Pith. sign in

REVIEW 3 major objections 5 minor 79 references

Learning to See More: UAS-Guided Super-Resolution of Satellite Imagery for Precision Agriculture

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This paper claims that fusing ordinary UAS RGB imagery with 10-meter Sentinel-2 data through a lightweight SRCNN can reproduce the accuracy of costly hyperspectral drone surveys for cover crop biomass and nitrogen monitoring.

desk verdict A genuinely useful applied pipeline, but the headline 18%/31% gains are probably inflated by training and evaluating against clean simulated Sentinel-2 targets without radiometric validation against real satellite reflectances. read the letter →

arxiv 2505.21746 v1 pith:OQF4BQSE submitted 2025-05-27 cs.CV cs.AIcs.LG

Arif Masrur , Peder A. Olsen , Paul R. Adler , Carlan Jackson , Matthew W. Myers , Nathan Sedghi , Ray R. Weil This is my paper · ORCID
classification cs.CVcs.AIcs.LG
keywords super-resolutionSRCNNspectralextensionSentinel-2UASimageryprecisionagriculturecovercropbiomassnitrogenestimation
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 show that a farmer does not need an expensive hyperspectral drone to get high-resolution crop health maps: fusing ordinary UAS RGB photos with free 10-meter Sentinel-2 satellite imagery, through a small super-resolution convolutional network, can recover the red-edge and near-infrared bands that matter for estimating cover crop biomass and nitrogen. In the study's case, the fused 0.125-meter reconstructions cut root-mean-square error by about 18% for biomass and 31% for nitrogen compared with models that used raw Sentinel-2 at 10 meters. The authors also claim that a spectral-extension model trained on targeted UAS data can be reused on other crops and regions, and that a spatial-extension model lets the farmer stop flying altogether once it is trained. The significance would be a low-cost, scalable alternative to hyperspectral sensors that can cost as much as $175,000, relying only on widely available satellite data and commodity drones.

What carries the argument

The engine is a super-resolution convolutional neural network (SRCNN) used in two modes. In spectral extension, the network takes co-registered UAS RGB plus upsampled Sentinel-2 bands as input and outputs 8 Sentinel-2 VNIR bands at 0.125 m resolution; the training targets are produced by non-negative least-squares fitting of 269 hyperspectral UAS bands to the published Sentinel-2 MSI spectral response functions, so the network only learns spatial super-resolution without spectral distortion. In spatial and temporal extension, the same SRCNN architecture sharpens 10 m Sentinel-2 to 1 m using ground truth from either UAS hyperspectral data or the spectral-extension output. Supporting machinery includes two-stage pixel alignment (reprojection plus corner alignment) and sub-pixel image registration by exhaustive translation search with linear-regression scoring, which reduces registration error from roughly 4.6 m to 0.81 m.

What would settle it

Take a field with coincident real Sentinel-2 and UAS hyperspectral coverage, train the spectral-extension SRCNN on clean simulations, and compare its 0.125 m reconstructions against real atmospherically corrected Sentinel-2 aggregated to the same grid; if the PSNR gap between simulated and real targets exceeds the roughly 0.2 dB statistical error reported in the model comparison, the clean-ground-truth premise fails.

Watch

Extended reading notes

Core claim

The central claim is that Sentinel-2's multispectral bands can be reconstructed at sub-meter resolution, and UAS RGB imagery can be spectrally extended into the red-edge and near-infrared, by a lightweight SRCNN that learns from high-resolution training targets synthesized by simulating the Sentinel-2 spectral response from UAS hyperspectral data. Using this, the paper's Random Forest models on reconstructed 8-band imagery outperform benchmark models on raw 10 m Sentinel-2 by approximately 18% in biomass RMSE and 31% in nitrogen RMSE, and also beat models built on raw UAS RGB for nitrogen prediction. The authors further claim that the spectral extension model generalizes across crops (wheat, corn, miscanthus) and regions in the Chesapeake Bay watershed, that a spatial extension model trained on a subset of fields yields better biomass and nitrogen predictions than raw UAS RGB data (allowing farmers to stop UAS flights), and that a temporal extension model improves predictions over a same-season Sentinel-2 benchmark. Finally, they show that spectral extension still works, at slightly lower fidelity, using only UAS RGB input when cloud-free Sentinel-2 is unavailable.

Load-bearing premise

The load-bearing premise is that Sentinel-2 bands simulated from clean UAS hyperspectral data are a trustworthy stand-in for real satellite reflectances, even though those simulations lack the atmospheric noise present in actual spaceborne observations.

Editorial extensions

If this is right

  • If the central claim holds, farmers can obtain ~0.125 m 8-band imagery for biomass and nitrogen monitoring using a commodity RGB camera plus Sentinel-2, bypassing multispectral and hyperspectral UAS sensors that cost up to $175,000.
  • A spectral-extension model trained on Maryland cover crops transfers to wheat, corn, and miscanthus fields in Pennsylvania and elsewhere in the region, so the training investment is reusable.
  • Once a spatial-extension model is trained for a farm, UAS flights can be stopped: its biomass and nitrogen predictions beat those from raw UAS RGB imagery on the same fields.
  • When cloud-free satellite imagery is unavailable, an RGB-only spectral extension model still produces usable 8-band reconstructions, only about 1.6 dB lower in PSNR on average than the version that uses satellite input.
  • Temporal extension recovers high-resolution imagery for dates without UAS coverage, improving biomass and nitrogen RMSE over a same-date Sentinel-2 benchmark by about 3% and 4.8% in MSE terms.

