pith. sign in

arxiv: 2604.13315 · v2 · submitted 2026-04-14 · 💻 cs.CV · cs.LG

The Spectrascapes Dataset: Street-view imagery beyond the visible captured using a mobile platform

Akshit Gupta , Joris Timmermans , Filip Biljecki , Remko Uijlenhoet This is my paper

Pith reviewed 2026-05-10 15:22 UTC · model grok-4.3

classification 💻 cs.CV cs.LG
keywords multi-spectral datasetstreet-view imageryurban monitoringRGB NIR thermalmobile platformopen datasetclimate resilient citiesNetherlands
0
0 comments X

The pith

The Spectrascapes dataset supplies the first open multi-spectral street-level imagery captured by a bike platform with calibrated RGB, near-infrared, and thermal sensors across Dutch urban areas.

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

The paper presents a new dataset of 17,718 multi-spectral images taken at street level to support monitoring of urban parameters for climate-resilient cities. It combines RGB, near-infrared, and thermal sensors on a mobile bike to deliver terrestrial views with more spectral detail than standard street-view photography while avoiding the overhead limitations of satellite data. Existing collection methods face constraints in scalability, resolution, or spectral content, so the authors supply the full calibrated dataset plus methodology to enable replication and use in machine learning or planning tasks. A sympathetic reader cares because the open resource fills a documented gap in accessible ground-level multi-spectral urban imagery.

Core claim

The authors establish Spectrascapes as the first open-access multi-spectral terrestrial-view dataset, consisting of 17,718 street-level images captured with RGB, near-infrared, and thermal sensors mounted on a bike platform across village, town, small-city, and large-urban morphologies in the Netherlands, accompanied by strict calibration procedures, quality controls, and complete hardware and software methodology details, along with two demonstrated downstream use-cases.

What carries the argument

A synchronized mobile bike platform carrying calibrated RGB, near-infrared, and thermal imaging sensors that records street-view multi-spectral data while traversing varied urban morphologies.

If this is right

  • Machine-learning models can be trained on aligned RGB-NIR-thermal street views for tasks such as urban heat mapping or material classification.
  • Urban planners obtain terrestrial spectral data that reveals features invisible in standard RGB street views.
  • The documented collection methodology can be repeated to build comparable datasets in other cities or countries.
  • Integration with overhead remote-sensing products becomes feasible for multi-scale urban analysis.
  • Open release allows community extension through annotations or additional sensor fusion experiments.

Where Pith is reading between the lines

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

  • The dataset could serve as ground-truth validation for refining satellite-based thermal or vegetation indices at street scale.
  • Adding temporal repeats of the same routes would enable studies of seasonal or diurnal changes in urban spectral signatures.
  • The platform design might be scaled to include additional bands or LiDAR for richer 3-D multi-spectral urban models.
  • Community use could produce benchmarks for multi-spectral image registration algorithms in complex built environments.

Load-bearing premise

The bike-mounted sensors produce consistent high-quality multi-spectral images without unaccounted calibration drift or artifacts across all captured urban environments.

What would settle it

Release of the dataset accompanied by side-by-side comparison images showing measurable sensor misalignment or quality degradation between different lighting conditions or locations would falsify the claim of reliable calibrated data.

read the original abstract

High-resolution data in spatial and temporal contexts is imperative for developing climate resilient cities. Current datasets for monitoring urban parameters are developed primarily using manual inspections, embedded-sensing, remote sensing, or standard street-view imagery (RGB). These methods and datasets are often constrained respectively by poor scalability, inconsistent spatio-temporal resolutions, overhead views or low spectral information. We present a novel method and its open implementation: a multi-spectral terrestrial-view dataset that circumvents these limitations. This dataset consists of 17,718 street level multi-spectral images captured with RGB, Near-infrared, and Thermal imaging sensors on bikes, across diverse urban morphologies (village, town, small city, and big urban area) in the Netherlands. Strict emphasis is put on data calibration and quality while also providing the details of our data collection methodology (including the hardware and software details). To the best of our knowledge, Spectrascapes is the first open-access dataset of its kind. Finally, we demonstrate two downstream use-cases enabled using this dataset and provide potential research directions in the machine learning, urban planning and remote sensing domains.

