Pith. sign in

REVIEW 4 major objections 4 minor 32 references

PoleStack: Robust Pole Estimation of Irregular Objects from Silhouette Stacking

T0 review · 4 major / 4 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read PoleStack estimates the rotation pole of an irregular space object by stacking silhouette images and finding the axis of maximum reflective symmetry in the Fourier amplitude spectrum.

desk verdict New, simple pole-estimation method whose symmetry insight and DFT trick are sound and honestly presented, but the experiments are too thin to fully back the degree-level claims; it deserves a referee, with revision focused on validation and a proof gap. read the letter →

arxiv 2502.02907 v1 pith:ASMHAPTG submitted 2025-02-05 cs.CV

Jacopo Villa , Jay W. McMahon , Issa A. D. Nesnas This is my paper · ORCID
classification cs.CV
keywords poleestimationsilhouettestackingreflectivesymmetrydiscreteFouriertransformtranslationinvariancespacecraftnavigationsmallbodiesprincipal-axisrotation
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 presents PoleStack, an algorithm for estimating the rotation pole of a principal-axis rotator—an object spinning about one of its principal moments of inertia—from silhouette images taken at multiple camera poses. The central assertion is that when silhouettes collected over a range of camera longitudes are co-added into a single image, that stacked image is approximately mirror-symmetric about the direction of the pole projected onto the image plane. The projected-pole direction is therefore recovered by searching for the axis of maximum reflective symmetry in the Discrete Fourier Transform amplitude spectrum of the stack, which preserves the symmetry while making the search invariant to where the object's center appears in the frame. Combining two or more such projected-pole measurements from different camera orientations yields the full 3D pole direction through a simple least-squares triangulation. The paper reports degree-level pole estimates from low-resolution, heavily shadowed, and centroid-aligned synthetic imagery of Bennu and 67P, which matters because early pole knowledge during spacecraft approach enables body-fixed navigation and shape reconstruction.

What carries the argument

The central object is the silhouette-stack image: a co-addition of binary silhouette masks of the rotating body, all expressed in a common camera frame. Two properties of the stack carry the argument. First, its symmetry-preserving component is reflectively symmetric about the projected pole direction (Theorem III.1), so the pole projection is identifiable as the angle of maximum symmetry. Second, the Discrete Fourier Transform amplitude spectrum of the stack preserves that reflective symmetry while being translation invariant (Theorem III.3), so the symmetry search no longer requires knowing the object's center-of-mass pixel location. A low-pass circular crop plus a log compression of the squared amplitude spectrum sharpens the symmetry peak before the per-angle mirror-correlation score is computed.

What would settle it

Render a synthetic body with a known pole and add a strong, straight shadow terminator whose image-plane orientation is fixed at, say, 30 degrees from the true pole projection but rotated relative to it, then run Algorithm 1 on silhouette stacks with and without that terminator. If the estimated symmetry axis moves toward the terminator orientation by more than the 1-degree query step, the noise model in Assumption III.2 is violated and the maximum-symmetry estimate is biased by structured shadowing.

Watch

Extended reading notes

Core claim

The discovery the paper tries to establish is geometric: a silhouette-stack image, defined as the integral or sum of binary silhouette occupancy masks observed across a camera-longitude interval, contains a component that is exactly reflectively symmetric about the projected pole direction, with all deviations from partial longitude coverage, surface shadowing, and registration offsets appearing as an additive error term. Theorem III.1 proves the symmetry for the perfect-silhouette case; Theorem III.2 folds shadowing and registration errors into a symmetric signal plus error; and Theorem III.4 shows that in the squared DFT amplitude spectrum the expected symmetry score is maximized at the true pole projection as long as the error spectrum has no preferred reflection axis. The algorithm then estimates the in-plane pole angle by rotating the log-power amplitude spectrum and scoring each rotation by normalized correlation with its mirror image, and it estimates the full 3D pole by solving a linear least-squares system that triangulates in-plane angles from known camera orientations.

Load-bearing premise

The result depends on the noise in the silhouette stack not having a preferred mirror axis of its own; if shadows or alignment errors line up along some direction that is not the pole, the maximum-symmetry search can lock onto that wrong direction instead.

