3BrainAIHistory and coverage
Repeated observations and existing archives put many locations into context, without commissioning a separate flight for each one.
SATELLITE ANALYSIS
See the change. Understand the evidence.
Satellite time series, geodata and analytical methods put change into context.
A location through time
Optical observations
t1
t2
t3
t4
t5
t6
Radar · a complementary signal
r1
r2
r3
r4
r5
r6
A closer look at one source
COPERNICUS SENTINEL · EUROPEAN EARTH OBSERVATION PROGRAMME
01 Context
152 × 105 px
02 Selected area
40 × 40 px
03 Pixels
8 × 8 px
Munich–Neufreimann · 15 Aug 2026 · Sentinel-2 · TCI source preview, not unprocessed sensor data.
Contains modified Copernicus Sentinel data (2026). Earth Search / AWS Sentinel COGs. Source and limits →
Why satellites matter.
Beyond visible colour
Optical bands and radar measure different properties. Time series can reveal patterns that a single photograph cannot show.
The right scale
Area-level change can be measurable while small details remain unresolved. Resolution must match the review question.
Drones, aerial imagery, VHR and LiDAR add local detail and geometry where needed. Satellites provide repeated measurements across large areas. The task determines the useful combination.
What does that mean for resolution?
Sentinel-2 measures 13 spectral bands at resolutions of 10, 20 or 60 metres. These measurements answer different questions from a close-up of individual building components. Enlarging an image adds no new detail.
CRI DEVELOPMENT PRINCIPLE
Multiple inputs. A traceable review.
- 01
Connect the inputs
Satellite time seriesGeodata and project recordsWhere needed: VHR, LiDAR, aerial or drone data - 02
Prepare and analyse
Check quality and alignment. Apply suitable specialist methods and AI models. - 03
Check for the task
Compare with independent reference evidence. Record limits and review with specialists.
Evidence Pack: sources, findings, limits and next questions. When evidence is insufficient, the question remains open.
We are developing CRI around the bank’s review question, combining specialist geospatial analysis and AI with project records. Our focus is source suitability, analytical validation and the Evidence Pack workflow. The institution retains the decision.
Development principle, not evidence of a fully implemented or validated CRI analysis.
Explore an example from research
AN EXAMPLE FROM EXTERNAL RESEARCH
A parking lot can reveal a pattern.
In DLR’s Sat4Ec research, repeated Sentinel-1 radar measurements of automotive parking areas were used to derive an indicator of production activity. Individual cars did not need to be resolved in the image. The indicator was compared with production statistics; performance varied by manufacturer.
The useful information came from the changing signal and its validation, not from a sharper photograph.
Read the DLR case on CDSE ↗External research example. This is not a CRI result or a demonstrated CRI capability.
A EUROPEAN DATA AND RESEARCH ECOSYSTEM
Satellite data. European infrastructure.
UP TO30 TBof Earth observation data per day1
Satellite imagery and measurements from Copernicus.
CDSE
Copernicus Data Space Ecosystem
European Union · distributed infrastructure
DLR
German Aerospace Center
CDSE infrastructure
- Data and services in the EU
- Geographically separated data copies
- Data-product integrity checks
These statements concern CDSE. They do not describe the entire CRI infrastructure or imply a partnership with the named organisations.
SECURITY REQUIREMENTS FOR CRI PILOTS
Controlled access. Traceable processing.
Controls to agree and verify with each institution before a pilot.
Data boundaries
Separate bank data. Role-based access and encryption.
Processing trail
Document sources, transformations, versions and retention.
Independent review
Arrange a scoped external assessment of security and AI governance.
The intended EU processing locations and providers are to be documented for the agreed pilot.
Assessment references
- ISO/IEC 27001 Information security
- ISO/IEC 27017 Cloud security controls
- ISO/IEC 42001 AI management
Proposed assessment references, not certifications held by 3BrainAI. The scope remains subject to agreement.
What CRI demonstrates today.
The Brief is available now. A separate CRI demonstrator is in final development, with a demonstration planned for October 2026. Institution-specific configuration, calibration and evaluation require an agreed scope and responsible human review. The public Evidence Pack illustrates the review record and its boundaries; this page explains the analytical principle.
Who develops CRI and what we are preparing.
3BrainAI’s internal team is responsible for the technical infrastructure, development of the product core and its integration with bank processes.
For the satellite component, we are preparing data acquisition, combination and preprocessing, together with the subsequent analytical methods. We are assessing which available components can be usefully adopted and adapted, and which parts require our own development. We are refining the division of work between the internal team and external specialists according to technical and economic requirements, the demonstrator and planned bank pilots.
For each published case, we identify the actual data source, access route and processing applied.
Start with the review record.
See how a prepared example brings together dated observations, sources, limits and the next review question.