Cross-site analysis of 5 UK infrastructure assets using Sentinel-2 optical and Sentinel-1 SAR satellite data. Same-season methodology with anomaly-filtered change detection. All findings verified against real satellite imagery.
| Site | Anomalous | Confidence | Red Zones | Key Finding |
|---|---|---|---|---|
| ☢ Sizewell C | 19.5% | 91% | 2 | 19.5% of the Sizewell monitoring area shows genuine year-on-year land-use change... |
| ☢ Sellafield | 33.1% | 88% | 3 | 33.1% anomalous change at Sellafield — the NDA’s primary condition monitoring ta... |
| ☢ Dounreay | 11% | 92% | 1 | 11.0% anomalous change at Dounreay — stable decommissioning site confirms baseli... |
| ☢ Hinkley Point C | 51.2% | 78% | 4 | 51.2% anomalous change at Hinkley Point C — the most active nuclear construction... |
| ⚡ East Anglia Grid | 22.9% | 91% | 3 | 22.9% of the East Anglia corridor shows anomalous vegetation change between summ... |
Satellite-based monitoring across the UK nuclear and energy estate could deliver an estimated £500k–£1.2m annual savings while increasing monitoring frequency from quarterly to every 12 days — a 7.5x improvement in temporal resolution at a 60–85% cost reduction.
Sentinel-2 optical and Sentinel-1 SAR satellites capture the UK every 5–12 days at 10m resolution. Free, open data from the EU Copernicus programme.
Prithvi-EO-2.0 foundation model (600M parameters, trained on 4.2M satellite samples) processes imagery. Same-season comparison eliminates seasonal noise. Only genuine year-on-year change is flagged.
Each monitoring zone receives a traffic-light risk rating based on the proportion and magnitude of anomalous change. RED zones get priority inspection timelines.
Site managers receive specific recommendations: which zones to inspect, when, and why. Exportable reports for procurement and compliance documentation.
600M parameter geospatial AI model pre-trained on 4.2 million satellite time-series samples. Fine-tunable for site-specific detection tasks.
IBM’s open-source framework for geospatial foundation models. Standardised training, inference, and evaluation pipelines built on PyTorch Lightning.
Free, open satellite data. Sentinel-2 provides optical imagery (10m, 13 bands). Sentinel-1 SAR provides all-weather radar. Combined for multi-sensor analysis.
Meno is designed to integrate into existing infrastructure asset management workflows, not replace them. Satellite intelligence augments manual inspection by providing continuous screening between site visits, directing ground teams to where they're needed most.
Sentinel satellites acquire imagery every 12 days. No tasking required — data is freely available from the EU Copernicus programme.
Same-season NDVI comparison runs automatically when new imagery arrives. Anomalous change is flagged and scored per zone.
RED zones trigger immediate notification to site management. AMBER zones are queued for scheduled review. GREEN zones confirm stability.
Ground inspection teams are directed to specific zones with satellite evidence. No more blanket surveys — every visit has a purpose.
Field teams confirm or dismiss satellite findings. Results feed back to improve detection thresholds. Continuous learning loop.
REST API delivers risk scores, change metrics, and zone-level findings directly to Maximo, SAP PM, Ellipse, or custom asset databases. Each 12-day cycle updates condition records automatically.
GeoTIFF outputs slot into ArcGIS, QGIS, or any OGC-compliant platform. Change maps overlay on existing site drawings and infrastructure plans. WMS/WMTS tile serving available.
Configurable alert rules per zone and risk level. RED alerts push to email, SMS, or Teams/Slack. AMBER alerts roll into weekly digests. Integrates with ServiceNow or PagerDuty.
Automated PDF reports generated per analysis cycle. Audit-ready output with full methodology documentation. Designed for AS9100, ONR, and environmental compliance frameworks.
All satellite source data is freely available under the EU Copernicus Open Access licence. Processed outputs and analysis results are owned by the client. The platform can be deployed on-premises or in a client-managed cloud environment for data sovereignty requirements. No third-party data dependencies beyond the open Sentinel programme.
Current capabilities use Sentinel-2 optical data. Phase 2 expands the sensor suite and analysis depth:
Sentinel-1 interferometric SAR measures ground movement to millimetre precision. Detects subsidence, structural settlement, and ground instability around nuclear facilities.
Landsat 8/9 thermal bands detect surface temperature variations. Cooling water plumes, equipment heat signatures, and operational status visible from space.
Landsat archive back to 1985 enables 40-year change analysis. Track decommissioning evolution across decades, not just months.
SAR-based flood extent mapping and coastal erosion monitoring around nuclear sites. Critical for climate resilience assessments required by ONR.
This prototype demonstrates satellite AI capability against real UK infrastructure sites and active procurement pipelines. A Discovery Phase engagement would validate the platform against your specific portfolio — tailored site selection, ground-truth calibration, and integration with existing asset management systems.