
• PyTorch 1.11.0.
• CUDA 11.3

🏋️‍♂️ Preparation

To get the training set, validation set and testing set, refer to SSPSR to download the mcodes for cropping the hyperspectral image.

Training

To train SDANet, run the following command from the project root:

./run_train.sh 0

Use another GPU by changing the index, e.g. ./run_train.sh 2.

You can also pass the option file directly (same style as manual command):

./run_train.sh options/train/HSI/SDAnetChux4.yml 0

Equivalent direct command:

CUDA_VISIBLE_DEVICES=0 python ./basicsr/train.py -opt ./options/train/HSI/SDAnetChux4.yml

Testing

To test SDANet, run the following command from the project root:

./run_test.sh 0

Use another GPU by changing the index, e.g. ./run_test.sh 2.

You can also pass the option file directly:

./run_test.sh options/test/HSI/test_SDAnetChux4.yml 0

Equivalent direct command:

CUDA_VISIBLE_DEVICES=0 python ./basicsr/test.py -opt ./options/test/HSI/test_SDAnetChux4.yml

📚 Citation

If you find this repository useful, please cite:

@ARTICLE{zty26grsl,
  author={Zhang, Tengya and Gao, Feng and Qi, Lin and Dong, Junyu and Du, Qian},
  journal={IEEE Geoscience and Remote Sensing Letters}, 
  title={Spectral Dynamic Attention Network for Hyperspectral Image Super-Resolution}, 
  year={2026},
  volume={23},
  pages={1-5}}

📬 Contact

If you have any questions, feel free to contact us via email:

• 📧 Feng Gao: gaofeng@ouc.edu.cn

We hope SDANet helps your research ⭐ If you find our work useful, please consider citing the paper and starring this repository.