pith. sign in

arxiv: 2605.19213 · v1 · pith:FIPJBGH5new · submitted 2026-05-19 · 💻 cs.CV

Smartphone-based Circular Plot Sampling for Forest Inventory

Su Sun , Jui-Cheng Chiu , Nabin Khanal , Songlin Fei , Yingjie Victor Chen This is my paper

Pith reviewed 2026-05-20 07:44 UTC · model grok-4.3

classification 💻 cs.CV
keywords smartphoneforest inventorycircular plot samplingDBH measurementmonocular depth estimationSLAMtree instance segmentationvideo-based mapping
0
0 comments X

The pith

A single smartphone video can measure tree diameters and locations in circular forest plots with accuracy matching traditional field methods.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops a pipeline that records a short walkthrough video inside a circular sample plot using only a consumer smartphone on a stand. It combines monocular depth estimation to recover distances, instance segmentation to isolate individual trees, and SLAM to track camera movement and fuse the data into real-world positions and diameters, with scale set by one reference measurement. This matters because conventional forest inventory either requires expensive LiDAR scanners or slow manual caliper-and-compass work, both of which limit how many plots can be sampled. If the approach holds, accurate tree data becomes feasible for far more users and at much larger scales without specialized equipment.

Core claim

Tree diameters at breast height and spatial positions inside circular plots are recovered from a single consumer smartphone video by fusing SLAM-derived camera poses with depth maps produced by pretrained monocular depth estimation and tree instance segmentation models, anchored to absolute scale with a calibrated reference length, yielding mean absolute errors of 1.51 cm in managed plots and 2.30 cm in natural plots.

What carries the argument

The joint refinement of camera trajectories and per-pixel depths through integration of pretrained monocular depth estimation, tree instance segmentation, and a SLAM framework applied to a single walkthrough video sequence.

If this is right

  • Complete circular plot sampling becomes possible with only a smartphone and stand instead of LiDAR or manual instruments.
  • Measurements remain consistent when the same plot is recorded from different starting directions or positions.
  • Equipment cost and operational complexity drop enough to make routine inventory feasible for non-expert users.
  • Accuracy levels stay comparable to established field methods across both managed and natural forest conditions.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same video-based fusion could be tested on non-circular sampling designs or in mixed urban-forest edges where vegetation density varies.
  • If model drift proves low, repeated videos of the same plot over seasons could track growth without resetting the reference scale each time.
  • Integration with consumer apps might allow immediate on-device estimates, though the paper does not demonstrate real-time processing.

Load-bearing premise

Pretrained depth and segmentation models produce sufficiently complete and accurate outputs in dense vegetation, and the SLAM system fuses them without large drift or scale error.

What would settle it

A direct comparison in a new dense natural forest plot showing mean absolute DBH error substantially larger than 2.5 cm against manual caliper measurements taken at the same trees.

Figures

Figures reproduced from arXiv: 2605.19213 by Jui-Cheng Chiu, Nabin Khanal, Songlin Fei, Su Sun, Yingjie Victor Chen.