Editorial extensions

If this is right

  • A single hovering-camera batch is enough to recover the projected-pole direction, with no knowledge of the object's center-of-mass location beyond silhouette extraction.
  • Two or more batches from different camera latitudes yield the full 3D pole through linear least squares; simulations show degree-level accuracy once camera boresights are separated by about 10 degrees.
  • The method holds up at a 90-degree sun phase, where self-shadowing heavily corrupts silhouettes, and with brightness-centroid registration, reducing the need for precise attitude or center knowledge.
  • Reducing image resolution from 1024 by 1024 to 256 by 256 and limiting longitude coverage to a hemisphere leaves pole-projection errors at 0 to 1 degree for the tested bodies.
  • Adding more camera views lowers variance and outlier rate, with estimates exceeding 5 degrees of error dropping from 1 percent of trials to 0.005 percent as the view count grows from 2 to 4.

Reading between the lines

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

  • The authors leave implicit that the symmetry principle could extend to tumbling objects viewed over short batches, yielding piecewise axis estimates that turn a single approach pass into a continuous pole track.
  • A direct extension would be to fuse centroid-based frame centering with the translation-invariant amplitude-spectrum scoring, letting residual mis-centering be absorbed in the frequency domain rather than corrected to pixel precision.
  • Testable extension: the error model predicts that pole bias should grow with the area and coherence of shadow-generated edges in the stack, so a controlled sweep of sun angle and terminator orientation would map the method's domain of validity.
  • The symmetry search could also serve as an initialization step for shape-from-silhouette or landmark-tracking pipelines, because it produces a body-fixed reference frame earlier than high-resolution feature tracks become available.
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 / 4 minor

Summary. The paper proposes PoleStack, a two-stage algorithm for estimating the rotation pole of a principal-axis rotator from silhouette images. Stage 1 stacks silhouettes collected from a hovering camera, computes the DFT amplitude spectrum, and estimates the projected-pole direction as the orientation maximizing a normalized-correlation symmetry score on the log-power spectrum. Stage 2 triangulates several projected-pole measurements from different camera views by solving a linear least-squares problem. The theoretical section derives a reflective-symmetry property of silhouette stacks, models shadowing and registration errors as additive asymmetric terms, and proves, under Assumption III.2, that the expected symmetry score of the noisy amplitude spectrum is maximized at the true projected pole. Experiments use simulated images of Bennu and 67P at a 90-degree sun phase for in-plane estimation, and a Monte Carlo study with synthetic angle noise for 3D triangulation.

Significance. If validated, the method addresses a practical need in small-body approach navigation: an early, landmark-free pole estimate from low-resolution silhouettes that is robust to shadowing and to unknown center-of-mass registration. The paper's strengths include an explicit formalization of the symmetry mechanism, a translation-invariant DFT formulation, and a very simple, precisely specified Algorithm 1. The falsifiable empirical claims, namely degree-level accuracy at a 90-degree sun phase and robustness to centroid alignment, are the right kind of claims for this community. However, the current validation is not yet strong enough for the stated conclusions, and one step of the triangulation derivation needs correction.

