pith. sign in

arxiv: 2606.30286 · v1 · pith:TWBDEVR6new · submitted 2026-06-29 · 🌌 astro-ph.IM

Streak detection in the VST/OmegaCAM archive using deep learning

Elisabeth Rachith , Stephan Hellmich , Vincent Fiszbin , Bel\'en Yu Irureta-Goyena , Andrew Price , Jean-Paul Kneib This is my paper

Pith reviewed 2026-06-30 03:43 UTC · model grok-4.3

classification 🌌 astro-ph.IM
keywords streak detectiondeep learningOmegaCAMVSTspace debrissatellite streaksastronomical archivesimage contamination
0
0 comments X

The pith

A deep learning pipeline identifies 25,335 streaks in one year of VST/OmegaCAM frames, with more than 20 percent uncatalogued and 16.9 percent of images contaminated.

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

The paper develops and tests an automated pipeline that combines an adapted Hough transform lookup convolutional network for initial streak detection on raw frames with a VGG6 classifier to remove false positives. The work augments a set of 384,000 manually labeled patches with physically simulated streaks, then applies the trained system to more than 1.2 million real OmegaCAM CCD exposures. On held-out 2023 data the classifier raises precision to 0.99 while keeping 97 percent of true detections, and the full run uncovers thousands of streaks not listed in existing catalogs. This approach converts incidental satellite and debris trails already present in wide-field survey archives into a systematic record of resident space objects and their effect on astronomical imaging.

Core claim

The HT-LCNN detector followed by the VGG6 classifier achieves F1 scores above 0.95 on augmented validation and test sets, detects more than 95 percent of artificial streaks above S/N of 4, and on real 2023 frames reaches 0.99 precision after classification while retaining 97 percent of true positives; when run on the full year of 1,246,048 frames it returns 25,335 streaks of which over 20 percent have no match in the space-track catalogue, and 16.9 percent of the images show some level of streak contamination.

What carries the argument

HT-LCNN detector for initial streak finding on raw images followed by VGG6-based CNN classifier that rejects false positives after astrometric calibration and catalogue cross-match.

If this is right

  • The pipeline turns existing survey archives into a source of large-scale statistics on resident space objects without requiring new telescope time.
  • More than one in six OmegaCAM images contains detectable streak contamination that can affect photometric or astrometric measurements.
  • Over 20 percent of detected streaks lack catalogue counterparts, indicating that the method can locate previously unknown faint debris.
  • Astrometric calibration performed after detection supplies positions that can be fed directly into orbit-determination routines for the uncatalogued objects.

Where Pith is reading between the lines

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

  • The same two-stage architecture could be retrained on other wide-field instruments to produce comparable contamination statistics across multiple surveys.
  • Repeated application to future archive releases would allow tracking of changes in streak density as satellite constellations grow.
  • Cross-matching the detected streaks against independent optical or radar catalogues would quantify the fraction of objects missed by current tracking systems.

Load-bearing premise

Detection and classification performance measured on the augmented training patches and the 2023 test images will remain stable across the entire multi-year archive despite uncharacterized changes in image quality, background levels, and streak appearance.

What would settle it

A measured drop below 0.9 precision or below 0.9 recall when the same trained pipeline is applied to a new set of OmegaCAM frames taken under different observing conditions or in a different year.

Figures

Figures reproduced from arXiv: 2606.30286 by Andrew Price, Bel\'en Yu Irureta-Goyena, Elisabeth Rachith, Jean-Paul Kneib, Stephan Hellmich, Vincent Fiszbin.

