Pith. sign in

REVIEW 4 major objections 6 minor 63 references

A Novel Large Vision Foundation Model (LVFM)-based Approach for Generating High-Resolution Canopy Height Maps in Plantations for Precision Forestry Management

T0 review · 4 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A vision foundation model with a self-supervised detail-enhancement module can turn 1-meter RGB imagery into canopy height maps that match lidar within 0.24 m RMSE in plantation settings.

desk verdict A useful architectural twist—DINOv2 plus a self-supervised CARAFE upsampler—gives genuinely strong CHM numbers on one plot, but the evaluation has a gap: the Google Earth tile for the validation plot is never dated, so the headline accuracy could be confounded by canopy growth between image and lidar. read the letter →

arxiv 2506.20388 v1 pith:BY2PD2CW submitted 2025-06-25 cs.CV

Shen Tan , Xin Zhang , Liangxiu Han , Huaguo Huang , Han Wang This is my paper · ORCID
classification cs.CV
keywords canopyheightmaplargevisionfoundationmodelDINOv2self-supervisedfeatureenhancementCARAFEplantationabovegroundbiomassRGBsatelliteimagerylidarvalidation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper tries to establish that a large vision foundation model, paired with a self-supervised module that restores spatial detail lost by ViT patch tokenization, can generate plantation-scale canopy height maps from ordinary 1-meter RGB satellite imagery with accuracy close to lidar. The authors report a mean absolute error of 0.09 m, an RMSE of 0.24 m, and an $R^2$ of 0.78 against UAV-lidar references on a held-out plantation plot, beating a U-Net CNN, a DPT baseline, and published global CHM products. If true, this would make wall-to-wall, frequently updated height and aboveground-biomass monitoring of small fragmented plantations feasible without repeated lidar flights. The same CHMs support individual-tree detection with over 90% success and biomass estimates with strong correlation, and the model transfers to a non-training conifer forest.

What carries the argument

The load-bearing mechanism is the self-supervised feature enhancement module, which counters the resolution loss inherent in ViT patch tokenization. DINOv2's 37×37 low-resolution features are upscaled with CARAFE; the upsampler is trained without extra labels by applying augmentations to the input, downsampling the enhanced features with a learned blurring kernel, and enforcing a multiview consistency loss that includes a spatially varying uncertainty parameter. This reconstruction preserves crown edges and fine plantation structure, giving the lightweight CNN height estimator the detail it needs to output individual-tree-level heights. The paper credits this module, rather than the LVFM alone, for the improvement over naive ViT baselines.

What would settle it

Obtain the actual acquisition date of the Plot B Google Earth tiles and re-run the pixel-wise error against the November 2020 lidar CHM; if the tiles predate the lidar by a year or more, tree growth inflates apparent accuracy and the error should scale with elapsed time. A second decisive check is applying the trained network to natural-forest lidar with heterogeneous canopy; if $R^2$ falls far below the reported 0.71 and crowns blur, the spatial-detail claim does not transfer beyond uniform plantations.

Watch

Extended reading notes

Core claim

The paper's central claim is that a ViT-based LVFM can serve as a high-resolution CHM generator when the spatial detail it discards is reconstructed before height regression. The network combines DINOv2-Large as the feature extractor, a self-supervised feature enhancement module built on CARAFE that upsamples low-resolution features under a multiview consistency loss with learned uncertainty, and a lightweight three-layer CNN height estimator. On the held-out Plot B, pixel-wise evaluation against the lidar CHM gives MAE = 0.09 m, RMSE = 0.24 m, and $R^2 = 0.78$, compared with U-Net (0.14 m, 0.33 m, 0.76) and DPT-Large (0.14 m, 0.36 m, 0.71). Applied to the ecologically different Saihanba forest with only about 5% of tiles used to calibrate the height projection, it reports RMSE = 0.30 m and $R^2 = 0.71$. The predicted CHMs also yield tree-detection success of 92 ± 6% across 212 lidar-referenced plantations and species-level AGB estimates with $R^2$ values from 0.71 to 0.98.

