PSFormer-hyperspectral-image-classification

Demo code of "PSFormer: Pyramid Superpixel Transformer for Hyperspectral Image Classification"

Step 1: prepare dataset

Salinas and Pavia University: https://www.ehu.eus/ccwintco/index.php/Hyperspectral_Remote_Sensing_Scenes

Houston University (without cloud): https://github.com/danfenghong/IEEE_TGRS_SpectralFormer

WHU-Hi-HongHu: http://rsidea.whu.edu.cn/resource_WHUHi_sharing.htm

QUH-Tangdaowan: https://github.com/Hang-Fu/QUH-classification-dataset

Step 2: train and test

python main.py --cuda 0

Step 3: record classification result

The quantitative evaluation results will be recorded in the '/results' folder.

The high-definition qualitative evaluation results can be generated with the codes in the '/visualization_code' folder.

(you can generate the full classification maps or classification maps without background category with custom palette 🫡🫡)

Citation

If you find this work interesting in your research, please kindly cite:

@ARTICLE{10695122,
  author={Zou, Jiaqi and He, Wei and Zhang, Hongyan},
  journal={IEEE Transactions on Geoscience and Remote Sensing}, 
  title={PSFormer: Pyramid Superpixel Transformer for Hyperspectral Image Classification}, 
  year={2024},
  volume={62},
  number={},
  pages={1-16},
  doi={10.1109/TGRS.2024.3468876}}

Thank you very much! (^▽^)

This code is constructed based on MSSG-UNet, SuperpixelHierarchy, and AM-GCN, thanks~💕.

If you have any questions, please feel free to contact me (Jiaqi Zou, immortal@whu.edu.cn).