Figure 1
Figure 1. Figure 1: OmegaCAM detector layout with a visible satellite streak. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Simulated streak (top) and real streak (bottom) of objects [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Sample input data used by the classification algorithm. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Examples of common false-positive detections made by [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Streak alignment procedure. (a) Original image cutout showing a detection from the network. (b) Rotated cutout with the [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Schematic representation of the CNN used for binary [PITH_FULL_IMAGE:figures/full_fig_p006_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Detection algorithm training loss (a) and validation loss [PITH_FULL_IMAGE:figures/full_fig_p006_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Training and validation loss for the detection network. [PITH_FULL_IMAGE:figures/full_fig_p006_8.png] view at source ↗
Figure 10
Figure 10. Figure 10: Detection algorithm precision-recall curve, evaluated [PITH_FULL_IMAGE:figures/full_fig_p007_10.png] view at source ↗
Figure 14
Figure 14. Figure 14: Classifier F1-score on the validation set as a function of the weight decay. As with the detection algorithm, we first evaluated the fi￾nal network on the validation dataset [PITH_FULL_IMAGE:figures/full_fig_p008_14.png] view at source ↗
Figure 13
Figure 13. Figure 13: Classifier F1-score on the validation set as a function of [PITH_FULL_IMAGE:figures/full_fig_p008_13.png] view at source ↗
Figure 17
Figure 17. Figure 17: Permutation-based feature importance for the MLP. Bars [PITH_FULL_IMAGE:figures/full_fig_p009_17.png] view at source ↗
Figure 16
Figure 16. Figure 16: Classifier precision-recall curve. the discarded detections suggests that the MLP has a strong ten￾dency to retain streaks that cross the full detector while rejecting detections that do not span the entire CCD. Although this be￾haviour helps remove many false positives, it also eliminates a number of streaks with an endpoint inside the image, which are particularly valuable for the photometric analysis p… view at source ↗
Figure 18
Figure 18. Figure 18: Sample detections from the OmegaCAM archive. All streaks were correlated with known objects. [PITH_FULL_IMAGE:figures/full_fig_p011_18.png] view at source ↗
Figure 19
Figure 19. Figure 19: Distribution of correlated objects in 2023 by size and object type for each orbital regime. The number in parentheses next [PITH_FULL_IMAGE:figures/full_fig_p011_19.png] view at source ↗
Figure 20
Figure 20. Figure 20: Same as Fig. 19 but for all space-track catalogued objects on orbit on 31 December 2023. [PITH_FULL_IMAGE:figures/full_fig_p011_20.png] view at source ↗
Figure 21
Figure 21. Figure 21: Relative difference in detection ratios between the OmegaCAM 2023 detections and the space-track catalogue. Categories are the same as in [PITH_FULL_IMAGE:figures/full_fig_p012_21.png] view at source ↗
Figure 22
Figure 22. Figure 22: Fraction of images affected by at least one streak for all pointing directions of OmegaCAM images acquired in 2023. The concentric circles represent the altitude in degrees, while the radial axis shows the azimuth in degrees. ing such objects using large field-of-view telescope archives is therefore a promising direction for future works. Although the seeing effect produces a visible wobble in OmegaCAM st… view at source ↗
read the original abstract

Ground-based astronomical surveys inadvertently capture streaks from satellites and space debris crossing their fields of view. These incidental observations from wide-field instruments such as OmegaCAM on the VST offer valuable opportunities to characterise resident space objects without the need for dedicated observing time. We developed an automated deep-learning pipeline to detect and classify streaks in the OmegaCAM archive, enabling large-scale analyses of space object populations and their impact on astronomical data. The pipeline combines an adapted Hough transform lookup-based convolutional neural network (HT-LCNN) for initial streak detection on raw images with a VGG6-based CNN classifier to reject false positives. We augmented a manually annotated dataset of 384 000 patches from archive images with physically simulated streaks. Following a detection, we applied astrometric calibration and cross-matched the results with the space-track catalogue. We find the detector achieves F1-scores of 0.966 (validation) and 0.958 (test) on the augmented dataset, detecting > 95% of artificial streaks with a signal-to-noise ratio of S/N > 4. On real 2023 data, the precision drops to 0.783 due to image variability, but the classifier boosts it to 0.990, while retaining 97% of true positives and rejecting > 96% of false positives. Applied to one year of VST observations (1 246 048 OmegaCAM CCD frames), the pipeline identified 25 335 streaks, including more than 20% uncorrelated with catalogue entries; finally, 16.9% of images revealed some level of contamination. The pipeline demonstrates robust performance on real archival data and successfully uncovers faint uncatalogued objects, highlighting the potential of survey archives for debris monitoring.

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 / 2 minor

Summary. The manuscript presents a deep-learning pipeline (HT-LCNN detector followed by VGG6 classifier) for identifying satellite and space-debris streaks in OmegaCAM/VST archival images. On an augmented dataset of 384000 patches it reports F1 scores of 0.966 (validation) and 0.958 (test), with >95% detection of artificial streaks at S/N>4. On real 2023 data the raw detector precision is 0.783 but rises to 0.990 after classification while retaining 97% of true positives. When applied to 1246048 frames from one year the pipeline yields 25335 streaks (>20% uncatalogued) and finds 16.9% of images contaminated.