major comments (4)
  1. [III.G, Theorem III.5, Eqs. (87)-(92)] The triangulation matrix is derived incorrectly. With the paper's Definition III.6, tan(alpha_j) = (omega^T i_Cj)/(omega^T j_Cj), so each view yields one scalar equation sin(alpha_j)(omega^T j_Cj) - cos(alpha_j)(omega^T i_Cj) = 0, i.e., a single row sin(alpha_j) j_Cj^T - cos(alpha_j) i_Cj^T. The two-row block M_j = [sin(alpha_j) i_Cj^T; -cos(alpha_j) j_Cj^T] in Eq. (92) instead requires, for generic alpha_j, both omega^T i_Cj = 0 and omega^T j_Cj = 0, which forces omega to be parallel to k_Cj and is not equivalent to the angle measurement. Since Section V.B uses Eqs. (87)-(88) as the estimator, the reported Monte Carlo results do not validate the stated triangulation problem; the formulation should be corrected to one row per view and the experiments rerun.
  2. [III.F.3, Assumption III.2, Theorem III.4] The central no-bias guarantee rests on Assumption III.2, which asserts that the squared-amplitude noise A_N^2 is uncorrelated with its own reflection about every axis. The paper itself notes in Section III.F.3 that terminator lines and high-contrast shadow edges can violate this condition. Because Algorithm 1 selects argmax_theta psi(A^2_theta) on a single noisy realization, while Theorem III.4 only proves that the expectation of psi is maximized at the true angle, a shadow pattern with a preferred orientation not aligned with the pole can shift the argmax. No finite-sample concentration bound, adversarial shadow experiment, or empirical check of Assumption III.2 is provided. This gap is load-bearing for the abstract's robustness-to-shadowing claim.
  3. [V.A, Figure 8] The in-plane accuracy claim is supported by only two shape models, a single sun phase of 90 degrees, a single camera latitude of 14 degrees, and one realization per alignment case. The reported 0-3 degree errors are point estimates with no repeated trials, standard deviations, or confidence intervals, so the phrase 'degree-level accuracy' in the abstract and conclusions is not statistically supported. At minimum, the authors should repeat the experiment over independent camera longitude offsets, pole orientations, and sun directions, and report error distributions.
  4. [V.B, Figures 11-13] The 3D validation injects zero-mean Gaussian noise directly into the projected-pole angles alpha_j and therefore does not exercise the image-based Algorithm 1. Errors produced by the symmetry maximization are not shown to be zero-mean, Gaussian, or independent across views; moreover, the sigma_alpha = 1 degree level used in Figure 11 is below the 3 degree worst-case error observed in Section V.A. An end-to-end test that feeds synthesized image sets from multiple camera views through both stages is needed before the 3D degree-level claim can be accepted.
minor comments (4)
  1. [III.F.3, Eq. (68)] In Assumption III.2, the image A^2_theta,N is not defined; the second factor in Eq. (68) appears to be the reflection of A^2_N,theta, but as printed both factors are identical. Please define the reflection operator explicitly.
  2. [III.B.2, Theorem III.1 proof] The substitution phi to tilde-phi in Eqs. (30)-(32) requires that the map from camera longitude to the paired longitude be one-to-one and differentiable; Definition III.12 only asserts existence of a matching longitude. State the additional regularity or give a measure-theoretic proof.
  3. [IV, Algorithm 1 and Definition VI.8] Algorithm 1 applies a circular crop to the amplitude spectrum but does not state whether the spectrum is centered via fftshift before cropping; since the DFT in Eq. (108) is not centered, the implementation should be specified.
  4. [Throughout] There are repeated typographical errors, including 'does notdoesnot' in Definition III.1, 'a the outlier rate' in Section V.B, and 'phi_f (lower)' in the proof of Theorem III.1, which should be corrected in revision.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the pole estimate is an argmax of a geometry-derived symmetry score, and the pipeline is validated against known ground truth.

full rationale

The central derivation is self-contained. The claim that silhouette stacks exhibit reflective symmetry about the projected pole is derived from the projective geometry of rotating surface points (Lemma III.1 and Theorem III.1), not assumed through the estimator. Algorithm 1 performs an argmax of a normalized-correlation symmetry score over the DFT amplitude spectrum, with no ground-truth pole angle or fitted parameter entering the objective. The only user-set quantity, the cutoff frequency tau, is a fixed preprocessing choice and is not tuned against reported errors. The 3D pole triangulation (Theorem III.5) is a linear least-squares reduction of measured in-plane angles and is validated in Monte Carlo simulations against known ground-truth pole directions. Assumption III.2, requiring the frequency-domain noise term to have no preferred reflective symmetry, is a genuine statistical premise and a potential correctness risk, but it is an input assumption rather than a conclusion built from the target result; the paper explicitly discusses conditions under which it can fail (Section III.F.3). Self-citations such as the rendering tool [29] and prior silhouette work [8,9] are used as experimental infrastructure or related work, not as the load-bearing argument. There is no step in which a prediction reduces by construction to a fitted input, a definitional tautology, or an authored prior claim.

