MillionTrees: A global benchmark for tree detection across geography, sensor, and annotation geometry

Airborne tree monitoring using remotely sensed data, combined with the growth of computer vision models for object detection and segmentation, can facilitate broad scale assessment of tree health, ecology and ecosystem function. However, datasets for airborne tree detection have been largely limited to a single geography, resolution, or annotation approach, making it difficult to develop general models and benchmark their utility for broad applications. We compile existing and newly published annotation datasets to create a comprehensive dataset for training and evaluating airborne tree detection in RGB imagery. The MillionTrees dataset has over a million individual trees from a broad range of forest types, tree species, and geographic locations. In addition, it has six million weakly supervised annotations from a range of sources including airborne LiDAR and photogrammetry, arborist surveys, and simulated landscapes. An evaluation standard is provided to facilitate rapid model development. We benchmark several existing tree detection algorithms and demonstrate that while current open-source models do not generalize across the diverse datasets in the benchmark, training on MillionTrees improves both within-domain and out-of-domain performance. Held-out validation datasets using ground truth derived from terrestrial LiDAR paired with overhead RGB show that overfitting to test data is common and hyperparameter tuning is an underappreciated challenge towards a global zero shot tree detection method. We discuss the limitations of the currently available data for airborne RGB tree detection, how to create more robust evaluation data and the need for better understanding of the uncertainty and utility of tree detection from airborne remote sensing.

Authors

Institutions

Publication Details

Journal
bioRxiv (Cold Spring Harbor Laboratory)
Published
2026-11-02
DOI
https://doi.org/10.64898/2026.10.30.762845
Primary Topic
Remote Sensing and LiDAR Applications
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
preprint

MillionTrees: A global benchmark for tree detection across geography, sensor, and annotation geometry

Taimur Khan, Henry Senyondo, David J. Russell, Stuart Grieve et al.
2 links
bioRxiv (Cold Spring Harbor Laboratory)
Remote Sensing and LiDAR Applications
preprint

MillionTrees: A global benchmark for tree detection across geography, sensor, and annotation geometry

Taimur Khan, Henry Senyondo, David J. Russell, Stuart Grieve, Myriam Cloutier, Camille C. Pawlak, Teja Kattenborn, Blas Mola‐Yudego, Moritz Lucas, Jonathan Ventura, Adison Altamirano, Oliver Schütte, John Brandt, Ryoungseob Kwon, Derek J. N. Young, Julian Frey, Sarah Graves, Sarah Anacleto, Maurice Mugabowindekwe, Harry Jon Foord Owen, Qinchuan Xin, Geoffrey A. Fricker, Rasmus Astrup, Kyaw Kyaw Htoo, Alejandro Miranda, Ryuichi Takeshige, Josh Veitch-Michaelis, Sebastian H. M. Hickman, Stefano Puliti, Sander de Haas, Sara Beery, Emily Rebecca Lines, Pedro Zamboni, Ethan P. White, Yihang She, Adnan Firoze, Arthur Ouaknine, Stephanie Bohlman, Dimitri Gominski, Étienne Laliberté, James Ball, Yusuke Onoda, Toby Jackson, Ben G. Weinstein, Jonas Troles, Masanori Onishi, Vicente Vásquez, Hugo Baudchon, Jessica Ertel, Qiue Xu, Simon-Olivier Duguay, Benjamin Wagner, Zhengpeng Feng, Matt J. Allen, Jaylen Nicholson, Md. Farhadur Rahman, Andrew Jansen, Isabelle Lefebvre, Ying Sun
preprint en

Abstract

Airborne tree monitoring using remotely sensed data, combined with the growth of computer vision models for object detection and segmentation, can facilitate broad scale assessment of tree health, ecology and ecosystem function. However, datasets for airborne tree detection have been largely limited to a single geography, resolution, or annotation approach, making it difficult to develop general models and benchmark their utility for broad applications. We compile existing and newly published annotation datasets to create a comprehensive dataset for training and evaluating airborne tree detection in RGB imagery. The MillionTrees dataset has over a million individual trees from a broad range of forest types, tree species, and geographic locations. In addition, it has six million weakly supervised annotations from a range of sources including airborne LiDAR and photogrammetry, arborist surveys, and simulated landscapes. An evaluation standard is provided to facilitate rapid model development. We benchmark several existing tree detection algorithms and demonstrate that while current open-source models do not generalize across the diverse datasets in the benchmark, training on MillionTrees improves both within-domain and out-of-domain performance. Held-out validation datasets using ground truth derived from terrestrial LiDAR paired with overhead RGB show that overfitting to test data is common and hyperparameter tuning is an underappreciated challenge towards a global zero shot tree detection method. We discuss the limitations of the currently available data for airborne RGB tree detection, how to create more robust evaluation data and the need for better understanding of the uncertainty and utility of tree detection from airborne remote sensing.

bioRxiv (Cold Spring Harbor Laboratory)
Australian National University (AU), Universidad de La Frontera (CL), University of Copenhagen (DK), World Resources Institute (US), University of Wisconsin–Madison (US), Helmholtz Centre for Environmental Research (DE), Seoul National University (KR), California Polytechnic State University (US), Sun Yat-sen University (CN), University of Freiburg (DE), University of California, Los Angeles (US), The University of Melbourne (AU), Queen Mary University of London (GB), Osnabrück University (DE), University of Eastern Finland (FI), Chinese Academy of Sciences (CN), Purdue University West Lafayette (US), Kyoto University (JP), University of Cambridge (GB), University of Florida (US), University of Bristol (GB), Institute of Subtropical Agriculture (CN), Norwegian Institute of Bioeconomy Research (NO), Conservation Leadership Programme (GB), Mila - Quebec Artificial Intelligence Institute (CA), Massachusetts Institute of Technology (US), Université de Montréal (CA), Technische Universität Dresden (DE), University of California, Davis (US), University of Bamberg (DE)
Life on land
Remote Sensing and LiDAR Applications
DOIOpen Access PDFOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon