CSA-CDGAN: Channel Self-Attention Based Generative Adversarial Network for Change Detection of Remote Sensing Images
A general framework for change detection of remote sensing images
paper link: https://link.springer.com/article/10.1007/s00521-022-07637-z
Python 3.7.0
Pytorch 1.6.0
Visdom 0.1.8.9
Torchvision 0.7.0
- CDD: Change detection in remote sensing images using conditional adversarial networks
- WHU-CD: Fully Convolutional Networks for Multisource Building Extraction From an Open Aerial and Satellite Imagery Data Set
- LEVIR-CD: A Spatial-Temporal Attention-Based Method and a New Dataset for Remote Sensing Image Change Detection
You also can download datasets after being processed by us. [Baiduyun] the password is hnbi. or [GoogleDrive]
Pretrained models for CDD, LEVIR-CD and WHU-CD are available. You can download them from the following link. [Baiduyun] the password is yudl. [GoogleDrive]
Before test, please download datasets and pretrained models. Revise the data-path in constants.py to your path. Copy pretrained models to folder './dataset_name/outputs/best_weights', and run the following command:
cd CSA-CDGAN_ROOT
python make_dataset.py
python test.py
make_dataset.py can generate .txt files for training, validation and test. Not that the dataset structure should be the same as following:
Custom dataset
|--train
|--file1
|--t0.jpg, t1.jpt, label.jpg
|--file2
|--t0.jpg, t1.jpt, label.jpg
...
|--test
|--file1
|--t0.jpg, t1.jpt, label.jpg
|--file2
|--t0.jpg, t1.jpt, label.jpg
...
|--validation
|--file1
|--t0.jpg, t1.jpt, label.jpg
|--file2
|--t0.jpg, t1.jpt, label.jpg
...
cd CSA-CDGAN_ROOT
python make_dataset.py
python -m visdom.server
python train.py
To display training processing, copy 'http://localhost:8097' to your browser.
If you use this repository or would like to refer the paper, please use the following BibTex entry.
@article{wang2022csa,
title={CSA-CDGAN: channel self-attention-based generative adversarial network for change detection of remote sensing images},
author={Wang, Zhixue and Zhang, Yu and Luo, Lin and Wang, Nan},
journal={Neural Computing and Applications},
pages={1--15},
year={2022},
publisher={Springer}
}
-Akcay, Samet, Amir Atapour-Abarghouei, and Toby P. Breckon. "Ganomaly: Semi-supervised anomaly detection via adversarial training." Asian conference on computer vision. Springer, Cham, 2018.