Assumptions & free parameters 1 free parameters · 5 assumptions · 0 invented entities

The method's central claim rests on several domain assumptions about the observation geometry (principal-axis rotation, hovering camera, known calibration, extracted silhouettes) and on two explicit modeling assumptions (discrete-continuous equivalence, noise symmetrylessness). The only hand-set algorithmic parameter is the cutoff frequency.

free parameters (1)
  • Cutoff frequency tau = 100 px (full resolution), 126 px (reduced data volume)
    This user-set low-pass filter radius on the DFT amplitude spectrum trades noise suppression against retained symmetry information. It is chosen by hand in each experiment, not fitted to ground truth, but it is a tunable parameter that affects results.
assumptions (5)
  • domain assumption The observed object is a principal-axis rotator with a fixed pole aligned with a principal axis.
    Invoked throughout Section II and the problem statement; the entire symmetry argument relies on rotation about a fixed pole.
  • domain assumption The camera hovers at constant range and latitude while the object rotates (hovering-camera model).
    Section II.B, Eq. 4; under this model the projected pole direction is constant within an image batch.
  • domain assumption The inertial camera attitude and intrinsic calibration are known, and the silhouette has been extracted.
    Assumptions II.D.1, II.D.4, II.D.5; required to form the stack and later to triangulate.
  • domain assumption The discrete silhouette-stack image approximates the continuous integral model (Assumption III.1, Eq. 46).
    Needed to apply the continuous Theorem III.2 to digital images; validated only empirically.
  • ad hoc to paper The amplitude-spectrum noise term has no reflective symmetry with respect to any axis (Assumption III.2, Eq. 68).
    This assumption is essential for Theorem III.4, which guarantees the maximum of the expected symmetry score is the true pole angle. It is plausible for irregular shapes but not proven; shadowing or terminator lines could violate it.

how reviews work

0 comments
Cite this review

Pith. "Pith review of PoleStack: Robust Pole Estimation of Irregular Objects from Silhouette Stacking." pith.science (2026). https://pith.science/paper/ASMHAPTG

@misc{pith2026250202907,
  author       = {Pith},
  title        = {Pith review of: PoleStack: Robust Pole Estimation of Irregular Objects from Silhouette Stacking},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ASMHAPTG}},
  note         = {Machine review of arXiv:2502.02907}
}
read the original abstract

We present an algorithm to estimate the rotation pole of a principal-axis rotator using silhouette images collected from multiple camera poses. First, a set of images is stacked to form a single silhouette-stack image, where the object's rotation introduces reflective symmetry about the imaged pole direction. We estimate this projected-pole direction by identifying maximum symmetry in the silhouette stack. To handle unknown center-of-mass image location, we apply the Discrete Fourier Transform to produce the silhouette-stack amplitude spectrum, achieving translation invariance and increased robustness to noise. Second, the 3D pole orientation is estimated by combining two or more projected-pole measurements collected from different camera orientations. We demonstrate degree-level pole estimation accuracy using low-resolution imagery, showing robustness to severe surface shadowing and centroid-based image-registration errors. The proposed approach could be suitable for pole estimation during both the approach phase toward a target object and while hovering.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 32 canonical work pages

  1. [1]

    Characterizing and navigating small bodies with imaging data,

    Gaskell, R., Barnouin-Jha, O., Scheeres, D. J., Konopliv, A., Mukai, T., Abe, S., Saito, J., Ishiguro, M., Kubota, T., Hashimoto, T., et al., “Characterizing and navigating small bodies with imaging data,”Meteoritics & Planetary Science, Vol. 43, No. 6, 2008, pp. 1049–1061

    work page 2008

  2. [2]

    Practical stereophotoclinometry for modeling shape and topography on planetary missions,

    Palmer, E. E., Gaskell, R., Daly, M. G., Barnouin, O. S., Adam, C. D., and Lauretta, D. S., “Practical stereophotoclinometry for modeling shape and topography on planetary missions,”The Planetary Science Journal, Vol. 3, No. 5, 2022, p. 102

    work page 2022

  3. [3]

    Stereophotoclinometry for OSIRIS-REx Spacecraft Navigation,

    Adam, C. D., McCarthy, L. K., Leonard, J. M., Gaskell, R., Antreasian, P. G., Liounis, A. J., Getzandanner, K., Moreau, M. C., Palmer, E. E., Weirich, J., et al., “Stereophotoclinometry for OSIRIS-REx Spacecraft Navigation,”The Planetary Science Journal, Vol. 4, No. 9, 2023, p. 167

    work page 2023

  4. [4]

    Pose estimation for spacecraft relative navigation using model-based algorithms,

    Opromolla, R., Fasano, G., Rufino, G., and Grassi, M., “Pose estimation for spacecraft relative navigation using model-based algorithms,”IEEE Transactions on Aerospace and Electronic Systems, Vol. 53, No. 1, 2017, pp. 431–447

    work page 2017

  5. [5]

