Vink Intelligence

Services

GeoAI and spatial intelligence consulting

We combine geospatial data, satellite and aerial imagery and machine learning to detect physical assets, model spatial patterns and support infrastructure and planning decisions. Estimates come with their uncertainty and are placed on the map, so they can be checked against reality. When the problem needs more than a map, we connect the results to knowledge graphs and predictive models.

Applied across our ventures Vink Energy, Vink Farming and Vink Timber. See GeoAI for energy infrastructure →

When GeoAI helps

  • Location changes the answer.
  • The data arrives as imagery, maps or sensor networks.
  • Records of physical assets are incomplete or out of date.
  • Predictions need to be placed on the ground, not just in a table.

What we build

01

Earth observation and remote sensing analysis

Turning satellite and aerial imagery into measurements and signals.

02

Asset detection and estimation

Finding physical assets in imagery and estimating their properties, with the uncertainty stated.

03

Spatial modelling

Models that account for where things are and how they relate in space.

04

Geodata pipelines

Combining spatial data with weather, network, parcel or customer data.

05

Map interfaces

Ways to explore results spatially and check them against reality.

Use cases

Energy infrastructure

Estimate distributed assets such as rooftop solar from imagery and combine them with weather and network context, as in Grid Intelligence from Vink Energy.

Preparing grid reinforcement

Identify routes, constraints and obstacles from spatial and imagery data before crews go into the field, as in Grid Build Intelligence.

Crop development and yield

Use Earth observation, parcel and weather data for earlier signals on crop development and yield, as in Vink Farming.

Forest and timber

Use spatial forest data for standing volume, harvest planning and supply questions, as researched in Vink Timber.

Buildings and physical assets

Estimate where drone cleaning is economically attractive for a building, as in FlyingVink.

How it works

  1. 01

    Bring the data together

    Combine imagery, Earth observation, maps and public spatial information such as the BAG with your weather, network, parcel or customer data.

  2. 02

    Detect and estimate

    Detect assets in imagery and estimate their properties, stating the uncertainty.

  3. 03

    Model spatially

    Use models that account for where things are and how they relate in space, and predict what happens next.

  4. 04

    Check against reality

    Show results on a map and compare them with known records and field observations.

Every project follows the same four steps. Understand → Build → Validate → Deploy. About our way of working →

Where we apply it

Pilot

Vink Energy

Geospatial data, imagery and network information for energy infrastructure visibility and planning.

See Vink Energy ↗

Pilot

Vink Farming

Earth observation and parcel data for earlier crop and yield signals.

See Vink Farming ↗

Research

Vink Timber

Spatial forest data for standing volume, harvest and supply questions.

See Vink Timber ↗

Frequently asked questions

What is GeoAI?

GeoAI applies machine learning, especially computer vision and spatial modelling, to geospatial data such as satellite and aerial imagery, maps and sensor networks. It is used to detect and measure things on the ground and to predict spatial patterns.

Which geospatial data sources can be used?

Satellite and aerial imagery, Earth observation data, public spatial information such as the BAG, weather data, and network, parcel or customer data that you provide. Each engagement uses the layers that are relevant and available.

Can GeoAI analyse satellite and aerial imagery?

Yes. Typical tasks are detecting assets such as rooftop solar, measuring objects and turning imagery into signals. Estimates from imagery are approximate, so we state the uncertainty and check results against known records.

What business problems are suitable for GeoAI?

Problems where location changes the answer and records of physical assets are incomplete or out of date, for example energy infrastructure planning, crop and yield forecasting or forestry supply. If location does not matter, a general machine learning approach is usually enough.

Can GeoAI be combined with existing operational data?

Yes, and it usually should be. Spatial estimates become more useful when combined with weather, network, parcel or customer data. A bounded pilot compares the result with how the work is done today.

Does location change your problem?