Figure 1
Figure 1. Figure 1: Overview of the proposed smartphone-based forest inventory pipeline. A monocular video is captured by rotating a smartphone-mounted stand within a circular sample plot (top). Dense depth estimation and semantic segmentation are performed for each frame (upper middle). A SLAM module refines the prior depth and reconstructs a globally consistent 3D point cloud and estimates camera poses (lower middle). Tree … view at source ↗
Figure 2
Figure 2. Figure 2: End-to-end dense SLAM pipeline. A frozen monocular depth estimator supplies per-frame depth priors to a keyframe buffer. Dense correspondences are computed via feature extraction, correlation, and a convolutional GRU, which jointly updates poses, depth, and prior scale/offset through local and global bundle adjustment. A loop closure module injects long-range edges to suppress drift. Right: recovered camer… view at source ↗
Figure 3
Figure 3. Figure 3: Reconstructed point cloud of an individual tree [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Metric scale recovery. Left: top-down view of the reconstructed point cloud with the estimated camera trajectory radius 𝑟est (cyan). Right: the known physical tripod arm length 𝑟real and the corresponding circular scanning path. The ratio 𝑠 = 𝑟real∕𝑟est converts all reconstructed quantities from the SLAM coordinate system to metric units [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Dense point cloud reconstruction from a single monocular video. Left: top-down view showing the recovered camera trajectory (cyan ellipse) and surrounding tree stems. Right: two front-view perspectives of the same reconstruction. 5 m Scene 4 Scene 1 Scene 2 Scene 3 [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Left: Spatial distribution of the four circular sample plots used for system evaluation within Martell Forest, Purdue University. Each red circle indicates a circular inventory plot where smartphone-based circular video acquisition was performed (Scene 1–Scene 4, radius ∼10 m). Right: Examples of field data collection. Trees were labeled with numbered markers to ensure consistent identity across videos and… view at source ↗
Figure 7
Figure 7. Figure 7: Spatial layouts of the four evaluation plots. Blue points represent trees with ground-truth DBH measurements. The red marker indicates the primary position used for circular smartphone video acquisition. Green and purple markers denote additional offset acquisition positions used to evaluate repeatability under small variations in camera placement. These offset positions are located approximately 1 m from … view at source ↗
Figure 8
Figure 8. Figure 8: Comprehensive DBH error analysis. Top row, left to right: predicted vs. ground truth scatter plot with linear fit; error distribution stratified by DBH range; effect of camera-to-stem distance on DBH error, colored by stem size. Bottom row, left to right: per-tree prediction spread across videos for each scene; RMSE, MAE, and bias stratified by DBH range; overall error histogram with mean indicated. locali… view at source ↗
Figure 9
Figure 9. Figure 9: Cross-video tree localization consistency across the four plots. For each scene, the left panel overlays predicted tree positions from the three same-center videos, and the right panel overlays all five videos including two captured from offset positions. Colored dots represent predicted tree centers from individual videos; overlapping clusters indicate consistent localization [PITH_FULL_IMAGE:figures/ful… view at source ↗
Figure 10
Figure 10. Figure 10: Natural forest evaluation site at Martell Forest. Left: aerial overview with candidate sampling locations marked. Right: representative on-site views illustrating the stem distribution, understory density, and lighting conditions encountered during acquisition. the cross-center videos (V4–V7) span a similar range, indi￾cating that the reconstruction is not strongly biased by the choice of acquisition cent… view at source ↗
Figure 11
Figure 11. Figure 11: DBH estimation analysis on the natural forest plot. (a) Predicted vs. ground truth DBH across seven videos, with linear fit (𝑦 = 1.02𝑥 − 1.2, 𝑅2 = 0.906). (b) Per-video MAE, RMSE, and bias. (c) Per-tree prediction spread: individual estimates (circles), mean prediction (cross), and ground truth (diamond) [PITH_FULL_IMAGE:figures/full_fig_p013_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Tree position maps recovered on the natural forest plot. Left: overlay of the three same-center videos initiated from different directions. Middle, Right: overlays of pairs of videos captured from different centers within the same plot. Red circles denote estimated camera trajectories and colored dots denote detected tree positions. Stem arrangements are consistent across panels under both initialization … view at source ↗
read the original abstract

Circular sample plots are a cornerstone of forest inventory, yet accurate measurement of tree diameter at breast height (DBH) and spatial location within such plots remains challenging. Conventional approaches rely either on costly terrestrial LiDAR systems or labor-intensive manual methods involving calipers and compass bearings, limiting their scalability and accessibility in large scale environments. We present a lightweight, smartphone-based pipeline that enables complete plot sampling based tree measurement from a single walkthrough video, requiring no specialized hardware beyond a consumer smartphone mounted on a portable stand. The proposed method integrates pretrained monocular depth estimation and tree instance segmentation with a simultaneous localization and mapping (SLAM) framework to jointly refine camera trajectories and depth across the video sequence. Tree positions and DBH estimates are recovered by fusing SLAM-derived camera poses with segmented depth maps, with absolute real-world scale anchored via a calibrated reference length. The system was evaluated in both managed forest plots and natural forest plot, achieving a mean absolute error of 1.51 cm (MARE 3.98%) and 2.30 cm (MARE 5.69%) respectively, with consistent performance across varying starting directions and positions. Cross-video consistency analysis further demonstrated stable and reproducible tree localization across measurements initiated from different starting positions. The proposed approach achieves accuracy comparable to established field methods while substantially reducing equipment cost and operational complexity, making it accessible to both professional researchers and non-expert forest managers in diverse operational settings.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The paper proposes a smartphone-based pipeline for circular plot sampling in forest inventory. From a single walkthrough video captured with a consumer smartphone on a portable stand, the method fuses pretrained monocular depth estimation and tree instance segmentation with a SLAM framework to recover tree positions and diameter at breast height (DBH). Absolute scale is anchored using a single calibrated reference length. Evaluation on managed and natural forest plots reports mean absolute errors of 1.51 cm (MARE 3.98%) and 2.30 cm (MARE 5.69%), with cross-video consistency checks across different starting positions.

