unthinkable
Case study · Green City Watch · geospatial urban ecology firm, TreeTect

City foresters in Amsterdam and Boston see where every tree stands and how green it is, from satellite images.

Green City Watch wanted municipalities and urban foresters to keep city trees alive longer, since urban trees live just 13 years against 100 or more in the countryside, by knowing their whole tree inventory.

What changed

90%precision (as reported)
11machine learning models trained (as reported)
30+cities mapped (as reported)
What stood in the way

What stood in the way.

What stood in the way
  • Foresters had no up-to-date inventory of each tree's location, crown area and greenness (NDVI).
  • Different species show up in different bands of 8-band satellite images, so a single model could not find them all.
What changed
  • 90%precision (as reported)
  • 11machine learning models trained (as reported)
  • 30+cities mapped (as reported)
What we built

What we built.

  • Object detection on GeoTIFF images from the WorldView-1 and WorldView-3 satellites at 30 cm resolution.
  • 11 machine learning models, each on a combination of 3 bands, combined in an ensemble to find trees, their species, coordinates and crown area.
  • NDVI calculation for each tree as a measure of greenness and growth.
  • A setup on AWS SageMaker, ECR, Lambda, SES and S3 storage, with QGIS, Leaflet and TensorFlow.
  • TensorFlow
  • QGIS
  • Leaflet
  • GBDX
  • AWS SageMaker
  • AWS Lambda
What it shows

What this work shows about us.

The method follows the data

Because species reflect differently in each band, several smaller models in an ensemble did what one model could not.

Built to move to new cities

Each new geography gets its own band combination on the same pipeline.

Want to build similar satellite image AI for mapping trees? Tell us what you want to change. Our AI assistant, Ask Unthinkable, helps you define your requirements, then connects you with the right team.

Start your plan ↗
If we built it today

If we built it today, with AI.

What we would add now, moment by moment.

Planning care

An agent would list the trees whose greenness is falling, with their location and history, and draft the inspection round for the forester. Foresters decide every treatment.

Answering residents

An agent would answer a resident's question about the trees on their street from the inventory, and log reports of damaged trees for the team. The council decides any action.

Related

Where this work fits.

    Facts and figures from the project records; nothing added.

    Ask Unthinkable

    Want this for your business?

    Tell us what you run or what you're building. Our AI assistant, Ask Unthinkable, helps you define your requirements, then connects you with a senior AI engineer.

    Send us a note instead

    Required

    Thanks. A senior AI engineer will reply to you.
    u
    Ask UnthinkableClarity first, then the right team
    Your conversation

    An AI agent. This prototype uses prepared answers and sends nothing.