System Operational · 3 Active Satellites

EARTH-GUARD

AI-Powered Satellite Environmental Monitoring — EU AI Act Compliant

SAT-EG01 · SAT-EG02 · SAT-EG03  |  ORBIT #14,228  |  REG. (EU) 2024/1689

Live Dashboard → EU AI Act Compliance
7
Active Alerts
98.4%
Area Coverage
2.1h
Revisit Time
97%
AI Accuracy
—
UTC Time
↓ scroll
🔥
Wildfire Detection
Thermal IR and NDVI fusion detects active fires within minutes of ignition. 97% detection accuracy across Mediterranean terrain.
💧
Flood Monitoring
SAR backscatter analysis maps inundation extent in real time, even through cloud cover. Soil moisture forecasting included.
🌫
Air Quality
Sentinel-5P NO₂, PM2.5 and AOD retrievals fused with ground sensors deliver city-level pollution maps at 1 km resolution.
🌊
Marine & Water
Chlorophyll-a, sea surface temperature, and turbidity derived from optical multispectral imagery. HABs detected early.
⚖️
EU AI Act Ready
Classified as High-Risk under Annex III. Full documentation, human oversight, and audit logging per Articles 9–15.
🔗
Open Standards
OGC-compliant APIs, INSPIRE-compatible metadata, WMS/WFS/WMTS endpoints. FAIR data principles throughout.
Live Demo

Operational Monitoring Dashboard

Real-time environmental surveillance across the Hellenic region. Three alert zones active — click any alert for details.

EARTH-GUARD OPERATIONAL CONTROL · earthguard.eu/dashboard · —
LIVE
ACTIVE ALERTS 3 CRITICAL
🔴 WILDFIRE · CRITICAL —
Active Fire — Penteli, Attica
📍 38.06°N, 23.91°E · Attica Region
Thermal anomaly 340 ha. Fire front advancing NE at 12 km/h. Extreme spread risk. Civil Protection notified.
FRP 2,840 MW RH 18% Wind 34 km/h
AI Confidence
97%
🟠 FLOOD RISK · HIGH —
Soil Saturation — Thessaly Plain
📍 39.36°N, 22.41°E · Larissa Prefecture
SAR confirms soil moisture 94% field capacity. Pinios River at 87% bankfull. 48h rain forecast 52 mm.
SM 94% River 87% BF 48h 52 mm
AI Confidence
84%
🟡 AIR QUALITY · HIGH —
PM2.5 Exceeded — Athens Basin
📍 37.97°N, 23.73°E · Greater Athens Area
PM2.5 at 162 μg/m³ — 3.2× WHO limit. Urban heat island and low wind causing poor dispersion.
PM2.5 162 μg/m³ AQI 214 NO₂ 89 ppb
AI Confidence
91%
⬡ SENTINEL COMPOSITE · HELLENIC REGION 34.8°N–41.8°N · 19.4°E–28.3°E
41°N 39°N 37°N THESSALY FLOOD 🟠 ATTICA FIRE 🔴 CRITICAL ATHENS AQI 214 🟡 THESSALONIKI NOMINAL CRETE · NOMINAL 🛰 N 150 km
Wildfire · Attica
Flood Risk · Thessaly
Air Quality · Athens
Nominal Sensor
EU AI ACT STATUS CERTIFIED
⚠
Classification
HIGH-RISK AI
Art. 6 · Annex III · Category 8
Reg. (EU) 2024/1689
Mandatory Requirements
✓
Art. 9
Risk Management
COMPLIANT
✓
Art. 10
Data Governance
COMPLIANT
✓
Art. 11
Technical Docs
COMPLIANT
✓
Art. 13
Transparency
COMPLIANT
✓
Art. 14
Human Oversight
COMPLIANT
!
Art. 17
Quality Mgmt
IN REVIEW
90%
SCORE
Satellite Network
2/3
Sats Online
98%
Coverage
2.1h
Revisit
7
Alerts
⬡ SATELLITE DATA STREAM
INGESTING
1.84 GB/s · 142ms latency
Regulatory Compliance

EU AI Act Compliant
High-Risk Classification

EARTH-GUARD is classified as a High-Risk AI System under Annex III, Category 8 of Regulation (EU) 2024/1689 — the EU Artificial Intelligence Act. This classification reflects the critical nature of environmental monitoring in public safety decision-making.

Full compliance with Articles 6–17 has been achieved through systematic implementation of risk management, human oversight mechanisms, and transparent documentation. Certification issued by TÜV HELLAS, Certificate No. EG-2026-0441.

