The STF-LST Dataset is a robust foundation for developing and evaluating innovative spatio-temporal fusion techniques, specifically designed to address the challenges of land surface temperature estimation. This dataset includes 51 paired MODIS/Landsat images, each with a resolution of 950 x 950 pixels. They were collected between March 18, 2013, and October 15, 2024, and they cover the Orleans Métropole in the Centre-Val de Loire region of France.
Below is a video showcasing diverse samples from the STF-LST Dataset (click on the picture and download the video) :
Features | Tutorial | Guide of use | Paper | ArXiv | How to cite us ?
The STF-LST Dataset offers the following features:
- 51 paired MODIS-Landsat images covering a wide range of time periods.
- Various preprocessing techniques were applied, including linear, spatial, and bicubic interpolation methods.
- A fully reproducible codebase that can be adapted for different regions and time periods by simply adjusting the parameters.
To generate the STF-LST dataset, run the following command in your terminal:
python3 run.pyThe code utilizes the Google Earth Engine platform, so you will need a valid account for authentication before downloading the data.
Please note that this process may take some time.
STF-LST dataset has been generated using the following versions:
- Python (v3.9.19).
- Torch (v2.4.1+cu121).
- Scipy (v1.13.1).
- Earth Engine (v1.1.2).
- Geemap (v0.34.5).
- Rasterio (v1.3.10).
- NumPy (v1.26.4).
- Pandas (V2.2.2).
STF-LST dataset has been developed by Sofiane Bouaziz, Adel Hafiane, Raphaël Canals and Rachid Nedjai.
In case you are using STF-LST dataset for your research, please consider citing our work:
@article{bouaziz2024deep,
title={Deep Learning for Spatio-Temporal Fusion in Land Surface Temperature Estimation: A Comprehensive Survey, Experimental Analysis, and Future Trends},
author={Bouaziz, Sofiane and Hafiane, Adel and Canals, Raphael and Nedjai, Rachid},
journal={arXiv preprint arXiv:2412.16631},
year={2024}
}