Load-bearing premise

The model's reported accuracy assumes the Google Earth RGB tiles used for training and validation were captured close in time to the November 2020 lidar survey; the paper gives a 2013-2020 range for its image corpus and seasonal caveats for some years, but never states the acquisition date of the specific tiles used for the held-out validation plot.

Editorial extensions

If this is right

  • Plantation-scale CHMs can be produced over large regions from 1-meter RGB with only a small lidar calibration set, replacing repeated airborne lidar campaigns.
  • Annual growth tracking for carbon projects can be updated from historical and current satellite images, because the same network maps different-year RGB tiles into a common height scale.
  • Individual-tree detection in small fragmented plantations can be automated with a simple local-maximum filter on the predicted CHM, meeting the 90% success criterion for most parcels.
  • Existing global CHM products at 1 to 30 meters miss plantation-scale height variability, so locally fine-tuned LVFM-based maps are the appropriate input for plantation AGB accounting.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The same self-supervised upsampling idea could be ported to other ViT-based dense prediction tasks, such as crown segmentation or species mapping from satellite imagery, wherever patch tokenization blurs edges.
  • The reported accuracy hinges on the validation RGB tiles being close in date to the November 2020 lidar; a natural stress test is to sort Google Earth tiles by known acquisition month and check whether error grows with time lag.
  • If the 0.24 m RMSE holds regionally, the method could support carbon-credit audits with lidar reduced to a sparse network of calibration plots rather than wall-to-wall coverage.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper proposes a large vision foundation model (LVFM)-based network for generating 1-m canopy height maps (CHMs) from high-resolution Google Earth RGB imagery. The architecture combines a DINOv2 feature extractor, a self-supervised feature enhancement module based on multiview reconstruction, and a lightweight CNN height estimator. The model is trained on two lidar-covered plots and evaluated on a third held-out plot in the Fangshan District of Beijing, where it reportedly achieves MAE = 0.09 m, RMSE = 0.24 m, and R2 = 0.78. The paper also compares the model with U-Net, DPT, and three existing CHM products, applies the model to a conifer forest in Saihanba to test generalizability, and demonstrates downstream applications in individual-tree detection, plantation AGB estimation, and growth tracking.

Significance. If the reported accuracy is robust, the method would be practically valuable for cost-effective plantation monitoring, because it combines a pretrained foundation model with a relatively small lidar training set and produces CHMs at a resolution relevant for individual-tree analysis. The paper includes a useful comparison with three established CHM products (Tolan et al., Lang et al., Potapov et al.) and provides a plausible downstream pipeline for tree detection and AGB estimation. Strengths include the use of independent lidar reference data, clear visualizations, and an explicit attempt at cross-region generalization. However, the evidence base is currently thin: the central claim rests on a single held-out plot, the acquisition dates of the RGB tiles are not reported, and the Saihanba experiment uses a portion of the target lidar for calibration, so the support for the headline claims is weaker than the abstract suggests.

major comments (4)
  1. [Section 2.2.2 and Section 3.1.2] The acquisition dates of the specific Google Earth RGB tiles used for training and for the Plot B evaluation are not reported. The RGB corpus is described as spanning 2013 to 2020, with known seasonal anomalies in 2015 and 2018, while the lidar reference was acquired in November 2020. Because the network receives no date information, any time gap between a tile and the lidar reference introduces unmodeled canopy growth into the reported MAE = 0.09 m and RMSE = 0.24 m. The authors themselves state in Section 4.1.3 that early-stage annual growth is typically less than 1 m, so even a one-year offset could be several times the RMSE. The paper should report the exact acquisition dates of all tiles used in training and validation, or otherwise demonstrate that the results are insensitive to plausible growth offsets.
  2. [Section 2.3.2 and Section 3.1.3] The Saihanba generalizability experiment is not a clean out-of-distribution test as presented. The text says that "only a limited subset of lidar observations (approximately 5% of the total tiles) was utilized to project the extracted features into continuous tree height estimates." This step appears to use target-domain lidar to calibrate or map the features to heights, so the reported R2 = 0.71 on Saihanba reflects a degree of adaptation to the target site rather than pure zero-shot transfer. The authors should specify exactly what this projection involves, whether any parameters are fit on Saihanba data, and ideally report performance with and without this calibration step.
  3. [Section 3.1.2 and Table 3] The headline accuracy is based on a single held-out plot (Plot B, 1813 x 1672 pixels). No error bars, confidence intervals, or repeated cross-validation results are provided, and the comparison with U-Net shows only a small R2 advantage (0.78 versus 0.76) with a larger MAE/RMSE difference. Given the spatial autocorrelation inherent in canopy height fields and the small number of independent test sites, the paper should quantify the uncertainty of the metrics, for example through block bootstrap or multiple held-out plots, before claiming state-of-the-art performance.
  4. [Section 4.2 and Section 5] The paper's own limitations section acknowledges that only a "more rigorous evaluation is necessary" for growth tracking and that the current method for tracking height and AGB is based on a qualitative relationship with stand age rather than rigorous validation. Nevertheless, the abstract and conclusion state that the model enables "effective tracking of plantation growth." These claims should be tempered to match the evidence, or the growth-tracking experiment should be strengthened with quantitative validation against field measurements.
minor comments (6)
  1. [Section 2.2.2] The text says "red dots in Fig. 2a" but the plantation parcels are shown in Fig. 5; please correct the cross-reference.
  2. [Table 2] The column header "MAS" should be "MAE", and the "Bias" rows should be defined with units and a description of how bias is computed.
  3. [Section 2.1.2, Equation (1)] The notation in the reconstruction loss is confusing: n appears both as the number of augmentations and as a subscript, and the relationship between s, the linear network, and the L2 term is not explained. Please clarify the derivation and define all symbols.
  4. [Section 4.1.3] The text refers to "Section 3.3" for the annual AGB estimations, but the relevant results are in Section 3.2.3; please update the cross-reference.
  5. [Section 2.2.2] The description of the RGB corpus says that about 20% of the area in 2015 and 30% in 2018 were affected by cloud or winter imagery. It is not stated whether these affected tiles are excluded from training and evaluation; please clarify the handling of these tiles.
  6. [Throughout] There are several typographical and grammatical errors, including "recued cost" (Section 1), "dimentional" (Section 1), "overperformance" (Section 3.1.2), and "the resulting R2" where R-squared is intended. A careful proofreading pass is recommended.

Circularity Check

0 steps flagged · score 0.0 of 10

No material circularity: the central CHM claim is supervised directly against external lidar data, and downstream AGB/growth analyses use external allometry without feeding back into the CHM result.

full rationale

The paper's central claim is a supervised CHM regression from 1 m RGB imagery to lidar-derived CHM: the model is trained on Plots A and C and evaluated on the held-out Plot B using pixel-wise MAE, RMSE, and R². This is an external benchmark; no predicted quantity is defined by a fitted value from this paper, and no equation in Sections 2.1 or 3.1 reduces to its own input. The LVFM feature extractor (DINOv2), CARAFE feature enhancement, and CNN height head are standard external components trained against lidar references. Downstream Experiment II uses a local-maximum tree detector on the predicted CHM and applies external allometric relationships (DBH = 1.117H + 5.38 from Qin et al. [10], and CCER species-specific AGB functions) identically to both model-derived and lidar-derived CHMs; those R² values are therefore transformed comparisons of CHM fidelity rather than self-referential predictions. The only study-area prior work cited, Qin et al. [10], supplies lidar collection details and a field-fitted DBH-height relation; even if considered group-internal, it is not load-bearing for the CHM accuracy claim and does not define the target quantity. The authors also explicitly acknowledge in Section 4.2 that growth tracking is based on a qualitative relationship with stand age rather than rigorous field validation, which is an honest limitation rather than a circular derivation. The potential mismatch between Google Earth image acquisition dates (2013–2020) and the November 2020 lidar reference is a data-contemporaneity risk, not a circularity, and does not make the derivation self-referential.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

