Biology subjects

Ball, J. G. C.

Publications and source records attributed to Ball, J. G. C..

7 recordsLinked to original sources

bioRxiv2026

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.

ecology↗

bioRxiv2026

Phenological regularity, not functional traits, determines whether tropical tree species can be mapped from imaging spectroscopy

AO_SCPLOWBSTRACTC_SCPLOWO_LIAirborne imaging spectroscopy enables species-level classification in hyperdiverse tropical forests, but accuracy varies enormously among species. We asked which ecological and evolutionary attributes make a tropical tree species spectrally separable. C_LIO_LIUsing 3,256 field-verified crowns spanning 169 species in a hyperdiverse moist forest in French Guiana, we tested seven hypothesised determinants of classification accuracy at species, pairwise, and individual-crown scales using random forest, beta regression, elastic net, and binomial GLMM analyses. C_LIO_LIPhenological regularity - the strength and consistency of seasonal leaf-cycling - was the single strongest predictor of separability, emerging as the top-ranked variable across all analyses. The presence of congeneric species in the classification pool also reduced accuracy, while broader phylogenetic isolation contributed in multivariate models. At the crown level, crown area was the strongest predictor of correct classification, while liana infestation reduced odds of correct identification by 38%. Leaf chemical traits did not predict separability. C_LIO_LIIt is the consistency of a species ecological signal - its phenological rhythm, spatial sampling, and freedom from canopy contamination - rather than any single functional trait, that determines whether it can be reliably mapped from imaging spectroscopy. C_LI

ecology↗

bioRxiv2026

Geospatial foundation models enable data-efficient tree species mapping in temperate montane forests

Accurate mapping of tree species from satellite data remains challenging in heterogeneous mountain forests due to environmental gradients, mixed stands, limited availability of high-purity training labels, and strong illumination-angle effects. Recent geospatial foundation models offer a new approach by learning generic, cloud-agnostic, information-rich representations from large multi-sensor archives suitable for a range of downstream tasks, but their ecological utility for species-level mapping remains incompletely understood. Here, we evaluate two geospatial foundation-model embeddings, AlphaEarth and Tessera, for tree species classification in the Trentino region of northern Italy, using parcel-level forest inventories as reference data (18 species and species groups). We compare their performance against conventional Sentinel-1+2 satellite composites across a series of controlled experiments examining classification accuracy, label efficiency, classifier complexity, robustness to label impurity, and temporal transferability. Foundation-model embeddings consistently outperform composite-based multispectral satellite baselines (weighted F1 = 0.83 vs. 0.80; macro F1 = 0.55 vs. 0.50), reaching near-asymptotic accuracy with as few as 5% of available training parcels and preserving ecologically meaningful structure aligned with functional and taxonomic groupings. However, realising this advantage requires a nonlinear classifier: a compact neural network provides better results than classic machine learning (i.e. Random Forest) and performs as well as deeper neural networks, while a linear classifier on foundation-model embeddings underperforms a neural network on conventional composites. Ancillary environmental covariates offer no additional classification benefit when added to embedding-based models. Classification accuracy remains robust to moderate levels of label impurity, allowing mixed parcels to be retained in the training dataset without substantial penalties, while training with parcel-level species proportions as soft labels achieves higher peak performance (macro F1 = 0.586 for Tessera, 0.589 for AlphaEarth) and lower Proportion L1 error than hard labels without requiring purity filtering, maximising the value of the full range of input data. However, temporal transfer across years reveals performance degradation, with weighted F1 declining by 9% for Tessera and 15% for AlphaEarth, and disproportionate losses for rare species. Overall, our results show that geospatial foundation models shift a primary bottleneck in species mapping from feature engineering toward the availability, quality, and temporal alignment of ecological reference data, while opening new opportunities for scalable biodiversity monitoring and the analysis of ecological change.

ecology↗

bioRxiv2024

Harnessing temporal and spectral dimensionality to map and identify species of individual trees in diverse tropical forests

Species-level canopy maps underpin tropical forest biodiversity monitoring, conservation planning, and carbon accounting, yet high species richness and structural complexity make remote classification challenging. Here we evaluate how far a two-step mapping approach can be taken, and where it breaks down, in hyperdiverse moist forest at the Paracou Field Station, French Guiana. First, we delineate individual tree crowns from ten repeat uncrewed aerial vehicle (UAV) RGB surveys with Mask R-CNN, fusing predictions across dates by temporal consensus: mean segmentation F1 rose from 0.68 (single date) to 0.78 (ten dates), covering approximately 86% of test-region canopy area. Second, we classify each crown from a single airborne hyperspectral acquisition (416-2500 nm, 1 m) using machine learning classifiers trained and tested on 3,186 field-verified crowns spanning 169 species. Linear Discriminant Analysis performed best (weighted F1 = 0.75), outperforming more flexible models, but unevenly: across repeated cross-validation (20 x 5-fold), on average 50 species (95% CI: 41-63) attained F1 [≥] 0.7 in a given fold and only 15 did so reliably, with many rare species unclassifiable (macro-average F1 = 0.48). Combining both steps, we estimate approximately 70% of the landscapes canopy area was correctly mapped to species. Band-importance and ablation analyses identified the far-red edge (748-775 nm) as the most informative spectral region. These results advance on studies limited to 20 or fewer species and set out spectral-resolution and training-data requirements for airborne and forthcoming spaceborne imaging spectrometers, while showing that accuracy remains strongly conditioned by training-data availability and the single-site, single-acquisition design.

