Pith. sign in

REVIEW 3 major objections 5 minor 34 references

Estimating forest carbon stocks from high-resolution remote sensing imagery by reducing domain shift with style transfer

T0 review · 3 major / 5 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read A Swin-Transformer-based style transfer model estimates forest carbon stock more accurately than existing regression methods.

desk verdict A useful ablation study with an unvalidated multi-year trend and a possible feature-selection leak; the 2020 results are plausible but the headline conclusion is not. read the letter →

arxiv 2502.00784 v1 pith:7NLYOCQU submitted 2025-02-02 cs.CV cs.AIeess.IV

Zhenyu Yu , Jinnian Wang This is my paper · ORCID
classification cs.CVcs.AIeess.IV
keywords carbonstockforeststyletransferremotesensingdeeplearningSwinTransformerPix2Pixdomainshift
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

The paper tries to show that forest carbon stock can be estimated more accurately from high-resolution satellite images if the images are first put into a common visual domain by style transfer. It proposes a model family based on Swin Transformer and Pix2Pix, converting carbon stock estimation into an image-to-image translation task. On a mountainous county in Yunnan, China, the proposed MSwin-Pix2Pix model achieves the best accuracy among tested statistical, machine-learning, and deep-learning baselines, and it is used to map carbon stock changes from 2005 to 2020. The central practical payoff would be a way to monitor forest carbon over long time periods using only one field campaign, by translating older or cloudy images into the style of a reference year.

What carries the argument

The load-bearing mechanism is the style-transfer step: Swin-Pix2Pix, a conditional GAN whose generator is a Swin Transformer U-Net, translates source-domain images (Landsat TM 2005, GF-1 WFV 2013 and 2015) into the target domain of the 2020 GF-1 WFV image by matching global features via shifted-window self-attention. For estimation, the same generator backbone is extended with an NDVI-derived mask to exclude non-forest areas and a median filter to remove high-intensity anomalies, producing the MSwin-Pix2Pix model.

What would settle it

Measure carbon stock from historical field plots in the study area for 2005, 2013, or 2015 and compare them to the style-transfer model's maps for those years; if the errors are much larger than the 2020 MAE and RMSE, the transfer does not preserve carbon-relevant features.

Watch

Extended reading notes

Core claim

The central claim is that replacing the generator of Pix2Pix with a Swin-Transformer-based U-Net, adding a vegetation mask and a median filter, yields a model whose carbon stock estimates are accurate enough for regional monitoring (MAE = 16.2891, RMSE = 29.3763, R2 = 0.7105, SSIM = 0.7510) and that the same style-transfer step aligns images from different years and sensors so that a model trained on 2020 field data can be applied to 2005, 2013, and 2015 imagery. Applying this pipeline, the paper reports that from 2005 to 2020 carbon stock increased in 44.04% of the study area, decreased in 10.22%, and stayed unchanged in 45.74%, implying an overall rise.

Load-bearing premise

The whole multi-year trend rests on the assumption that translating historical images into the 2020 reference style preserves exactly the spectral and textural features that predict carbon stock, so a model trained only on 2020 field data remains valid for 2005, 2013, and 2015 without any historical ground truth to check it.

Editorial extensions

If this is right

  • If the style-transfer step works as claimed, multi-year carbon stock mapping no longer requires ground truth for every year; one field campaign can anchor a time series.
  • The reported accuracy gaps (MSwin-Pix2Pix MAE 16.29 vs Pix2Pix 21.12, RF 52.98) suggest that deep global-feature extraction and targeted post-filtering matter more than traditional spectral regression.
  • Combining mask and median filter improved accuracy more than either alone, pointing to boundary handling as a key remaining source of error.
  • The reported trend, with 44.04% of the area gaining carbon and 45.74% unchanged, is the kind of result that could inform regional carbon accounting and ecological policy.
  • De-clouding through style transfer extends usable optical imagery in persistently cloudy highland regions.