The central CHM claim rests on supervised training with lidar as ground truth and on the assumption that RGB and lidar are temporally aligned. The downstream AGB and growth claims rest on externally fitted allometric equations and a DBH-H relationship from prior work. No new physical entities are introduced; the feature enhancement module is an algorithmic component.

free parameters (4)
  • DBH-height relationship coefficients = DBH = 1.117H + 5.38
    Fitted for the study area in Qin et al. [10] and used to convert CHM heights to DBH for AGB estimation; no uncertainty is reported.
  • Species-specific AGB allometric coefficients = Table 1: e.g., P. tabulaeformis AGB = 0.92H^2 - 0.46H + 5.03
    Used to convert tree heights to AGB per species in Experiment II; source and fit statistics are not reported.
  • Local maximum tree detection radius = 5 m
    Hand-chosen radius in the local-maxima tree detection step; detection success depends on this value.
  • Learned network weights of feature enhancer, downsampler, and CNN head = not released
    Trained on lidar CHM references; the paper reports total parameter counts but no final weights, so these fitted parameters are not independently auditable.
assumptions (6)
  • domain assumption Lidar-derived CHM is a bias-free ground truth for canopy height.
    Used throughout training and evaluation; lidar errors, grid interpolation, and point density effects are not modeled.
  • domain assumption The RGB imagery and the lidar CHM describe the same canopy state at the same time.
    The paper does not report acquisition dates for the tiles used in the main experiment; the RGB corpus spans 2013-2020 while lidar is November 2020.
  • ad hoc to paper Multiview consistency of downsampled high-resolution features is sufficient to train an upsampler that preserves spatial details.
    This is the central training signal for the feature enhancement module; no independent evaluation of the learned upsampling is given.
  • domain assumption Plot-level train/test split removes spatial autocorrelation.
    Training on Plots A and C and testing on Plot B is asserted to avoid spatial autocorrelation, but residual spatial correlation is not quantified.
  • ad hoc to paper A roughly 5% calibration subset of Saihanba lidar is enough to transfer the estimator to a new ecosystem.
    The generalizability experiment uses a small slice of the target lidar to project features to heights, so this is not a zero-shot transfer.
  • domain assumption The DBH-H and AGB equations from Qin et al. and Table 1 apply across plantation species in the study area.
    AGB results depend on these externally fitted equations, whose uncertainty is not propagated.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Novel Large Vision Foundation Model (LVFM)-based Approach for Generating High-Resolution Canopy Height Maps in Plantations for Precision Forestry Management." pith.science (2026). https://pith.science/paper/BY2PD2CW

@misc{pith2026250620388,
  author       = {Pith},
  title        = {Pith review of: A Novel Large Vision Foundation Model (LVFM)-based Approach for Generating High-Resolution Canopy Height Maps in Plantations for Precision Forestry Management},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BY2PD2CW}},
  note         = {Machine review of arXiv:2506.20388}
}
read the original abstract

Accurate, cost-effective monitoring of plantation aboveground biomass (AGB) is crucial for supporting local livelihoods and carbon sequestration initiatives like the China Certified Emission Reduction (CCER) program. High-resolution canopy height maps (CHMs) are essential for this, but standard lidar-based methods are expensive. While deep learning with RGB imagery offers an alternative, accurately extracting canopy height features remains challenging. To address this, we developed a novel model for high-resolution CHM generation using a Large Vision Foundation Model (LVFM). Our model integrates a feature extractor, a self-supervised feature enhancement module to preserve spatial details, and a height estimator. Tested in Beijing's Fangshan District using 1-meter Google Earth imagery, our model outperformed existing methods, including conventional CNNs. It achieved a mean absolute error of 0.09 m, a root mean square error of 0.24 m, and a correlation of 0.78 against lidar-based CHMs. The resulting CHMs enabled over 90% success in individual tree detection, high accuracy in AGB estimation, and effective tracking of plantation growth, demonstrating strong generalization to non-training areas. This approach presents a promising, scalable tool for evaluating carbon sequestration in both plantations and natural forests.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

63 extracted references · 63 canonical work pages

  1. [1]

