Reliable Tree Recruitment from Scarce Annotations for Forest Inventory Expansion
Authors: Luiz Fernando Souza Luz, Pedro J. Soto Vega, Matheus Pinheiro Ferreira, Laura Elena Cué la Rosa, Hudson Franklin Pessoa Veras, Ernandes Macedo da Cunha Neto, Evandro Orfanó Figueiredo, Ana Paula Dalla Corte, Carlos Roberto Sanquetta, José Matheus Segre Moneva Viveiros, Matheus Santos Fuza, Heitor Eduardo Ferreira Campos Morato Filpi, Cristiano Vinícius Vidal, Vinícius Moura Costa, Débora Nascimento Rosário, Danilo Roberti Alves de Almeida, Paulo Guilherme Molin, Dário Augusto Borges Oliveira
Organizations: School of Applied Mathematics, Getulio Vargas Foundation, Rio de Janeiro, RJ, Brazil. · Department of Forest Sciences, “Luiz de Queiroz” College of Agriculture, University of São Paulo (USP/ESALQ), Piracicaba, SP, Brazil. · Laboratory of Geo-Information Science and Remote Sensing, Wageningen University and Research, Wageningen, The Netherlands. · Instituto Federal do Acre, Rio Branco, AC, Brazil. · Department of Forest Science, Federal University of Paraná (UFPR), Curitiba, PR, Brazil. · Embrapa Acre, Rio Branco, AC, Brazil. · Department of Forest Sciences, University of São Paulo (USP/ESALQ), Piracicaba, SP, Brazil. · Bioflore, Belo Horizonte, MG, Brazil. · Department of Agricultural Engineer, Federal University of Viçosa, Viçosa, Brazil. · Center for Natural Sciences, Universidade Federal de São Carlos, Buri, Brazil.
Species-level inventories of individual tree crowns (ITCs) from UAV imagery are limited by annotation cost: in tropical forests, many species are represented by only a handful of field-validated crowns. We propose a reliable tree recruitment framework that expands a small annotated inventory by iteratively self-training a multi-task DeepLabV3 network, combining semantic segmentation with a crown-distance regression head, on RGB UAV orthomosaics. At each iteration, candidate crowns enter the training inventory only if they pass a confidence filter on class probability and distance map and a morphology filter on crown area and solidity, with at most five new crowns per species, keeping recruited samples consistent with the reference inventory. We evaluate the framework at two Brazilian sites: an Amazonian rainforest in Acre (17 species, 355 annotated crowns, 4 cm GSD) and a forest site in the Cerrado of Maranhão (4 species, 51 annotated crowns, 1.8 cm GSD). Over 20 iterations, the number of retained crowns grew from about 500 to 960 in Acre and from about 700 to over 3100 in Maranhão, while crown-level F1 exceeded single- and multi-task DeepLabV3 baselines (53.5% vs. 50.0% and 45.2%; 92.2% vs. 79.7% and 83.6%). The framework offers a low-cost route to expanding species inventories with consumer-grade UAVs and limited fieldwork.
Figures & tables
Fig. 1 : Overview of the proposed framework. (1) Training the ML model with the labeled samples. (2) Predicted semantic segmentation and the distance map by the multi-task model. (3) Prediction refined with a confidence threshold above 70% and distance map value above 0.25. (4) Final Individual Tree Crowns (ITCs) selected and prepared for the tree recruitment scheme’s sample inventory, incorporating morphological validation.
Fig. 2 : Overview of the proposed multitask DeepLabV3 architecture. It features a shared encoder, built upon a ResNet-50 backbone, that branches into two specialized decoders: one for semantic segmentation (supervised by LPCFL ) and another for distance transform map prediction (supervised by LPMSE ).
Fig. 3 : Example of the distance transform generation. (a) RGB image. (b) Binary mask of the ITC. (c) Euclidean distance transform with Gaussian filter
Fig. 4 : Study areas location. (A) Embrapa experimental forest in Rio Branco, Acre (AC) state, Brazil. (B) Baixao Preto Farm in Riachao, Maranhao (MA) state, Brazil, showing 3 of the 15 surveyed regions.
Fig. 5 : Examples of individual tree crowns of the 17 target species in the Amazon AC dataset (UAV RGB imagery, 4 cm GSD).
Fig. 6 : Examples of individual tree crowns of the four target species in the Bioflore MA dataset (UAV RGB imagery, 1.8 cm GSD).
Fig. 7 : Examples of patches used in the training stage. Red polygons refer to individual tree crowns.
Fig. 8 : F1-score (%) for each specie in both datasets
Fig. 9 : Visual analysis of predicted ITCs by each evaluated method on Amazon AC dataset. In the figure, TL , and PL refer to True Label and Predicted Label, respectively.
