Pith. sign in

REVIEW 3 major objections 4 minor 49 references

Geographical Context Matters: Bridging Fine and Coarse Spatial Information to Enhance Continental Land Cover Mapping

T0 review · 3 major / 4 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read Adding latitude, longitude, and region labels to a land-cover model lifts EU-wide mapping accuracy, with the largest gains when both levels of geography are used together.

desk verdict The full model beats standard baselines, but the paper's own ablation contradicts its central claim about joint geospatial integration. read the letter →

arxiv 2504.12368 v1 pith:WCVARBQH submitted 2025-04-16 cs.CV cs.LG

Babak Ghassemi , Cassio Fraga-Dantas , Raffaele Gaetano , Dino Ienco , Omid Ghorbanzadeh , Emma Izquierdo-Verdiguier , Francesco Vuolo This is my paper · ORCID
classification cs.CVcs.LG
keywords landcovermappinggeospatialmetadatapositionalencodingfeaturedisentanglementbiogeographicalregionsdeeplearningcroptypeclassificationEarthObservation
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

BRIDGE-LC is a lightweight neural network for land cover classification that learns from two levels of geographic context alongside satellite features: the exact latitude and longitude of each sample, and the broad ecological zone, one of eight across the EU, that contains it. A second branch, active only during training, predicts the zone from a separate embedding while a contrastive training signal keeps zone-specific and zone-invariant information apart, so at inference the zone branch can be discarded and only coordinates are needed. The paper's central claim is that integrating this geospatial context improves continental land cover mapping, and that the largest gains come from using both levels together: 80.30% F1 on seven broad land cover classes and 64.01% F1 on nineteen crop types when all regions are present in training, and 73.59% and 54.44% when a whole region is held out, ahead of Random Forest, SVM, and XGBoost throughout. A sympathetic reader would care because continental land cover maps underpin agricultural monitoring and climate policy, and the result suggests that a cheap metadata signal, where a pixel sits rather than just what it reflects, can be folded into standard classifiers without adding inference cost.

What carries the argument

The central mechanism is bi-level spatial conditioning paired with feature disentanglement. Coordinates are mapped through a fixed sinusoidal positional encoding, then reshaped by a small learned MLP into a 128-dimensional location embedding that is concatenated to the spectral feature vector; this is the fine-grained channel. The coarse channel is a training-only branch that predicts the sample's ecological zone from its own embedding, while a supervised contrastive loss over land-cover classes plus zone classes pulls matching embeddings together and pushes different classes and zones apart, enforcing that the land-cover branch stays region-invariant. At inference the zone branch is removed, leaving a model that needs only coordinates, not zone labels.

What would settle it

Re-run the extrapolation and leave-one-region-out scenarios after tuning all baselines on a separate validation split with the same computational budget; if tuned RF or XGBoost closes the gap to within a fraction of a point, the claim that geospatial integration drives the improvement would be falsified. A second check: replace the learned positional encoder with the fixed sinusoidal encoding alone, since the paper itself reports degraded performance in that configuration, and a match would weaken the mechanism story.

Watch

Extended reading notes

Core claim

The discovery the paper argues for is that geospatial metadata can be made a structured part of the learning objective rather than a passive input. BRIDGE-LC encodes coordinates with a sinusoidal positional encoding followed by a learned MLP, concatenates the result to 109 satellite and auxiliary features, and passes it through two parallel encoders: one produces a region-invariant embedding for land cover classification, the other a region-specific embedding for predicting the eight ecological zones, with a supervised contrastive loss keeping the two spaces apart. The paper reports that this design outperforms all compared baselines in both the extrapolation scenario and the leave-one-region-out scenario, that ablations attribute gains to both fine- and coarse-grained geospatial information, and that the learned location embeddings visibly recover the biogeographical structure supplied during training. In the authors' reading, the model is not memorizing coordinates but learning which appearance patterns travel across regions and which are local.

Load-bearing premise

The load-bearing premise is that Random Forest, SVM, and XGBoost run with package-default settings are strong enough baselines, because no validation set was used and BRIDGE-LC's own architecture was not held to the same no-tuning standard.

