Pith. sign in

REVIEW 4 major objections 5 minor 18 references

The Under-Water Dark-Room Experimental Facility at the University of Winnipeg

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read A newly built underwater dark-room facility demonstrates that camera intrinsic parameters can be calibrated in water, reporting focal lengths near 3,290 pixels and reprojection errors mostly below five pixels.

desk verdict A genuinely useful facility paper whose camera-calibration section overclaims and is internally inconsistent; the facility description is the real contribution. read the letter →

arxiv 2505.18395 v1 pith:TJH4J2Y3 submitted 2025-05-23 physics.ins-det hep-ex

Ajmi Ali , Blair Jamieson , Lyndsay Green , Tapendra BC , Rituparna Banerjee , Mahnoor Mansoor , Andrea Mayorga , Anna Harms
show 9 more authors
Fabio Castellanos Lenes Brijesh Sharma Flora Easter David Ostapchuk Shomi Ahmed Kyle Macdonald Craig Wood Marshall Kirton Gonzalo Paz
This is my paper · ORCID
classification physics.ins-dethep-ex
keywords underwatercameracalibrationdark-roomtestfacilityfive-axisgantryintrinsicparametersfisheyelensmodelwaterCherenkovdetectorphotogrammetry
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 reports the construction of an optically isolated, water-filled test facility consisting of a 1000-gallon tank, a five-axis gantry, a pan-tilt head, and water purification and temperature-control loops. It then uses the facility to perform what it describes as the first underwater camera calibration, placing a checkerboard target on the gantry arm and scanning it through spherical arcs around a sealed camera at the tank bottom. The calibration returns focal lengths of about $3.292 \times 10^3$ and $3.290 \times 10^3$ pixels and a principal point close to the center of the $9504 \times 6336$ pixel array, with reprojection errors mostly below five pixels. The sympathetic reading is that underwater calibration through a water–dome–air–lens optical path is feasible with minimal introduced error, and that the facility can support photogrammetry development for large water Cherenkov neutrino detectors.

What carries the argument

The load-bearing object is the five-axis motion system: a Cartesian gantry with three stepper-driven axes and a two-motor pan-tilt head on the vertical arm. The gantry moves a flat printed checkerboard through spherical and cylindrical scan patterns around a fixed underwater camera, while the pan-tilt adds polar and azimuthal rotation. Corner positions extracted by a standard computer-vision routine are fitted with a generic fish-eye lens model, and the intrinsic and extrinsic parameters are refined by minimizing reprojection error over many non-repetitive views.

What would settle it

Independently measure the checkerboard positions underwater with a laser tracker or an external camera system; if that measurement finds target-position errors comparable to or larger than the reported 2.5–25 pixel reprojection deviations, the calibration claim is not separately established.

Watch

Extended reading notes

Core claim

The central discovery, stated on the paper's own terms, is that a camera's intrinsic parameters—the focal length and the principal point, the image location where the optical axis lands—can be determined while the camera is underwater. The effective optical system includes water, an acrylic dome, a thin air layer, and the camera's own lenses, and the paper reports that this compound system was calibrated by moving a checkerboard with the gantry and fitting a fish-eye lens model. The resulting focal lengths were $f_x = 3.292 \times 10^3$ and $f_y = 3.290 \times 10^3$ pixels, with the principal point at $(x = 4.662 \times 10^3, y = 3.092 \times 10^3)$ pixels, close to the center of the $9504 \times 6336$ array. The authors take this near-centered principal point as evidence of a spherically symmetric effective lens system, with reprojection errors mostly under five pixels across the field of view.

Load-bearing premise

The gantry and pan-tilt position readouts are treated as true coordinates for the checkerboard, and the board is assumed rigid and flat underwater; if these assumptions fail, the reprojection errors include positioning and shape error and are not purely camera calibration error.

Editorial extensions

If this is right

  • Underwater camera calibration can be carried out in situ, so the effective lens system that will actually operate in water is calibrated rather than inferred from air measurements.
  • The near-centered principal point implies that the dome and housing introduce no large decentering distortion, supporting the use of a single rotationally symmetric model.
  • The same gantry-driven scan-and-fit procedure can be repeated with more views or different target paths, since the calibrated coordinate space already provides the needed geometry.
  • The elevated 10–25 pixel deviations in an intermediate radial band identify where the current fish-eye model fits least well and where future calibration data should be concentrated.