    Tumbling asteroids,

    Harris, A. W., “Tumbling asteroids,”Icarus, Vol. 107, No. 1, 1994, pp. 209–211

    work page 1994

  6. [6]

    Optimization methods for asteroid lightcurve inversion: II. The complete inverse problem,

    Kaasalainen, M., Torppa, J., and Muinonen, K., “Optimization methods for asteroid lightcurve inversion: II. The complete inverse problem,”Icarus, Vol. 153, No. 1, 2001, pp. 37–51

    work page 2001

  7. [7]

    Globally optimal shape and spin pole determination with light-curve inversion,

    Chng, C.-K., Sasdelli, M., and Chin, T.-J., “Globally optimal shape and spin pole determination with light-curve inversion,” Monthly Notices of the Royal Astronomical Society, Vol. 513, No. 1, 2022, pp. 311–332

    work page 2022

  8. [8]

    Light-robust pole-from-silhouette algorithm and visual-hull estimation for autonomous optical navigation to an unknown small body,

    Bandyopadhyay, S., Villa, J., Osmundson, A., Hockman, B., Morrell, B., Lubey, D., Bhaskaran, S., Bayard, D., McMahon, J., and Nesnas, I. A., “Light-robust pole-from-silhouette algorithm and visual-hull estimation for autonomous optical navigation to an unknown small body,”Proc. 31st AAS/AIAA Space Flight Mechanics Meeting, 2021, pp. 1–13

    work page 2021

Show all 32 references
  1. [9]

    Autonomous exploration of small bodies toward greater autonomy for deep space missions,

    Nesnas, I. A., Hockman, B. J., Bandopadhyay, S., Morrell, B. J., Lubey, D. P., Villa, J., Bayard, D. S., Osmundson, A., Jarvis, B., Bersani, M., et al., “Autonomous exploration of small bodies toward greater autonomy for deep space missions,”Frontiers in Robotics and AI, Vol. ...

    2021

  2. [10]

    Distinctive image features from scale-invariant keypoints,

    Lowe, D. G., “Distinctive image features from scale-invariant keypoints,”International journal of computer vision, Vol. 60, 2004, pp. 91–110

    2004

  3. [11]

    Vision-based estimation of small body rotational state,

    Panicucci, P., Lebreton, J., Brochard, R., Zenou, E., and Delpech, M., “Vision-based estimation of small body rotational state,” Acta Astronautica, Vol. 213, 2023, pp. 177–196

    2023

  4. [12]

    Optical navigation for autonomous approach of small unknown bodies,

    Villa, J., Bandyopadhyay, S., Morrell, B., Hockman, B., Lubey, D., Harvard, A., Chung, S.-J., Bhaskaran, S., and Nesnas, I. A., “Optical navigation for autonomous approach of small unknown bodies,”2020 AAS Guidance, Navigation and Control Conference, AAS, 2020

    2020

  5. [13]

    Autonomous feature tracking for autonomous approach to a small body,

    Morrell, B. J., Villa, J., Bandyopadhyay, S., Lubey, D., Hockman, B., Bhaskaran, S., Bayard, D., and Nesnas, I., “Autonomous feature tracking for autonomous approach to a small body,”ASCEND 2020, 2020, p. 4151

    2020

  6. [14]

    Initial Pole Axis and Spin Direction Estimation of Asteroids Using Infrared Imagery,

    Kuppa, K., McMahon, J. W., and Dietrich, A. B., “Initial Pole Axis and Spin Direction Estimation of Asteroids Using Infrared Imagery,”Journal of Guidance, Control, and Dynamics, Vol. 47, No. 6, 2024, pp. 1055–1071

    2024

  7. [15]

    Pole Estimation and Optical Navigation Using Circle of Latitude Projections,

    Christian, J. A., “Pole Estimation and Optical Navigation Using Circle of Latitude Projections,”Journal of Guidance, Control, and Dynamics, Vol. 47, No. 3, 2024, pp. 407–416

    2024

  8. [16]

    The OpenCV Library,

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

    2000

  9. [17]

    Hartley, R., and Zisserman, A.,Multiple view geometry in computer vision, Cambridge university press, 2003

    2003

  10. [18]

