Our solutions

Explore our Solutions

From Research to Real-World Solutions

We combine high-level scientific research with advanced technology to deliver innovative services. Discover our high-tech geoinformation services, powered by remote sensing and AI algorithms.

Earth Observation for Climate Resilience

Our primary solution is the design, creation and management of systems and services for environmental, territorial and infrastructural surveillance. This includes providing up-to-date risk and impact assessments associated with natural disasters and extreme events, including those caused by climate change.

Multi-Source Satellite Data

We utilize multispectral, hyperspectral, and SAR (Synthetic Aperture Radar) satellite data to extract actionable insights that support environmental protection.

The Power of Earth Observation

Our approach harnesses the full potential of Earth Observation, from global coverage and long-term time-series to high spatial, temporal, and spectral resolutions.

Data Fusion & Validation

Our strategy also integrates in-situ measurements and modeled data with satellite imagery. These complementary information are crucial for the validation of our products and for developing data fusion pipelines.

Earth Observation for Climate Resilience

Multi-Source Satellite Data

We utilize multispectral, hyperspectral, and SAR (Synthetic Aperture Radar) satellite data to extract actionable insights that support environmental protection.

The Power of Earth Observation

Our approach harnesses the full potential of Earth Observation, from global coverage and long-term time-series to high spatial, temporal, and spectral resolutions.

Data Fusion & Validation

Lightweight On-board AI

Lightweight
On-board AI

We develop lightweight AI algorithms to bring Earth Observation analysis directly on-board satellites. By adapting models to limited computing resources, we aim to process data close to acquisition and deliver useful information with less dependence on ground processing.

Model optimization and deployment

We work with optimization approaches, such as quantization and pruning, to reduce model size and computational demand while assessing their impact on predictive accuracy.

On-board processing

Our research targets tasks such as image classification, object detection, and parameter estimation. The aim is to extract relevant information in orbit and transmit compact data products to the ground.

Lightweight On-board AI

Model optimization and deployment

We work with optimization approaches, such as quantization and pruning, to reduce model size and computational demand while assessing their impact on predictive accuracy.

On-board processing

Explainable AI

Given the high complexity of state-of-the-art AI models and the need to move beyond the "black box" paradigm, we research and develop explainable AI (xAI) solutions to guarantee algorithmic and scientific interpretability.

Algorithmic explainability

We explore xAI tools, including feature importance, activation maps, and saliency maps, to understand model behaviour and identify the factors that influence predictions.

Scientific interpretability

We implement methods to support the physical consistency of model predictions and their interpretation through mechanisms such as scattering and absorption. We also use physics-informed neural networks (PINNs) to integrate physical knowledge into model development.

Training and Dissemination

We deliver high-level training activities that combine theoretical foundations with hands-on sessions. In addition, our nature as a research-driven spin-off ensures that our courses are constantly updated with the latest emerging technologies.

EO data & Machine Learning

We offer introductory programs covering the fundamentals of Earth Observation, satellite data processing, and machine learning approaches. Our courses utilize both open-source softwares and Python environment.