Reading between the lines

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

  • An inference beyond the paper: because no independent check of the gantry's underwater positioning accuracy is reported, the stated reprojection errors should be read as an upper bound on calibration quality, not separated from positioning error.
  • A further inference: the observed radial error pattern could be tested by fitting the same images with a model that explicitly includes refraction at the water–dome–air interfaces; a systematic improvement there would point to compound optics rather than target motion as the source of the intermediate-ring deviations.
  • Another testable extension: the five-axis motion could also be used to calibrate the relative pose of multiple cameras by moving a single target through their overlapping field of view, a next step the paper does not discuss.
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

4 major / 5 minor

Summary. The paper describes the design, construction, and commissioning of the Under-Water Dark-Room Test Facility (UWDTF) at the University of Winnipeg: a 1000-gallon HDPE water tank inside an optically isolated room, a Cartesian gantry with a pan-tilt head providing five axes of motion, a water-circulation and purification system, and auxiliary infrastructure for underwater detector R&D. It also reports a preliminary underwater camera calibration performed with a checkerboard target mounted on the gantry, analyzing images with the OpenCV fisheye-lens model. The authors report focal lengths of fx = 3.292e3 and fy = 3.290e3 pixels, a principal point at (4.662e3, 3.092e3) pixels, and conclude that "first of its kind" underwater calibration was "successfully accomplished with minimal error introduction." The facility description is detailed, but the calibration result is presented with only a qualitative and internally inconsistent error statement.

Significance. If the facility performs as described, it provides a useful and relatively low-cost testbed for underwater optical and detector R&D relevant to WCTE, IWCD, and Hyper-Kamiokande. The paper's strengths are the concrete engineering details: tank selection, gantry calibration plots, water-purification loop, dark-room construction, and payload positioning. These are valuable for other groups building similar facilities. The underwater camera calibration, however, is not yet established. The reported reprojection residuals contradict the summary statement, no uncertainty estimates are given for the fitted parameters, and the validation is purely internal to the calibration fit. The central claim of a successful minimal-error underwater calibration therefore needs substantive revision or re-analysis, though the facility description itself could support a publishable paper after major corrections.

major comments (4)
  1. [§5, Figs. 16–17, and Conclusion] The paper states that corner reprojection deviations are "mostly less than 5 pixels at the central and extreme regions... peaking at around 2.5 pixels," then immediately states that "many points in the intermediate region tend to have a higher deviation in the range ∼10-25 pixels." The Conclusion nevertheless calls the calibration "successfully accomplished with minimal error introduction." These statements are contradictory. The manuscript must provide a quantitative error budget: RMS, median, and 95th-percentile reprojection error, the number of images and corners used, and a clear statement of whether the 10–25 pixel deviations are outliers, systematic model mismatch, or a separate population of frames. Without such a budget, "minimal error" is undefined, and the central claim is unsupported.
  2. [§5, fitted parameters paragraph] The reported intrinsic parameters fx = 3.292e3, fy = 3.290e3, principal point (4.662e3, 3.092e3) are presented without any uncertainties. Since these are fitted values, the claim that the principal point "matches close to the mid-point" of the 9504×6336 array cannot be evaluated without standard errors or confidence intervals. The paper should report at least the covariance of the intrinsic parameters or a bootstrap/leave-one-out estimate.
  3. [§5, validation method] The reprojection error is computed from the same fit that determined the intrinsic and extrinsic parameters, so it measures internal consistency, not absolute calibration accuracy. To support the claim of minimal error introduction, the authors should provide an independent check: for example, holding out a subset of images from the fit and computing reprojection error on those held-out views, or comparing gantry-commanded target positions with positions reconstructed from the calibrated camera. The gantry and pan-tilt calibrations shown in Figs. 3–8 give linear relationships between counts and commanded coordinates, but the paper does not report positioning uncertainties, and the underwater rigidity and flatness of the checkerboard target are not demonstrated. Without such ground truth, the residuals cannot be attributed to the camera model alone.
  4. [Introduction and Conclusion] The claim that underwater camera calibration is "first of its kind" and "has never been presented before" is not supported by a literature survey. Underwater camera calibration for refraction-aware and housing-based systems exists in the computer vision and marine robotics literature. The manuscript should either cite and differentiate from prior underwater calibration work or moderate the novelty claim to "a calibration of this specific housing and gantry system." This is a load-bearing point for the paper's stated significance, though it is secondary to the internal error inconsistency.
minor comments (5)
  1. [Abstract/Title] The keyword list is empty; the authors should provide standard keywords for indexing.
  2. [Throughout] There are numerous typographical errors, e.g., "CP volation" in §1, "Fortuantely" in §2, "di fferent" in §2, "purspose" in §2, "instrinsic" in §5, and "realisized" in the Acknowledgements. The manuscript needs a careful proofreading pass.
  3. [Figures 13 and 15] The axis labels in the gantry-coordinate plots are difficult to read and the units are not specified. The captions should state what the coordinate values represent (millimeters? encoder counts?) and the relationship to the plots in Figs. 3–8.
  4. [§4] The abbreviations RO, UV, and mPMT are used without full expansion at first use (RO is explained only in passing, mPMT is expanded in a footnote, UV is not expanded). Please define all abbreviations at first occurrence.
  5. [§5] The paper mentions that the camera was rotated by 90 degrees after each set of scans and the process repeated four times, but does not state how many images were acquired in total or how many were used in the calibration. This information is needed to assess the reliability of the fit.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the paper reports a standard camera-calibration fit and facility description, with no derivation chain that reduces to its inputs.

