UNICORN

Project Description

UNICORN develops next-generation Copernicus Emergency Management applications that leverage Earth Observation (EO) technologies and data to address the escalating frequency and intensity of extreme natural events — wildfires, flash floods, and volcanic eruptions — and their cascading impacts on society, the economy, and the environment.

The project responds to a critical gap in European disaster risk management: despite significant advances in EO capabilities, early warning and detection systems remain under-exploited, leading to prolonged, costly, and recurrent emergency responses. Europe’s Mediterranean region in particular has witnessed a sharp increase in natural disasters causing substantial human casualties and economic losses, a trend directly linked to accelerating climate change. UNICORN’s answer is a suite of scalable, transferable, and user-centred services that push the technological frontier for precise, timely, and actionable intelligence from Copernicus data and knowledge graphs.

The project is structured around four use cases drawn from different European regions, hazard types, and target stakeholder communities, ranging from local civil protection authorities and first responders to industries and policymakers. UNICORN’s end-user-validated approach ensures that tools are co-designed with operational end users — including the Region of Attica (Greece) and the fire service of Haute-Corse (France) — and that outputs directly strengthen emergency preparedness, forecasting capacity, short-term recovery, and business continuity planning for climate-impacted sectors. The project specifically leverages knowledge graph-based data integration to transform heterogeneous EO and hazard data streams into structured, interpretable outputs for decision makers at all levels.

LINKS Foundation Activities

LINKS Foundation participates as a full partner within UNICORN, contributing expertise in AI-driven Earth Observation, deep learning for satellite imagery analysis, crowdsourcing, human-computer interaction, and emergency management platform development — areas in which LINKS has built a recognised track record through projects including I-REACT, SAFERS, FASTER, and SHELTER.

AI-Based EO Processing for Wildfire and Flood Hazard Applications

Development and validation of deep learning models applied to Sentinel-1 and Sentinel-2 satellite imagery for the automated delineation of wildfire-affected areas and flood extents, building on LINKS’ established research in semantic segmentation and sparse annotation:

• training and deployment of convolutional neural network (CNN) and transformer-based architectures for burned area mapping, land cover segmentation, and flood delineation from multi-temporal EO data
• integration of outputs within UNICORN’s Copernicus Emergency Management application stack, contributing to near real-time hazard monitoring and post-event damage assessment services
• validation against four use case scenarios across Mediterranean regions, generating operationally relevant geospatial products for local authorities and civil protection services

Knowledge Graph Development for Multi-Hazard Data Integration

Design and implementation of semantic knowledge graph architectures to interconnect heterogeneous data sources — Copernicus EO data, historical hazard databases, meteorological forecasts, and socioeconomic exposure datasets — into structured, queryable representations for decision support:

• development of ontological frameworks and knowledge representation models capturing relationships between hazard events, affected assets, population exposure, and emergency response actions
• enabling policy makers, emergency managers, and industry stakeholders to query integrated multi-hazard knowledge bases for risk profiling, preparedness planning, and business continuity assessment
• contribution to UNICORN’s goal of transforming raw EO and hazard data into actionable intelligence aligned with Copernicus Emergency Service standards and EU policy frameworks

Human-Centred Interface Design and Stakeholder Engagement

Contribution of expertise in user experience (UX) design, crowdsourcing, and citizen engagement to ensure UNICORN’s tools meet the operational needs of diverse end users — from first responders and civil protection professionals to industry and citizens:

• co-design and usability evaluation of UNICORN’s decision support dashboards and early warning interfaces through structured end-user validation activities with local authorities and emergency services
• development of crowdsourcing and citizen science components to enrich UNICORN’s situational awareness layer with real-time ground-truth observations during and after extreme events
• ensuring UNICORN’s outputs are accessible and effectively adopted by non-specialist stakeholders, including local businesses assessing continuity risk and policymakers designing climate adaptation strategies

Funding & Dates

SDGs — Sustainable Development Goals

🌡️ Goal 13 — Climate Action

🌿 Goal 15 — Life on Land

Key Technologies

🛰️ Copernicus EMS & Sentinel Earth Observation

Multi-temporal exploitation of Copernicus Emergency Management Service data and Sentinel-1/2/3 satellite imagery for near real-time detection, monitoring, and post-event mapping of wildfires, flash floods, and volcanic hazards — providing the foundational EO data layer for all UNICORN services.

🤖 Deep Learning for Satellite Image Analysis

Application of convolutional neural networks (CNNs), transformer architectures, and sparse annotation training techniques to EO imagery for automated burned area delineation, flood extent mapping, land cover segmentation, and damage assessment — directly building on LINKS’ published research contributions.

🕸️ Knowledge Graphs for Multi-Hazard Intelligence

Semantic knowledge graph frameworks interconnecting heterogeneous hazard, exposure, and socioeconomic datasets into structured, queryable representations — enabling complex multi-hazard risk queries and supporting business continuity and policy planning across sectors affected by extreme events.

⚠️ Early Warning & Hazard Forecasting Systems

Integration of Copernicus EO data streams with numerical weather prediction models, historical event databases, and ground-based sensor networks to produce probabilistic early warnings and multi-day hazard forecasts for wildfires, flash floods, and volcanic lava flows, tailored to operational end-user needs.

👥 Crowdsourcing & Citizen Engagement Platforms

Citizen science and crowdsourcing tools that enrich UNICORN’s situational awareness layer with real-time ground-truth observations during emergencies — improving the spatial and temporal resolution of hazard intelligence and fostering inclusive, community-driven disaster resilience.

🖥️ Human-Centred Decision Support Dashboards

UX-designed interfaces and operational dashboards for civil protection authorities, first responders, industries, and policymakers — co-developed through structured end-user validation to ensure trust, usability, and effective adoption of UNICORN’s Copernicus emergency applications at scale.

Project Partners