Reading between the lines

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

  • Reading beyond the paper: if the clean-simulation premise survives real-atmosphere tests, the same spectral-extension recipe could be applied to historical satellite archives such as Landsat, enabling sub-meter vegetation-index time series long before drone records exist.
  • Reading beyond the paper: the cost argument implies a break-even calculation for spatial extension based on the number of flights saved per season; a farmer-facing comparison of $162/hour pilot costs against the measured RMSE gains would help adoption decisions.
  • Reading beyond the paper: the robustness of the RGB-only model on cloudy dates suggests an operational service that predicts spectral bands from RGB time series alone, using satellite imagery only as optional seasonal context; this is not evaluated in the paper but follows from its cloudy-date results.
  • Reading beyond the paper: the 18% and 31% improvements should reproduce on independent farms with real Sentinel-2 acquisitions if the simulation-to-reality gap is small; the paper's quadrat-level validation is a first step, but on-farm trials with varied soil moisture and atmospheric conditions remain open.
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 presents an end-to-end framework for fusing UAS and Sentinel-2 imagery using super-resolution CNNs. Three use cases are developed: spectral extension (UAS RGB plus 10 m Sentinel-2 bands to produce 0.125 m 8-band multispectral imagery), spatial extension (Sentinel-2 only to 1 m imagery, trained with ground truth from either hyperspectral UAS or spectral-extension output), and temporal extension (reconstructing imagery for dates without UAS data). The downstream application is random-forest prediction of cover crop biomass and nitrogen content. The headline result is that the spectral extension model reduces biomass RMSE by about 18% and nitrogen RMSE by about 31% relative to a 10 m Sentinel-2 benchmark (Table 8, M27 vs. M23/M24). The reconstruction quality is measured by PSNR/RMSE against simulated Sentinel-2 ground truth generated from UAS hyperspectral data.

Significance. If the core claims are valid, the framework is practically significant: it would let farmers replace expensive hyperspectral UAS surveys with low-cost RGB UAS plus freely available satellite data while improving biomass and nitrogen predictions. The paper is also honest about several limitations and provides an open-source code link. Methodologically, the site-level 8-fold cross-validation for the SRCNN reconstruction is a careful step that avoids direct leakage between reconstruction training and test imagery. The systematic comparison of model architectures and the operational focus are valuable. However, the central quantitative claim depends on a radiometric consistency that is never validated, and the downstream random-forest evaluation uses random folds that may leak spatial autocorrelation. These issues must be addressed before the headline numbers can be accepted as evidence for the framework's real-world performance.

major comments (3)
  1. [§2.2.2, §3.4, Table 8] The Spectral-SRCNN is trained and evaluated entirely on Sentinel-2 reflectances simulated from UAS hyperspectral data via the published S2 spectral response functions, and the paper explicitly states that these simulations do not contain atmospheric noise present in real spaceborne observations. Reconstruction quality in Tables 4–6 is measured against the same clean simulations, and the downstream comparison of M27 (reconstructed predictors) against M23/M24 (real Sentinel-2 predictors) inherits this confound: M27's predictors are clean simulated reflectances, while M23/M24 use real Sentinel-2 reflectances with atmospheric and BRDF effects. No experiment compares the simulated 8-band reflectances, aggregated to 10 m, against real Sentinel-2 L2A reflectances over the same pixels and dates. Without such a radiometric validation, the claimed 18% and 31% improvements may reflect the removal of atmospheric noise rather than the spatial and spectral extension that is the paper's central contribution. This is a load-bearing omission and should be addressed, for example by adding a validation experiment with real Sentinel-2 data and by re-running the benchmark models on simulated versus real inputs.
  2. [§2.3.2, Table 8] The random-forest regression models for all extension scenarios are evaluated with random 5-fold cross-validation on quadrat-level samples, with no blocking by field or date. Since quadrats within the same field and date are highly spatially autocorrelated, random assignment allows near-duplicate samples to appear in both training and test folds, inflating R² and deflating RMSE for all entries in Table 8. Because the headline 18%/31% figures are derived from these RF models, the quantitative validity of the central claim is affected. The authors should use field-blocked or spatial leave-one-field-out cross-validation, or at minimum report the number of quadrats per field and demonstrate that the conclusions are robust to clustered evaluation.
  3. [§3.2, §4] The paper asserts that 'exceptional fidelity' of the reconstructed images is demonstrated by PSNR values around 30–36, but these values are computed against the simulated ground truth, not against an independent reference. No analysis links reconstruction PSNR/RMSE to downstream biomass and N prediction accuracy, and the spatial extension model, which has lower PSNR (~29), still produces RF predictions that outperform the Sentinel-2 benchmark. This suggests that the reconstruction quality metrics are not a reliable proxy for downstream utility, and the authors should either provide a direct link or temper the claims that image fidelity drives the observed agronomic improvements.
minor comments (5)
  1. [Table 5] The statement 'The statistical significance of the numbers is approximately 0.2 DB' should read '0.2 dB', and the method for estimating this uncertainty should be stated.
  2. [§3.3.1, Table 7b] The text says 'the RGB only model outperforms the competition', but for sites B and H the model using Sentinel-2 from 11/2/19 gives a higher PSNR than the RGB-only model; the conclusion holds only on average or for site E. Please make this precise.
  3. [Figure 12 caption] The caption references M22–M24 as spatial-resolution models, but the spatial-resolution experiments are M20–M22 in Table 3; also M12–M18 are not all 'spectral resolution', as M10–M14 vary spectral range. The caption should be corrected and aligned with the table tags.
  4. [§2.2.2] The phrase 'nonnegative linear squares' should be 'non-negative least squares'.
  5. [Table 8] The test sets differ across extension scenarios (as noted), but a reader can easily misread M25 and M29 as comparable models; a note in the caption stating that each column uses a different test set would improve clarity.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the super-resolution targets are simulated from independent hyperspectral data and the biomass/N gains are measured against independent field quadrats.

