Toward Consistent Canopy Characterization From Low‐Cost UAV Imagery
Abstract
Monitoring tropical forest structure at landscape scale requires cost-effective methods capable of bridging the gap between field inventories and satellite remote sensing products. However, the diversity of available individual tree crown (ITC) segmentation algorithms raises questions about the consistency and reliability of derived structural metrics-a critical issue for any ecological application relying on these outputs. We evaluated three ITC segmentation algorithms-Detectree2, SAM, and the hybrid Detectree2SAM (D2S)-applied to very high-resolution RGB orthomosaics (5 cm resolution) acquired over ~200 ha of the Luki Biosphere Reserve (Democratic Republic of Congo) using a low-cost drone. Algorithms were validated at the individual scale against a photointerpretation reference of 1882 manually delineated crowns and at the plot scale against a field inventory of 360 trees across 18 plots. At the individual scale, IoU-based F1 scores ranged from 0.57 (D2S) to 0.67 (SAM), revealing clear precision-recall trade-offs. At the plot scale, D2S provided the most accurate crown area estimates (RMSD = 26%-29%), while SAM best reproduced canopy density (RMSD = 22%). All three algorithms systematically overestimated aggregated metrics; a linear correction substantially reduced prediction errors and improved cross-algorithm convergence, yielding consistent estimates-median crown area of the 20 largest trees (~148 m 2 ), total crown area (~3850 m 2 per plot), and canopy density (~80 ind/ha). Agreement maps revealed systematic spatial divergences, including edge artifacts, SAM's tendency to underestimate canopy density, and localized overestimation of crown areas in disturbed stands. Raw ITC outputs carry substantial systematic biases requiring explicit field-calibrated correction before ecological use. We propose a three-step operational pipeline-(i) plot-scale field validation, (ii) per-algorithm linear bias correction, and (iii) landscapescale aggregation-that reduces inter-algorithm divergence and delivers consistent, ecologically interpretable structural estimates from low-cost UAV imagery.
Related
- Pre-season bare-soil spectral proxies complement pod-filling canopy indices for field-scale soybean-yield prediction from UAV imagery
- Monitoring seasonal and long-term forest canopy dynamics using Sentinel-2 imagery and Random Forest on the Google Earth Engine platform
- Effects of Shading-Induced Senescence and Retention of Lower-Canopy Needles on Photosynthetic Dynamics in Upper-Canopy Needles of Cunninghamia lanceolata
- GEDI canopy cover assessment using airborne LiDAR across eco-climate regions of the United States
- Retrieval of Soybean SPAD at Different Vertical Canopy Scales Based on Fractional-Order Derivatives and Spectral Indices
BibTeX
@article{plumacker2026toward,
title = {{Toward Consistent Canopy Characterization From Low‐Cost UAV Imagery}},
author = {Antoine Plumacker and Bhely Angoboy Ilondea and Nicolas Barbier and Thibauld Collet and Hugo de Lame and Pauline Depoortere and Alexia Favaro and Adeline Fayolle and Marjane Kaddouri and Carmel Mounziegou‐mounziegou and Pierre Ploton and Zoé Rousseau and Edmond Saint Denis and Paul Tresson and Hadrien Tulet and Arthur Vander Linden and Jean‐françois Bastin},
journal = {Ecology and Evolution},
year = {2026},
month = sep,
doi = {10.1002/ece3.74260},
url = {https://hal.inrae.fr/hal-05766644v1},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Toward Consistent Canopy Characterization From Low‐Cost UAV Imagery
AU - Antoine Plumacker
AU - Bhely Angoboy Ilondea
AU - Nicolas Barbier
AU - Thibauld Collet
AU - Hugo de Lame
AU - Pauline Depoortere
AU - Alexia Favaro
AU - Adeline Fayolle
AU - Marjane Kaddouri
AU - Carmel Mounziegou‐mounziegou
AU - Pierre Ploton
AU - Zoé Rousseau
AU - Edmond Saint Denis
AU - Paul Tresson
AU - Hadrien Tulet
AU - Arthur Vander Linden
AU - Jean‐françois Bastin
PY - 2026
DA - 2026/09/28
JO - Ecology and Evolution
DO - 10.1002/ece3.74260
UR - https://hal.inrae.fr/hal-05766644v1
AB - Monitoring tropical forest structure at landscape scale requires cost-effective methods capable of bridging the gap between field inventories and satellite remote sensing products. However, the diversity of available individual tree crown (ITC) segmentation algorithms raises questions about the consistency and reliability of derived structural metrics-a critical issue for any ecological application relying on these outputs. We evaluated three ITC segmentation algorithms-Detectree2, SAM, and the hybrid Detectree2SAM (D2S)-applied to very high-resolution RGB orthomosaics (5 cm resolution) acquired over ~200 ha of the Luki Biosphere Reserve (Democratic Republic of Congo) using a low-cost drone. Algorithms were validated at the individual scale against a photointerpretation reference of 1882 manually delineated crowns and at the plot scale against a field inventory of 360 trees across 18 plots. At the individual scale, IoU-based F1 scores ranged from 0.57 (D2S) to 0.67 (SAM), revealing clear precision-recall trade-offs. At the plot scale, D2S provided the most accurate crown area estimates (RMSD = 26%-29%), while SAM best reproduced canopy density (RMSD = 22%). All three algorithms systematically overestimated aggregated metrics; a linear correction substantially reduced prediction errors and improved cross-algorithm convergence, yielding consistent estimates-median crown area of the 20 largest trees (~148 m 2 ), total crown area (~3850 m 2 per plot), and canopy density (~80 ind/ha). Agreement maps revealed systematic spatial divergences, including edge artifacts, SAM's tendency to underestimate canopy density, and localized overestimation of crown areas in disturbed stands. Raw ITC outputs carry substantial systematic biases requiring explicit field-calibrated correction before ecological use. We propose a three-step operational pipeline-(i) plot-scale field validation, (ii) per-algorithm linear bias correction, and (iii) landscapescale aggregation-that reduces inter-algorithm divergence and delivers consistent, ecologically interpretable structural estimates from low-cost UAV imagery.
SN - 2045-7758
ER -