pith. sign in

arxiv: 1906.10413 · v1 · pith:NGJ5NPYCnew · submitted 2019-06-25 · 💻 cs.CV · eess.IV

A CNN-Based Super-Resolution Technique for Active Fire Detection on Sentinel-2 Data

Massimiliano Gargiulo , Domenico Antonio Giuseppe Dell'Aglio , Antonio Iodice , Daniele Riccio , Giuseppe Ruello This is my paper

Pith reviewed 2026-05-25 16:55 UTC · model grok-4.3

classification 💻 cs.CV eess.IV
keywords super-resolutionCNNSentinel-2active fire detectionSWIRremote sensingdata fusion
0
0 comments X

The pith

CNN-based super-resolution enhances Sentinel-2 SWIR bands to 10 meters for active fire detection.

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

The paper proposes using a convolutional neural network to increase the spatial resolution of Sentinel-2's short-wave infrared bands from 20 meters to 10 meters. This aims to produce more detailed maps of active fires by applying standard detection indices to the finer data. The approach is shown to outperform other super-resolution methods on accuracy measures and is demonstrated on fires that occurred near Mount Vesuvius in 2017. A sympathetic reader would care because Sentinel-2 offers frequent coverage but its coarser SWIR bands limit the precision of fire monitoring applications.

Core claim

The proposed CNN-based super-resolution data fusion method achieves better results than alternative methods in terms of some accuracy metrics when moving the SWIR bands of Sentinel-2 toward 10-m spatial resolution. When these super-resolved bands are used to monitor active fire through classic indices, the method provides advantages and limits that are validated on the mount Vesuvius area damaged by fires in summer 2017.

What carries the argument

A convolutional neural network for super-resolution data fusion that upsamples the 20-meter SWIR bands to 10-meter resolution while aiming to preserve spectral properties.

If this is right

  • The super-resolved bands enable more detailed active fire detection maps.
  • The CNN method outperforms alternatives on accuracy metrics.
  • The approach is validated through application to real fire events using standard indices.
  • Advantages and limits are identified for the specific geographical test area.

Where Pith is reading between the lines

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

  • This technique could potentially improve fire detection in other regions with similar Sentinel-2 data.
  • Integration with other resolution enhancement methods might further refine results.
  • Real-world deployment could lead to earlier or more precise fire response if spectral fidelity holds.

Load-bearing premise

The CNN super-resolution must preserve the spectral fidelity of the SWIR bands so that active fire indices remain reliable without artifacts that distort detection on actual fire events.

What would settle it

Ground-truth fire locations from independent high-resolution imagery or field reports where the fire indices from CNN-super-resolved bands show no improvement or degradation compared to native 20m bands.

Figures

Figures reproduced from arXiv: 1906.10413 by Antonio Iodice, Daniele Riccio, Domenico Antonio Giuseppe Dell'Aglio, Giuseppe Ruello, Massimiliano Gargiulo.