full rationale

The paper's derivation chain is: (1) simulate high-resolution Sentinel-2 reflectances from UAS hyperspectral data using the published Sentinel-2 spectral response functions (Section 2.2.2); (2) train a Spectral-SRCNN to reconstruct those high-resolution bands from co-aligned UAS RGB and real Sentinel-2 inputs; (3) feed the reconstructions, on held-out folds, into random forest models whose targets are measured quadrat-level biomass and nitrogen (Sections 2.3.2 and 3.4). The simulated reflectances are functions only of hyperspectral radiance and the published S2-SRF, not of the biomass/N measurements, so the 18% and 31% improvements over the Sentinel-2 benchmark are not fitted constants or self-defined quantities. The SRCNN test reconstructions come from site-level eight-fold splits that exclude the test site's ground truth from training, and the RF models use an additional five-fold split, so there is no leakage that would force the reported gains. The admitted absence of atmospheric noise in the simulated targets ('the hyperspectral-derived MSI simulations do not contain the atmospheric noise present in actual spaceborne observations, resulting in cleaner ground-truth') is an external-validity limitation: PSNR/RMSE reconstruction metrics are self-consistency checks against the simulation, and no experiment validates the simulation against real Sentinel-2 reflectances. That is a serious validation gap, but it is not circularity, because the final biomass/N evaluation uses independent field data. Self-citations (Adigun et al. 2022, Sedghi et al. 2023, FarmVibes repository) are background, data-source, or implementation references and are not load-bearing for the central claim. No equation or parameter is defined in terms of the quantity it is claimed to predict, and no uniqueness theorem or prior self-citation is invoked to force the chosen architecture. Therefore the central claims are not circular, though their operational transferability depends on the untested radiometric fidelity of the simulated Sentinel-2 targets.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The central claims depend on two hand-chosen target resolutions and three domain assumptions about the validity of simulated satellite ground truth, the generalizability of small quadrat samples, and the learnability of VRE/NIR from RGB context. No new physical entities are introduced; the SRCNN variants are method components, not invented observables.

free parameters (2)
  • UAS working resolution for spectral extension = 0.125 m (downsampled from ~3 cm)
    The target resolution is chosen by the authors (Section 2.2.3: 'we match our chosen UAS resolution (12.5 cm in our dataset)'). All spectral extension results and downstream RF improvements depend on this scale.
  • Output resolution for spatial and temporal extension = 1 m
    Spatial and temporal extension SRCNN models produce 1 m Sentinel-2 reconstructions; downstream RF results (M31-M34) are specific to this output resolution.
assumptions (3)
  • domain assumption Simulated Sentinel-2 bands from UAS hyperspectral data are faithful high-resolution ground truth for SRCNN training and evaluation.
    Section 2.2.2: 'the hyperspectral-derived MSI simulations do not contain the atmospheric noise present in actual spaceborne observations, resulting in cleaner ground-truth.' The entire SRCNN training pipeline and downstream RF evaluation depend on this proxy.
  • domain assumption Random forest models trained on 72 to 96 quadrats per scenario generalize to held-out fields and seasons within the Chesapeake Bay region.
    Section 2.3.2: 'A variable number of quadrats were used as data samples, ranging from 72 to 96... Model testing was performed on a randomly selected fold out of five folds.' The claimed 18% and 31% improvements rest on the representativeness of these small samples.
  • domain assumption VRE and NIR reflectance can be functionally inferred from UAS RGB and co-located Sentinel-2 context by the SRCNN.
    The spectral extension scenario assumes that the missing red-edge and NIR bands are learnable from RGB plus coarse satellite bands; this is the premise of the RGB-only cloudy-sky model in Section 3.3.1.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Learning to See More: UAS-Guided Super-Resolution of Satellite Imagery for Precision Agriculture." pith.science (2026). https://pith.science/paper/OQF4BQSE

@misc{pith2026250521746,
  author       = {Pith},
  title        = {Pith review of: Learning to See More: UAS-Guided Super-Resolution of Satellite Imagery for Precision Agriculture},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OQF4BQSE}},
  note         = {Machine review of arXiv:2505.21746}
}
read the original abstract

Unmanned Aircraft Systems (UAS) and satellites are key data sources for precision agriculture, yet each presents trade-offs. Satellite data offer broad spatial, temporal, and spectral coverage but lack the resolution needed for many precision farming applications, while UAS provide high spatial detail but are limited by coverage and cost, especially for hyperspectral data. This study presents a novel framework that fuses satellite and UAS imagery using super-resolution methods. By integrating data across spatial, spectral, and temporal domains, we leverage the strengths of both platforms cost-effectively. We use estimation of cover crop biomass and nitrogen (N) as a case study to evaluate our approach. By spectrally extending UAS RGB data to the vegetation red edge and near-infrared regions, we generate high-resolution Sentinel-2 imagery and improve biomass and N estimation accuracy by 18% and 31%, respectively. Our results show that UAS data need only be collected from a subset of fields and time points. Farmers can then 1) enhance the spectral detail of UAS RGB imagery; 2) increase the spatial resolution by using satellite data; and 3) extend these enhancements spatially and across the growing season at the frequency of the satellite flights. Our SRCNN-based spectral extension model shows considerable promise for model transferability over other cropping systems in the Upper and Lower Chesapeake Bay regions. Additionally, it remains effective even when cloud-free satellite data are unavailable, relying solely on the UAS RGB input. The spatial extension model produces better biomass and N predictions than models built on raw UAS RGB images. Once trained with targeted UAS RGB data, the spatial extension model allows farmers to stop repeated UAS flights. While we introduce super-resolution advances, the core contribution is a lightweight and scalable system for affordable on-farm use.

Figures

Figures reproduced from arXiv: 2505.21746 by the authors.

Figure 1
Figure 1. Application scenarios of super-resolution in cost-effective precision farming. In all these scenarios we present [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. A comparison of common super-resolution modeling frameworks in terms of the input and output targets’ [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. (A) Cover cropping sites - two farms are near Chestertown, MD and two farms near Easton, MD, studied over [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (14 more)
Figure 4
Figure 4. Figure 4: Hyperspectral flights locations. The study sites are marked as "Maryland cover crops" in light green, while [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Measured normalized spectral response function (SRF) for Sentinel-2A and Sentinel-2B (S2-SRF) for VNIR [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Simulated Spectral response average of 8 Sentinel-2 VNIR bands using the Headwall Nano-Hyperspectral [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: (A) Transforming and aligning the UAS image to the Sentinel-2 image. The middle image shows that simply [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]
Figure 8
Figure 8. Figure 8: Optimal shift values found for each image-to-image registration. Sentinel-2 tiles cover more than 100 km [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]
Figure 9
Figure 9. Figure 9: An example image-to-image registration for site I. The invalid (no-data) hyperspectral pixels are also masked [PITH_FULL_IMAGE:figures/full_fig_p011_9.png]
Figure 10
Figure 10. Figure 10: The proposed end-to-end system for data preparation, super-resolution fusion and predictive modeling [PITH_FULL_IMAGE:figures/full_fig_p013_10.png]
Figure 11
Figure 11. Figure 11: The SRCNN network architecture. Other Spectral Extension Super-resolution Networks (VGGExt, ResNetExt, and SRDB): We have deviated from the original SRCNN architecture by not using batch normalization, and instead of the ReLU activation we use the LeakyReLU activation…
Figure 12
Figure 12. Figure 12: RMSEs of biomass (top row) and N estimations (bottom row) as functions of spectral range (M7-M11), [PITH_FULL_IMAGE:figures/full_fig_p017_12.png]
Figure 13
Figure 13. Figure 13: The figure shows a UAS image from a winter wheat field in Washington state in the center surrounded by [PITH_FULL_IMAGE:figures/full_fig_p021_13.png]
Figure 15
Figure 15. Figure 15: We have little evidence to assess how well the super-resolution methods will generalize outside of the location [PITH_FULL_IMAGE:figures/full_fig_p022_15.png]
Figure 14
Figure 14. Figure 14: Effectiveness of different sensors for super-resolution across spectral, spatial, and temporal domains in [PITH_FULL_IMAGE:figures/full_fig_p023_14.png]
Figure 15
Figure 15. Figure 15: The spectral extension model compared to the ground-truth generated by the hyperspectral camera for site I. [PITH_FULL_IMAGE:figures/full_fig_p023_15.png]
Figure 16
Figure 16. Figure 16: The spatial extension model M32 compared to the ground-truth generated by the hyperspectral camera for site E. The first column is the Sentinel-2 input to the model, while the second column is the super-resolved prediction and the third column is the desired ground-tr…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

79 extracted references · 76 canonical work pages

  1. [1]

    A., and Chandra, R

    Adigun, O., Olsen, P. A., and Chandra, R. (2022). Location aware super-resolution for satellite data fusion. In IGARSS 2022-2022 IEEE International Geoscience and Remote Sensing Symposium , pages 3758--3761. IEEE

    work page 2022

  2. [2]

    R., Nguyen, H., Rau, B

    Adler, P. R., Nguyen, H., Rau, B. M., and Dell, C. J. (2024). Modeling n2o emissions with remotely sensed variables using machine learning. Environmental Research Communications , 6(9):091004

    work page 2024

  3. [3]

