DeepCube platform

The different technological components of DeepCube, namely the functionalities of the Earth System Data Cube, the Semantic Web technology Ontop-spatial, the visualization tools for EO and non-EO data, and the functionalities of the Hopsworks distributed Deep Learning platform, will be extended to better scale to big Copernicus datasets and will be integrated into an open, interoperable TRL 7 platform – the DeepCube platform – that will be deployed on several cloud infrastructures and High Performance Computing, including Copernicus DIAS environments, with the aim to address the Earth Observation value chain as a whole.

DeepCube complements DIAS in creating wrappers and ingestion modules to systematically access EO and non-EO sources, extract metadata and populate Data Cubes with Analysis Ready Data from heterogeneous sources.

The DeepCube platform users will be able to use existing or create their own data wrappers to ingest multi-modal Earth Observation and non-Earth Observation data to the platform, pre-process the data and store them in an Analysis Ready Format, explore time-series analytics using Data Cubes, perform feature engineering, feed well known Machine Learning algorithms or create new Deep Learning architectures that are trained in a distributed environment, use tools to explain and interpret the behavior of an AI model and understand the importance and relation of covariates (Explainable Artificial Intelligence), semantically enrich AI outputs and combine them with Linked Open Data to create new value chains, and finally allow visual analytics of spatio-temporal annotated layers.

Interested in learning more? Contact us!
Christophe Demange, [email protected]