top of page
Screenshot from 2026-03-09 21-22-10.png

Remote Sensing · Machine Learning · Geography

Discovering the world's trees

We combine satellite imagery and deep learning to map trees — their location, cover, biomass, and canopy height — from local to continental scale.

PLANETSCOPE ·  SENTINEL-2  ·  GAOFEN  ·  LANDSAT ·  SKYSAT

THE TEAM

International researchers, one shared vision

Based at the University of Copenhagen and funded by DNRF, ESA, ERC, DFF and Schmidt Sciences. We collaborate with institutions across Europe, Africa and Asia.

WhatsApp Image 2025-04-18 at 9.25.24 PM.jpeg

KEY PUBLICATIONS

Screenshot from 2026-03-09 21-10-17.png

OUR VISION

01

HIGHLY DETAILED

We detect and map trees as objects.

​

​

02

HOMOGENEOUS

Multi-sensor fusion ensures temporal and spatial consistency.

​

03

LARGE SCALE

Our methods scale seamlessly from a single farm to continental and global extents.

04

LONG TERM

Time series analysis going back 25 years to track change over decades.

​​

HOW WE WORK

From pixels to trees

Our pipeline fuses multi-resolution imagery with deep learning to extract tree related information at scales that were previously impossible.

  • PlanetScope, Sentinel-2, Landsat, Skysat, Maxar, GaoFen imagery​

​

  • Deep learning object detection tools

​

  • Tree location, cover, biomass, canopy height outputs

​

  • Validation against ground truth and lidar reference data

2026-03-30_12-26.png

GEOGRAPHIC COVERAGE

Our published studies span every major inhabited continent — combining the same methods across dramatically different environments, from hyper-arid drylands to tropical forests and dense European farmland.

2026 Martin Brandt

bottom of page