Reading between the lines

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

  • A direct test of the transfer assumption would be to acquire a small set of historical field plots and compare the model's 2005 and 2013 estimates against them; the paper does not report such a validation.
  • Because the mask is based on NDVI of the reference year, areas that changed from non-vegetation to vegetation between 2005 and 2020 may be systematically included or excluded depending on which year's mask is used, and this could bias the reported increase.
  • The method's reliance on a single reference domain suggests it could be extended to other regions by training on one contemporary image and field campaign, but sensor differences beyond Landsat and GF-1 would need separate testing.
  • The median filter removes pixels above intensity 240; if large carbon stock values genuinely occur in those bright pixels, the filter could truncate real extremes rather than only noise.
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

3 major / 5 minor

Summary. The paper proposes a deep-learning pipeline for estimating forest carbon stocks in Huize County, China, from multispectral satellite imagery (GF-1 WFV and Landsat TM). The authors first use a Swin-Transformer-based Pix2Pix variant (Swin-Pix2Pix) for style transfer to harmonize multi-temporal images into a 2020 GF-1 WFV reference domain, then apply a masked SUNet-based regression model (MSwin-Pix2Pix) to map the harmonized images to carbon stock values. The reported results include a best 2020 accuracy of MAE = 16.2891, RMSE = 29.3763, R2 = 0.7105, SSIM = 0.7510, and a claimed 2005–2020 trend of 44.04% of the area increasing, 10.22% decreasing, and 45.74% unchanged in carbon stock. The central claims are (i) that MSwin-Pix2Pix outperforms several statistical, machine-learning, and deep-learning baselines on the 2020 data, and (ii) that the style transfer enables reliable multi-temporal carbon-stock trend analysis without historical ground truth.

Significance. If validated, the 2020 carbon-stock estimation result would be a useful addition to the growing literature on optical-image-based forest carbon mapping, and the cross-domain style-transfer idea is conceptually appealing for dealing with multi-sensor, multi-temporal imagery. The paper includes a fairly broad comparison with OLS, GWR, RF, SVR, CNN, and Pix2Pix, and the ablation study for the proposed modules is detailed. However, the paper's most advertised contribution is the 2005–2020 trend analysis, and that claim is not supported by the presented evidence. The feature-selection procedure also raises a leakage concern that affects even the 2020 accuracy numbers. These issues make the central claims unreliable as they stand.

major comments (3)
  1. [Section 3.1 and Section 3.3] The 2005–2020 trend percentages rest on an unvalidated transferability assumption. The style-transfer model Swin-Pix2Pix is trained with the 2020 GF-1 WFV image as ground truth, and the 2020-trained MSwin-Pix2Pix estimator is then applied to the translated 2005, 2013, and 2015 images. No quantitative fidelity metric is reported for the style transfer, and no comparison against historical field measurements is made. Because the style-transfer objective is to make historical images resemble the 2020 domain, genuine temporal changes in forest structure, reflectance, or carbon stock can be erased by the translation itself. The trend percentages in Section 3.3 therefore cannot be interpreted as evidence of an actual carbon-stock increase. A concrete remedy would be to validate the pipeline against any historical inventory data (e.g., the 2005 standing-volume statistics cited in Section 3.3) or to demonstrate on held-out temporal data that the style transfer preserves carbon-stock-relevant features.
  2. [Section 2.3.1 and Section 3.2] The feature-screening step selects the three GLCM-Mean bands (19, 35, 27) using correlations with measured carbon stock computed across the full dataset, but the five-fold cross-validation in Section 3.2 is described only after this selection. If the band selection is not nested inside the cross-validation folds, the test-fold carbon-stock labels have influenced the choice of predictor bands, which biases the reported MAE, RMSE, R2, and SSIM optimistically. The feature selection must be performed independently on each training fold, or the band set must be fixed a priori without access to the test labels. Without this correction, Table 4 does not provide an unbiased estimate of generalization performance.
  3. [Section 2.2 and Section 3.1] The spatial-resolution mismatch between the 2005 Landsat TM (30 m) and the 2020 GF-1 WFV (16 m) images is not addressed in the style-transfer description. The paper says only that the four common bands are selected and that Swin-Pix2Pix is used for style transfer; it is not stated whether the 30 m images are resampled to 16 m before translation or whether the 2020-trained model is applied to 30 m imagery directly. If the resolution remains 30 m, the feature scale presented to the 2020-trained estimator is inconsistent, and the trend analysis in Section 3.3 is further compromised.
minor comments (5)
  1. [Section 2.3.5 and Section 2.3.6] Equation numbering restarts at (1) in both Section 2.3.5 and Section 2.3.6, which makes cross-references confusing; the equations should be numbered continuously throughout the paper.
  2. [Abstract and Introduction] The term 'high-resolution' is used for 16 m and 30 m imagery; these are more commonly called medium resolution in the remote-sensing literature, and the wording should be adjusted to avoid overstatement.
  3. [Table 3] The R2 values for the L2-only and SmoothL1-only rows (0.4197 and 0.3609) are notably low relative to their MAE and RMSE values; the authors should verify these numbers or provide an explanation, as the inconsistency is distracting.
  4. [Section 2.3.1] The NDVI mask threshold is computed from the 2020 image (M = μ − 2σ) and then applied to the other years without re-estimation; sensor and illumination differences between Landsat TM and GF-1 WFV may make a fixed threshold inappropriate across years.
  5. [Section 3.1] The style-transfer results are evaluated only visually (Figure 5); reporting a quantitative metric such as FID, SSIM, or a vegetation-index-preservation statistic would strengthen the claim that Swin-Pix2Pix outperforms Pix2Pix.