full rationale

This paper is primarily a facility description plus a preliminary camera-calibration study. The camera intrinsics (fx, fy, principal point) are obtained by an OpenCV fisheye-model fit to checkerboard corner observations, and the corner reprojection deviations shown in Figs. 16 and 17 are residuals of that same fit. There is no claim that these residuals are independent predictions, no parameter fitted to one subset and validated on another, no self-citation invoked as a load-bearing theorem, and no renaming of a known result as a new derivation. The conclusion that the calibration was 'successfully accomplished with minimal error introduction' is an internal consistency statement based on the fit itself, which is weak evidence and is in tension with the reported 10-25 pixel intermediate-region deviations; however, that is a correctness and validation concern, not circularity in the sense of a derived quantity reducing by construction to its input. The gantry and pan-tilt readouts are treated as ground truth, but this is an experimental-assumption issue, not a circular step. No enumerated circular step can be exhibited from the paper's own equations or citations, so the circularity score is 0.

Assumptions & free parameters 1 free parameters · 3 assumptions · 0 invented entities

The central numerical results are camera intrinsic parameters obtained by fitting, and the interpretation of success relies on unverified assumptions about target rigidity and gantry accuracy.

free parameters (1)
  • Camera intrinsic parameters (fx, fy, cx, cy and distortion coefficients) = fx=3292, fy=3290, cx=4662, cy=3092 pixels; distortion coefficients not reported
    Obtained by fitting the Kannala-Brandt fisheye model to checkerboard images via OpenCV; these are the reported results and are fitted, not predicted.
assumptions (3)
  • domain assumption The checkerboard target remains flat and rigid when submerged in water.
    The target is attached to the gantry arm; any bending or deformation would appear as reprojection error but is not quantified.
  • domain assumption The gantry and pan-tilt coordinate readouts provide accurate ground-truth positions for the target.
    Calibration assumes known target poses from gantry counts; no accuracy specification for the gantry is given in the calibration section.
  • domain assumption The Kannala-Brandt fisheye model in OpenCV accurately describes the combined optical system of water, acrylic dome, air gap, and camera lens.
    The model is fitted without testing alternative refraction-aware models; the 10-25 pixel residuals may indicate model mismatch.

how reviews work

0 comments
Cite this review

Pith. "Pith review of The Under-Water Dark-Room Experimental Facility at the University of Winnipeg." pith.science (2026). https://pith.science/paper/TJH4J2Y3

@misc{pith2026250518395,
  author       = {Pith},
  title        = {Pith review of: The Under-Water Dark-Room Experimental Facility at the University of Winnipeg},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TJH4J2Y3}},
  note         = {Machine review of arXiv:2505.18395}
}
read the original abstract

