This is a official implementation for our paper: "Oil spill contextual and boundary-supervised detection network based on marine SAR images" has been published on IEEE Transactions on Geoscience and Remote Sensing by Qiqi Zhu , Yanan Zhang , Ziqi Li , Xiaorui Yan , Qingfeng Guan , Yanfei Zhong , Liangpei Zhang and Deren Li.
Deep-SAR Oil Spill (SOS) dataset
Baidu Drive (extraction code: urs6)：http://cugurs5477.mikecrm.com/QaXx0Rw
Google Drive：http://cugurs5477.mikecrm.com/5tk5gyO
The dataset consists of two study areas and is mainly used for scientific purposes.
The dataset contains the Gulf of Mexico oil spill area and the Persian Gulf oil spill area, acquired from the ALOS and Sentinel-1A satellites, respectively. It is a pixel-level dataset of oil spills and non-oil spills collected and produced by the ECHO research team.
Among them, 3101 samples from the Mexican oil spill area were used for model training and 776 samples were used for testing.
From the Persian Gulf oil spill area, 3354 samples were used for training and 839 samples were used for testing.
train.py run: python train.py
1.Dataset path (palsar, Sentinel)
ROOT = '.datasets/train/palsar/'
2.CBDNet is the network, dice_bce_loss is the loss function, and the learning rate is 2e-4
solver = MyFrame(CBDNet, dice_bce_loss, 2e-4)
3.Name of weights document
NAME = 'palsar_CBDNet'
test.py run: python test.py
1.Dataset path
source = '.datasets/test/palsar/sat/'
2.Loading Network
solver = TTAFrame(CBDNet)
3.Loading the weights file
solver.load('.weights/palsar_CBDNet.th')
4.Test Results Documentation
target = '.submits/palsar_CBDNet/'
t1p3-Iou.py
run: python t1p3-Iou.py
1.truth label
name_truth = '.datasets/test/palsar/gt/'
2.Predicted results
name_pred = '.submits/palsar_CBDNet/'
If you find our work useful for your research, please consider citing our paper:
@article{zhu2021oil,
title={Oil spill contextual and boundary-supervised detection network based on marine SAR images},
author={Zhu, Qiqi and Zhang, Yanan and Li, Ziqi and Yan, Xiaorui and Guan, Qingfeng and Zhong, Yanfei and Zhang, Liangpei and Li, Deren},
journal={IEEE Transactions on Geoscience and Remote Sensing},
volume={60},
pages={1--10},
year={2021},
publisher={IEEE}
}
You can contact the e-mail [email protected] if you have further questions about the usage of codes and datasets. For any possible research collaboration, please contact Prof. Qiqi Zhu ([email protected]). The homepage of our academic group is: http://grzy.cug.edu.cn/zhuqiqi/en/index.htm. Date: Dec 4, 2023