EO4EU
AI-augmented ecosystem for Earth Observation data accessibility with Extended reality User Interfaces for Service and data exploitation
A user-friendly platform to access, analyse and visualise Earth Observation data, powered by Machine Learning and extended reality.
The project
The EO4EU Horizon Europe project, funded by the European Commission, aims to facilitate the use of Earth Observation (EO) data for environmental, governmental, and business forecasting, through the use of a user-friendly platform.
The EO4EU platform connects major EO data sources (including the DestinE platform) enabling users to access, analyze, and visualize data in one place. Powered by Machine Learning (ML) and a high-performance cloud computing infrastructure, it handles large volumes and demanding processing workloads. Intuitive interfaces, including extended reality, make EO data freely accessible to any user.
MEEO’s role
MEEO, together with SISTEMA and the Euro-Mediterranean Center on Climate Change (CMCC) led Use Case 3 (UC3), focused on food security. The core of MEEO's technical contribution is a machine learning pipeline integrated into the EO4EU platform.
- Pre-processing: heterogeneous sources (ERA5 climate data, NDVI vegetation indices derived from Sentinel-2 imagery, the Copernicus Digital Elevation Model, SoilGrids pedological information, CMCC very high-resolution climate projections, and production data) are merged and transformed into 84 climate indicators, computed across different temporal windows of the crop growing cycle. These indicators form the feature set for training the model, which is based on a neural network architecture.
- Inference: the model can be queried through the EO4EU platform to generate georeferenced forecasts for user-defined areas and future climate scenarios, producing a yield classification output in four classes.
The methodology is designed to be scalable: initially developed for tomatoes in the Emilia-Romagna region and Spain, it has been successfully extended to corn, demonstrating the transferability of the approach to other crops and geographical contexts.
The goal is to provide producers, trade associations, and institutional decision-makers with predictive information that supports climate-resilient agricultural management.
- 84
- Climate indicators feeding the model
- 4
- Yield classes in the forecast output