| --- |
| datasets: |
| - PanCollection |
| language: en |
| license: gpl-2.0 |
| size_categories: |
| - 1K<n<10K |
| tags: |
| - Pytorch |
| --- |
| |
| # ✨ PanCollection |
|
|
| 🤗 To get started with PanCollection benchmark (training, inference, etc.), we recommend reading [Google Colab](https://colab.research.google.com/drive/1KpWWj1lVUGllZCws01zQfd6CeURuGL2O#scrollTo=k53dsFhAdp6n)! |
|
|
|
|
| ## Recommendations |
|
|
| We recommend users to use the code-toolbox [DLPan-Toolbox](https://github.com/liangjiandeng/DLPan-Toolbox/tree/main/02-Test-toolbox-for-traditional-and-DL(Matlab)) + the dataset [PanCollection](https://drive.google.com/drive/folders/15VXUjqPybtqUN_spKfJbw40W05K4nDdY?usp=sharing) for fair training and testing! |
|
|
| ### Deploy |
|
|
| PanCollection has provided complete packages. |
| ``` |
| pip install pancollection --upgrade |
| ``` |
|
|
| ## How to Get Started with the Model |
|
|
|
|
| ```python |
| import pancollection as pan |
| cfg = pan.TaskDispatcher.new(task='pansharpening', mode='entrypoint', arch='FusionNet', |
| dataset_name="gf2", use_resume=False, |
| dataset={'train': 'gf2', 'test': 'test_gf2_multiExm1.h5'}) |
| print(pan.TaskDispatcher._task) |
| pan.trainer.main(cfg, pan.build_model, pan.getDataSession) |
| ``` |
|
|
| ## Training Details |
|
|
| See [Google Colab](https://colab.research.google.com/drive/1KpWWj1lVUGllZCws01zQfd6CeURuGL2O) for quick start. |
|
|
| See [Github Project](https://github.com/XiaoXiao-Woo/PanCollection) for coding details. |
|
|
| ## Evaluation |
|
|
| See the [Leaderboard](https://paperswithcode.com/dataset/worldview-3-pancollection) for model results. |
|
|
| See the [PanCollection Paper](https://liangjiandeng.github.io/papers/2022/deng-jig2022.pdf) for early results. |
|
|
|
|
|
|
| | **Satellite** | **Value** | **Comment** | |
| |--------------------|-----------|----------------------------------------| |
| | WorldView-3 | 2047 | | |
| | QuickBird | 2047 | | |
| | GaoFen-2 | 1023 | | |
| | WorldView-2 | 2047 | | |
|
|
|
|
| ## Citation |
|
|
| To learn more about the PanCollection dataset, see the [Github Pages](https://github.com/liangjiandeng/PanCollection). |
|
|
| ``` |
| @ARTICLE{dengjig2022, |
| author={邓良剑，冉燃，吴潇，张添敬}, |
| journal={中国图象图形学报}, |
| title={遥感图像全色锐化的卷积神经网络方法研究进展}, |
| year={2022}, |
| volume={}, |
| number={9}, |
| pages={}, |
| doi={10.11834/jig.220540} |
| } |
| ``` |
|
|
| ``` |
| @ARTICLE{deng2022vivone, |
| author={L. -J. Deng, G. Vivone, M. E. Paoletti, G. Scarpa, J. He, Y. Zhang, J. Chanussot, and A. Plaza}, |
| journal={IEEE Geoscience and Remote Sensing Magazine}, |
| title={Machine Learning in Pansharpening: A Benchmark, from Shallow to Deep Networks}, |
| year={2022}, |
| volume={10}, |
| number={3}, |
| pages={279-315}, |
| doi={10.1109/MGRS.2022.3187652} |
| } |
| ``` |
|
|
|
|
| ## License |
|
|
| PanCollection is made available under the GPLv2.0 license. |
|
|
| ## Contact |
| wxwsx1997@gmail.com |
|
|
| liangjiandeng@uestc.edu.cn |
