MISGNet: Multilevel Intertemporal Semantic Guidance Network for Remote Sensing Images Change Detection
Here, we provide the pytorch implementation of the paper: MISGNet: Multilevel Intertemporal Semantic Guidance Network for Remote Sensing Images Change Detection.
albumentations>=1.3.0
numpy>=1.20.2
opencv_python>=4.7.0.72
opencv_python_headless>=4.7.0.72
Pillow>=9.4.0
Pillow>=9.5.0
scikit_learn>=1.0.2
torch>=1.9.0
torchvision>=0.10.0
Clone this repo:
git clone https://github.com/JackLiu-97/MISGNet.git
cd MISGNet
Firstly, you can download our MISGNet pretrained model
LEVIR-CD: baidu drive, code: itrs .
SYSU-CD: baidu drive, code: itrs .
After downloaded the pretrained model, you can put it in output
.
Then, run a demo to get started as follows:
python demo.py --ckpt_url ${model_path} --data_path ${sample_data_path} --out_path ${out_data_path}
To train a model from scratch, use
python train.py --data_path ${train_data_path} --val_path ${val_data_path} --lr ${lr} --batch_size ${-batch_size}
To evaluate a model on the test subset, use
python predict.py --ckpt_url ${model_path} --data_path ${test_data_path}
In order to make it more convenient for readers to compare with our model, we also provide the inference results of our model.
LEVIR-CD: baidu drive, code: itrs .
SYSU-CD: baidu drive, code: itrs .
WHU-CD :The WHU Building Change Detection Dataset :The dataconsists of two aerial images of two different time phases and the exact location, which contains
SYSU-CD : The dataset contains
Dataset | Name | Link |
---|---|---|
LEVIR-CD building change detection dataset | LEVIR-CD |
website |
SYSU-CD building change detection dataset | SYSU-CD |
website |
WHU building change detection dataset | WHU-CD |
website |
Code is released for non-commercial and research purposes only. For commercial purposes, please contact the authors.