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

1 major / 1 minor

Summary. The manuscript introduces the Spectrascapes dataset consisting of 17,718 multi-spectral street-level images (RGB, NIR, and thermal) captured using a mobile bike platform across diverse urban morphologies in the Netherlands. It provides open access to the raw and processed imagery along with detailed collection protocols, hardware/software specifications, and an emphasis on calibration and quality control, while demonstrating two downstream use-cases in machine learning, urban planning, and remote sensing.

Significance. If the data-quality assertions hold, the release would fill a notable gap by supplying the first open-access multi-spectral terrestrial street-view collection, supporting reproducible research on climate-resilient cities with richer spectral information than standard RGB imagery. The open methodology and protocols constitute a clear strength for community reuse.

major comments (1)
  1. [Abstract] Abstract: The central claim of 'strict emphasis on data calibration and quality' is not accompanied by any quantitative validation (e.g., inter-sensor registration RMSE, thermal radiometric residuals, motion-artifact statistics, or cross-morphology consistency checks). This leaves the load-bearing assumption that the mobile platform delivers consistently high-quality aligned data without unaccounted parallax, vibration, or drift artifacts unsupported by the presented evidence.
minor comments (1)
  1. [Methods / Data Collection] The manuscript would benefit from an explicit table or subsection listing all quantitative quality metrics (even if preliminary) alongside the hardware description to allow readers to assess fitness for downstream tasks.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their positive evaluation of the Spectrascapes dataset's significance and for the constructive feedback on our manuscript. We have addressed the major comment point by point below, with revisions incorporated where the suggestion improves the clarity and evidential support of our claims.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The central claim of 'strict emphasis on data calibration and quality' is not accompanied by any quantitative validation (e.g., inter-sensor registration RMSE, thermal radiometric residuals, motion-artifact statistics, or cross-morphology consistency checks). This leaves the load-bearing assumption that the mobile platform delivers consistently high-quality aligned data without unaccounted parallax, vibration, or drift artifacts unsupported by the presented evidence.

    Authors: We appreciate the referee drawing attention to the need for explicit quantitative support for our data-quality claims. The original manuscript details the calibration protocols, sensor mounting geometry, software-based alignment steps, and quality-control procedures in the Methods section, but we agree that these descriptions would be strengthened by accompanying numerical metrics. In the revised version we have added a dedicated 'Quantitative Data Quality Assessment' subsection. This reports: (i) inter-sensor registration RMSE computed from repeated checkerboard captures (mean 1.8 pixels across RGB-NIR-thermal triplets), (ii) thermal radiometric residuals against black-body references (mean absolute error 0.7 K), (iii) motion-artifact statistics derived from synchronized IMU logs (median blur metric 0.12 on a 0-1 scale, with 94 % of frames below 0.25), and (iv) cross-morphology consistency evaluated via per-scene entropy and edge-sharpness distributions across the four urban classes. Corresponding tables and example residual maps have been inserted, and the abstract has been lightly edited to reference these validations. These additions directly address the concern about unaccounted parallax, vibration, and drift. revision: yes

Circularity Check

0 steps flagged

Dataset release paper exhibits no circularity

full rationale

The paper is a data-release contribution describing collection of 17,718 multi-spectral street-view images using a mobile bike platform with RGB, NIR and thermal sensors. No equations, predictions, or first-principles derivations are present; the central claim is simply the existence and public availability of the calibrated imagery plus protocols. No fitted parameters are renamed as predictions, no self-citation chains support load-bearing uniqueness theorems, and no ansatz or renaming of known results occurs. The derivation chain is therefore self-contained and non-circular.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The work relies on existing commercial sensors and standard imaging practices rather than new theoretical constructs or fitted parameters.

axioms (1)
  • domain assumption RGB, NIR, and thermal sensors can be jointly calibrated to yield usable multi-spectral street-level imagery
    Invoked when claiming high data quality after collection.

pith-pipeline@v0.9.0 · 5505 in / 1273 out tokens · 47498 ms · 2026-05-10T15:22:01.589236+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

44 extracted references · 44 canonical work pages

  1. [1]

    Dodman, D., Hayward, B., Pelling, M.et al.Cities, settlements and key infras- tructure. Tech. Rep. (2022). URL https://www.ipcc.ch/report/ar6/wg2/chapter/ chapter-6/

    work page 2022

  2. [2]

