The repository is for the article "cross-supervised learning for cloud detection" on GIScience&Remote Sensing.
The current version on the website does't include Supplementary. Download the Supplementary at here.
configs/MyNet_GF1.py
python train.py
X
for the index of GPUs):
CUDA_VISIBLE_DEVICES=X,X,X python train.py
model_name
, dataset
, pth1
, pth2
and exp_id
in the file: test.py
.python test.py
X
for the index of GPUs):
CUDA_VISIBLE_DEVICES=X,X,X python test.py
HY1C-UPC Dataset is built from images of Chinese HY1-C satellite. The coastal zone imager (CZI) on the HY1-C satellite has a 50-m spatial resolution with four multi-spectral bands.
The HY1C-UPC dataset includes 25 scenes from September 2021 to February 2022. The main scenes are collected from the coastal zones as shown bellow. The observation width of the CZI is large, hence, the dataset includes various terrains, e.g., city, snow, forest, ocean, etc., as shown bellow. The HY1C-UPC dataset contains 8 manually labeled scenes that are labeled by experts with the Photoshop software and 17 unlabeled scenes.
HY1C-UPC dadaset is avaliable at: aliyundrive. (key: uu49)
If you use this project in your research please cite:
@article{CSL:WU2023,
doi = {10.1080/15481603.2022.2147298},
author = {Kang Wu and Zunxiao Xu and Xinrong Lyu and Peng Ren},
title = {Cross-supervised learning for cloud detection},
journal = {GIScience \& Remote Sensing},
volume = {60},
number = {1},
pages = {2147298},
year = {2023},
publisher = {Taylor & Francis},
doi = {10.1080/15481603.2022.2147298},
URL = {https://doi.org/10.1080/15481603.2022.2147298},
eprint = {https://doi.org/10.1080/15481603.2022.2147298}
}
Some implementations are built based on segmentation_models.pytorch.
This repository can only be used for personal/research/non-commercial purposes. If you have any questions about this work, please raise an issue or contact me at kang_wu#foxmail.com
. (Please replace #
with @
)
此处可能存在不合适展示的内容,页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容,可点击提交进行申诉,我们将尽快为您处理。