Article 6 Classification Basis
EARTH-GUARD meets the Annex III threshold as an AI system used by public authorities for environmental risk assessment affecting public safety, critical infrastructure protection, and emergency response. The system's outputs directly inform decisions by Civil Protection Services and local authorities across the Hellenic region.
Human Oversight — Article 14
All AI-generated alerts are reviewed by certified environmental analysts before dispatch to authorities. The system is designed to be interpretable and auditable. No automated enforcement actions — humans remain in the decision loop at all times, consistent with HLEG-AI (2019) principles.
✓ TÜV HELLAS Certified Art. 6 Annex III Cat. 8 Reg. (EU) 2024/1689
Requirements Status — Articles 8–17
Mandatory obligations for high-risk AI systems
✓
Art. 9
Risk Management System
Continuous risk identification, evaluation and mitigation. Updated quarterly.
COMPLIANT
✓
Art. 10
Data & Data Governance
Training on Copernicus open data, bias assessment completed, data lineage documented.
COMPLIANT
✓
Art. 11
Technical Documentation
Full technical dossier maintained per Annex IV. Version-controlled, publicly accessible summary.
COMPLIANT
✓
Art. 12
Record-Keeping
Immutable audit log of all inferences, alert generations and human review decisions.
COMPLIANT
✓
Art. 13
Transparency & Information
Instructions for use, capability and limitation disclosure to all deploying authorities.
COMPLIANT
✓
Art. 14
Human Oversight
Override mechanisms, analyst review mandatory for CRITICAL/HIGH alerts, explainability layer.
COMPLIANT
✓
Art. 15
Accuracy, Robustness & Security
Validated at 97% fire detection accuracy. Adversarial robustness tested. ISO 27001 security.
COMPLIANT
!
Art. 17
Quality Management System
QMS documentation under review. Expected completion Q3 2026.
IN REVIEW
System Architecture

How EARTH-GUARD Works

From satellite sensor to authorised alert in under 8 minutes, with human oversight at every critical decision point.

01 / ACQUIRE
🛰
Satellite Acquisition
Three-satellite constellation (SAT-EG01–03) collects multispectral, thermal IR, and SAR imagery over the Hellenic region every 2.1 hours.
→
02 / PROCESS
⚙️
AI Processing
Deep learning models analyse 1.8 GB/s of raw sensor data. Fire, flood, air quality and water contamination classifiers run in parallel on GPU clusters.
→
03 / REVIEW
👁
Human Oversight
Per Article 14 EU AI Act, all HIGH and CRITICAL alerts are reviewed by certified environmental analysts before dispatch. Humans remain in the loop.
→
04 / ALERT
🚨
Authorised Alert
Validated alerts reach Civil Protection, local authorities, and emergency services via secure API within the 8-minute pipeline target.
→
05 / AUDIT
📋
Audit & Improve
Every inference, decision, and human override is recorded in an immutable audit log. Model performance is reviewed and updated monthly.
3
Active Satellites
LEO · 520 km orbit
10m
Resolution
Optical multispectral
2.1h
Revisit Time
Avg. Hellenic region
97%
Fire Accuracy
Validated on test set
<8 min
Alert Pipeline
Sensor to authority
1.8 GB/s
Throughput
Real-time ingestion
Research Origin

Developed at Aristotle University of Thessaloniki

EARTH-GUARD was conceived and developed as a research project within the MSc programme in ICT Law and Policy at the Aristotle University of Thessaloniki (AUTH), one of the largest and most prestigious universities in Greece and southeast Europe.

The project addresses a critical intersection of space technology, artificial intelligence, and EU regulatory law — demonstrating how AI Act compliance frameworks can be designed into satellite-based environmental monitoring systems from the ground up.

The research received a grade of 10/10 and was submitted in February 2026. It has since been developed into a functional prototype and deployed at earthguard.eu.

AUTH · Thessaloniki MSc ICT Law & Policy Grade 10/10 February 2026
Principal Researcher
Pelagia Eleftheriadou
MSc ICT Law & Policy · AUTH
Research focus: AI regulatory compliance frameworks for satellite-based environmental monitoring systems. Expertise in EU AI Act implementation, space law, and human-centred AI design.
Dissertation: "Designing EU AI Act Compliant High-Risk AI Systems for Satellite Environmental Monitoring: The EARTH-GUARD Framework"
10 /10
Final Grade Aristotle University of Thessaloniki · February 2026