Editorial extensions

If this is right

  • Existing continental land cover pipelines can be upgraded by adding coordinates and region labels during training only; the deployed model still takes the usual features plus coordinates.
  • Crop-type mapping is where the method pays off most: in the extrapolation scenario the F1 advantage over the best baseline is about 2.4 points at Level 2 versus under one point at Level 1.
  • Holding out a complete biogeographical region still leaves BRIDGE-LC ahead of all baselines, so the geospatial conditioning contributes to spatial generalization rather than only to fitting familiar regions.
  • The learned location embeddings are spatially coherent and recover the biogeographical structure, meaning the model builds an interpretable map-like prior as a byproduct of training.

Reading between the lines

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

  • Testable extension: the same dual-branch recipe should transfer to other coarse partitions, such as continents or agro-ecological zones; applying it to a North American ecoregion partition would show whether the result depends on the EU's eight-zone scheme.
  • Testable extension: because the ablation shows fixed positional encodings underperform learned ones, the learned PE is doing real work; comparing BRIDGE-LC against a model that feeds raw coordinates through an MLP without sinusoidal pre-encoding would isolate where that work happens.
  • Caution: part of the gain may come from spatial autocorrelation in the labels, so the margin could shrink on years or landscapes with strong distribution shift; retraining on a later survey campaign would separate a true geospatial prior from interpolation between nearby training points.
  • Not claimed by the paper: the coarse partition is highly imbalanced, with two regions covering a tiny share of samples, so results in those regions are noisy; a finer or differently chosen partition could shift the balance between region-specific and invariant features.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper proposes BRIDGE-LC, a lightweight MLP classifier for EU-27 land cover and crop type mapping that combines Sentinel-1/2 and auxiliary features with a learned sinusoidal positional encoding of latitude/longitude and a training-time branch supervised by biogeographical region labels through a contrastive disentanglement loss. The authors evaluate the method in a random-split extrapolation scenario and a leave-one-biogeographical-region-out scenario, compare against RF, SVM, and XGBoost, and inspect the learned embeddings. The headline claims are that both fine-grained and coarse-grained geospatial information improve accuracy and that the largest gains come from jointly using both.

Significance. If the claims held, the architecture would be a useful and computationally modest contribution to large-scale land cover mapping, since inference requires only coordinates and the method outperforms classical baselines in two scenarios. The use of a public in-situ benchmark with external labels, the availability of code, and the per-class breakdown are strengths. However, the central ablation evidence is internally inconsistent, and the reported gains over the no-geo baseline are small single-run differences. As presented, the paper does not establish that coarse information helps or that the joint configuration is best, which is the central novelty claim.

major comments (3)
  1. [§4.4, Table 12] The ablation table contradicts the text and the central claim. The first data row (no geoinformation) reports Level-1 Acc/F1 74.49/73.17 and Level-2 Acc/F1 57.76/53.93. Among the ablation rows, the row with Level-1 Acc/F1 71.90/70.20 and Level-2 Acc/F1 55.70/52.14 is below the no-geo baseline on all four metrics, and the row with Level-2 F1 54.88 outperforms the full model's 54.44. Consequently, the statement in §4.4 that 'incorporating either latitude/longitude ... or biogeographical region information enhances classification performance across all tasks' is not supported by the table. The specific claims in §5.3 that coarse information raises Level-1 F1 from 73.17 to 73.26 and that the joint model is best are also not established: the 73.17-to-73.26 comparison is baseline-versus-full, not a controlled coarse-only effect, and the joint-model claim is contradicted by the 54.88 Level-2 F1 row.
  2. [Tables 6-12, §4.2-§4.4] All results are single-run point estimates with no error bars, seeds, confidence intervals, or significance tests. The model-vs-baseline and ablation differences that support the paper's thesis are small; for example, in Table 12 the full model versus the no-geo baseline is Level-1 F1 73.59 versus 73.17 and Level-2 F1 54.44 versus 53.93. These gaps are within the range of plausible run-to-run noise, so the paper does not demonstrate that the improvements are real. This is load-bearing because the central claim is precisely that small gains arise from spatial-information integration.
  3. [§4.1.1] The comparison against RF, SVM, and XGBoost is asymmetric. The authors state that no validation set was used and therefore no hyperparameter optimization was performed for any method, but BRIDGE-LC's architecture and training choices (500 epochs, learning rate 10^-4, batch size 256, dropout 50%, and the contrastive temperature tau in Eq. (6)) are specific selections made by the authors. RF, SVM, and XGBoost are run with package defaults. The reported gains over these baselines may therefore partly reflect more favorable settings for the proposed method; a sensitivity analysis or a matched tuning budget is needed before attributing the gains to the geospatial information.
minor comments (4)
  1. [§3.4, Eq. (4)-(5)] Equations (4) and (5) as written do not define the usual categorical cross-entropy loss because they omit the minus sign; the standard form is -sum 1(condition) log(predicted_probability). The notation should be corrected.
  2. [§2.2] The sample-count narrative is unclear: 133,813 samples after filtering, 81,427 cloud-free samples balanced to 86,831, yet 139,217 labeled samples are used in the classification procedure. The relationship between these numbers should be explained explicitly.
  3. [§1, §4.4] The manuscript contains typos and formatting inconsistencies, including 'hese data characteristics' in the Introduction, 'bioRegionOut' in the Section 4.4 heading, and irregular spacing in words like 'di fferent'. A careful copy-edit is needed.
  4. [§4.6.1-§4.6.2, Figures 10-11] The t-SNE and PCA visualizations are qualitative and based on 50 samples per class in Figure 10. They should be framed as illustrative rather than as quantitative evidence of the claimed gains.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the central comparison is against external LUCAS labels and baselines, and the geospatial-information claim is an empirical result rather than a construction.