Circularity Check

1 steps flagged · score 5.0 of 10

Partial circularity in the reported accuracy due to label-based feature screening outside cross-validation; the 2005–2020 trend is unsupported but not formally circular.

  1. fitted input called prediction [Section 2.3.1 (Feature screening) and Section 3.2 (Ablation study for carbon stock estimation)]
    "To avoid the limitation of a single correlation index, we selected bands that showed the highest correlation to carbon stock. ... Using this approach, the three bands that showed the highest correlation were 19th (GLCM-Mean_Band1), 35th (GLCM-Mean_Band3), and 27th (GLCM-Mean_Band2). ... To obtain stable evaluation, we utilized a five-fold cross-validation method with the dataset divided into training, validation, and tests sets."

    The three input bands are selected by their correlation with the measured carbon stock, which is the prediction target, and this screening is described before the five-fold cross-validation with no statement that the selection is nested inside the training folds. The held-out test folds therefore influence which features enter the model, so the reported MAE=16.2891, RMSE=29.3763, R2=0.7105, and SSIM=0.7510 are partly fitted to the test labels rather than independent predictions. The 'prediction' is thus partially a restatement of the feature-target correlation used to choose the inputs.

full rationale

The paper contains one concrete circularity-like step: feature screening uses the measured carbon stock to select the three GLCM-Mean bands, and the subsequent cross-validation is not described as nested, so the headline accuracy is partly fitted to the test data. This is a fitted-input-called-prediction pattern and justifies a moderate circularity score. No load-bearing self-citations or imported uniqueness theorems are present; Swin Transformer, Pix2Pix, and SUNet are external, standard components. The 2005–2020 trend claim is a serious validity gap because style transfer is trained with the 2020 image as ground truth and no historical field measurements are used, but that is an unvalidated transferability assumption rather than a formal equivalence between inputs and outputs, so it does not further raise the circularity score. The architecture comparison and ablation results retain independent content beyond the feature-selection leakage.