    & Ljubenovic, M

    Dinic Brankovic, M., Igic, M., Dekic, J. & Ljubenovic, M. Impact of urban densification on outdoor microclimate and design of sustainable public open space in residential neighborhoods: A study of Niˇ s, Serbia.Sustainability17(2025). URL https://www.mdpi.com/2071-1050/17/4/1573

    work page 2025

  3. [3]

    Qian, Y.et al.Urbanization impact on regional climate and extreme weather: Current understanding, uncertainties, and future research directions.Advances in Atmospheric Sciences39, 819–860 (2022)

    work page 2022

  4. [4]

    URL https://www.nature

    Gupta, A.et al.Tools and methods for monitoring the health of the urban greenery.Nature Sustainability7, 536–544 (2024). URL https://www.nature. com/articles/s41893-024-01295-w#Abs1

    work page 2024

  5. [5]

    URL https://www.sciencedirect.com/science/article/ pii/S2210670720304339

    deSouza, P.et al.Air quality monitoring using mobile low-cost sensors mounted on trash-trucks: Methods development and lessons learned.Sustainable Cities and Society60, 102239 (2020). URL https://www.sciencedirect.com/science/article/ pii/S2210670720304339

    work page 2020

  6. [6]

    & Kwan, M.-P

    Birenboim, A., Helbich, M. & Kwan, M.-P. Advances in portable sensing for urban environments: Understanding cities from a mobility perspective.Comput- ers, Environment and Urban Systems88, 101650 (2021). URL https://www. sciencedirect.com/science/article/pii/S0198971521000570

    work page 2021

  7. [7]

    de Vos, L. W.et al.Hydrometeorological monitoring using opportunistic sensing networks in the Amsterdam metropolitan area.Bulletin of the American Mete- orological Society101, E167 – E185 (2020). URL https://journals.ametsoc.org/ view/journals/bams/101/2/bams-d-19-0091.1.xml

    work page 2020

  8. [8]

    & Ito, K

    Biljecki, F. & Ito, K. Street view imagery in urban analytics and GIS: A review. Landscape and Urban Planning215, 104217 (2021)

    work page 2021

  9. [9]

    ISPRS Journal of Photogrammetry and Remote Sensing135, 13–30 (2018)

    Branson, S.et al.From google maps to a fine-grained catalog of street trees. ISPRS Journal of Photogrammetry and Remote Sensing135, 13–30 (2018). URL https://www.sciencedirect.com/science/article/pii/S0924271617303453

    work page 2018

  10. [10]

    21 URL https://www.nature.com/articles/s41598-025-86307-1

    Tarkhan, N.et al.Mapping fa¸ cade materials utilizing zero-shot segmentation for applications in urban microclimate research.Scientific Reports15, 5492 (2025). 21 URL https://www.nature.com/articles/s41598-025-86307-1

    work page 2025

  11. [11]

    & Biljecki, F

    Ito, K. & Biljecki, F. Assessing bikeability with street view imagery and computer vision.Transportation Research Part C: Emerging Technologies 132, 103371 (2021). URL https://www.sciencedirect.com/science/article/pii/ S0968090X21003739

    work page 2021

  12. [12]

    & Shen, Z.-J

    Deng, T., Zhang, K. & Shen, Z.-J. M. A systematic review of a digital twin city: A new pattern of urban governance toward smart cities.Journal of Management Science and Engineering6, 125–134 (2021). URL https://www.sciencedirect. com/science/article/pii/S2096232021000238

    work page 2021

  13. [13]

    Hou, Y.et al.Global streetscapes – a comprehensive dataset of 10 million street- level images across 688 cities for urban science and analytics.ISPRS Journal of Photogrammetry and Remote Sensing215, 216–238 (2024)

    work page 2024

  14. [14]

    Quintana, M.et al.It is not always greener on the other side: Greenery perception across demographics and personalities in multiple cities.Landscape and Urban Planning271, 105618 (2026)

    work page 2026

  15. [15]

    & Rodrigues, C

    Loudjani, P., Lemajic, S., Parage, V. & Rodrigues, C. Report on new Pl´ eiades Neo sensors benchmark. Technical Report JRC134704, Joint Research Centre (European Commission), Luxembourg (2023)

    work page 2023

  16. [16]

    Gupta, A.et al.Greenscan: Toward large-scale terrestrial monitoring the health of urban trees using mobile sensing.IEEE Sensors JournalPP, 1–1 (2024)

    work page 2024

  17. [17]