    Alvarez-Vanhard, E., Corpetti, T., and Houet, T. (2021). Uav & satellite synergies for optical remote sensing applications: A literature review. Science of remote sensing , 3:100019

    work page 2021

  4. [4]

    Argento, F., Anken, T., Abt, F., Vogelsanger, E., Walter, A., and Liebisch, F. (2021). Site-specific nitrogen management in winter wheat supported by low-altitude remote sensing and soil data. Precision Agriculture , 22:364--386

    work page 2021

  5. [5]

    J., Myers-Smith, I

    Assmann, J. J., Myers-Smith, I. H., Kerby, J. T., Cunliffe, A. M., and Daskalova, G. N. (2020). Drone data reveal heterogeneity in tundra greenness and phenology not captured by satellites. Environmental Research Letters , 15(12):125002

    work page 2020

  6. [6]

    Belgiu, M., Marshall, M., Boschetti, M., Pepe, M., Stein, A., and Nelson, A. (2023). Prisma and sentinel-2 spectral response to the nutrient composition of grains. Remote Sensing of Environment , 292:113567

    work page 2023

  7. [7]

    Berger, K., Verrelst, J., F \'e ret, J.-B., Wang, Z., Wocher, M., Strathmann, M., Danner, M., Mauser, W., and Hank, T. (2020). Crop nitrogen monitoring: Recent progress and principal developments in the context of imaging spectroscopy missions. Remote Sensing of Environment , 242:111758

    work page 2020

  8. [8]

    Brook, A., De Micco, V., Battipaglia, G., Erbaggio, A., Ludeno, G., Catapano, I., and Bonfante, A. (2020). A smart multiple spatial and temporal resolution system to support precision agriculture from satellite images: Proof of concept on aglianico vineyard. Remote Sensing of Environment , 240:111679

    work page 2020

Show all 79 references
  1. [9]

    Clevers, J. G. and Gitelson, A. A. (2013). Remote estimation of crop and grass chlorophyll and nitrogen content using red-edge bands on sentinel-2 and-3. International Journal of Applied Earth Observation and Geoinformation , 23:344--351

    2013

  2. [10]

    P., Pearse, G

    Dash, J. P., Pearse, G. D., and Watt, M. S. (2018). Uav multispectral imagery can complement satellite data for monitoring forest health. Remote Sensing , 10(8):1216

    2018

  3. [11]

    de Oca, A. M. and Flores, G. (2021). The agriq: A low-cost unmanned aerial system for precision agriculture. Expert Systems with Applications , 182:115163

    2021

  4. [12]

    M., Guan, K., Lopez, B., Zhou, Q., White, C

    Deines, J. M., Guan, K., Lopez, B., Zhou, Q., White, C. S., Wang, S., and Lobell, D. B. (2023). Recent cover crop adoption is associated with small maize and soybean yield losses in the united states. Global change biology , 29(3):794--807

    2023

  5. [13]

    do Nascimento Bendini, H., Fieuzal, R., Carrere, P., Clenet, H., Galvani, A., Allies, A., and Ceschia, \'E . (2024). Estimating winter cover crop biomass in france using optical sentinel-2 dense image time series and machine learning. Remote Sensing , 16(5):834

    2024

  6. [14]

    C., He, K., and Tang, X

    Dong, C., Loy, C. C., He, K., and Tang, X. (2014). Learning a deep convolutional network for image super-resolution. In European conference on computer vision , pages 184--199. Springer

    2014

  7. [15]

    C., He, K., and Tang, X

    Dong, C., Loy, C. C., He, K., and Tang, X. (2015). Image super-resolution using deep convolutional networks. IEEE transactions on pattern analysis and machine intelligence , 38(2):295--307

    2015

  8. [16]

    C., and Tang, X

    Dong, C., Loy, C. C., and Tang, X. (2016). Accelerating the super-resolution convolutional neural network. In European conference on computer vision , pages 391--407. Springer

    2016

  9. [17]

    Small unmanned aircraft systems (uas) regulations

    FAA (2020). Small unmanned aircraft systems (uas) regulations. Accessed: November 25, 2024

    2020

  10. [18]

    Fan, X., Vrieling, A., Muller, B., and Nelson, A. (2020). Winter cover crops in dutch maize fields: Variability in quality and its drivers assessed from multi-temporal sentinel-2 imagery. International Journal of Applied Earth Observation and Geoinformation , 91:102139

    2020

  11. [19]

    Galar, M., Sesma, R., Ayala, C., Albizua, L., and Aranda, C. (2020). Super-resolution of sentinel-2 images using convolutional neural networks and real ground truth data. Remote Sensing , 12(18):2941

    2020

  12. [20]

    C., and Hively, W

    Gao, F., Anderson, M. C., and Hively, W. D. (2020). Detecting cover crop end-of-season using ven s and sentinel-2 satellite imagery. Remote Sensing , 12(21):3524

    2020

  13. [21]

    Goffart, D., Curnel, Y., Planchon, V., Goffart, J.-P., and Defourny, P. (2021). Field-scale assessment of belgian winter cover crops biomass based on sentinel-2 data. European Journal of Agronomy , 126:126278

    2021

  14. [22]

    Gommers, R., Virtanen, P., Burovski, E., Weckesser, W., Oliphant, T. E., Haberland, M., Cournapeau, D., Reddy, T., Alexbrc , Peterson, P., Nelson, A., Wilson, J., Endolith , Mayorov, N., Ilhan Polat , Van Der Walt, S., Roy, P., Brett, M., Laxalde, D., Larson, E., Millman, J., ...

    2022

  15. [23]

    K., Bagavathiannan, M., Mowrer, J., Jagannadham, P

    Govindasamy, P., Muthusamy, S. K., Bagavathiannan, M., Mowrer, J., Jagannadham, P. T. K., Maity, A., Halli, H. M., GK, S., Vadivel, R., TK, D., et al. (2023). Nitrogen use efficiency—a key to enhance crop productivity under a changing climate. Frontiers in Plant Science , 14:1121073

    2023

  16. [24]

    Gr \"u ner, E., Astor, T., and Wachendorf, M. (2021). Prediction of biomass and n fixation of legume--grass mixtures using sensor fusion. Frontiers in Plant Science , 11:603921

    2021

  17. [25]

    He, K., Zhang, X., Ren, S., and Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 770--778

    2016

  18. [26]

    D., Lee, S., Sadeghi, A

    Hively, W. D., Lee, S., Sadeghi, A. M., McCarty, G. W., Lamb, B. T., Soroka, A., Keppler, J., Yeo, I.-Y., and Moglen, G. E. (2020). Estimating the effect of winter cover crops on nitrogen leaching using cost-share enrollment data, satellite remote sensing, and soil and water a...

    2020

  19. [27]

    a biger, T., Rose, T., Kage, H., and K \

    Holzhauser, K., R \"a biger, T., Rose, T., Kage, H., and K \"u hling, I. (2022). Estimation of biomass and n uptake in different winter cover crops from uav-based multispectral canopy reflectance data. Remote Sensing , 14(18):4525

    2022

  20. [28]

    Huang, G., Liu, Z., Van Der Maaten, L., and Weinberger, K. Q. (2017). Densely connected convolutional networks. In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 4700--4708

    2017

  21. [29]

    Hunt Jr, E. R. and Daughtry, C. S. (2018). What good are unmanned aircraft systems for agricultural remote sensing and precision agriculture? International journal of remote sensing , 39(15-16):5345--5376

    2018

  22. [30]

    D., Samadzadegan, F., Mehravar, S., Toosi, A., Khatami, R., and Stein, A

    Javan, F. D., Samadzadegan, F., Mehravar, S., Toosi, A., Khatami, R., and Stein, A. (2021). A review of image fusion techniques for pan-sharpening of high-resolution satellite imagery. ISPRS journal of photogrammetry and remote sensing , 171:101--117

    2021

  23. [31]

    Jonak, M., Mucha, J., Jezek, S., Kovac, D., and Cziria, K. (2024). Spagri-ai: Smart precision agriculture dataset of aerial images at different heights for crop and weed detection using super-resolution. Agricultural Systems , 216:103876

    2024

  24. [32]

    KC, K., Zhao, K., Romanko, M., and Khanal, S. (2021). Assessment of the spatial and temporal patterns of cover crops using remote sensing. Remote Sensing , 13(14):2689

    2021

  25. [33]

    Kong, J., Ryu, Y., Jeong, S., Zhong, Z., Choi, W., Kim, J., Lee, K., Lim, J., Jang, K., Chun, J., et al. (2023). Super resolution of historic landsat imagery using a dual generative adversarial network (gan) model with cubesat constellation imagery for spatially enhanced long-...

    2023

  26. [34]

    Lanaras, C., Bioucas-Dias, J., Galliani, S., Baltsavias, E., and Schindler, K. (2018). Super-resolution of sentinel-2 images: Learning a globally applicable deep neural network. ISPRS Journal of Photogrammetry and Remote Sensing , 146:305--319

    2018

  27. [35]

    Lawson, C. L. and Hanson, R. J. (1995). Solving least squares problems . SIAM

    1995

  28. [36]

    Li, J., Hong, D., Gao, L., Yao, J., Zheng, K., Zhang, B., and Chanussot, J. (2022). Deep learning in multimodal remote sensing data fusion: A comprehensive review. International Journal of Applied Earth Observation and Geoinformation , 112:102926

    2022

  29. [37]

    Li, S., Kang, X., Fang, L., Hu, J., and Yin, H. (2017). Pixel-level image fusion: A survey of the state of the art. information Fusion , 33:100--112

    2017

  30. [38]

    M., and Ghamkhar, K

    Li, Y., Al-Sarayreh, M., Irie, K., Hackell, D., Bourdot, G., Reis, M. M., and Ghamkhar, K. (2021). Identification of weeds based on hyperspectral imaging and machine learning. Frontiers in Plant Science , 11:611622

    2021

  31. [39]

    Liu, H., Fu, Z., Han, J., Shao, L., and Liu, H. (2018). Single satellite imagery simultaneous super-resolution and colorization using multi-task deep neural networks. Journal of Visual Communication and Image Representation , 53:20--30

    2018

  32. [40]

    Ma, J., Yu, W., Chen, C., Liang, P., Guo, X., and Jiang, J. (2020). Pan-GAN : An unsupervised pan-sharpening method for remote sensing image fusion. Information Fusion , 62:110--120