full rationale

The paper's central claim—that jointly using fine- and coarse-grained geospatial information improves land cover mapping—is an empirical claim tested against external LUCAS 2022 in-situ labels and compared with RF, SVM, and XGBoost baselines. The method is not defined in terms of its own outputs: the fixed sinusoidal positional encoding (Eqs. 1–2), the learned positional-encoding MLP (Eq. 3), the land-cover cross-entropy loss (Eq. 4), the region-classification loss (Eq. 5), and the supervised contrastive loss (Eq. 6) are all standard supervised-learning components, and no reported metric is obtained by reinserting the target result into the loss. The self-citations (Ghassemi et al. 2024 for the dataset and 109-feature subset; Dantas et al. 2024 for feature disentanglement) are data and method provenance rather than load-bearing derivations: the dataset is a published Scientific Data artifact based on the external LUCAS Copernicus survey, and the disentanglement idea is an architectural choice, not an unverified uniqueness theorem imported to force the result. The ablation in Table 12 does not support the text's claim that coarse-only information 'enhances classification performance'—the coarse-only row is worse than baseline—and the full model's Level-2 F1 (54.44) is below the learned-PE-only variant (54.88), but this is an internal-consistency and statistical-robustness problem (no error bars or significance tests), not a circularity pattern: it does not make the reported prediction equivalent to the model's inputs by construction. Consequently, under the hard rules requiring a specific equation-level or definitional reduction, no circular step is established.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The method relies on standard building blocks (MLP, sinusoidal encoding, contrastive loss) and on external labels from LUCAS/EEA. The main non-standard dependencies are the authors' own prior dataset/feature selection and the hand-chosen hyperparameters, which are documented above.

free parameters (2)
  • Supervised contrastive temperature tau = Not reported
    Used in Eq. 6 to scale cosine similarities; its value affects embedding geometry and final performance, but the manuscript never states it.
  • Architecture hyperparameters (learning rate, batch size, dropout, hidden width) = LR=1e-4, batch=256, dropout=0.5, width=256
    Chosen by the authors without a validation set; this is a hand-selected configuration, while baselines use package defaults.
assumptions (4)
  • domain assumption EEA biogeographical region labels are a meaningful supervisory signal for land cover and crop type variability.
    The method's disentanglement branch is trained to predict these 8 region labels (Section 3.4), assuming they capture ecologically relevant variation.
  • domain assumption LUCAS 2022 in-situ labels are accurate ground truth for land cover and crops.
    All models are trained and evaluated against these labels (Section 2.1.1), so label errors transfer directly to reported scores.
  • domain assumption The RF-based feature selection and balancing from Ghassemi et al. 2024 are valid for the new method and do not leak test information.
    The 109-feature subset and the balanced sample set are taken from the authors' prior work (Section 2.2); any leakage in that pipeline would affect all methods.
  • standard math Sinusoidal positional encoding and cosine similarity are appropriate for encoding geographic coordinates.
    Equations (1)-(2) define the fixed encoding; standard definitions from Transformer literature are assumed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Geographical Context Matters: Bridging Fine and Coarse Spatial Information to Enhance Continental Land Cover Mapping." pith.science (2026). https://pith.science/paper/WCVARBQH