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

The central claim rests on several unstated domain assumptions about label quality, cross-year transferability of the mask and the learned model, and on a feature-selection step that may have used test labels. No new physical entities are introduced.

free parameters (4)
  • NDVI forest-mask threshold (M = mu - 2*sigma) = 0.3951 for 2020, applied to 2005/2013/2015
    Eq. 4 defines the vegetation/non-vegetation threshold from the mean and standard deviation of NDVI. The paper sets the threshold to 0.3951 for 2020 and reuses it for all years. It directly controls which pixels are included in the carbon stock maps and in the reported evaluation, and transferability across sensors and years is assumed.
  • Selected feature bands (19, 35, 27) = GLCM-Mean_Band1, GLCM-Mean_Band3, GLCM-Mean_Band2
    Section 2.3.1 selects three bands from 50 features using seven correlation/distance measures against measured carbon stock. If this selection is done on the full dataset before cross-validation, it leaks test labels into the feature set and inflates the reported accuracy.
  • Median filter high-intensity removal threshold = 240 (8-bit intensity)
    Section 2.3.3 removes only image elements above intensity 240 after median filtering. The threshold is chosen empirically to remove noise without distorting the image, and it affects the final maps and metrics.
  • cGAN loss weight lambda = 100
    Eqs. 10 and 11 weigh the L1 or SmoothL1 loss by lambda=100. This is standard Pix2Pix practice, but it is still a hyperparameter chosen by the authors and not tuned or justified for this task.
assumptions (4)
  • domain assumption Third National Land Survey small-plot carbon stock values are accurate, unit-consistent ground truth and align spatially with GF-1 WFV pixels.
    The model is trained and evaluated against these measured values (Section 2.2), but the paper does not describe how carbon stock was computed from survey attributes or how plots were matched to image pixels.
  • domain assumption Style transfer from Landsat TM and GF-1 WFV to the 2020 GF-1 WFV reference domain preserves carbon-relevant spectral and textural information.
    Section 3.1 and Figure 5 use a 2020 image as style reference. The whole multi-year analysis assumes that translated images still encode vegetation and carbon signals, but no carbon validation is done on translated historical images.
  • domain assumption The NDVI-based mask computed for 2020 remains valid for other years after style transfer.
    Section 2.3.1 says this extraction method was used to apply the mask calculation for the remaining years, without re-estimating the threshold from historical data.
  • domain assumption The learned image-to-carbon relationship is stationary from 2005 to 2020.
    A model trained only on 2020 data is applied to 2005, 2013, and 2015 images to produce the temporal trend in Section 3.3. Forest structure, species composition, and sensing conditions may change over 15 years, but the model assumes a fixed relationship.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Estimating forest carbon stocks from high-resolution remote sensing imagery by reducing domain shift with style transfer." pith.science (2026). https://pith.science/paper/7NLYOCQU

@misc{pith2026250200784,
  author       = {Pith},
  title        = {Pith review of: Estimating forest carbon stocks from high-resolution remote sensing imagery by reducing domain shift with style transfer},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7NLYOCQU}},
  note         = {Machine review of arXiv:2502.00784}
}
read the original abstract

Forests function as crucial carbon reservoirs on land, and their carbon sinks can efficiently reduce atmospheric CO2 concentrations and mitigate climate change. Currently, the overall trend for monitoring and assessing forest carbon stocks is to integrate ground monitoring sample data with satellite remote sensing imagery. This style of analysis facilitates large-scale observation. However, these techniques require improvement in accuracy. We used GF-1 WFV and Landsat TM images to analyze Huize County, Qujing City, Yunnan Province in China. Using the style transfer method, we introduced Swin Transformer to extract global features through attention mechanisms, converting the carbon stock estimation into an image translation.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

34 extracted references · 22 canonical work pages

  1. [1]

    D., Happ, P

    Bermudez, J. D., Happ, P. N., Oliveira, D. A. B., & Feitosa, R. Q. (2018). Sar to Optical Image Synthesis for Cloud Removal with Generative Adversarial Networks. ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, IV–1, 5–11. https://doi.org/10.5194/isprs-annals-IV-1-5-2018

    work page doi:10.5194/isprs-annals-iv-1-5-2018 2018

  2. [2]

    P., Moalafhi, D

    Byakatonda, J., Parida, B. P., Moalafhi, D. B., & Kenabatho, P. K. (2018). Analysis of long term drought severity characteristics and trends across semiarid 35 Botswana using two drought indices. Atmospheric Research, 213, 492–508. https://doi.org/10.1016/j.atmosres.2018.07.002

    work page doi:10.1016/j.atmosres.2018.07.002 2018

  3. [3]

    Cao, H., Wang, Y., Chen, J., Jiang, D., Zhang, X., Tian, Q., & Wang, M. (2023). Swin-Unet: Unet-like Pure Transformer for Medical Image Segmentation. In Computer Vision–ECCV 2022 Workshops, 205-218. http://arxiv.org/abs/2105.05537

    arXiv 2023

  4. [4]

    C., Innes, J

    Cao, L., Coops, N. C., Innes, J. L., Sheppard, S. R. J., Fu, L., Ruan, H., & She, G. (2016). Estimation of forest biomass dynamics in subtropical forests using multi-temporal airborne LiDAR data. Remote Sensing of Environment, 178, 158–171.https://doi.org/10.1016/j.rse.2016.03.012

    work page doi:10.1016/j.rse.2016.03.012 2016

  5. [5]

    M., Crowther, T

    Chen, Y., Feng, X., Fu, B., Ma, H., Zohner, C. M., Crowther, T. W., Huang, Y., Wu, X., & Wei, F. (2023). Maps with 1 km resolution reveal increases in above- and belowground forest biomass carbon pools in China over the past 20 years. Earth System Science Data, 15(2),897–910.https://doi.org/10.5194/essd-15-897-2023

    work page doi:10.5194/essd-15-897-2023 2023

  6. [6]

    A., & Duchesne, R

    Chopping, M., Wang, Z., Schaaf, C., Bull, M. A., & Duchesne, R. R. (2022). Forest aboveground biomass in the southwestern United States from a MISR multi-angle index, 2000–2015. Remote Sensing of Environment, 275, 112964. https://doi.org/10.1016/j.rse.2022.112964

    work page arXiv 2022

  7. [7]

    E., Shimabukuro, M

    Christovam, L. E., Shimabukuro, M. H., Galo, M. de L. B. T., & Honkavaara, E. (2022). Pix2pix Conditional Generative Adversarial Network with MLP Loss Function for Cloud Removal in a Cropland Time Series. Remote Sensing, 14(1), 36 Article1.https://doi.org/10.3390/rs14010144 Dosovitskiy,A.,Beyer,L.,Kolesnikov,A.,Weissenborn,D.,Zhai,X.,Unterthiner,T., Dehgh...

    arXiv doi:10.48550/arxiv.2010.11929 2022

  8. [8]

    Fan, B., Dai, Y., & He, M. (2021). Sunet: symmetric undistortion network for rolling shuttercorrection. In Proceedings of the IEEE/CVF International Conference on Computer Vision(pp.4541-4550).https://doi.org/10.48550/arXiv.2108.04775 Fan,C.M.,Liu,T.J.,&Liu,K.H.(2022).SUNet:SwinTransformerUNetforImage Denoising. 2022 IEEE International Symposium on Circui...

    work page Pith review arXiv doi:10.48550/arxiv.2108.04775 2021

Show all 34 references
  1. [9]

    S., & Ge, J

    Gao, W., Shen, F., Tan, K., Zhang, W., Liu, Q., Lam, N. S., & Ge, J. (2021). Monitoring terrain elevation of intertidal wetlands by utilising the spatial-temporal fusion of multi-source satellite data: A case study in the 37 Yangtze (Changjiang) Estuary. Geomorphology, 383, 10...

    2021

  2. [10]

    Gogoi, A., Ahirwal, J., & Sahoo, U. K. (2022). Evaluation of ecosystem carbon storage in major forest types of Eastern Himalaya: Implications for carbon sink management. Journal of Environmental Management, 302, 113972. https://doi.org/10.1016/j.jenvman.2021.113972

    2022

  3. [11]

    & Bengio, Y

    Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., ... & Bengio, Y. (2014). Generative adversarial nets in advances in neural information processing systems (NIPS). Curran Associates, Inc. Red Hook, NY, USA,2672-2680

    2014

  4. [12]

    J., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., & Bengio, Y

    Goodfellow, I. J., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., & Bengio, Y. (2014). Generative Adversarial Networks. Advances in neural information processing systems, 1–9. arXiv:1406.2661. https://doi.org/10.48550/arXiv.1406.2661