Significance. If the accuracy claims hold after addressing component-level validation, the work could substantially lower barriers to forest inventory by replacing expensive LiDAR or labor-intensive manual caliper/compass methods with accessible smartphone video. The integration of off-the-shelf models with SLAM and scale anchoring is a pragmatic engineering contribution, though the manuscript provides no evidence of reproducible code, parameter-free derivations, or falsifiable predictions beyond the reported MAEs.

major comments (3)
  1. [Abstract / Methods] Abstract and Methods: The central accuracy claims (MAE 1.51 cm / 2.30 cm) rest on the assumption that off-the-shelf monocular depth and instance segmentation models produce unbiased depth maps and complete stem masks at breast height in occluded forest scenes. No per-component metrics (depth error at 1.3 m height, mask IoU on trunks, or failure rates under foliage) are reported, nor is any domain adaptation or forest-specific fine-tuning described. This is load-bearing because systematic underestimation through gaps or clipped masks would directly propagate into the fused horizontal diameter calculation even if SLAM poses are locally consistent.
  2. [Results] Results: The manuscript reports cross-video consistency and MAEs but provides no baseline comparisons against established field methods (e.g., caliper + compass or terrestrial LiDAR) on the same plots, nor any ablation on the contribution of each pipeline stage. Without these, it is impossible to determine whether the reported errors are competitive or whether post-hoc choices in reference length or starting position affect the numbers.
  3. [Evaluation] Evaluation: The two plot types (managed and natural) are mentioned, but dataset details (number of trees, plot radii, video lengths, occlusion levels) and error analysis (e.g., dependence on distance from camera path or stem visibility) are absent. This limits assessment of whether the method generalizes beyond the tested conditions.
minor comments (2)
  1. [Abstract] The abstract states 'complete plot sampling based tree measurement' — consider rephrasing for grammatical clarity.
  2. [Abstract] Notation for MARE and MAE should be defined on first use with explicit formulas if they deviate from standard definitions.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed comments, which have helped us identify areas where the manuscript can be strengthened. We address each major comment below and describe the corresponding revisions.

read point-by-point responses

  1. Referee: [Abstract / Methods] Abstract and Methods: The central accuracy claims (MAE 1.51 cm / 2.30 cm) rest on the assumption that off-the-shelf monocular depth estimation and instance segmentation models produce unbiased depth maps and complete stem masks at breast height in occluded forest scenes. No per-component metrics (depth error at 1.3 m height, mask IoU on trunks, or failure rates under foliage) are reported, nor is any domain adaptation or forest-specific fine-tuning described. This is load-bearing because systematic underestimation through gaps or clipped masks would directly propagate into the fused horizontal diameter calculation even if SLAM poses are locally consistent.

    Authors: We agree that component-level validation would increase transparency. The manuscript presents an end-to-end pipeline using publicly available pretrained models without forest-specific fine-tuning, as our contribution centers on the integration with SLAM and scale anchoring rather than model retraining. In the revised manuscript we have added a dedicated paragraph in the Methods section that reports qualitative observations of depth and mask quality at breast height, cites the original model papers for their reported error characteristics, and includes a brief failure-mode analysis for heavily occluded stems. We now also explicitly discuss the absence of domain adaptation as a limitation and its potential effect on systematic bias. revision: partial

  2. Referee: [Results] Results: The manuscript reports cross-video consistency and MAEs but provides no baseline comparisons against established field methods (e.g., caliper + compass or terrestrial LiDAR) on the same plots, nor any ablation on the contribution of each pipeline stage. Without these, it is impossible to determine whether the reported errors are competitive or whether post-hoc choices in reference length or starting position affect the numbers.