@misc{pith2026250412368,
  author       = {Pith},
  title        = {Pith review of: Geographical Context Matters: Bridging Fine and Coarse Spatial Information to Enhance Continental Land Cover Mapping},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WCVARBQH}},
  note         = {Machine review of arXiv:2504.12368}
}
read the original abstract

Land use and land cover mapping from Earth Observation (EO) data is a critical tool for sustainable land and resource management. While advanced machine learning and deep learning algorithms excel at analyzing EO imagery data, they often overlook crucial geospatial metadata information that could enhance scalability and accuracy across regional, continental, and global scales. To address this limitation, we propose BRIDGE-LC (Bi-level Representation Integration for Disentangled GEospatial Land Cover), a novel deep learning framework that integrates multi-scale geospatial information into the land cover classification process. By simultaneously leveraging fine-grained (latitude/longitude) and coarse-grained (biogeographical region) spatial information, our lightweight multi-layer perceptron architecture learns from both during training but only requires fine-grained information for inference, allowing it to disentangle region-specific from region-agnostic land cover features while maintaining computational efficiency. To assess the quality of our framework, we use an open-access in-situ dataset and adopt several competing classification approaches commonly considered for large-scale land cover mapping. We evaluated all approaches through two scenarios: an extrapolation scenario in which training data encompasses samples from all biogeographical regions, and a leave-one-region-out scenario where one region is excluded from training. We also explore the spatial representation learned by our model, highlighting a connection between its internal manifold and the geographical information used during training. Our results demonstrate that integrating geospatial information improves land cover mapping performance, with the most substantial gains achieved by jointly leveraging both fine- and coarse-grained spatial information.

Figures

Figures reproduced from arXiv: 2504.12368 by the authors.

Figure 1
Figure 1. Database samples, distributed over the EU-27 countries [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Subset of the biogeographical regions produced by the European Environment Agency (EEA). information provided by European Environment Agency (EEA)3 , comprising 8 unique values (Alpine, Atlantic, Black Sea, Boreal, Continental, Mediterranean, Pannonian, and Steppic) in the study area, were utilized as supervision sig￾nals on the proposed approach. The distribution of the available labelled data in terms of the clima… view at source ↗
Figure 3
Figure 3. Proposed framework. In addition to the standard land cover classification branch (yellow), which processes reflectance-based input features, we introduce two complementary components: (1) a fine-grained spatial informa￾tion branch (green) utilizing latitude-longitude coordinates, and (2) a coarse-grained spatial information branch (red) based on biogeographical region labels. During training, the red branch guides t… view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: Confusion matrices (row-normalized) for Level-1 classification, extrapolation scenario. that BRIDGE-LC’s integration of multi-level geospatial information still provides benefits even when calibration data from certain regions is unavailable. These findings highlight o…
Figure 5
Figure 5. Figure 5: Confusion matrices (row-normalized) for Level-2 classification, extrapolation scenario. performance drop when using fixed positional encodings alone, especially compared to the learned PE variant. This degradation indicates that directly incorporating fine-grained geol…
Figure 6
Figure 6. Figure 6: LC maps Spain area - Mediterranean region. 20 [PITH_FULL_IMAGE:figures/full_fig_p020_6.png]
Figure 7
Figure 7. Figure 7: LC maps France area - Atlantic region. 21 [PITH_FULL_IMAGE:figures/full_fig_p021_7.png]
Figure 8
Figure 8. Figure 8: LC maps Austria area - Continental region. 22 [PITH_FULL_IMAGE:figures/full_fig_p022_8.png]
Figure 9
Figure 9. Figure 9: LC maps Romania area - Steppic region. 23 [PITH_FULL_IMAGE:figures/full_fig_p023_9.png]
Figure 10
Figure 10. Figure 10: t-SNE visualization of the embeddings extracted by the proposed approach (left column) and its ablated version without coarse and fine-grained spatial information (right column) considering the Leave-One-bioRegion-Out scenario on the Alpine (top row) and the Pannonian…
Figure 11
Figure 11. Figure 11: Visualization of the learned location embeddings when both coarse and fine-grained spatial information are exploited (left column) and when only the fine-grained information is used (right column). 26 [PITH_FULL_IMAGE:figures/full_fig_p026_11.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

49 extracted references · 28 canonical work pages

  1. [7]

