Global Initiative
Global climate change is affecting forest ecosystems at an accelerating rate, putting pressure on forest resilience and biodiversity. Monitoring tree species distribution at scale is critical to designing policies that protect ecosystem services and guide forest management under a changing climate.
TreeAI addresses this challenge by developing deep learning models for automatic tree species detection in high-resolution aerial and drone imagery. We are building the first open, community-driven global database of annotated tree crowns, providing the training foundation these models require. Contributions come from research groups worldwide and span five continents.
The outputs of this initiative, including the database, trained models, and benchmark evaluations, are released openly to support improved forest inventory, targeted management planning, and research on forest resilience. The goal is the first generalizable model for individual tree species identification that scales across geographic regions.
Team
Placeholder Lead Name
Principal Investigator
Institution, Department
Placeholder bio describing relevant expertise in forest remote sensing or machine learning for ecological applications.
contact@institution.eduPlaceholder Lead Name
Data Coordination Lead
Institution, Department
Placeholder bio describing relevant expertise in geospatial data management or biodiversity monitoring.
contact@institution.eduPlaceholder Lead Name
ML Research Lead
Institution, Department
Placeholder bio describing relevant expertise in computer vision or deep learning for species identification.
contact@institution.eduPartners
ETH Zurich
Switzerland
WSL Swiss Federal Institute for Forest, Snow and Landscape Research
Switzerland
Wuhan University
China
University College London
United Kingdom
Norwegian Institute of Bioeconomy Research (NIBIO)
Norway
University of Freiburg
Germany
University of Leipzig
Germany
University of Copenhagen
Denmark
Attribution
If you use the TreeAI dataset or any contributing sub-dataset in your research, please cite the following.
Full record, version history, and file downloads on Zenodo.
Collaborate
Interested in contributing but not yet ready to submit a dataset? We welcome early conversations, whether you are exploring eligibility, have questions about data format, or want to discuss a collaboration. Reach out and we will get back to you.
You can also help by sharing the data call with colleagues who work with high-resolution tree imagery.
Community
Research groups and individuals who have contributed annotated datasets to the initiative.
| Surname | First Name | Institute | Country | Research Area |
|---|