This work was partly presented at the EGU 2018 in the session Learning from spatial data: unveiling the geo-environment through quantitative approaches.
Here, you will find some Python codes. A step-by-step implementation using two Python libraries: TenserFlow and Keras with the TensorFlow back-end.
The benchmark datasets used in this work can be easily downloaded to a local workstation. They are freely available from open-source repositories.
Address
Swiss Geocomputing Centre
Quartier UNIL-Mouline, Building Géopolis
1015 Lausanne, Switzerland.
A special thank to Ludovic Räss for his assistance and input regarding
the efficient use of GPU clusters .