    2020

  33. [41]

    M., Erkbol, H., and Fritschi, F

    Maimaitijiang, M., Sagan, V., Sidike, P., Daloye, A. M., Erkbol, H., and Fritschi, F. B. (2020). Crop monitoring using satellite/uav data fusion and machine learning. Remote Sensing , 12(9):1357

    2020

  34. [42]

    Marx, A., McFarlane, D., and Alzahrani, A. (2017). Uav data for multi-temporal landsat analysis of historic reforestation: a case study in costa rica. International journal of remote sensing , 38(8-10):2331--2348

    2017

  35. [43]

    Meng, C., Shu, L., Han, R., Chen, Y., Yi, L., and Deng, D.-J. (2024). Farmsr: Super-resolution in precision agriculture field production scenes. In 2024 IEEE 22nd International Conference on Industrial Informatics (INDIN) , pages 1--6. IEEE

    2024

  36. [44]

    and Anderson, J

    Olson, D. and Anderson, J. (2021). Review on unmanned aerial vehicles, remote sensors, imagery processing, and their applications in agriculture. Agronomy Journal , 113(2):971--992

    2021

  37. [45]

    Ouhami, M., Hafiane, A., Es-Saady, Y., El Hajji, M., and Canals, R. (2021). Computer vision, iot and data fusion for crop disease detection using machine learning: A survey and ongoing research. Remote Sensing , 13(13):2486

    2021

  38. [46]

    Puliti, S., Saarela, S., Gobakken, T., St hl, G., and N sset, E. (2018). Combining uav and sentinel-2 auxiliary data for forest growing stock volume estimation through hierarchical model-based inference. Remote sensing of environment , 204:485--497

    2018

  39. [47]

    Radoglou-Grammatikis, P., Sarigiannidis, P., Lagkas, T., and Moscholios, I. (2020). A compilation of uav applications for precision agriculture. Computer Networks , 172:107148

    2020

  40. [48]

    Rossi, M., Candiani, G., Nutini, F., Gianinetto, M., and Boschetti, M. (2023). Sentinel-2 estimation of cnc and lai in rice cropping system through hybrid approach modelling. European Journal of Remote Sensing , 56(1):2117651

    2023

  41. [49]

    and Streit, B

    Roth, L. and Streit, B. (2018). Predicting cover crop biomass by lightweight uas-based rgb and nir photography: an applied photogrammetric approach. Precision Agriculture , 19:93--114

    2018

  42. [50]

    Spectral responses (s2a)

    S2A (2024). Spectral responses (s2a). Accessed: November 25, 2024

    2024

  43. [51]

    G., and Fritschi, F

    Sagan, V., Maimaitijiang, M., Sidike, P., Maimaitiyiming, M., Erkbol, H., Hartling, S., Peterson, K., Peterson, J., Burken, J. G., and Fritschi, F. (2019). Uav/satellite multiscale data fusion for crop monitoring and early stress detection. In Int. Arch. Photogramm. Remote Sen...

    2019

  44. [52]

    J., and Norouzi, M

    Saharia, C., Ho, J., Chan, W., Salimans, T., Fleet, D. J., and Norouzi, M. (2021). Image super-resolution via iterative refinement

    2021

  45. [53]

    J., and Norouzi, M

    Saharia, C., Ho, J., Chan, W., Salimans, T., Fleet, D. J., and Norouzi, M. (2022). Image super-resolution via iterative refinement. IEEE transactions on pattern analysis and machine intelligence , 45(4):4713--4726

    2022

  46. [54]

    Salehin, S. M. U., Poudyal, C., Rajan, N., and Bagavathiannan, M. (2025). Cover crop types influence biomass estimation using unmanned aerial vehicle-mounted multispectral sensors. Remote Sensing , 17(8):1471

    2025

  47. [55]

    Salgueiro Romero, L., Marcello, J., and Vilaplana, V. (2020). Super-resolution of sentinel-2 imagery using generative adversarial networks. Remote Sensing , 12(15):2424

    2020

  48. [56]

    Samadzadegan, F., Toosi, A., and Dadrass Javan, F. (2025). A critical review on multi-sensor and multi-platform remote sensing data fusion approaches: current status and prospects. International Journal of Remote Sensing , 46(3):1327--1402

    2025

  49. [57]

    V., Atkinson Amorim, J

    Schreiber, L. V., Atkinson Amorim, J. G., Guimar \ a es, L., Motta Matos, D., Maciel da Costa, C., and Parraga, A. (2022). Above-ground biomass wheat estimation: deep learning with uav-based rgb images. Applied Artificial Intelligence , 36(1):2055392

    2022

  50. [58]

    Sedghi, N., Fox, R., Sherman, L., Gaudlip, C., and Weil, R. (2023). Aerial interseeding and planting green to enhance nitrogen capture and cover crop biomass carbon. Journal of Soil and Water Conservation , 78(3):282--298

    2023

  51. [59]

    K., Laguette, S., Casady, G

    Seelan, S. K., Laguette, S., Casady, G. M., and Seielstad, G. A. (2003). Remote sensing applications for precision agriculture: A learning community approach. Remote sensing of environment , 88(1-2):157--169

    2003

  52. [60]

    Sharifi, A. (2020). Using sentinel-2 data to predict nitrogen uptake in maize crop. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 13:2656--2662

    2020

  53. [61]

    and Zisserman, A

    Simonyan, K. and Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556

    2014 arXiv

  54. [62]

    Simula, A. (2021). Establishing drone technology to agriculture as a service provider. Master's thesis, Jyväskylä University of Applied Sciences

    2021

  55. [63]

    C., Petropoulos, G

    Singh, P., Pandey, P. C., Petropoulos, G. P., Pavlides, A., Srivastava, P. K., Koutsias, N., Deng, K. A. K., and Bao, Y. (2020). Hyperspectral remote sensing in precision agriculture: Present status, challenges, and future trends. In Hyperspectral remote sensing , pages 121--1...