    2014 doi arXiv

  5. [13]

    & Bengio, Y

    Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., ... & Bengio, Y. (2020). Generative adversarial networks. Communications of the ACM, 63(11),139-144.https://doi.org/10.1145/3422622

    2020 doi

  6. [14]

    Hamedianfar, A., Mohamedou, C., Kangas, A., & Vauhkonen, J. (2022). Deep learning for forest inventory and planning: A critical review on the remote sensing approaches so far and prospects for further applications. Forestry: An International Journal of Forest Research, 95(4), ...

    2022 doi

  7. [15]

    Huang, Y., Lu, Z., Shao, Z., Ran, M., Zhou, J., Fang, L., & Zhang, Y. (2019). Simultaneous denoising and super-resolution of optical coherence tomography images based on generative adversarial network. Optics Express, 27(9), 12289. https://doi.org/10.1364/OE.27.012289

    2019 doi

  8. [16]

    Hui, Z., Li, J., Gao, X., & Wang, X. (2021). Progressive perception-oriented network for single image super-resolution. Information Sciences, 546, 769–786. https://doi.org/10.1016/j.ins.2020.08.114 Isola,P.,Zhu,J.Y.,Zhou,T.,&Efros,A.A.(2018).Image-to-imagetranslationwith condi...

    2021 doi arXiv

  9. [17]

    Jiang, P., Deng, F., Wang, X., Shuai, P., Luo, W., & Tang, Y. (2023). Seismic First Break Picking Through Swin Transformer Feature Extraction. IEEE Geoscience and Remote Sensing Letters, 20, 1–5. https://doi.org/10.1109/LGRS.2023.3248233

    2023

  10. [18]

    Lang, N., Kalischek, N., Armston, J., Schindler, K., Dubayah, R., & Wegner, J. D. (2022). Global canopy height regression and uncertainty estimation from GEDI LIDAR waveforms with deep ensembles. Remote Sensing of Environment, 268, 112760.https://doi.org/10.1016/j.rse.2021.112...

    2022

  11. [19]

    Korhonen, L., Mammarella, I., & Vesala, T. (2022). Does growing atmospheric CO 2 explain increasing carbon sink in a boreal coniferous forest? Global Change Biology, 28(9),2910–2929.https://doi.org/10.1111/gcb.16117

    2022 doi

  12. [20]

    J., Yim, J., Nam, J.,Choi, K., & Khim, J

    Lee, J., Kim, B., Noh, J., Lee, C., Kwon, I., Kwon, B.-O., Ryu, J., Park, J., Hong, S., Lee,S., Kim, S.-G.,Son, S., Yoon,H. J., Yim, J., Nam, J.,Choi, K., & Khim, J. S. (2021). The first national scale evaluation of organic carbon stocks and sequestration rates of coastal sedi...

    2021

  13. [21]

    Liu, Z., Zhao, F., Liu, X., Yu, Q., Wang, Y., Peng, X., Cai, H., & Lu, X. (2022). Direct estimation of photosynthetic CO2 assimilation from solar-induced 40 chlorophyll fluorescence (SIF). Remote Sensing of Environment, 271, 112893. https://doi.org/10.1016/j.rse.2022.112893 Lu...

    2022

  14. [22]

    Pathak, D., Krähenbühl, P., Donahue, J., Darrell, T., & Efros, A. A. (2016). Context Encoders: Feature Learning by Inpainting. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2536–2544. https://doi.org/10.1109/CVPR.2016.278

    2016 doi

  15. [23]

    Pei, Z., Jin, M., Zhang, Y., Ma, M., & Yang, Y.-H. (2021). All-in-focus synthetic apertureimagingusinggenerativeadversarialnetwork-basedsemanticinpainting. Pattern Recognition, 111,107669.https://doi.org/10.1016/j.patcog.2020.107669 Puliti,S.,Breidenbach,J.,Schumacher,J.,Haugl...