ecology↗

bioRxiv2022

Tree segmentation in airborne laser scanning data is only accurate for canopy trees

Individual tree segmentation from airborne laser scanning data is a longstanding and important challenge in forest remote sensing. There are a number of segmentation algorithms but robust intercomparison studies are rare due to the difficulty of obtaining reliable reference data. Here we provide a benchmark data set for temperate and tropical broadleaf forests generated from labelled terrestrial laser scanning data. We compare the performance of four widely used tree segmentation algorithms against this benchmark data set. All algorithms achieved reasonable accuracy for the canopy trees, but very low accuracy for the understory trees. The point cloud based algorithm AMS3D (Adaptive Mean Shift 3D) had the highest overall accuracy, closely followed by the 2D raster based region growing algorithm Dalponte2016+. This result was consistent across both forest types. This study emphasises the need to assess tree segmentation algorithms directly using benchmark data. We provide the first openly available benchmark data set for tropical forests and we hope future studies will extend this work to other regions.

ecology↗

bioRxiv2022

Accurate tropical forest individual tree crown delineation from RGB imagery using Mask R-CNN

Tropical forests are a major component of the global carbon cycle and home to two-thirds of terrestrial species. Upper-canopy trees store the majority of forest carbon and can be vulnerable to drought events and storms. Monitoring their growth and mortality is essential to understanding forest resilience to climate change, but in the context of forest carbon storage, large trees are underrepresented in traditional field surveys, so estimates are poorly constrained. Aerial photographs provide spectral and textural information to discriminate between tree crowns in diverse, complex tropical canopies, potentially opening the door to landscape monitoring of large trees. Here we describe a new deep convolutional neural network method, Detectree2, which builds on the Mask R-CNN computer vision framework to recognise the irregular edges of individual tree crowns from airborne RGB imagery. We trained and evaluated this model with 3,797 manually delineated tree crowns at three sites in Malaysian Borneo and one site in French Guiana. As an example application, we combined the delineations with repeat lidar surveys (taken between 3 and 6 years apart) of the four sites to estimate the growth and mortality of upper-canopy trees. Detectree2 delineated 65,000 upper-canopy trees across 14 km2 of aerial images. The skill of the automatic method in delineating unseen test trees was good (F1 score = 0.64) and for the tallest category of trees was excellent (F1 score = 0.74). As predicted from previous field studies, we found that growth rate declined with tree height and tall trees had higher mortality rates than intermediate-size trees. Our approach demonstrates that deep learning methods can automatically segment trees in widely accessible RGB imagery. This tool (provided as an open-source Python package) has many potential applications in forest ecology and conservation, from estimating carbon stocks to monitoring forest phenology and restoration. Python package available to install at https://github.com/PatBall1/Detectree2

ecology↗

bioRxiv2021

Using deep convolutional neural networks to forecast spatial patterns of Amazonian deforestation

AO_SCPLOWBSTRACTC_SCPLOWO_LITropical forests are subject to diverse deforestation pressures but their conservation is essential to achieve global climate goals. Predicting the location of deforestation is challenging due to the complexity of the natural and human systems involved but accurate and timely forecasts could enable effective planning and on-the-ground enforcement practices to curb deforestation rates. New computer vision technologies based on deep learning can be applied to the increasing volume of Earth observation data to generate novel insights and make predictions with unprecedented accuracy. C_LIO_LIHere, we demonstrate the ability of deep convolutional neural networks to learn spatiotemporal patterns of deforestation from a limited set of freely available global data layers, including multispectral satellite imagery, the Hansen maps of historic deforestation (2001-2020) and the ALOS JAXA digital surface model, to forecast future deforestation (2021). We designed four original deep learning model architectures, based on 2D Convolutional Neural Networks (2DCNN), 3D Convolutional Neural Networks (3DCNN), and Long Short-Term Memory (LSTM) Recurrent Neural Networks (RNN) to produce spatial maps that indicate the risk to each forested pixel (~30 m) in the landscape of becoming deforested within the next year. They were trained and tested on data from two ~80,000 km2 tropical forest regions in the Southern Peruvian Amazon. C_LIO_LIWe found that the networks could predict the likely location of future deforestation to a high degree of accuracy. Our best performing model - a 3DCNN - had the highest pixel-wise accuracy (80-90%) when validated on 2020 deforestation based 2014-2019 training. Visual examination of the forecasts indicated that the 3DCNN network could automatically discern the drivers of forest loss from the input data. For example, pixels around new access routes (e.g. roads) were assigned high risk whereas this was not the case for recent, concentrated natural loss events (e.g. remote landslides). C_LIO_LICNNs can harness limited time-series data to predict near-future deforestation patterns, an important step in using the growing volume of satellite remote sensing data to curb global deforestation. The modelling framework can be readily applied to any tropical forest location and used by governments and conservation organisations to prevent deforestation and plan protected areas. C_LI

ecology↗