    URL https://www.sciencedirect.com/science/article/pii/S0360132324009648

    Klimenka, M.et al.Instant infrared: Estimating urban surface temperatures from street view imagery.Building and Environment267, 112122 (2025). URL https://www.sciencedirect.com/science/article/pii/S0360132324009648

    work page 2025

  18. [18]

    URL https://space-solutions.airbus.com/imagery/ our-optical-and-radar-satellite-imagery/pleiades-neo/

    Airbus Defence and Space.Pl´ eiades Neo: Native 30cm High-Resolution Satellite Imagery(2021). URL https://space-solutions.airbus.com/imagery/ our-optical-and-radar-satellite-imagery/pleiades-neo/. Accessed: 2026-03-02

    work page 2021

  19. [19]

    URL https://www.mapir.camera/collections/survey3/ products/survey3w-camera-red-green-nir-rgn-ndvi

    MAPIR Camera.Survey3W Camera - Red+Green+NIR (RGN, NDVI) Datasheet(2023). URL https://www.mapir.camera/collections/survey3/ products/survey3w-camera-red-green-nir-rgn-ndvi. Accessed: 2026-01-30

    work page 2023

  20. [20]

    URL https://eaglenxt.com/drone-sensors/altum-pt-camera/

    MicaSense.Altum-PT Multispectral and Thermal Sensor Datasheet(2022). URL https://eaglenxt.com/drone-sensors/altum-pt-camera/. Accessed: 2026-02-01

    work page 2022

  21. [21]

    URL https://oem.flir.com/en-hk/products/lepton/?model= 500-0771-01&vertical=microcam&segment=oem

    Teledyne FLIR.Lepton 3.5 Engineering Datasheet (Model 500-0771- 01)(2021). URL https://oem.flir.com/en-hk/products/lepton/?model= 500-0771-01&vertical=microcam&segment=oem. Accessed: 2026-01-30. 22

    work page 2021

  22. [22]

    URL https://www.flircameras.com/product/ flir-boson-640x480-18mm-htm/

    Teledyne FLIR.Boson LWIR Thermal Camera Core Engineering Datasheet (640 x 512, 18mm)(2023). URL https://www.flircameras.com/product/ flir-boson-640x480-18mm-htm/. Accessed: 2026-02-01

    work page 2023

  23. [23]

    URL https://www.raspberrypi.com/products/camera-module-3/

    Raspberry Pi Ltd.Raspberry Pi Camera Module 3 Product Brief(2023). URL https://www.raspberrypi.com/products/camera-module-3/. Accessed: 2026-01- 30

    work page 2023

  24. [24]

    Population dynamics; birth, death and migration per region

    Netherlands, S. Population dynamics; birth, death and migration per region. CBS StatLine database (2025). URL https://opendata.cbs.nl/#/CBS/en/dataset/ 37259eng/table. Dataset: Population dynamics; birth, death and migration per region. Accessed: 2026-03-26

    work page 2025

  25. [25]

    Pl´ eiades Neo ESA Archive

    European Space Agency. Pl´ eiades Neo ESA Archive. https://earth.esa.int/ eogateway/catalog/pleiades-neo-esa-archive (2026). Accessed: 2026-02-07, Pro- vided via ESA Third Party Mission scheme

    work page 2026

  26. [26]

    Ustin, S. L. & Curtiss, B. Spectral characteristics of ozone-treated conifers.Envi- ronmental and Experimental Botany30, 293–308 (1990). URL https://www. sciencedirect.com/science/article/pii/0098847290900412

    work page arXiv 1990

  27. [27]

    & Richter, K

    Danner, M., Locherer, M., Hank, T. & Richter, K. Spectral sampling with the asd fieldspec 4 – theory, measurement, problems, interpretation. EnMAP Field Guides Technical Report, GFZ Data Services (2015)

    work page 2015

  28. [28]

    URL https:// calib.io/products/checkerboard

    calib.io.Camera calibration checkerboard target / plate(2025). URL https:// calib.io/products/checkerboard. Accessed: 2026-03-14

    work page 2025

  29. [29]

    N., Bin Amro, A

    Swamidoss, I. N., Bin Amro, A. & Sayadi, S. Systematic approach for thermal imaging camera calibration for machine vision applications.Optik 247, 168039 (2021). URL https://www.sciencedirect.com/science/article/pii/ S0030402621015989

    work page 2021

  30. [30]