    Earth System Science Data 16, 5723–5735

    Advances in lucas copernicus 2022: enhancing earth observations with comprehensive in situ data on eu land cover and use. Earth System Science Data 16, 5723–5735. doi: 10.5194/essd-16-5723-2024 . Dantas, C.F., Gaetano, R., Paris, C., Ienco, D.,

    work page doi:10.5194/essd-16-5723-2024 2022

  2. [8]

    IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 1–15doi:10.1109/JSTARS.2024

    Reuse out-of-year data to enhance land cover mapping via feature disentanglement and contrastive learning. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 1–15doi:10.1109/JSTARS.2024. 3503756. Drusch, M., Del Bello, U., Carlier, S., Colin, O., Fernandez, V ., Gascon, F., Hoersch, B., Isola, C., Laberinti, P., Martimort, ...

    doi:10.1109/jstars.2024 2024

  3. [11]

    d’Andrimont, R., Verhegghen, A., Lemoine, G., Kempeneers, P., Meroni, M., van der Velde, M.,

    doi:10.3390/rs11070888. d’Andrimont, R., Verhegghen, A., Lemoine, G., Kempeneers, P., Meroni, M., van der Velde, M.,

    work page doi:10.3390/rs11070888

  4. [12]

    Huete, A.,

    doi: 10.3390/rs12081253. Huete, A.,

    doi:10.3390/rs12081253

  5. [13]

    30 European Commission, Joint Research Centre (JRC),

    doi: 10.1038/s41597-020-00675-z . 30 European Commission, Joint Research Centre (JRC),

    doi:10.1038/s41597-020-00675-z

  6. [14]

    Arii, M., van Zyl, J.J., Kim, Y .,

    doi: 10.3390/rs14112654. Arii, M., van Zyl, J.J., Kim, Y .,

    work page doi:10.3390/rs14112654

  7. [15]

    d’Andrimont, R., Yordanov, M., Sedano, F., Verhegghen, A., Strobl, P., Zachariadis, S., Camilleri, F., Palmieri, A., Eiselt, B., Rubio Iglesias, J.M., van der Velde, M.,

    doi: 10.3390/rs15153859. d’Andrimont, R., Yordanov, M., Sedano, F., Verhegghen, A., Strobl, P., Zachariadis, S., Camilleri, F., Palmieri, A., Eiselt, B., Rubio Iglesias, J.M., van der Velde, M.,

    work page doi:10.3390/rs15153859

  8. [16]

    Ghassemi, B., Izquierdo-Verdiguier, E., Verhegghen, A., Yordanov, M., Lemoine, G., Moreno Mart ´ınez, ´A., De Marchi, D., van der Velde, M., Vuolo, F., d’Andrimont, R.,

    doi: 10.3390/rs14030541. Ghassemi, B., Izquierdo-Verdiguier, E., Verhegghen, A., Yordanov, M., Lemoine, G., Moreno Mart ´ınez, ´A., De Marchi, D., van der Velde, M., Vuolo, F., d’Andrimont, R.,

    work page doi:10.3390/rs14030541

Show all 49 references
  1. [17]

    Scientific Data 11,

    European union crop map 2022: Earth observation’s 10-meter dive into europe’s crop tapestry. Scientific Data 11,

    2022

  2. [19]

    URL: https:// lpdaac.usgs.gov/products/mod21c3v061/, doi:10.5067/MODIS/MOD21C3.061

    Modis /terra land surface temperature /3-band emissivity monthly l3 global 0.05deg cmg v061. URL: https:// lpdaac.usgs.gov/products/mod21c3v061/, doi:10.5067/MODIS/MOD21C3.061. Jacques, D.C., Kergoat, L., Hiernaux, P., Mougin, E., Defourny, P.,

    doi

  3. [21]

    Global Change Biology 26, 1754–1766

    A deep learning approach to conflating heterogeneous geospatial data for corn yield estimation: A case study of the us corn belt at the county level. Global Change Biology 26, 1754–1766. doi:https://doi.org/10.1111/gcb.14885. Jiang, Z., Huete, A.R., Didan, K., Miura, T.,

    doi

  4. [23]

    The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences XLI-B7, 269–276

    Comparison of filters dedicated to speckle suppression in sar images. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences XLI-B7, 269–276. doi: 10.5194/isprs-archives-XLI-B7-269-2016 . Kussul, N., Lavreniuk, M., Skakun, S., Shelestov, A.,

    doi

  5. [28]

    Biogeosciences 9, 179–191

    Mapping tropical forest biomass with radar and spaceborne lidar in lop ´e national park, gabon: overcoming problems of high biomass and persistent cloud. Biogeosciences 9, 179–191. URL: https://bg.copernicus.org/articles/9/179/2012/. Mustafa Abdullah, D., Mohsin Abdulazeez, A.,

    2012

  6. [29]

    Qubahan Academic Journal 1, 81–90

    Machine learning applications based on svm classification a review. Qubahan Academic Journal 1, 81–90. URL: https://journal.qubahan.com/index.php/qaj/article/view/50, doi:10.48161/qaj.v1n2a50. NASA/METI/AIST/Japan Spacesystems And U.S./Japan ASTER Science Team,

    doi

  7. [30]

    URL: https: //lpdaac.usgs.gov/products/astgtmv003/, doi:10.5067/ASTER/ASTGTM.003

    Aster global digital elevation model v003. URL: https: //lpdaac.usgs.gov/products/astgtmv003/, doi:10.5067/ASTER/ASTGTM.003. Panangian, D., Bittner, K.,

    doi

  8. [32]

    doi: 10.3390/s19040904. Paszke, A., Gross, S., Massa, F., Lerer, A., Bradbury, J., Chanan, G., Killeen, T., Lin, Z., Gimelshein, N., Antiga, L., Desmaison, A., Kopf, A., Yang, E., DeVito, Z., Raison, M., Tejani, A., Chilamkurthy, S., Steiner, B., Fang, L., Bai, J., Chintala, S.,

    doi

  9. [33]

    (Eds.), Advances in Neural Information Processing Systems, Curran Associates, Inc

    Pytorch: An imperative style, high-performance deep learning library, in: Wallach, H., Larochelle, H., Beygelzimer, A., d'Alch ´e-Buc, F., Fox, E., Garnett, R. (Eds.), Advances in Neural Information Processing Systems, Curran Associates, Inc. URL: https://proceedings.neurips.c...

    2019

  10. [35]

    Geocarto International 34, 1531–1551

    A new spectral index for the extraction of built-up land features from landsat 8 satellite imagery. Geocarto International 34, 1531–1551. doi: 10.1080/10106049.2018.1497094. Rajadel-Lambistos, C., Izquierdo-Verdiguier, E., Moreno-Mart´ınez, A., Maneta, M., Begueria, S., Kimbal...

    2018

  11. [36]

    arXiv preprint arXiv:2402.01444

    Mission critical–satellite data is a distinct modality in machine learning. arXiv preprint arXiv:2402.01444 . Rußwurm, M., Klemmer, K., Rolf, E., Zbinden, R., Tuia, D.,

    arXiv

  12. [38]

    International Journal of Remote Sensing 42, 7581–7609

    Vegetation cover monitoring in tropical regions using sar-c dual-polarization index: seasonal and spatial influences. International Journal of Remote Sensing 42, 7581–7609. doi: 10.1080/01431161.2021.1959955. Saralioglu, E., Gungor, O.,

    2021

  13. [40]

    IEEE Transactions on Geoscience and Remote Sensing 58, 14–33

    Monitoring and predicting agricultural droughts for a water-limited subcontinental region by integrating a land surface model and microwave remote sensing. IEEE Transactions on Geoscience and Remote Sensing 58, 14–33. doi:10.1109/TGRS.2019.2927342. Schwalbert, R.A., Amado, T.,...

  14. [41]

    Agricultural and Forest Meteorology 284, 107886

    Satellite-based soybean yield forecast: Integrating machine learning and weather data for improving crop yield prediction in southern brazil. Agricultural and Forest Meteorology 284, 107886. doi:https: //doi.org/10.1016/j.agrformet.2019.107886. Singh, G., Singh, S., Sethi, G.,...