    2021

  16. [24]

    Multi-temporal MODIS–Landsat data fusion for relative radiometric normalization, gap filling, and prediction of Landsat data

    Roy, D.P., Ju,J.,Lewis,P., Schaaf,C., Gao, F.,Hansen,M.,& Lindquist, E.(2008). Multi-temporal MODIS–Landsat data fusion for relative radiometric normalization, gap filling, and prediction of Landsat data. Remote Sensing of Environment, 112(6),3112–3130.https://doi.org/10.1016/...

    2008 doi

  17. [25]

    Salimi, S., Berggren, M., & Scholz, M. (2021). Response of the peatland carbon dioxide sink function to future climate change scenarios and water level management. Global Change Biology, 27(20), 5154–5168. https://doi.org/10.1111/gcb.15753 Santoro,M.,Cartus,O.,&Fransson,J.E.S....

    2021

  18. [26]

    E., Forkel, M., Camps-Valls, G., Jung, M., Miralles, D

    Teubner, I. E., Forkel, M., Camps-Valls, G., Jung, M., Miralles, D. G., Tramontana, G.,vanderSchalie,R.,Vreugdenhil,M.,Mösinger,L.,&Dorigo,W.A.(2019). A carbon sink-driven approach to estimate gross primary production from microwavesatelliteobservations. Remote Sensing of Envi...

    2019 doi

  19. [27]

    Polosukhin, I. (2017). Attention is all you need. Advances in neural information processing systems, 30.https://doi.org/10.48550/arXiv.1706.03762

    2017 doi arXiv

  20. [28]

    I., Liu, Y., Fang, S., Bösch, H., O’Dell, C

    Wang, J., Feng, L., Palmer, P. I., Liu, Y., Fang, S., Bösch, H., O’Dell, C. W., Tang, X., Yang, D., & Liu, L. (2020). Large Chinese land carbon sink estimated from atmospheric carbon dioxide data. Nature, 586(7831), 720–723. https://doi.org/10.1038/s41586-020-2849-9 Yadav,B.K....

    2020 doi

  21. [29]

    Yin, J., & Sun, S. (2022). Incomplete multi-view clustering with cosine similarity. Pattern Recognition, 123,108371.https://doi.org/10.1016/j.patcog.2021.108371

    2022

  22. [30]

    Yan, S. (2021). Tokens-to-Token ViT: Training Vision Transformers from Scratch on ImageNet. 2021 IEEE/CVF International Conference on Computer Vision (ICCV),538–547.https://doi.org/10.1109/ICCV48922.2021.00060

    2021

  23. [31]

    Zaitsev, D. A. (2017). A generalized neighborhood for cellular automata. Theoretical Computer Science, 666,21–35.https://doi.org/10.1016/j.tcs.2016.11.002 Zeng,Y.,Yang,X.,Pan,L.,Zhu,W.,Wang,D.,Zhao,Z.,Liu,J.,Sun,C.,&Zhou,C. (2023). Fish school feeding behavior quantification u...

    2017

  24. [32]

    Zhang, L., Su, G., Yin, J., Li, Y., Lin, Q., Zhang, X., & Shao, L. (2022). Bioinspired Scene Classification by Deep Active Learning with Remote Sensing Applications. IEEE Transactions on Cybernetics, 52(7), 5682–5694. https://doi.org/10.1109/TCYB.2020.2981480

    2022

  25. [33]

    Zhang, R., Zhou, X., Ouyang, Z., Avitabile, V., Qi, J., Chen, J., & Giannico, V. (2019). Estimating aboveground biomass in subtropical forests of China by integrating multisource remote sensing and ground data. Remote Sensing of Environment, 232,111341.https://doi.org/10.1016/...

    2019

  26. [34]

    Zhong, B., Wei, T., Luo, X., Du, B., Hu, L., Ao, K., Yang, A., & Wu, J. (2023). Multi-Swin Mask Transformer for Instance Segmentation of Agricultural Field Extraction. Remote Sensing, 15(3),Article3.https://doi.org/10.3390/rs15030549

    2023 doi

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.