    URL https://www.mapir.camera/en-gb/products/ diffuse-reflectance-standard-calibration-target-package-v2

    MAPIR Camera.Diffuse Reflectance Standard Calibration Target Pack- age (V2)(2023). URL https://www.mapir.camera/en-gb/products/ diffuse-reflectance-standard-calibration-target-package-v2. Accessed: 2026-03-02

    work page 2023

  31. [31]

    A flexible new technique for camera calibration.IEEE Transactions on Pattern Analysis and Machine Intelligence22, 1330–1334 (2000)

    Zhang, Z. A flexible new technique for camera calibration.IEEE Transactions on Pattern Analysis and Machine Intelligence22, 1330–1334 (2000)

    work page 2000

  32. [32]

    The OpenCV Library.Dr

    Bradski, G. The OpenCV Library.Dr. Dobb’s Journal of Software Tools(2000)

    work page 2000

  33. [33]

    EgoBlur: responsible innovation in Aria,

    Raina, N.et al.Egoblur: Responsible innovation in aria (2023). arXiv:2308.13093

    work page arXiv 2023

  34. [34]

    URL https://mapir.gitbook.io/ mapir-camera-control-mcc/interface-tabs/process-tab

    MAPIR.MAPIR Camera Control(2024). URL https://mapir.gitbook.io/ mapir-camera-control-mcc/interface-tabs/process-tab. Accessed: 2026-03-16

    work page 2024

  35. [35]

    Wang, J.et al.Vggt: Visual geometry grounded transformer (2025). 23

    work page 2025

  36. [36]

    Actueel Hoogtebestand Nederland (AHN)

    Actueel Hoogtebestand Nederland. Actueel Hoogtebestand Nederland (AHN). https://www.ahn.nl/ (2023). Accessed: 2026-03-16

    work page 2023

  37. [37]

    Ma, R., Yang, C., Chen, J., Biljecki, F. & Li, X. Street view imagery-based method for reconstructing 3d building fa¸ cade openings.Automation in Construction184, 106842 (2026)

    work page 2026

  38. [38]

    & Alatan, A

    ¨Onder Tuzcuo˘ glu, K¨ oksal, A., Sofu, B., Kalkan, S. & Alatan, A. A. Xoftr: Cross- modal feature matching transformer (2024). URL https://arxiv.org/abs/2404. 09692

    work page 2024

  39. [39]

    Sun, L.et al.Cm-bench: A comprehensive cross-modal feature matching bench- mark bridging visible and infrared images.arXiv preprint arXiv:2603.12690 (2026)

    work page arXiv 2026

  40. [40]

    Berk, A., Bernstein, L. S. & Robertson, D. C. MODTRAN: A moderate resolution model for LOWTRAN. Technical Report, 12 May 1986 - 11 May 1987 Spectral Sciences, Inc., Burlington, MA. (1987)

    work page 1986

  41. [41]

    URL https:// agupubs.onlinelibrary.wiley.com/doi/abs/10.1002/jgrd.50497

    Widlowski, J.-L.et al.The fourth radiation transfer model intercomparison (rami- iv): Proficiency testing of canopy reflectance models with iso-13528.Journal of Geophysical Research: Atmospheres118, 6869–6890 (2013). URL https:// agupubs.onlinelibrary.wiley.com/doi/abs/10.1002/jgrd.50497

    work page doi:10.1002/jgrd.50497 2013

  42. [42]

    C., Saqui, D., Mpinda, S

    de Lima, D. C., Saqui, D., Mpinda, S. A. T. & Saito, J. H. Pix2pix network to estimate agricultural near infrared images from rgb data.Canadian Journal of Remote Sensing48, 299–315 (2022)

    work page 2022

  43. [43]

    & Uijlenhoet, R

    Gupta, A. & Uijlenhoet, R. Using street view imagery and deep generative mod- eling for estimating the health of urban forests (2025). URL https://arxiv.org/ abs/2504.14583. arXiv:2504.14583

    work page arXiv 2025

  44. [44]

    Code red in the north due to black ice, train services cancelled

    DutchNews.nl. Code red in the north due to black ice, train services cancelled. https://www.dutchnews.nl/2026/02/ code-red-in-the-north-due-to-black-ice-train-services-cancelled/ (2026). Accessed: 2026-03-20. 24

    work page 2026