    Authors: Direct simultaneous measurements with LiDAR or additional manual teams on the identical plots were not feasible within the scope of this study. The revised Results section now contains a comparison table that places our reported MAEs alongside published accuracies for manual caliper/compass methods and terrestrial LiDAR from the recent literature, while noting the differing experimental conditions. We have also added a limited ablation using the existing cross-video consistency data to quantify the contribution of the SLAM refinement and reference-length scaling steps to the final diameter and position errors. revision: yes

  3. Referee: [Evaluation] Evaluation: The two plot types (managed and natural) are mentioned, but dataset details (number of trees, plot radii, video lengths, occlusion levels) and error analysis (e.g., dependence on distance from camera path or stem visibility) are absent. This limits assessment of whether the method generalizes beyond the tested conditions.

    Authors: We have expanded the Evaluation section to report the missing dataset statistics, including the number of trees measured, plot radii, approximate video durations, and a qualitative categorization of occlusion levels for each plot. We have further included new figures that plot diameter and position errors against distance from the camera trajectory and against a simple stem-visibility score derived from the segmentation masks, thereby addressing the requested dependence analysis. revision: yes

Circularity Check

0 steps flagged

No significant circularity; empirical pipeline validated on independent field data

full rationale

The paper describes an engineering pipeline that fuses off-the-shelf pretrained monocular depth estimation and instance segmentation models with a SLAM framework, then anchors absolute scale using a single calibrated reference length. Reported accuracies (MAE 1.51 cm / 2.30 cm) are obtained from direct comparison against manual measurements in managed and natural forest plots, with cross-start consistency checks performed on separate video sequences. No equations, fitted parameters, or uniqueness theorems are presented that reduce the output metrics to the input models or reference length by construction. The central claims rest on external empirical benchmarks rather than self-referential definitions or self-citation chains.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The central claim rests on the generalization of pretrained vision models to forest scenes and the ability of SLAM to produce usable trajectories and depths in vegetation; a single calibration length supplies scale.

free parameters (1)
  • calibrated reference length
    Single known length used to anchor absolute scale in the fused depth and pose output.
axioms (2)
  • domain assumption Pretrained monocular depth estimation models produce reliable depth maps in forest environments
    Invoked when the pipeline fuses segmented depth maps with SLAM poses.
  • domain assumption Tree instance segmentation masks are sufficiently complete and accurate for DBH recovery
    Required to extract per-tree depth values from the refined depth maps.

pith-pipeline@v0.9.0 · 5794 in / 1389 out tokens · 46580 ms · 2026-05-20T07:44:17.974789+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

64 extracted references · 64 canonical work pages

  1. [1]

    IEEE/CVF International Conference on Computer Vision , month =

    Berton, Gabriele and Trivigno, Gabriele and Caputo, Barbara and Masone, Carlo , title =. IEEE/CVF International Conference on Computer Vision , month =. 2023 , pages =

    work page 2023

  2. [2]

    Proceedings of the 35th International Conference on Neural Information Processing Systems , articleno =

    Teed, Zachary and Deng, Jia , title =. Proceedings of the 35th International Conference on Neural Information Processing Systems , articleno =. 2021 , isbn =

    work page 2021

  3. [3]

    Forestry: An International Journal of Forest Research , volume =

    Grondin, Vincent and Fortin, Jean-Michel and Pomerleau, François and Giguère, Philippe , title =. Forestry: An International Journal of Forest Research , volume =. 2022 , month =. doi:10.1093/forestry/cpac043 , url =

    work page doi:10.1093/forestry/cpac043 2022

  4. [4]

    Putree: A Photorealistic Large-Scale Virtual Benchmark for Forest Training , year=

    Lu, Yawen and Huang, Yunhan and Sun, Su and Fei, Songlin and Chen, Victor , booktitle=. Putree: A Photorealistic Large-Scale Virtual Benchmark for Forest Training , year=

    work page

  5. [5]