  15. [42]

    Remote Sensing of Environment 120, 9–24

    Gmes sentinel-1 mission. Remote Sensing of Environment 120, 9–24. doi: https://doi.org/10.1016/j.rse.2011.05.028. the Sentinel Missions - New Opportunities for Science. Van Deventer, A., Ward, A., Gowda, P., Lyon, J.,

    2011 doi

  16. [43]

    ISPRS Journal of Photogrammetry and Remote Sensing 193, 200–215

    Cross-phenological-region crop mapping framework using sentinel-2 time series imagery: A new perspective for winter crops in china. ISPRS Journal of Photogrammetry and Remote Sensing 193, 200–215. doi: https://doi.org/10. 1016/j.isprsjprs.2022.09.010. Wulf, H., Stuhler, S.,

  17. [44]

    IET Image Processing 16, 273–284

    Deep coordinate attention network for single image super-resolution. IET Image Processing 16, 273–284. doi:https://doi.org/10.1049/ipr2.12364. Xu, H.,

    doi

  18. [45]

    International Journal of Remote Sensing 27, 3025–3033

    Modification of normalised di fference water index (ndwi) to enhance open water features in remotely sensed imagery. International Journal of Remote Sensing 27, 3025–3033. doi: 10.1080/01431160600589179. Y . Zha, J.G., Ni, S.,

    doi

  19. [47]

    Ecological Informatics 74, 101955

    Application of machine learning approaches for land cover monitoring in northern cameroon. Ecological Informatics 74, 101955. doi: https://doi.org/10.1016/j.ecoinf.2022.101955. Zhang, Z., Xu, W., Qin, Q., Long, Z.,

    2022

  20. [48]

    IEEE Transactions on Geoscience and Remote Sensing 59, 1012–1028

    Downscaling solar-induced chlorophyll fluorescence based on convolutional neural network method to monitor agricultural drought. IEEE Transactions on Geoscience and Remote Sensing 59, 1012–1028. doi: 10.1109/TGRS.2020.2999371. Zhao, S., Tu, K., Ye, S., Tang, H., Hu, Y ., Xie, C.,

    2020

  21. [49]

    doi:10.3390/s23218966. 32

    doi

  22. [1048]

    Gislason, P.O., Benediktsson, J.A., Sveinsson, J.R.,

    doi: 10.1038/s41597-024-03884-y . Gislason, P.O., Benediktsson, J.A., Sveinsson, J.R.,

    doi

  23. [1994]

    Remote Sensing of Environment 48, 31 119–126

    A modified soil adjusted vegetation index. Remote Sensing of Environment 48, 31 119–126. doi: https://doi.org/10.1016/0034-4257(94)90134-1 . Rafik Bouhennache, Toufik Bouden, A.T.A., Cheddad, A.,

    doi

  24. [1996]

    International Journal of Remote Sensing 17, 1425–1432

    The use of the normalized difference water index (ndwi) in the delineation of open water features. International Journal of Remote Sensing 17, 1425–1432. doi: 10.1080/01431169608948714. Mitchard, E.T.A., Saatchi, S.S., White, L.J.T., Abernethy, K.A., Jeffery, K.J., Lewis, S.L....

    doi

  25. [2002]

    International Journal of Remote Sensing 23, 725–749

    An assessment of support vector machines for land cover classification. International Journal of Remote Sensing 23, 725–749. doi: 10.1080/01431160110040323. Campos-Taberner, M., Garc´ıa-Haro, F.J., Mart´ınez, B., Izquierdo-Verdiguier, E., Atzberger, C., Camps-Valls, G., Gilabe...

    doi

  26. [2003]

    International Journal of Remote Sensing 24, 583–594

    Use of normalized difference built-up index in automatically mapping urban areas from tm imagery. International Journal of Remote Sensing 24, 583–594. doi: 10.1080/01431160304987. Yuan, Y ., Lin, L.,

    doi

  27. [2006]

    Pattern Recognition Letters 27, 294–300

    Random forests for land cover classification. Pattern Recognition Letters 27, 294–300. doi:https://doi.org/10.1016/j.patrec.2005.08.011. pattern Recognition in Remote Sensing (PRRS 2004). Gitelson, A.A., Kaufman, Y .J., Merzlyak, M.N.,

    2005 doi

  28. [2008]

    Remote Sensing of Environment 112, 3833–3845

    Development of a two-band enhanced vegetation index without a blue band. Remote Sensing of Environment 112, 3833–3845. doi:https://doi.org/10.1016/j.rse.2008.06.006. Jo, S., Yoon, S.W.,

    2008 doi

  29. [2010]