Figure 1
Figure 1. Figure 1: (a) false colour composite (ρ12, ρ11 and ρ8 bands) and (b) RGB image of Vesuvius . 3. METHODOLOGY 3.1. Proposed CNN-based Super-Resolution Fusion Our goal is to improve the spatial resolution of SWIR bands using a Convolutional Neural Network (CNN). CNNs have attracted an increasing interest in many remote sensing applications, like object detection [10], classification [11], pansharpening [12], and others… view at source ↗
Figure 2
Figure 2. Figure 2: Active Fire Indices related to Vesuvius. 3.3. Results Accuracy Metrics To evaluate the performance when the target image is available (in our case, at 20-m spatial resolution), the proposed method is compared to alternative methods using four reference metrics, commonly used for pansharpening [20]: - Spectral Angular Mapper (SAM) the spectral distortion between pixel of reference image and estimated one [2… view at source ↗
Figure 3
Figure 3. Figure 3: Detail of the study area obtained by several super-resolution techniques and our proposal to underline the improvement in terms of spectral distortion. In the middle of the first row: z is only composed by RGB bands. 4.2. Comparison between Super-Resolution Proposal and SISR/SRDF techniques In this section, SRNN+ is compared to a pre-trained CNN-based method (SRNN), three popular SRDFs adapted to the Senti… view at source ↗
Figure 4
Figure 4. Figure 4: Detail of the area under investigation obtained by several super-resolution techniques and our proposal to underline the improvement in terms of spectral distortion. In the middle of the first row: z is only composed by RGB bands that are affected by smoke presence (in the CNN input z is also composed by ρ8 band. 4.3. Comparison between Different AFIs and Maps Once active fire (AF) is detected by consideri… view at source ↗
Figure 5
Figure 5. Figure 5: In the first row the RGB image in which we can observe the presence of the smoke and the ground truth. Then, from the second row to the bottom: in the first column false-RGB, in the second AF I1 and AF I3, in the third the respective Maps. REFERENCES 1. Neha Joshi, Matthias Baumann, Andrea Ehammer, Rasmus Fensholt, Kenneth Grogan, Patrick Hostert, Martin Jepsen, Tobias Kuemmerle, Patrick Meyfroidt, Edward … view at source ↗
Figure 6
Figure 6. Figure 6: In the first row the RGB image in which we can observe the absense of the smoke and the ground truth. Then, from the second row to the bottom: in the first column false-RGB, in the second AF I2, and in the third the respective Map. 11. Alex Krizhevsky, Ilya Sutskever, and Geoffrey E. Hinton, “Imagenet classification with deep convolutional neural networks,” pp. 1106–1114, 2012. 12. G. Scarpa, S. Vitale, an… view at source ↗
read the original abstract

Remote Sensing applications can benefit from a relatively fine spatial resolution multispectral (MS) images and a high revisit frequency ensured by the twin satellites Sentinel-2. Unfortunately, only four out of thirteen bands are provided at the highest resolution of 10 meters, and the others at 20 or 60 meters. For instance the Short-Wave Infrared (SWIR) bands, provided at 20 meters, are very useful to detect active fires. Aiming to a more detailed Active Fire Detection (AFD) maps, we propose a super-resolution data fusion method based on Convolutional Neural Network (CNN) to move towards the 10-m spatial resolution the SWIR bands. The proposed CNN-based solution achieves better results than alternative methods in terms of some accuracy metrics. Moreover we test the super-resolved bands from an application point of view by monitoring active fire through classic indices. Advantages and limits of our proposed approach are validated on specific geographical area (the mount Vesuvius, close to Naples) that was damaged by widespread fires during the summer of 2017.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 1 minor

Summary. The manuscript proposes a CNN-based super-resolution data fusion method to enhance the spatial resolution of Sentinel-2 SWIR bands from 20m to 10m for more detailed active fire detection. It claims better results than alternative methods in some accuracy metrics and validates the approach by testing super-resolved bands with classic active fire indices on the 2017 Vesuvius fire events.

Significance. If the quantitative results support the claims, this could be a useful contribution to remote sensing by improving the utility of Sentinel-2 data for fire monitoring applications. The application point of view validation on real events is a strength that goes beyond pure metric comparison.

major comments (2)
  1. Abstract: the claim of superior accuracy metrics and successful index-based monitoring supplies no quantitative values, error bars, dataset sizes, or validation details, making the central claim impossible to assess from the provided text.
  2. Validation section on Vesuvius events: the central application claim requires explicit evidence that super-resolution preserves SWIR spectral fidelity (e.g., before/after comparison of fire index values or detection maps on the 2017 events) to rule out artifacts that could distort results.
minor comments (1)
  1. Methods: expand the CNN architecture description with details on the number of input bands, loss function, and training dataset size to support reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed review and constructive suggestions. We address the two major comments below and will incorporate revisions to improve clarity and evidence in the manuscript.

read point-by-point responses

  1. Referee: Abstract: the claim of superior accuracy metrics and successful index-based monitoring supplies no quantitative values, error bars, dataset sizes, or validation details, making the central claim impossible to assess from the provided text.