    A photogrammetric system for dendrometric feature estimation of individual trees , year=

    Collazos, Daniel Tobón and Cano, Victor Romero and Villota, Juan Carlos Perafan and Toro, Walter Mayor , booktitle=. A photogrammetric system for dendrometric feature estimation of individual trees , year=

    work page

  6. [6]

    Forest Science , volume=

    A review of past research on dendrometers , author=. Forest Science , volume=

    work page

  7. [7]

    and Scott, Charles T

    Bechtold, William A. and Scott, Charles T. , title =. The Enhanced Forest Inventory and Analysis Program---National Sampling Design and Estimation Procedures , editor =. 2005 , series =

    work page 2005

  8. [8]

    Forest Inventory and Analysis National Core Field Guide: Field Data Collection Procedures for Phase 2 Plots, Version 9.3 , year =

    work page

  9. [9]

    Terrestrial laser scanning in forest inventories , journal =

    Liang, Xinlian and Kankare, Ville and Hyypp. Terrestrial laser scanning in forest inventories , journal =. 2016 , publisher=

    work page 2016

  10. [10]

    Forest Inventory with Terrestrial LiDAR: A Comparison of Static and Hand-Held Mobile Laser Scanning , journal =

    Bauwens, S. Forest Inventory with Terrestrial LiDAR: A Comparison of Static and Hand-Held Mobile Laser Scanning , journal =. 2016 , publisher=

    work page 2016

  11. [11]

    Annals of Botany , volume=

    Terrestrial laser scanning: a new standard of forest measuring? , author=. Annals of Botany , volume=. 2021 , doi=

    work page 2021

  12. [12]

    Comparing positioning accuracy of mobile laser scanning systems under a forest canopy , journal =

    Jesse Muhojoki and Teemu Hakala and Antero Kukko and Harri Kaartinen and Juha Hyyppä , keywords =. Comparing positioning accuracy of mobile laser scanning systems under a forest canopy , journal =. 2024 , issn =. doi:https://doi.org/10.1016/j.srs.2024.100121 , url =

    work page doi:10.1016/j.srs.2024.100121 2024

  13. [13]

    Mapping tropical forest trees across large areas with lightweight cost-effective terrestrial laser scanning , journal =

    Tao, Shengli and Labri. Mapping tropical forest trees across large areas with lightweight cost-effective terrestrial laser scanning , journal =. 2021 , volume =. doi:10.1007/s13595-021-01113-9 , url =

    work page doi:10.1007/s13595-021-01113-9 2021

  14. [14]

    and White, Joanne C

    Luoma, Ville and Saarinen, Ninni and Wulder, Michael A. and White, Joanne C. and Vastaranta, Mikko and Holopainen, Markus and Hyyppä, Juha , TITLE =. Forests , VOLUME =. 2017 , NUMBER =

    work page 2017

  15. [15]

    2004 , publisher=

    Tree and Forest Measurement , author=. 2004 , publisher=

    work page 2004

  16. [16]

    Southern Forests: a Journal of Forest Science , volume =

    Horst W Kassier , title =. Southern Forests: a Journal of Forest Science , volume =. 2011 , publisher =. doi:10.2989/20702620.2011.574816 , URL =

    work page doi:10.2989/20702620.2011.574816 2011

  17. [17]

    and Delitti, Welington B.C

    Chave, Jérôme and Réjou-Méchain, Maxime and Búrquez, Alberto and Chidumayo, Emmanuel and Colgan, Matthew S. and Delitti, Welington B.C. and Duque, Alvaro and Eid, Tron and Fearnside, Philip M. and Goodman, Rosa C. and Henry, Matieu and Martínez-Yrízar, Angelina and Mugasha, Wilson A. and Muller-Landau, Helene C. and Mencuccini, Maurizio and Nelson, Bruce ...

    work page doi:10.1111/gcb.12629

  18. [18]

    2015 , publisher=

    Forest Measurements: Fifth Edition , author=. 2015 , publisher=

    work page 2015

  19. [19]