A completely new under-water dark-room test facility (UWDTF) has been built at the University of Winnipeg during 2021-2023, for the testing of the equipments, optical components and detectors before they might be used in different underwater experiments, like the Hyper-Kamiokande (Hyper-K), and others. The Facility is designed for Research and Development activities primarily related to the different calibration systems, which are/will be used in the Water Cherenkov Test Experiment (WCTE) at CERN, the Intermediate Water Cherenkov Detector (IWCD) at Tokai, Japan and the Hyper-Kamiokande Far Detector at Kamioka, Japan. The facility houses a large tank of water (1000 gallons) in an optically isolated room, and is equipped with a gantry that provides for the 3D motion of a maximum of 50 lbs of load inside the tank. A customized pan-tilt system has also been devised to accommodate further degrees of freedom of motion to the payload in the polar and azimuthal direction. The facility is primarily used for testing of the under-water camera housings designed for the Hyper-K experiment, besides many other research and development activities. The preliminary results of the camera calibration done in this multi-purpose underwater-darkroom facility are presented here, starting with the description of the vital features of this facility.

Figures

Figures reproduced from arXiv: 2505.18395 by the authors.

Figure 1
Figure 1. Structural features of the water tank used in the facility. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 4
Figure 4. Calibration along the Y-axis motion of the gantry. [PITH_FULL_IMAGE:figures/full_fig_p003_4.png] view at source ↗
Figure 5
Figure 5. Calibration along the Z-axis motion of the gantry. [PITH_FULL_IMAGE:figures/full_fig_p003_5.png] view at source ↗
Figures from the paper (9 more)
Figure 6
Figure 6. Figure 6: The pantilt system casing on the left, and the blown-up [PITH_FULL_IMAGE:figures/full_fig_p003_6.png]
Figure 7
Figure 7. Figure 7: Calibration along the theta-axis or motion in polar direction [PITH_FULL_IMAGE:figures/full_fig_p004_7.png]
Figure 9
Figure 9. Figure 9: Water Circulation Loop, designed to avoid any microbial [PITH_FULL_IMAGE:figures/full_fig_p004_9.png]
Figure 10
Figure 10. Figure 10: Set-up used for checking the water-circulation system in [PITH_FULL_IMAGE:figures/full_fig_p004_10.png]
Figure 11
Figure 11. Figure 11: The Detector mounting structure, with two potential [PITH_FULL_IMAGE:figures/full_fig_p005_11.png]
Figure 12
Figure 12. Figure 12: Top: Schematics of the top-view of the UWDTF facility. [PITH_FULL_IMAGE:figures/full_fig_p005_12.png]
Figure 13
Figure 13. Figure 13: Cylindrically scanned coordinates by the Gantry arm [PITH_FULL_IMAGE:figures/full_fig_p006_13.png]
Figure 15
Figure 15. Figure 15: Spherically scanned coordinates by the Gantry arm around [PITH_FULL_IMAGE:figures/full_fig_p006_15.png]
Figure 16
Figure 16. Figure 16: Corner deviations measured at each point of the calibration [PITH_FULL_IMAGE:figures/full_fig_p007_16.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

18 extracted references · 15 canonical work pages

  1. [1]

    Hyper-Kamiokande Design Report

    K. Abe et al. “Hyper-Kamiokande Design Report”, (May 2018). arXiv: 1805.04163

    arXiv 2018

  2. [2]

    Status of the Hyper-Kamiokande Experiment

    Ali Ajmi (for the Hyper-Kamiokande Collaboration), “Status of the Hyper-Kamiokande Experiment”, (Apr 8, 2024), Published in: PoS HQL2023 (2024) 098

    work page 2024

  3. [3]

    G. L. Fogli, E. Lisi, A. Marrone, D. Montanino, A. Palazzo and A. M. Rotunno, Phys. Rev. D 86, 013012 (2012) doi:10.1103/PhysRevD.86.013012 [arXiv:1205.5254 [hep-ph]]

    work page Pith review arXiv 2012

  4. [4]

