TreeDGS: Aerial Gaussian Splatting for Distant DBH Measurement

Belal Shaheen, Minh-Hieu Nguyen, Bach-Thuan Bui, Shubham, Tim Wu, Michael Fairley, Matthew David Zane, Michael Wu, James Tompkin, [2pt] Corresponding author: [email protected] · 2026

Read TreeDGS: Aerial Gaussian Splatting for Distant DBH Measurement by Belal Shaheen, Minh-Hieu Nguyen, Bach-Thuan Bui, Shubham, Tim Wu, Michael Fairley, Matthew David Zane, Michael Wu, James Tompkin, [2pt] Corresponding author: [email protected] (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-12823.txt.

Introduction

Belal Shaheen and a large team measure tree trunks from the air with Gaussian splatting. Foresters need DBH, the diameter of a trunk at breast height, for carbon and timber estimates. Aerial remote sensing surveys large areas efficiently. Accurate direct measurement of single objects in complex natural scenes stays difficult, though.

Learned radiance field methods, like NeRF and 3D Gaussian Splatting, have raised reconstruction fidelity. The team build TreeDGS, an aerial Gaussian splatting method for distant DBH measurement. Denser geometry from posed imagery could let drones measure trees without ground crews. That would scale forest inventories to whole landscapes. The work links 3D vision with climate relevant forestry.

Catalog record

CREATORBelal Shaheen, Minh-Hieu Nguyen, Bach-Thuan Bui, Shubham, Tim Wu, Michael Fairley, Matthew David Zane, Michael Wu, James Tompkin, [2pt] Corresponding author: [email protected]
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE46 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-12823