    Sullivan, Martin J. P. and Lewis, Simon L. and Hubau, Wannes and Qie, Lan and Baker, Timothy R. and Banin, Lindsay F. and Chave, Jerôme and Cuni-Sanchez, Aida and Feldpausch, Ted R. and Lopez-Gonzalez, Gabriela and Arets, Eric and Ashton, Peter and Bastin, Jean-François and Berry, Nicholas J. and Bogaert, Jan and Boot, Rene and Brearley, Francis Q. and Br...

    work page doi:10.1111/2041-210x.12962 2041

  20. [20]

    The Forestry Chronicle , volume=

    A best practices guide for generating forest inventory attributes from airborne laser scanning data using an area-based approach , author=. The Forestry Chronicle , volume=. 2013 , publisher=

    work page 2013

  21. [21]

    Remote Sensing of Environment , volume=

    Area-based vs tree-centric approaches to mapping forest carbon in Southeast Asian forests from airborne laser scanning data , author=. Remote Sensing of Environment , volume=. 2017 , publisher=

    work page 2017

  22. [22]

    ISPRS journal of photogrammetry and remote sensing , volume=

    International benchmarking of terrestrial laser scanning approaches for forest inventories , author=. ISPRS journal of photogrammetry and remote sensing , volume=. 2018 , publisher=

    work page 2018

  23. [23]

    Calders, Kim and Adams, Jennifer and Armston, John and Bartholomeus, Harm and Bauwens, Sébastien and Bentley, Lisa and Chave, Jerome and Danson, Francis and Demol, Miro and Disney, M. and Gaulton, Rachel and Moorthy, Sruthi and Levick, Shaun and Saarinen, Ninni and Schaaf, Crystal and Stovall, Atticus and Terryn, Louise and Wilkes, Phil and Verbeeck, Hans...

    work page

  24. [24]

    2018 , issn =

    International benchmarking of terrestrial laser scanning approaches for forest inventories , journal =. 2018 , issn =. doi:https://doi.org/10.1016/j.isprsjprs.2018.06.021 , url =

    work page doi:10.1016/j.isprsjprs.2018.06.021 2018

  25. [25]

    ISPRS Journal of Photogrammetry and Remote Sensing , volume=

    Full-waveform topographic lidar: State-of-the-art , author=. ISPRS Journal of Photogrammetry and Remote Sensing , volume=. 2009 , doi=

    work page 2009

  26. [26]

    and Armston, John D

    Newnham, Glenn J. and Armston, John D. and Calders, Kim and Disney, Mathias I. and Lovell, Jenny L. and Schaaf, Crystal B. and Strahler, Alan H. and Danson, F. Mark , journal =. Terrestrial Laser Scanning for Plot-Scale Forest Measurement , url =

    work page

  27. [27]

    Global Change Biology , volume=

    Improved allometric models to estimate the aboveground biomass of tropical trees , author=. Global Change Biology , volume=. 2014 , doi=

    work page 2014

  28. [28]

    Scandinavian Journal of Forest Research , volume=

    Feasibility of mobile laser scanning for forest inventory , author=. Scandinavian Journal of Forest Research , volume=. 2021 , doi=

    work page 2021

  29. [29]

    Annals of Forest Science , volume=

    Mobile laser scanning for the assessment of tree structural attributes in forest inventories , author=. Annals of Forest Science , volume=. 2018 , doi=

    work page 2018

  30. [30]

    Remote Sensing , volume=

    Mobile laser scanning: State of the art and future perspectives , author=. Remote Sensing , volume=. 2020 , doi=

    work page 2020

  31. [31]

    Evaluation of a Smartphone App for Forest Sample Plot Measurements , journal =

    Vastaranta, Mikko and Gonz. Evaluation of a Smartphone App for Forest Sample Plot Measurements , journal =. 2015 , publisher=

    work page 2015

  32. [32]

    Forestry: An International Journal of Forest Research , volume =

    Marzulli, Maria Immacolata and Raumonen, Pasi and Greco, Roberto and Persia, Manuela and Tartarino, Patrizia , title =. Forestry: An International Journal of Forest Research , volume =. 2020 , publisher=

    work page 2020

  33. [33]