    Bastin, J.-F., et al., The global tree restoration potential. 2019. 365(6448): p. 76-79

    work page 2019

  2. [2]

    Liu, X., et al., Neural network guided interpolation for mapping canopy height of China's forests by integrating GEDI and ICESat-2 data. 2022. 269: p. 112844

    work page 2022

  3. [3]

    Rohatyn, S., et al., Limited climate change mitigation potential through forestation of the vast dryland regions. 2022. 377(6613): p. 1436-1439

    work page 2022

  4. [4]

    Xu, H., et al., Forestation at the right time with the right species can generate persistent carbon benefits in China. 2023. 120(41): p. e2304988120

    work page 2023

  5. [5]

    Grelle, A., et al., From source to sink –recovery of the carbon balance in young forests. 2023. 330: p. 109290

    work page 2023

  6. [6]

    Yang, H., et al., Global increase in biomass carbon stock dominated by growth of northern young forests over past decade. 2023. 16(10): p. 886-892

    work page 2023

  7. [7]

    China Certified Emission Reduction (CCER)

    China Certified Emission Reduction (CCER), 2022. China Certified Emission Reduction (CCER). Retrieved from https://www.ccer.com.cn/ (2022)

    work page 2022

  8. [8]

    Huang, W., et al., Review of recent progress of emission trading policy in China. 2022. 349: p. 131480

    work page 2022

Show all 63 references
  1. [9]

    Brockerhoff, E.G., et al., Plantation forests and biodiversity: oxymoron or opportunity? Biodiversity and Conservation, 2008. 17: p. 925-951

    2008

  2. [10]

    Qin, S., et al., Forest emissions reduction assessment using airborne LiDAR for biomass estimation. 2022. 181: p. 106224

    2022

  3. [11]

    Saatchi, S.S., et al., Benchmark map of forest carbon stocks in tropical regions across three continents

  4. [12]

    Gasparri, N.I., et al., Assessing multi-temporal Landsat 7 ETM+ images for estimating above -ground biomass in subtropical dry forests of Argentina. 2010. 74(10): p. 1262-1270

    2010

  5. [13]

    Im, and R

    Gleason, C.J., J.J.G. Im, and R. Sensing, A review of remote sensing of forest biomass and biofuel: options for small-area applications. 2011. 48(2): p. 141-170

    2011

  6. [14]

    Lu, D.J.I.j.o.r.s., The potential and challenge of remote sensing ‐based biomass estimation. 2006. 27(7): p. 1297-1328

    2006

  7. [15]

    Rao, and S.C.J.N.P

    Konings, A.G., K. Rao, and S.C.J.N.P. Steele‐Dunne, Macro to micro: microwave remote sensing of plant water content for physiology and ecology. 2019. 223(3): p. 1166-1172. Journal of Remote Sensing Manuscript Template Page 22 of 23

    2019

  8. [16]

    Fan, L., et al., Satellite-observed pantropical carbon dynamics. 2019. 5(9): p. 944-951

    2019

  9. [17]

    Li, X., et al., The first global soil moisture and vegetation optical depth product retrieved from fused SMOS and SMAP L-band observations. 2022. 282: p. 113272

    2022

  10. [18]

    Global change biology, 2017

    Tian, F., et al., Mapping gains and losses in woody vegetation across global tropical drylands. Global change biology, 2017. 23(4): p. 1748-1760

    2017

  11. [19]

    Mielcarek, M., et al., Testing and evaluating different LiDAR -derived canopy height model generation methods for tree height estimation. 2018. 71: p. 132-143

    2018

  12. [20]

    Migliavacca, M., et al., The three major axes of terrestrial ecosystem function. 2021. 598(7881): p. 468- 472

    2021

  13. [21]

    Lang, N., et al., A high-resolution canopy height model of the Earth. 2023. 7(11): p. 1778-1789

    2023

  14. [22]

    Picos, J., et al., Individual tree detection in a eucalyptus plantation using unmanned aerial vehicle (UAV)- LiDAR. 2020. 12(5): p. 885

    2020

  15. [23]

