The specific aim of the project is the creation of a Machine Learning model and/or a Convolutional Neural Network that processes RAW datasets relative to user-selected terrestrial, airborne, orbit scenarios and gives as a result a landform classification map. The more global aim is the realization of models, workflows, procedures for classifying landorms in real-time and further predicting landforms on unkown areas.

Landform classification maps generated autonomously by the models that correspond well to the maps generated and validated manually by users, is the minimal result. The perfect results are autonomously generated maps that correspond exactly to user-generated ones and the possibility to use predictions in real-time applications, e.g. rover and spacecraft data, on unkown areas with above-average to high accuracy and precision.