This repository includes code implementation for our article:

Binh Minh Nguyen, Ganglin Tian, Minh-Triet Vo, Aurélie Michel, Thomas Corpetti, et al.. Convolutional Neural Network Modelling for MODIS Land Surface Temperature Super-Resolution. 2022.

The brief description for each file is shown in the table below:

pip install -r requirements.txt

User should firstly register on NASA's website[https://urs.earthdata.nasa.gov/users/new]. User could use the following command to download automatically MODIS dataset with the user name and corresponding password, the downloaded data will be placed folder "MODIS" in current directory.

python modis_downloader.py --username <username> --password <password> 

By assigning values to '--year_begin' and '--year_end', user can select the time range for downloading data, from year_begin to year_end. Default 'year_begin' is 2020, 'year_end' is 2021.

python modis_downloader.py --year_begin <year_begin> --year_end <year_end> --username <username> --password <password> 

After downloading the raw data for the desired years, data need to be pre-processed, which involves cropping, cloud and sea pixels elimination and downsampling:

python modis_data_preprocessing.py --year_begin <year_begin> --year_end <year_end>

Our principle contribution in this project is to design and implement a new deep learning model based on U-net architecture called Multi-residual U-Net. The architecture of this model is shown as below:

To train the network with your tif data, please make sure to put all of the data into a folder and run the following command:

python train.py --datapath <path/to/training/tifs/directory> --model_name <name of the model> --lr <learning rate> --epochs <number of epochs> --batch_size <size of batch> --continue_train <True/False>

P/s: The checkpoint of training process lies in the same directory as the train.py.

To run the inference of the pretrained network on your tif data, please make sure to put all of the data into a folder and run the following command:

python run_inference.py --datapath <path/to/testing/tifs/directory> --pretrained <path/to/pretrained/model.pth> --savepath <path/to/save/directory> --max_val <normalization factor>

P/s:

1. The max_val is the maximum pixel value of the data used for training the pretrained model.
2. A pretrained Multi-residual U-Net can be found in Visualization benchmark.

To quantify the results, we employed some famous metrics in image super resolution: PSNR and SSIM. We also mentioned RMSE because it is meaningful in remote sensing.

Qualitative results:

Quantitative results:

You can find the notebook as well as the instruction for the benchmarking visualization in the folder Visualization benchmark. The notebook provides functions to plot and benchmark several techniques mentioned in the paper.