    Authors: We agree that the abstract is too terse and does not allow readers to evaluate the central claims. In the revised manuscript we will expand the abstract to include the key quantitative accuracy metrics (PSNR/SSIM values and the specific indices where improvement is observed), the size of the training and test sets, and a brief statement of the Vesuvius validation protocol. revision: yes

  2. Referee: Validation section on Vesuvius events: the central application claim requires explicit evidence that super-resolution preserves SWIR spectral fidelity (e.g., before/after comparison of fire index values or detection maps on the 2017 events) to rule out artifacts that could distort results.

    Authors: The current validation demonstrates that the super-resolved SWIR bands can be used with standard active-fire indices on the 2017 Vesuvius events, but we acknowledge that direct before/after comparisons of index values and detection maps are not presented. We will add these explicit comparisons (original 20 m vs. super-resolved 10 m index maps and quantitative differences) to the revised validation section to demonstrate spectral fidelity. revision: yes

Circularity Check

0 steps flagged

No significant circularity identified

full rationale

The paper presents a standard supervised CNN pipeline for super-resolving Sentinel-2 SWIR bands to 10 m, followed by empirical comparison against alternative methods on accuracy metrics and downstream validation via active-fire indices on the 2017 Vesuvius events. No load-bearing step reduces by construction to its own inputs, fitted parameters renamed as predictions, or self-citation chains; the central claims rest on data-driven training and external benchmarks rather than definitional equivalence or imported uniqueness theorems.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract supplies no explicit free parameters, axioms, or invented entities; the approach rests on standard CNN training assumptions not detailed here.

pith-pipeline@v0.9.0 · 5733 in / 1018 out tokens · 20268 ms · 2026-05-25T16:55:27.835895+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

26 extracted references · 26 canonical work pages · 2 internal anchors

  1. [1]

    A review of the application of optical and radar remote sensing data fusion to land use mapping and monitoring,

    Neha Joshi, Matthias Baumann, Andrea Ehammer, Rasmus Fensholt, Kenneth Grogan, Patrick Hostert, Martin Jepsen, Tobias Kuemmerle, Patrick Meyfroidt, Edward Mitchard, et al., “A review of the application of optical and radar remote sensing data fusion to land use mapping and monitoring,” Remote Sensing, vol. 8, no. 1, pp. 70, 2016

    work page 2016

  2. [2]

    Sentinel-2: Esa’s optical high-resolution mission for gmes operational services,

    M. Drusch et al., “Sentinel-2: Esa’s optical high-resolution mission for gmes operational services,” Remote Sensing of Environment, vol. 120, no. Supplement C, pp. 25 – 36, 2012, The Sentinel Missions - New Opportunities for Science

    work page 2012

  3. [3]

    The potential of Sentinel satellites for burnt area mapping and monitoring in the Congo Basin forests,

    Astrid Verhegghen, Hugh Eva, Guido Ceccherini, Frederic Achard, Valery Gond, Sylvie Gourlet-Fleury, and Paolo Cerutti, “The potential of Sentinel satellites for burnt area mapping and monitoring in the Congo Basin forests,” Remote Sensing, vol. 8, no. 12, pp. 986, 2016

    work page 2016

  4. [4]

    Landsat-8 and Sentinel-2 for fire monitoring at a local scale: A case study on Vesuvius,

    L Cicala, CV Angelino, N Fiscante, and SL Ullo, “Landsat-8 and Sentinel-2 for fire monitoring at a local scale: A case study on Vesuvius,” in 2018 IEEE International Conference on Environmental Engineering (EE). IEEE, 2018, pp. 1–6

    work page 2018

  5. [5]

    Remote sensing of burned areas: A review. A review of remote sensing methods for the study of large wildland fires,

    Jos´ e Pereira, Emilio Chuvieco, A Beudoin, and N Desbois, “Remote sensing of burned areas: A review. A review of remote sensing methods for the study of large wildland fires,”Departamento de Geografa, Universidad de Alcal , pp. 127–184, 01 1997

    work page 1997

  6. [6]

    Studies on land surface temperature over heteroge- neous areas using AVHRR data,

    Yogesh Kant and K. V. S. Badarinath, “Studies on land surface temperature over heteroge- neous areas using AVHRR data,” International Journal of Remote Sensing , vol. 21, no. 8, pp. 1749–1756, 2000

    work page 2000

  7. [7]

    Definitions and Architectures - Fusion of Images of Different Spatial Resolutions, Presses de l’Ecole, Ecole des Mines de Paris, Paris, France, 2002, ISBN 2-911762-38-X

    Lucien Wald, Data Fusion. Definitions and Architectures - Fusion of Images of Different Spatial Resolutions, Presses de l’Ecole, Ecole des Mines de Paris, Paris, France, 2002, ISBN 2-911762-38-X

    work page 2002

  8. [8]

    Vesuvius national park monitoring by COSMO-SkyMed PingPong data analysis,

    Lucio Mascolo, Maurizio Sarti, Ferdinando Nunziata, and Maurizio Migliaccio, “Vesuvius national park monitoring by COSMO-SkyMed PingPong data analysis,” in ESA Special Pub- lication, 2013, vol. 713

    work page 2013

  9. [9]

    Gli incendi boschivi stanno cambiando: cambiamo le strategie per governarli,

    G Bovio, M Marchetti, L Tonarelli, M Salis, G Vacchiano, R Lovreglio, M Elia, P Fiorucci, and D Ascoli, “Gli incendi boschivi stanno cambiando: cambiamo le strategie per governarli,” Foresta - Rivista di Selvicoltura ed Ecologia Forestale , , no. 4, pp. 202–205, 2017

    work page 2017

  10. [10]

    Geospatial Object Detection in High Resolution Satellite Images Based on Multi-Scale Convolutional Neural Network,

    Wei Guo, Wen Yang, Haijian Zhang, and Guang Hua, “Geospatial Object Detection in High Resolution Satellite Images Based on Multi-Scale Convolutional Neural Network,” Remote Sensing, vol. 10, no. 1, pp. 131, 2018. 7 (Ground-Truth) Bicubic GS2-GLP SRNN + (Proposed ) Figure 6: In the first row the RGB image in which we can observe the absense of the smoke and...

    work page 2018

  11. [11]

    Imagenet classification with deep convolutional neural networks,

    Alex Krizhevsky, Ilya Sutskever, and Geoffrey E. Hinton, “Imagenet classification with deep convolutional neural networks,” pp. 1106–1114, 2012

    work page 2012

  12. [12]

    Target-adaptive cnn-based pansharpening,

    G. Scarpa, S. Vitale, and D. Cozzolino, “Target-adaptive cnn-based pansharpening,” IEEE Transactions on Geoscience and Remote Sensing , vol. 56, no. 9, pp. 5443–5457, Sept 2018

    work page 2018

  13. [13]

    A CNN-Based Fusion Method for Super-Resolution of Sentinel-2 data,

    Massimiliano Gargiulo, Antonio Mazza, Raffaele Gaetano, Giuseppe Ruello, and Giuseppe Scarpa, “A CNN-Based Fusion Method for Super-Resolution of Sentinel-2 data,” IGARSS, 2018

    work page 2018

  14. [14]

    Pansharpening by convolutional neural networks,

    Giuseppe Masi, Davide Cozzolino, Luisa Verdoliva, and Giuseppe Scarpa, “Pansharpening by convolutional neural networks,” Remote Sensing, vol. 8, no. 7, pp. 594, 2016

    work page 2016

  15. [15]

    Neural Network Guidance for UAVs,

    Kyle D Julian and Mykel J Kochenderfer, “Neural Network Guidance for UAVs,” p. 1743, 2017

    work page 2017

  16. [16]