    IEEE Transactions on Geoscience and Remote Sensing 48, 3349–3357

    A general characterization for polarimetric scattering from vegetation canopies. IEEE Transactions on Geoscience and Remote Sensing 48, 3349–3357. doi: 10.1109/TGRS.2010.2046331. Baudoux, L., Inglada, J., Mallet, C.,

    2010

  30. [2012]

    Remote Sensing of Environment 120, 25–36

    Sentinel-2: Esa’s optical high-resolution mission for gmes operational services. Remote Sensing of Environment 120, 25–36. doi:https://doi.org/10.1016/j.rse.2011.11.026. the Sentinel Missions - New Opportunities for Science. Du, Z., Yang, J., Ou, C., Zhang, T.,

    2011 doi

  31. [2013]

    IEEE Transactions on Geoscience and Remote Sensing 51, 162–173

    Crop yield estimation based on unsupervised linear unmixing of multidate hyperspectral imagery. IEEE Transactions on Geoscience and Remote Sensing 51, 162–173. doi:10.1109/TGRS.2012.2198826. van der Maaten, L., Hinton, G.E.,

    2012

  32. [2014]

    Remote Sensing of Environment 153, 40–49

    Monitoring dry vegetation masses in semi-arid areas with modis swir bands. Remote Sensing of Environment 153, 40–49. doi: https://doi.org/10.1016/j.rse.2014.07.027. Jiang, H., Hu, H., Zhong, R., Xu, J., Xu, J., Huang, J., Wang, S., Ying, Y ., Lin, T.,

    2014 doi

  33. [2016]

    IEEE Transactions on Geoscience and Remote Sensing 54, 6563–6573

    Spiking neural networks for crop yield estimation based on spatiotemporal analysis of image time series. IEEE Transactions on Geoscience and Remote Sensing 54, 6563–6573. doi: 10.1109/TGRS.2016.2586602. Bourcier, J., Dashyan, G., Alahari, K., Chanussot, J.,

    2016

  34. [2017]

    IEEE Geoscience and Remote Sensing Letters 14, 778–782

    Deep learning classification of land cover and crop types using remote sensing data. IEEE Geoscience and Remote Sensing Letters 14, 778–782. doi: 10.1109/LGRS.2017.2681128. Loshchilov, I., Hutter, F.,

    2017

  35. [2019]

    Luo, B., Yang, C., Chanussot, J., Zhang, L.,

    Decoupled weight decay regularization, in: 7th International Conference on Learning Representations, ICLR 2019, New Orleans, LA, USA, May 6-9, 2019, OpenReview.net. Luo, B., Yang, C., Chanussot, J., Zhang, L.,

    2019

  36. [2020]

    Scientific Data

    Harmonised lucas in-situ land cover and use database for field surveys from 2006 to 2018 in the european union. Scientific Data

    2006

  37. [2021]

    Remote Sensing of Environment 266, 112708

    From parcel to continental scale – a first european crop type map based on sentinel-1 and lucas copernicus in-situ observations. Remote Sensing of Environment 266, 112708. doi:https://doi.org/10.1016/j.rse.2021.112708. d’Andrimont, R., Yordanov, M., Martinez-Sanchez, L., Eisel...

    2021

  38. [2022]

    Geocarto International 37, 657–677

    Semantic segmentation of land cover from high resolution multispectral satellite images by spectral-spatial convolutional neural network. Geocarto International 37, 657–677. doi: 10.1080/10106049.2020.1734871. Sawada, Y ., Koike, T., Ikoma, E., Kitsuregawa, M.,

    2020

  39. [2023]

    arXiv preprint arXiv:2310.06743

    Geographic location encoding with spherical harmonics and sinusoidal representation networks. arXiv preprint arXiv:2310.06743 . Erli Pinto dos Santos, D.D.D.S., do Amaral, C.H.,

    arXiv

  40. [2024]

    IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 17, 2980–2994

    End-to-end learning for land cover classification using irregular and unaligned sits by combining attention-based interpolation with sparse variational gaussian processes. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 17, 2980–2994. doi: 10.1...

  41. [2025]

    Pasqualotto, N., Delegido, J., Van Wittenberghe, S., Rinaldi, M., Moreno, J.,

    Can location embeddings enhance super-resolution of satellite imagery? URL: https://arxiv.org/abs/ 2501.15847, arXiv:2501.15847. Pasqualotto, N., Delegido, J., Van Wittenberghe, S., Rinaldi, M., Moreno, J.,

    arXiv

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.