In this work two unsupervised deep learning models were utilized for Super Resolving satellite imagery obtained from Sentinel-2 constellation.
The original implementation of this model was proposed by Ulyanov et al.
- python == 3.8
- earthpy
- numpy
- pytorch
- matplotlib
- scikit-image
- gdal
- rasterio
- jupyter notebook
The original implementation of this model was proposed by Shocher et al.
- python == 2.7
- numpy
- tensorflow
- matplotlib
- scikit-image
- opencv-python
- imageio
The original implementation of this modes was proposed by Wang et al.
- Python 3.6
- PyTorch == 1.1.0
- numpy
- skimage
- imageio
- matplotlib
- cv2
1.1 Download the DIV2K dataset and the Flickr2K dataset.
1.2 Combine the HR images from these two datasets in your_data_path/DF2K/HR to build the DF2K dataset.
Run ./main.sh to train on the DF2K dataset. Please update dir_data in the bash file as your_data_path.
Download benchmark datasets (e.g., Set5, Set14 and other test sets) and prepare HR/LR images in your_data_path/benchmark.
Run ./test.sh to test on benchmark datasets. Please update dir_data in the bash file as your_data_path.
Run ./quick_test.sh to test on an LR image. Please update img_dir in the bash file as your_img_path.