    Absolute triangulation algorithms for space exploration,

    Henry, S., and Christian, J. A., “Absolute triangulation algorithms for space exploration,”Journal of Guidance, Control, and Dynamics, Vol. 46, No. 1, 2023, pp. 21–46

    2023

  11. [19]

    A tutorial on horizon-based optical navigation and attitude determination with space imaging systems,

    Christian, J. A., “A tutorial on horizon-based optical navigation and attitude determination with space imaging systems,”IEEE Access, Vol. 9, 2021, pp. 19819–19853

    2021

  12. [20]

    Methods of optical navigation,

    Owen Jr, W. M., “Methods of optical navigation,” 2011

    2011

  13. [21]

    O.,The fast Fourier transform and its applications, Prentice-Hall, Inc., 1988

    Brigham, E. O.,The fast Fourier transform and its applications, Prentice-Hall, Inc., 1988

    1988

  14. [22]

    A signal processing approach to symmetry detection,

    Keller, Y., and Shkolnisky, Y., “A signal processing approach to symmetry detection,”IEEE Transactions on Image Processing, Vol. 15, No. 8, 2006, pp. 2198–2207

    2006

  15. [23]

    3-d symmetry detection and analysis using the pseudo-polar fourier transform,

    Bermanis, A., Averbuch, A., and Keller, Y., “3-d symmetry detection and analysis using the pseudo-polar fourier transform,” International journal of computer vision, Vol. 90, 2010, pp. 166–182

    2010

  16. [24]

    A., and Mullis, C

    Roberts, R. A., and Mullis, C. T.,Digital signal processing, Addison-Wesley Longman Publishing Co., Inc., 1987

    1987

  17. [25]

    Fast optical flow using dense inverse search,

    Kroeger, T., Timofte, R., Dai, D., and Van Gool, L., “Fast optical flow using dense inverse search,”Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11–14, 2016, Proceedings, Part IV 14, Springer, 2016, pp. 471–488. 41

    2016

  18. [26]

    An iterative image registration technique with an application to stereo vision,

    Lucas, B. D., and Kanade, T., “An iterative image registration technique with an application to stereo vision,”IJCAI’81: 7th international joint conference on Artificial intelligence, Vol. 2, 1981, pp. 674–679

    1981

  19. [27]

    Image-based lunar terrain relative navigation without a map: Measurements,

    Christian, J. A., Hong, L., McKee, P., Christensen, R., and Crain, T. P., “Image-based lunar terrain relative navigation without a map: Measurements,”Journal of Spacecraft and Rockets, Vol. 58, No. 1, 2021, pp. 164–181

    2021

  20. [28]

    H., and Van Loan, C

    Golub, G. H., and Van Loan, C. F.,Matrix computations, JHU press, 2013

    2013

  21. [29]

    Image Rendering and Terrain Generation of Planetary Surfaces Using Source-Available Tools,

    Villa, J., Mcmahon, J., and Nesnas, I., “Image Rendering and Terrain Generation of Planetary Surfaces Using Source-Available Tools,”Proceedings of the 46th Annual AAS Guidance, Navigation & Control Conference, Breckenridge, CO, USA, 2023, pp. 1–24

    2023

  22. [30]

    The unexpected surface of asteroid (101955) Bennu,

    Lauretta, D., DellaGiustina, D., Bennett, C., Golish, D., Becker, K., Balram-Knutson, S., Barnouin, O., Becker, T., Bottke, W., Boynton, W., et al., “The unexpected surface of asteroid (101955) Bennu,”Nature, Vol. 568, No. 7750, 2019, pp. 55–60

    2019

  23. [31]

    The morphological diversity of comet 67P/Churyumov-Gerasimenko,

    Thomas, N., Sierks, H., Barbieri, C., Lamy, P. L., Rodrigo, R., Rickman, H., Koschny, D., Keller, H. U., Agarwal, J., A’Hearn, M. F., et al., “The morphological diversity of comet 67P/Churyumov-Gerasimenko,”Science, Vol. 347, No. 6220, 2015, p. aaa0440

    2015

  24. [32]

    R.,Digital image processing, Prentice Hall Press, 1996

    Castleman, K. R.,Digital image processing, Prentice Hall Press, 1996. Appendix Image Symmetries and Transformations Definition VI.1.Let x∈ R2 be a 2D vector. The reflective transformation ofx with respect to the line passing by the origin and with an angle𝜃 with respect to the...

    1996

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.