    2020

  56. [64]

    J., Vaughan, I

    Stark, D. J., Vaughan, I. P., Evans, L. J., Kler, H., and Goossens, B. (2018). Combining drones and satellite tracking as an effective tool for informing policy change in riparian habitats: a proboscis monkey case study. Remote Sensing in Ecology and Conservation , 4(1):44--52

    2018

  57. [65]

    A., Valentino, G., Chen, M., Briffa, J

    Tarasiewicz, T., Nalepa, J., Farrugia, R. A., Valentino, G., Chen, M., Briffa, J. A., and Kawulok, M. (2023). Multitemporal and multispectral data fusion for super-resolution of sentinel-2 images. IEEE Transactions on Geoscience and Remote Sensing , 61:1--19

    2023

  58. [66]

    C., Fitz, J

    Thieme, A., Yadav, S., Oddo, P. C., Fitz, J. M., McCartney, S., King, L., Keppler, J., McCarty, G. W., and Hively, W. D. (2020). Using nasa earth observations and google earth engine to map winter cover crop conservation performance in the chesapeake bay watershed. Remote Sens...

    2020

  59. [67]

    Toosi, A., Samadzadegan, F., and Javan, F. D. (2025). Toward the optimal spatial resolution ratio for fusion of uav and sentinel-2 satellite imageries using metaheuristic optimization. Advances in Space Research

    2025

  60. [68]

    Wang, T., Liu, Y., Wang, M., Fan, Q., Tian, H., Qiao, X., and Li, Y. (2021). Applications of uas in crop biomass monitoring: A review. Frontiers in Plant Science , 12:616689

    2021

  61. [69]

    S., Heaphy, M., Dunningham, A., and Rolando, C

    Watt, M. S., Heaphy, M., Dunningham, A., and Rolando, C. (2017). Use of remotely sensed data to characterize weed competition in forest plantations. International Journal of Remote Sensing , 38(8-10):2448--2463

    2017

  62. [70]

    Weiss, M., Jacob, F., and Duveiller, G. (2020). Remote sensing for agricultural applications: A meta-review. Remote sensing of environment , 236:111402

    2020

  63. [71]

    Wu, M., Jin, X., Jiang, Q., Lee, S.-j., Liang, W., Lin, G., and Yao, S. (2021). Remote sensing image colorization using symmetrical multi-scale dcgan in yuv color space. The Visual Computer , 37:1707--1729

    2021

  64. [72]

    Xia, Y., Guan, K., Copenhaver, K., and Wander, M. (2021). Estimating cover crop biomass nitrogen credits with sentinel-2 imagery and sites covariates. Agronomy Journal , 113(2):1084--1101

    2021

  65. [73]

    Yuan, M., Burjel, J., Isermann, J., Goeser, N., and Pittelkow, C. (2019). Unmanned aerial vehicle--based assessment of cover crop biomass and nitrogen uptake variability. Journal of Soil and Water Conservation , 74(4):350--359

    2019

  66. [74]

    C., Martin, N

    Yuan, M., Burjel, J. C., Martin, N. F., Isermann, J., Goeser, N., and Pittelkow, C. M. (2021). Advancing on-farm research with uavs: Cover crop effects on crop growth and yield. Agronomy Journal , 113(2):1071--1083

    2021

  67. [75]

    and Kovacs, J

    Zhang, C. and Kovacs, J. M. (2012). The application of small unmanned aerial systems for precision agriculture: a review. Precision agriculture , 13:693--712

    2012

  68. [76]

    A., Mauzerall, D

    Zhang, X., Davidson, E. A., Mauzerall, D. L., Searchinger, T. D., Dumas, P., and Shen, Y. (2015). Managing nitrogen for sustainable development. Nature , 528(7580):51--59

    2015

  69. [77]

    Zhang, Y., Li, K., Li, K., Wang, L., Zhong, B., and Fu, Y. (2018a). Image super-resolution using very deep residual channel attention networks. In Proceedings of the European conference on computer vision (ECCV) , pages 286--301

    2018

  70. [78]

    Zhang, Y., Tian, Y., Kong, Y., Zhong, B., and Fu, Y. (2018b). Residual dense network for image super-resolution. In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 2472--2481

    2018

  71. [79]

    Zhu, H., Lin, C., Liu, G., Wang, D., Qin, S., Li, A., Xu, J.-L., and He, Y. (2024). Intelligent agriculture: Deep learning in uav-based remote sensing imagery for crop diseases and pests detection. Frontiers in Plant Science , 15:1435016

    2024

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.