    Hu, and C.J.I.J.o.R.S

    St‐Onge, B., Y. Hu, and C.J.I.J.o.R.S. Vega, Mapping the height and above‐ground biomass of a mixed forest using lidar and stereo Ikonos images. 2008. 29(5): p. 1277-1294

    2008

  16. [24]

    Ota, T., et al., Aboveground biomass estimation using structure from motion approach with aerial photographs in a seasonal tropical forest. 2015. 6(11): p. 3882-3898

    2015

  17. [25]

    Hong, Y., et al., Combining Multisource Data and Machine Learning Approaches for Multiscale Estimation of Forest Biomass. 2023. 14(11): p. 2248

    2023

  18. [26]

    Guo, Q., et al., Lidar boosts 3D ecological observations and modelings: A review and perspective. 2020. 9(1): p. 232-257

    2020

  19. [27]

    Tang, X., et al., Monitoring of monthly height growth of individual trees in a subtropical mixed plantation using UAV data. 2023. 15(2): p. 326

    2023

  20. [28]

    Environmental Research Letters, 2022

    Dubayah, R., et al., GEDI launches a new era of biomass inference from space. Environmental Research Letters, 2022. 17(9): p. 095001

    2022

  21. [29]

    Potapov, P., et al., Mapping global forest canopy height through integration of GEDI and Landsat data

  22. [30]

    Li, S., et al., Deep learning enables image -based tree counting, crown segmentation, and height prediction at national scale. 2023. 2(4): p. pgad076

    2023

  23. [31]

    Wagner, F.H., et al., Sub-meter tree height mapping of California using aerial images and LiDAR - informed U-Net model. 2024. 305: p. 114099

    2024

  24. [32]

    Oquab, M., et al., Dinov2: Learning robust visual features without supervision. 2023

    2023

  25. [33]

    Learning transferable visual models from natural language supervision

    Radford, A., et al. Learning transferable visual models from natural language supervision . in International conference on machine learning. 2021. PMLR

    2021

  26. [34]

    Bochkovskiy, and V

    Ranftl, R., A. Bochkovskiy, and V. Koltun. Vision transformers for dense prediction. in Proceedings of the IEEE/CVF international conference on computer vision. 2021

    2021

  27. [35]

    Monovit: Self-supervised monocular depth estimation with a vision transformer

    Zhao, C., et al. Monovit: Self-supervised monocular depth estimation with a vision transformer. in 2022 international conference on 3D vision (3DV). 2022. IEEE

    2022

  28. [36]

    and C.J.a.p.a

    Agarwal, A. and C.J.a.p.a. Arora, Depthformer: Multiscale vision transformer for monocular depth estimation with local global information fusion. 2022

    2022

  29. [37]

    Azad, R., et al., Advances in medical image analysis with vision transformers: a comprehensive review. 2023: p. 103000

    2023

  30. [38]

    Li, Y., et al., Localvit: Bringing locality to vision transformers. 2021

    2021

  31. [39]

    Lv, P., et al., SCViT: A spatial-channel feature preserving vision transformer for remote sensing image scene classification. 2022. 60: p. 1-12

    2022

  32. [40]

    Wang, G., et al., P 2fevit: Plug-and-play cnn feature embedded hybrid vision transformer for remote sensing image classification. 2023. 15(7): p. 1773

    2023

  33. [41]

    Fayad, I., et al., Vision Transformers, a new approach for high -resolution and large-scale mapping of canopy heights. 2023

    2023

  34. [42]

    Tolan, J., et al., Very high resolution canopy height maps from RGB imagery using self-supervised vision transformer and convolutional decoder trained on aerial lidar. 2024. 300: p. 113888

    2024

  35. [43]

    arXiv preprint arXiv:2411.14354, 2024

    Rolf, E., et al., Contrasting local and global modeling with machine learning and satellite data: A case Journal of Remote Sensing Manuscript Template Page 23 of 23 study estimating tree canopy height in African savannas. arXiv preprint arXiv:2411.14354, 2024

    2024 arXiv

  36. [44]

