TTXian/RemoteSensingChangeDetection-RSCD.HA2F

1## 🛠️ Requirements2 3### Environment4- **Linux system**,5- **Python** 3.8+, recommended 3.106- **PyTorch** 2.0 or higher, recommended 2.1.07- **CUDA** 11.7 or higher, recommended 12.18 9### Environment Installation10 11It is recommended to use Miniconda for installation. The following commands will create a virtual environment named `stnr` and install PyTorch. In the following installation steps, the default installed CUDA version is 12.1. If your CUDA version is not 12.1, please modify it according to the actual situation.12 13```bash14# Create conda environment15conda create -n stnr python=3.8 -y16conda activate stnr17 18# Install PyTorch19pip install -r requirements.txt20```21 22## 📁 Dataset Preparation23 24We evaluate our method on five remote sensing change detection datasets: **WHU-CD**, **LEVIR-CD**, **SYSU-CD**.25 26| Dataset | Link |27|---------|------|28| WHU-CD | [Download](https://aistudio.baidu.com/datasetdetail/251669) |29| LEVIR-CD | [Download](https://opendatalab.org.cn/OpenDataLab/LEVIR-CD) |30| SYSU-CD | [Download](https://mail2sysueducn-my.sharepoint.com/personal/liumx23_mail2_sysu_edu_cn/_layouts/15/onedrive.aspx?id=%2Fpersonal%2Fliumx23%5Fmail2%5Fsysu%5Fedu%5Fcn%2FDocuments%2FSYSU%2DCD&ga=1) |31 32 33 34### Example of Training on LEVIR-CD Dataset35 36```bash37python main.py --file_root LEVIR --max_steps 80000 --model_type small --batch_size 16 --lr 2e-4 --gpu_id 038```39 40### Example of Training on LEVIR-CD Dataset41 42```bash43python eval.py --file_root LEVIR --max_steps 80000 --model_type small --batch_size 16 --lr 2e-4 --gpu_id 044```45 46## 📂 DATA Structure47 48```49├─Train50      ├─A          jpg/png51      ├─B          jpg/png52      └─label      jpg/png53  ├─Val54      ├─A 55      ├─B56      └─label57  ├─Test58      ├─A59      ├─B60      └─label61```62 63## 🙏 Acknowledgement64 65We sincerely thank the following works for their contributions:66 67- [ChangeViT](https://arxiv.org/pdf/2406.12847) – A state-of-the-art method for remote sensing change detection that inspired and influenced parts of this work.68