    Forests , year =

    Ahamed, Aakash and Foye, John and Poudel, Sanjok and Trieschman, Erich and Fike, John , title =. Forests , year =

    work page

  34. [34]

    o r \"a l \

    Xinlian Liang and Yunsheng Wang and Jiri Py \"o r \"a l \"a and Matti Lehtom \"a ki and Xiaowei Yu and Harri Kaartinen and Antero Kukko and Eija Honkavaara and Issaoui, \ Aimad E. I.\ and Olli Nevalainen and Matti Vaaja and Juho-Pekka Virtanen and Masato Katoh and Songqiu Deng. Forest in situ observations using unmanned aerial vehicle as an alternative of...

    work page doi:10.1186/s40663-019-0173-3 2019

  35. [35]

    Methods in Ecology and Evolution , volume =

    Tatsumi, Shinichi and Yamaguchi, Keiji and Furuya, Naoyuki , title =. Methods in Ecology and Evolution , volume =. 2023 , doi =

    work page 2023

  36. [36]

    and De Stefano, Andrea , title =

    Howie, Noah A. and De Stefano, Andrea , title =. Forest Science , volume =. 2024 , publisher=

    work page 2024

  37. [37]

    Measuring tree diameter using a LiDAR-equipped smartphone: a comparison of smartphone- and caliper-based DBH , journal =

    G. Measuring tree diameter using a LiDAR-equipped smartphone: a comparison of smartphone- and caliper-based DBH , journal =. 2023 , doi =

    work page 2023

  38. [38]

    Remote Sensing , volume=

    Examining the accuracy of smartphone LiDAR for forestry applications , author=. Remote Sensing , volume=. 2021 , doi=

    work page 2021

  39. [39]

    and Szeliski, Richard , title =

    Snavely, Noah and Seitz, Steven M. and Szeliski, Richard , title =. Seminal Graphics Papers: Pushing the Boundaries, Volume 2 , articleno =. 2023 , isbn =

    work page 2023

  40. [40]

    Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , pages=

    Structure-from-Motion Revisited , author=. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , pages=. 2016 , doi=

    work page 2016

  41. [41]

    Foundations and Trends in Computer Graphics and Vision , volume=

    Multi-View Stereo: A Tutorial , author=. Foundations and Trends in Computer Graphics and Vision , volume=. 2015 , doi=

    work page 2015

  42. [42]

    IEEE Transactions on Robotics , volume=

    Mur-Artal, Ra. IEEE Transactions on Robotics , volume=. 2017 , publisher=

    work page 2017

  43. [43]

    IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=

    Direct Sparse Odometry , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=. 2018 , publisher=

    work page 2018

  44. [44]

    and Glocker, Ben and Davison, Andrew J

    Whelan, Thomas and Leutenegger, Stefan and Salas-Moreno, Renato F. and Glocker, Ben and Davison, Andrew J. , booktitle=. Elastic. 2016 , publisher=

    work page 2016

  45. [45]

    and Pollefeys, Marc , booktitle=

    Zhu, Zihan and Peng, Songyou and Larsson, Viktor and Xu, Weiwei and Bao, Hujun and Cui, Zhaopeng and Oswald, Martin R. and Pollefeys, Marc , booktitle=. 2022 , doi=

    work page 2022

  46. [46]

    Wang, Hengyi and Wang, Jingwen and Agapito, Lourdes , booktitle=. Co-

    work page

  47. [47]

    IEEE Transactions on Robotics , volume=

    Past, Present, and Future of Simultaneous Localization and Mapping: Toward the Robust-Perception Age , author=. IEEE Transactions on Robotics , volume=. 2016 , doi=

    work page 2016

  48. [48]

    IEEE Robotics & Automation Magazine , volume=

    Visual Odometry [Tutorial] , author=. IEEE Robotics & Automation Magazine , volume=. 2011 , doi=

    work page 2011

  49. [49]

    Proceedings of the IEEE and ACM International Symposium on Mixed and Augmented Reality (ISMAR) , pages=

    Parallel Tracking and Mapping for Small AR Workspaces , author=. Proceedings of the IEEE and ACM International Symposium on Mixed and Augmented Reality (ISMAR) , pages=. 2007 , doi=

    work page 2007

  50. [50]

    Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) , pages=

    iMAP: Implicit Mapping and Positioning in Real-Time , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) , pages=. 2021 , doi=

    work page 2021

  51. [51]

    Visual Appearance Analysis of Forest Scenes for Monocular SLAM , year=

    Garforth, James and Webb, Barbara , booktitle=. Visual Appearance Analysis of Forest Scenes for Monocular SLAM , year=

    work page

  52. [52]

    Forest tree diameter estimation using a monocular

    Yan, Zhenghao and others , booktitle=. Forest tree diameter estimation using a monocular

    work page

  53. [53]

    Remote Sensing , VOLUME =

    Piermattei, Livia and Karel, Wilfried and Wang, Di and Wieser, Martin and Mokroš, Martin and Surový, Peter and Koreň, Milan and Tomaštík, Julián and Pfeifer, Norbert and Hollaus, Markus , TITLE =. Remote Sensing , VOLUME =. 2019 , NUMBER =

    work page 2019

  54. [54]

    Vision Transformers for Dense Prediction , booktitle =

    Ranftl, Ren. Vision Transformers for Dense Prediction , booktitle =

    work page

  55. [55]

    Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) , pages =

    Yin, Wei and Zhang, Chi and Chen, Hao and Cai, Zhipeng and Yu, Gang and Wang, Kaixuan and Chen, Xiaozhi and Shen, Chunhua , title =. Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) , pages =

    work page

  56. [56]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , year =

    Yang, Lihe and Kang, Bingyi and Huang, Zilong and Xu, Xiaogang and Feng, Jiashi and Zhao, Hengshuang , title =. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , year =

    work page

  57. [57]

    Video Depth Anything: Consistent Depth Estimation for Super-Long Videos , year=

    Chen, Sili and Guo, Hengkai and Zhu, Shengnan and Zhang, Feihu and Huang, Zilong and Feng, Jiashi and Kang, Bingyi , booktitle=. Video Depth Anything: Consistent Depth Estimation for Super-Long Videos , year=

    work page

  58. [58]

    Proceedings of the 28th International Conference on Neural Information Processing Systems - Volume 2 , pages =

    Eigen, David and Puhrsch, Christian and Fergus, Rob , title =. Proceedings of the 28th International Conference on Neural Information Processing Systems - Volume 2 , pages =. 2014 , publisher =

    work page 2014

  59. [59]

    Forestry: An International Journal of Forest Research , volume=

    Comparing mobile and terrestrial laser scanning for measuring and modelling tree stem taper , author=. Forestry: An International Journal of Forest Research , volume=. 2023 , month=

    work page 2023

  60. [60]

    Remote Sensing of Environment , volume=

    A Three-stage Framework for Stand-level Automated Stem Volume Estimation in Temperate Forests Using Mobile Laser Scanning , author=. Remote Sensing of Environment , volume=. 2026 , doi=

    work page 2026

  61. [61]

    The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences , volume=

    General Framework for Georeferencing and Interpretation of Multi-Resolution LiDAR Data for Fine-Scale Forest Inventory , author=. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences , volume=

    work page

  62. [62]

    Sensors , volume=

    Single Shot High-accuracy Diameter at Breast Height Measurement with Smartphone Embedded Sensors , author=. Sensors , volume=. 2025 , doi=

    work page 2025

  63. [63]

    HI-SLAM2: Geometry-Aware Gaussian SLAM for Fast Monocular Scene Reconstruction , volume=

    Zhang, Wei and Cheng, Qing and Skuddis, David and Zeller, Niclas and Cremers, Daniel and Haala, Norbert , year=. HI-SLAM2: Geometry-Aware Gaussian SLAM for Fast Monocular Scene Reconstruction , volume=. doi:10.1109/tro.2025.3626627 , journal=

    work page doi:10.1109/tro.2025.3626627 2025

  64. [64]

    Remote Sensing of Environment , year=

    Integrated scan simultaneous trajectory enhancement and mapping (IS2-TEAM) for fine resolution forest inventory using backpack LiDAR , author=. Remote Sensing of Environment , year=

    work page