SDPNet

A Deep Network for Pan-sharpening with Enhanced Information Representation (TGRS 2020)

To test with the pretrained model:

run test.py

To train:

• Step 1: download the h5 file or create your own h5 file according to it.
• Step 2: run P2MS_main.py and MS2P_main.py for P2MS and MS2P models
• Step 3: run spec_main.py and spat_main.py for spatial and spectral encoder and decoders
• Step 4: run spec_diff.py and spat_diff.py for unique channels
• Step 5: run main.py for training (2 GPUs are needed)

If this work is helpful, please cite it as:

@article{xu2020sdpnet,
  title={SDPNet: A Deep Network for Pan-Sharpening With Enhanced Information Representation},
  author={Xu, Han and Ma, Jiayi and Shao, Zhenfeng and Zhang, Hao and Jiang, Junjun and Guo, Xiaojie},
  journal={IEEE Transactions on Geoscience and Remote Sensing},
  year={2020},
  publisher={IEEE}
}