Fig. 10 : Visual analysis of predicted ITCs by each evaluated method on Bioflore MA dataset. In the figure, TL , and PL refer to True Label and Predicted Label, respectively.
Fig. 11 : Illustration of component selection based on model predictions: (a) Predicted semantic segmentation by the model. (b) Prediction with confidence above 70%. (c) High-confidence prediction where the distance map value is above 0.25. (d) ITCs are selected to compose the tree recruitment scheme sample inventory.
Fig. 12 : Individual tree crowns (ITCs) of the target species (see Table I ) over the iterations of the tree recruitment process for Amazon AC dataset.
Fig. 13 : Comparison of the canopy area distribution of the species Abiorana rosa ( Micropholis venulosa ). Highlighted are the selection limits used in the first iteration of the model.
Fig. 14 : Number of individual tree crowns (ITCs) detected and macro-average F1-score across tree recruitment iterations.
Fig. 15 : Number of predicted individual tree crowns (ITCs) in the 1st, 8th, and 13th iterations of the tree recruitment cycle. A different bar color represents each species. The count of ITCs is displayed above each bar, with the corresponding F1-score in parentheses.
We present a highly detailed instance segmentation model for delineating individual tree crowns in natural broadleaf forests using aerial imagery acquired by unmanned aerial vehicles (UAVs). Tree crown delineation in broadleaf forests is more challenging than in other forest types due to diversity of crown shapes and the lack of clearly defined treetops. To address this issue, we developed a deep-learning-based crown segmentation model trained on high-quality annotated crown outlines. We manually delineated 18,507 crown polygons from orthomosaic images collected across seven forests in Japan by skilled annotators, and developed a model based on Mask2Former with multiple backbone architectures. The best model achieved high segmentation performance in structurally complex broadleaf forests using only RGB imagery. This performance was maintained when applied to geographically distinct forests within Japan, as well as to biologically distinct tropical rainforests in Borneo. These results demonstrate that using a large number of high-quality annotated datasets is critical for achieving detailed and generalizable crown segmentation across diverse forest ecosystems. The developed model has been integrated into DF Scanner Pro, a software that supports practical forest monitoring using UAVs, and this implementation is expected to enable a wide range of users to analyze tree-level information in broadleaf forest from UAVs.
Accurate classification of tropical tree species from unoccupied aerial vehicle (UAV) imagery remains challenging due to high species diversity and strong visual similarity among species at typical image resolutions (centimeters per pixel). In contrast, models trained on close-up citizen science photographs captured with smartphones achieve strong plant species classification performance. Recent advances in UAV data acquisition now enable the collection of close-up images that are spatially registered with top-view aerial imagery and approach the level of visual detail found in smartphone photographs, with the trade-off that such high-resolution photos cannot be acquired for many trees. In this work, we evaluate the performance of existing methods using paired top-view and close-up UAV imagery collected in a species-rich tropical forest. Through fine-tuning experiments, we quantify the performance gap between vision foundation models and in-domain generalist plant recognition models across both image types (high-resolution close-up versus coarser-resolution top-view imagery). We show that classification performance is consistently higher on close-up images than on top-view aerial imagery, and that this performance gap widens for rare species. Finally, we propose that self-supervised representation alignment across these two spatial scales offers a promising approach for integrating fine-grained visual information into canopy-level species classification models based on top-view UAV imagery. Leveraging high-resolution close-up UAV imagery to enhance canopy-level species classification could substantially improve large-scale monitoring of tropical forest biodiversity.
Sulagna Saha, Arthur Ouaknine, Etienne Laliberté +7
Mila – Quebec AI Institute · McGill University · Université de Montréal +4
Accurate estimation of forest biomass, a major carbon sink, relies heavily on tree-level traits such as height and species. Unoccupied Aerial Vehicles (UAVs) capturing high-resolution imagery from a single RGB camera offer a cost-effective and scalable approach for mapping and measuring individual trees. We introduce BIRCH-Trees, the first benchmark for individual tree height and species estimation from tree-centered UAV images, spanning three datasets: temperate forests, tropical forests, and boreal plantations. We also present DINOvTree, a unified approach using a Vision Foundation Model (VFM) backbone with task-specific heads for simultaneous height and species prediction. Through extensive evaluations on BIRCH-Trees, we compare DINOvTree against commonly used vision methods, including VFMs, as well as biological allometric equations. We find that DINOvTree achieves top overall results with accurate height predictions and competitive classification accuracy while using only 54% to 58% of the parameters of the second-best approach.
Jannik Endres, Etienne Laliberté, David Rolnick +1
Mila – Quebec AI Institute · Université de Montréal · McGill University