🌊 Deep coastal sea elements forecasting using U-Net based models

The official code of the following paper : https://arxiv.org/pdf/2011.03303.pdf

📊 Results

Some animation of the actual vs prediction of the AsymmInceptionRes-3DDR-UNet model, using a 48h ahead prediction:

Variable	Actual Vs Prediction
Sea Surface Height
Sea Water Salinity
Eastward Current Velocity
Northward Current Velocity

Note: the dark area is made of pixels that correspond to the land.

💻 Installation

The required modules can be installed via:

pip install -r requirements.txt

Quick Start

To launch the training, please run:

python train_selected_model.py

📜 Scripts

• The scripts contain the models, the data preprocessing, as well as the training files.

🔍 Models

We show here the schema related to the AsymmInceptionRes-3DDR-UNet model.

📂 Data

In order to download the data, please email to the following addresses:

j.garciafernandez@student.maastrichtuniversity.nl

i.alaouiabdellaoui@student.maastrichtuniversity.nl

siamak.mehrkanoon@maastrichtuniversity.nl

The data must be downloaded and unzipped inside the 'Data/' directory.

🔗 Citation

If you use our data and code, please cite the paper using the following bibtex reference:

@article{Fernández2020coastal,
    title={Coastal sea elements prediction using U-Net based models},
    author={García Fernández, Jesús and Alaoui Abdellaoui, Ismail and Mehrkanoon, Siamak},
    journal={arXiv preprint arXiv:2011.03303},
    year={2020}
}