    Adam: A Method for Stochastic Optimization

    Diederik P Kingma and Jimmy Ba, “Adam: A method for stochastic optimization,” arXiv preprint arXiv:1412.6980, 2014

    work page internal anchor Pith review Pith/arXiv arXiv 2014

  17. [17]

    Separability Analysis of Sentinel-2A Multi-Spectral Instrument (MSI) Data for Burned Area Discrimination,

    Haiyan Huang, David Roy, Luigi Boschetti, Hankui Zhang, L Yan, Sanath Kumar, Jose Gomez- Dans, and Jian Li, “Separability Analysis of Sentinel-2A Multi-Spectral Instrument (MSI) Data for Burned Area Discrimination,” Remote Sensing, vol. 8, 11 2016. 8

    work page 2016

  18. [18]

    Active fire detection using Landsat-8/OLI data,

    Wilfrid Schroeder, Patricia Oliva, Louis Giglio, Brad Quayle, Eckehard Lorenz, and Fabiano Morelli, “Active fire detection using Landsat-8/OLI data,” Remote Sensing of Environment , vol. 185, 09 2015

    work page 2015

  19. [19]

    Infrared detection of active fires and burnt areas: theory and observations,

    A Barducci, D Guzzi, P Marcoionni, and I Pippi, “Infrared detection of active fires and burnt areas: theory and observations,” Infrared physics & technology, vol. 43, no. 3-5, pp. 119–125, 2002

    work page 2002

  20. [20]

    A review of quality metrics for fused image,

    P Jagalingam and Arkal Vittal Hegde, “A review of quality metrics for fused image,” Aquatic Procedia, vol. 4, pp. 133–142, 2015

    work page 2015

  21. [21]

    Comparison of pansharpening algorithms: Outcome of the 2006 GRS- S data-fusion contest,

    Luciano Alparone, Lucien Wald, Jocelyn Chanussot, Claire Thomas, Paolo Gamba, and Lori Mann Bruce, “Comparison of pansharpening algorithms: Outcome of the 2006 GRS- S data-fusion contest,” IEEE Transactions on Geoscience and Remote Sensing , vol. 45, no. 10, pp. 3012–3021, 2007

    work page 2006

  22. [22]

    A universal image quality index,

    Zhou Wang and A. C. Bovik, “A universal image quality index,” IEEE Signal Processing Letters, vol. 9, no. 3, pp. 81–84, March 2002

    work page 2002

  23. [23]

    Convolutional neural networks for medical image analysis: Full training or fine tuning?,

    Nima Tajbakhsh, Jae Y Shin, Suryakanth R Gurudu, R Todd Hurst, Christopher B Kendall, Michael B Gotway, and Jianming Liang, “Convolutional neural networks for medical image analysis: Full training or fine tuning?,” IEEE transactions on medical imaging , vol. 35, no. 5, pp. 1299–1312, 2016

    work page 2016

  24. [24]

    An overview of gradient descent optimization algorithms

    Sebastian Ruder, “An overview of gradient descent optimization algorithms,” arXiv preprint arXiv:1609.04747, 2016

    work page internal anchor Pith review Pith/arXiv arXiv 2016

  25. [25]

    A critical comparison among pan- sharpening algorithms,

    Gemine Vivone, Luciano Alparone, Jocelyn Chanussot, Mauro Dalla Mura, Andrea Garzelli, Giorgio A Licciardi, Rocco Restaino, and Lucien Wald, “A critical comparison among pan- sharpening algorithms,” IEEE Transactions on Geoscience and Remote Sensing , vol. 53, no. 5, pp. 2565–2586, 2015

    work page 2015

  26. [26]

    Comparison of three different methods to merge multires- olution and multispectral data: Landsat TM and SPOT panchromatic,

    P.S. Chavez and J.A. Anderson, “Comparison of three different methods to merge multires- olution and multispectral data: Landsat TM and SPOT panchromatic,” Photogramm. Eng. Remote Sens., vol. 57, no. 3, pp. 295–303, 1991

    work page 1991