    Han, K., et al., A survey on vision transformer. 2022. 45(1): p. 87-110

    2022

  37. [45]

    Masked autoencoders are scalable vision learners

    He, K., et al. Masked autoencoders are scalable vision learners . in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2022

    2022

  38. [46]

    Carafe: Content -aware reassembly of features

    Wang, J., et al. Carafe: Content -aware reassembly of features . in Proceedings of the IEEE/CVF international conference on computer vision. 2019

    2019

  39. [47]

    Shelhamer, and W.T.J.a.p.a

    Hamilton, M., E. Shelhamer, and W.T.J.a.p.a. Freeman, It is likely that your loss should be a likelihood. 2020

    2020

  40. [48]

    Liu, X., et al., Comparison between three convolutional neural networks for local climate zone classification using Google Earth Images: A case study of the Fujian Delta in China. 2023. 148: p. 110086

    2023

  41. [49]

    Frantz, D., et al., Improvement of the Fmask algorithm for Sentinel -2 images: Separating clouds from bright surfaces based on parallax effects. 2018. 215: p. 471-481

    2018

  42. [50]

    and S.J.I

    Howard, J. and S.J.I. Gugger, Fastai: a layered API for deep learning. 2020. 11(2): p. 108

    2020

  43. [51]

    Roussel, J.-R., et al., lidR: An R package for analysis of Airborne Laser Scanning (ALS) data. 2020. 251: p. 112061

    2020

  44. [52]

    IEEE transactions on pattern analysis and machine intelligence, 2022

    Han, K., et al., A survey on vision transformer. IEEE transactions on pattern analysis and machine intelligence, 2022. 45(1): p. 87-110

    2022

  45. [53]

    Hao, Z., et al., Automated tree-crown and height detection in a young forest plantation using mask region-based convolutional neural network (Mask R-CNN). 2021. 178: p. 112-123

    2021

  46. [54]

    https://verra.org/methodologies/vm0047-afforestation-reforestation-and-revegetation-v1-0/, 2023

    Verra, VM0047 Afforestation, Reforestation, and Revegetation, v1.0. https://verra.org/methodologies/vm0047-afforestation-reforestation-and-revegetation-v1-0/, 2023

    2023

  47. [55]

    Liu, and R

    Zhu, X., D.J.I.J.o.P. Liu, and R. Sensing, Improving forest aboveground biomass estimation using seasonal Landsat NDVI time-series. 2015. 102: p. 222-231

    2015

  48. [56]

    Martínez-Fernández, J., et al., Tracking tree growth through satellite soil moisture monitoring: A case study of Pinus halepensis in Spain. 2019. 235: p. 111422

    2019

  49. [57]

    and C.J.B.D

    Humphrey, V. and C.J.B.D. Frankenberg, Continuous ground monitoring of vegetation optical depth and water content with GPS signals. 2022. 2022: p. 1-44

    2022

  50. [58]

    Forest Ecosystems, 2015

    Zeng, W., et al., The national forest inventory in China: history -results-international context. Forest Ecosystems, 2015. 2: p. 1-16

    2015

  51. [59]

    Brandt, M., et al., An unexpectedly large count of trees in the West African Sahara and Sahel. Nature,

  52. [60]

    Shi, and M

    Amirkolaee, H.A., M. Shi, and M. Mulligan, TreeFormer: a Semi -Supervised Transformer -based Framework for Tree Counting from a Single High Resolution Image. IEEE Transactions on Geoscience and Remote Sensing, 2023

    2023

  53. [61]

    Remote Sensing of Environment, 2024

    Blickensdörfer, L., et al., National tree species mapping using Sentinel -1/2 time series and German National Forest Inventory data. Remote Sensing of Environment, 2024. 304: p. 114069

    2024

  54. [62]

    Earth System Science Data Discussions, 2024

    Grabska-Szwagrzyk, E.M., et al., Map of forest tree species for Poland based on Sentinel -2 data. Earth System Science Data Discussions, 2024. 2024: p. 1-21

    2024

  55. [2011]

    9899-9904

    108(24): p. 9899-9904

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.