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

TreeDGS: Aerial Gaussian Splatting for Distant DBH Measurement Belal Shaheen, Minh-Hieu Nguyen, Bach-Thuan Bui, Shubham, Tim Wu, Michael Fairley, Matthew David Zane, Michael Wu, James Tompkin, [2pt] Corresponding author: [email protected] Abstract Aerial remote sensing enables efficient large-area surveying, but accurate direct object-level measurement remains difficult in complex natural scenes. Recent advancements in 3D vision, particularly learned radiance-field representations such as NeRF and 3D Gaussian Splatting, have begun to raise the ceiling on reconstruction fidelity and densifiable geometry from posed imagery. Nevertheless, direct aerial measurement of important natural attributes such as tree diameter at breast height (DBH) remains challenging. Trunks in aerial forest scans are distant and sparsely observed in image views: at typical operating altitudes, stems may span only a few pixels. With these constraints, conventional reconstruction methods leave breast-height trunk geometry weakly constrained. We present TreeDGS, an aerial image reconstruction method that leverages 3D Gaussian Splatting as a continuous, densifiable scene representation for trunk measurement. After SfM–MVS initialization and Gaussian optimization, we extract a dense point set from the Gaussian field using RaDe-GS’s depth-aware cumulative-opacity integration and associate each sample with a multi-view opacity reliability score. Then, we estimate DBH from trunk-isolated points using opacity-weighted solid-circle fitting. Evaluated on 10 plots with field-measured DBH, TreeDGS reaches 4.79,cm RMSE (about 2.6 pixels at this GSD) and outperforms a state-of-the-art LiDAR baseline (7.91,cm RMSE). This shows that TreeDGS can enable accurate, low-cost aerial DBH measurement (Fig. ). 1 Introduction Aerial remote sensing has quickly become a cornerstone of modern environmental monitoring for its ability to survey large areas rapidly and cost efficiently [ 1 , 7 , 36 ] . Yet, despite major gains in sensor resolution, platform stability, and reconstruction pipelines, extracting direct, object-level measurements from aerial imagery remains difficult in structurally complex natural scenes [ 8 , 15 ] . Forested environments are challenging: heterogeneous geometry, self-similar textures, frequent occlusion, and strong appearance variability across lighting conditions all weaken the visual cues that aerial reconstruction methods rely upon [ 15 ] . A prominent example is tree diameter at breast height (DBH)—a foundational forestry variable that aerial remote sensing struggles to recover reliably. Its importance for forestry cannot be understated: DBH measures an individual tree’s size and growth stage, and is a primary input to standard allometric models and reporting workflows that estimate wood volume, biomass, and carbon [ 2 , 17 , 5 , 18 ] . In turn, these allometric models underpin decision making used across forestry operations and policy, ranging from timber inventory and silvicultural planning to carbon offset quantification, wildfire risk mitigation, and long-term ecosystem management [ 41 , 3 ] . DBH estimation from above-canopy aerial sensing is fundamentally constrained by observability. In typical UAV image surveys flown at operational altitudes, stems are not “large objects.” In our data, captured at \sim 70 m above ground ( \mathrm{GSD}\approx 1.84 cm/px), stems in the evaluation plots span only \sim 15 pixels across per view on average (Tab. 1 ), and can be just a few pixels wide for smaller trees (Fig. 2 ). This pixel scarcity is compounded by view scarcity: the breast-height band ( h_{\mathrm{BH}}\approx 1.37 m in our field protocol) is frequently occluded by crowns, branches, and understory vegetation, so each stem may be seen only in a handful of oblique glimpses through noisy canopy gaps. Therefore, the core geometric challenge of measurement is recovering enough faithful breast-height trunk surface to support stable circle or cylinder fitting under a regime of few pixels, few views, and partial visibility. Active sensing can mitigate some of these constraints by providing explicit range measurements rather than relying on image texture. Terrestrial laser scanning (TLS), for example, is well known to support robust DBH measurements because it can produce dense, well-populated stem cross-sections [ 30 , 38 , 20 ] . However, the mostly downward-looking trajectories of airborne or UAV laser scanning do not typically yield equally dense returns on the trunk at breast height: even with high point density, trunk hits near h_{\mathrm{BH}} can be sparse, uneven, and contaminated by surrounding vegetation, complicating per-tree stem isolation and reliable fitting in cluttered stands [ 25 , 35 , 24 ] . These limitations help explain why aerial LiDAR has seen broader operational adoption for canopy structure and terrain products rather than for direct DBH measurement [ 13 , 31 , 39 , 47 , 6 ] . RGB-only UAV photogrammetry has nevertheless made substantial progress for other forest attributes, including canopy height models and structural mapping [ 8 , 15 , 45 , 44 ] . But DBH remains difficult from aerial imagery alone, even under careful survey design with aggressive overlap and strong oblique views [ 34 ] . In the pixel- and view-limited regime described above, conventional SfM–MVS pipelines often fail to recover a coherent breast-height surface. Sparse SfM relies on repeatable keypoints and long feature tracks [ 40 ] , but distant trunks provide weak texture and are frequently interrupted by foliage occlusions, yielding unstable or missing correspondences. View-dependent MVS densification [ 9 ] then tends to drop out on trunks when visibility is intermittent and texture cues are weak. Figure 2 : Pixel-limited trunk observations from distant aerial imagery. At \sim 70 m altitude ( \mathrm{GSD}\approx 1.84 cm/px), a typical pine stem can occupy only \sim 13 pixels across in a single RGB image, making per-image diameter cues highly quantized, sensitive to occlusion, and difficult to measure precisely. Recent radiance-field representations, such as Neural Radiance Fields (NeRF) [ 33 ] and 3D Gaussian Splatting (3DGS) [ 22 ] , can improve reconstruction fidelity from posed imagery and offer a promising path forward for thin structures that are only weakly observed in individual frames. Early forestry applications have begun exploring radiance fields for forest monitoring and tree-scale reconstruction [ 14 , 23 , 42 , 27 ] , but the problem of direct DBH measurement from stand-off, above-canopy UAV surveys remains largely unaddressed. Moreover, translating a “fluffy” radiance-field reconstruction into a reliable diameter requires a measurement-oriented extraction step: a sufficiently dense sampling at breast height and a way to prioritize samples that are consistently supported across views. Most forestry applications of radiance fields to date have focused on close-range terrestrial [ 27 ] imagery which have a different set of constraints, leaving aerial DBH measurement largely unaddressed by radiance field methods. We introduce TreeDGS: an aerial reconstruction pipeline that uses 3D Gaussian Splatting as a measurement-oriented scene representation for estimating DBH from commodity UAV RGB imagery under pixel- and view-limited trunk observations. Starting from standard SfM camera poses and an MVS initialization, we optimize a RaDe-GS model [ 50 ] to obtain a continuous Gaussian field whose covariances and opacities are refined by multi-view photometric consistency. Then, we extract a dense point set using depth-aware cumulative-opacity integration, retaining samples whose accumulated opacity indicates a reliably occupied surface under a fixed threshold. To mitigate spurious samples from foliage and partial occlusions, we compute a per-point multi-view opacity support score as a reliability signal and retain trunk-isolated points via 3D semantic segmentation. Finally, we estimate DBH with opacity-weighted, slice-wise circle fitting, yielding diameter estimates even when the trunk is only weakly observed in individual images. Our

Read the complete text