    M. C. Gonzalez-Garcia, M. Maltoni and T. Schwetz, JHEP 11, 052 (2014) doi:10.1007 /JHEP11(2014)052 [arXiv:1409.5439 [hep-ph]]

    arXiv 2014

  5. [5]

    G. C. Branco et. al. Leptonic CP violation, Rev. Mod. Phys. 84, 515, April 2012 ying

    work page 2012

  6. [6]

    G. C. Branco et al. Flavour Physics and CP Violation in the StandardModel and Beyond, arXiv: 1402.4068v2. 7

    work page Pith review arXiv

  7. [7]

    Dick et al

    K. Dick et al. CP-Violation in Neutrino Oscillations, arXiv: hep- ph/9903308v3

    work page arXiv

  8. [8]

    Abe et al

    K. Abe et al. [Hyper-Kamiokande Proto-], PTEP 2015 (2015), 053C02 doi:10.1093/ptep/ptv061 [arXiv:1502.05199 [hep-ex]]

    arXiv 2015

Show all 18 references
  1. [9]

    Igarashi, K

    S. Igarashi, K. Satou, C. Ohmori, Y . Arakaki, M. Furusawa, K. Hara, K. Hasegawa, Y . Hashimoto, Y . Hori and H. Hotchi,et al. PTEP 2021 (2021) no.3, 033G01 doi:10.1093/ptep/ptab011

    2021 doi

  2. [10]

    Scott [NuPRISM and Hyper-K], JPS Conf

    M. Scott [NuPRISM and Hyper-K], JPS Conf. Proc. 12 (2016), 010039 doi:10.7566 /JPSCP.12.010039 [arXiv:1603.01251 [physics.ins-det]]

    2016 arXiv

  3. [11]

    D. C. Rodriguez [WCTE], PoS ICHEP2024 (2025), 173 doi:10.22323/1.476.0173

    2025 doi

  4. [12]

    Camera calibration method with focus-related intrinsic parameters based on the thin-lens model

    Zhangji Lu and Lilong Cai, “Camera calibration method with focus-related intrinsic parameters based on the thin-lens model”, Opt. Express 28, 20858-20878 (2020)

    2020

  5. [13]

    A comparative review of camera calibrating methods with accuracy evaluation

    Joaquim Salvi, Xavier Armangué, Joan Batlle, “A comparative review of camera calibrating methods with accuracy evaluation”, Pattern Recognition, V ol- ume 35, Issue 7, 2002, Pages 1617-1635, ISSN 0031- 3203, https: //doi.org/10.1016/S0031-3203(01)00126-1. (https://www.scienced...

    2002 doi

  6. [14]

    A generic camera model and calibration method for conventional, wide-angle, and fish- eye lenses

    Kannala, Juho, and Sami S. Brandt. “A generic camera model and calibration method for conventional, wide-angle, and fish- eye lenses”. IEEE transactions on pattern analysis and machine intelligence 28.8 (2006): 1335-1340

    2006

  7. [15]

    Calibration of omnidirectional cameras in practice: A comparison of methods

    Puig, Luis, et al. “Calibration of omnidirectional cameras in practice: A comparison of methods”. Computer Vision and Im- age Understanding 116.1 (2012): 120-137

    2012

  8. [16]

    A flexible technique for accurate omnidirectional cam- era calibration and structure from motion

    Scaramuzza, Davide, Agostino Martinelli, and Roland Sieg- wart. “A flexible technique for accurate omnidirectional cam- era calibration and structure from motion”. Fourth IEEE Inter- national Conference on Computer Vision Systems (ICVS’06). IEEE, 2006

    2006

  9. [17]

    Improved wide-angle, fisheye and omnidirectional camera cali- bration

    Ste ffen Urban, Jens Leitlo ff, Stefan Hinz, “Improved wide-angle, fisheye and omnidirectional camera cali- bration”, ISPRS Journal of Photogrammetry and Re- mote Sensing, V olume 108, 2015, Pages 72-79, ISSN 0924-2716, https: //doi.org/10.1016/j.isprsjprs.2015.06.005. (https:...

    2015 doi

  10. [18]

    The opencv library

    Bradski, Gary. “The opencv library”. Dr. Dobb’s Journal: Soft- ware Tools for the Professional Programmer 25.11 (2000): 120- 123. 8

    2000

Pith tools

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