Significance. If the reported generalization holds, the work supplies a practical, scalable method for mining existing wide-field archives for resident-space-object statistics without dedicated telescope time. The concrete numerical results on both augmented and real data, together with the large-scale application, constitute a useful empirical contribution to astro-imaging and space-situational-awareness studies.

major comments (2)
  1. [Abstract / archive application] Abstract and the section describing the archive application: the central claim that the pipeline reliably produces the archive-wide counts (25335 streaks, 16.9% contamination) rests on the untested assumption that performance measured on the 2023 test set generalizes to the full one-year dataset. No stratified evaluation, cross-validation by year/seeing/airmass/CCD, or performance breakdown on non-2023 subsets is reported, despite the explicit statement that image variability already lowers precision on real 2023 data.
  2. [Methods] Methods (dataset and augmentation description): the train/validation/test split ratios, exact augmentation parameters, and any quantitative comparison between the distribution of simulated streaks and real streaks (length, orientation, S/N histograms) are not provided. These details are load-bearing for interpreting the quoted F1 scores and the >95% detection claim for S/N>4 artificial streaks.
minor comments (2)
  1. [Abstract] The abstract states a manually annotated set of 384000 patches but does not indicate the fraction allocated to training versus testing or the number of real versus simulated examples; adding these numbers would improve reproducibility assessment.
  2. [Results] The S/N>4 threshold used for the detection-rate claim should be explicitly tied to the free parameter listed in the methods; its effect on the final catalogue statistics should be quantified.

Simulated Author's Rebuttal

2 responses · 1 unresolved

We thank the referee for the constructive comments on our manuscript. We address the major points below, indicating where revisions will be made and where limitations prevent further analysis.

read point-by-point responses

  1. Referee: [Abstract / archive application] Abstract and the section describing the archive application: the central claim that the pipeline reliably produces the archive-wide counts (25335 streaks, 16.9% contamination) rests on the untested assumption that performance measured on the 2023 test set generalizes to the full one-year dataset. No stratified evaluation, cross-validation by year/seeing/airmass/CCD, or performance breakdown on non-2023 subsets is reported, despite the explicit statement that image variability already lowers precision on real 2023 data.

    Authors: We agree that the archive-wide statistics rely on the assumption that the 2023 real-data test performance is representative of the full year. The 2023 subset was chosen specifically to evaluate on recent, real observations where variability is present. Without ground-truth labels for the remaining data, stratified evaluation by year or observing conditions is not possible. We will revise the abstract and application section to explicitly caveat the generalization assumption, discuss the observed impact of variability, and note the limitation for future work. revision: partial

  2. Referee: [Methods] Methods (dataset and augmentation description): the train/validation/test split ratios, exact augmentation parameters, and any quantitative comparison between the distribution of simulated streaks and real streaks (length, orientation, S/N histograms) are not provided. These details are load-bearing for interpreting the quoted F1 scores and the >95% detection claim for S/N>4 artificial streaks.

    Authors: These details were inadvertently omitted. In the revised manuscript we will report the exact train/validation/test split ratios, list the precise augmentation parameters (including ranges for rotation, scaling, noise, etc.), and add quantitative comparisons (histograms and summary statistics) of streak length, orientation, and S/N between the simulated and real distributions to support the reported metrics. revision: yes

standing simulated objections not resolved
  • Ground-truth labels exist only for the 2023 test subset; therefore no stratified performance evaluation or cross-validation on non-2023 portions of the archive is feasible.

Circularity Check

0 steps flagged

No circularity; purely empirical ML pipeline with held-out metrics

full rationale

The paper describes a standard supervised deep-learning pipeline (HT-LCNN detector + VGG6 classifier) trained on manually annotated plus simulated patches, with F1/precision numbers obtained from explicit validation and test splits on augmented data and 2023 real images. These metrics are measured outcomes, not quantities derived from equations that reduce to fitted parameters by construction. No self-citation chains, uniqueness theorems, or ansatzes are invoked to justify the architecture or results; the archive-wide counts follow directly from running the trained model on the target data. The derivation chain is therefore self-contained and externally falsifiable via the reported data splits.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central performance claims rest on the representativeness of the manually annotated plus simulated-streak training set and on the assumption that the two-stage architecture generalizes across the archive's image variability; no new physical entities or ad-hoc constants are introduced.

free parameters (1)
  • S/N > 4 threshold
    Reporting threshold used to claim >95% detection of artificial streaks; chosen post hoc for the quoted statistic.
axioms (1)
  • domain assumption Simulated streaks accurately capture the appearance and noise properties of real satellite streaks in OmegaCAM images
    Invoked when augmenting the 384,000-patch dataset and when claiming generalization to real 2023 data.

pith-pipeline@v0.9.1-grok · 5869 in / 1424 out tokens · 50915 ms · 2026-06-30T03:43:33.166415+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

218 extracted references · 174 canonical work pages · 11 internal anchors

  1. [1]

    2001, MNRAS, 327, 799, doi: 10.1046/j.1365-8711.2001.04769.x

    The effects of seeing on. MNRAS , author =. 2001 , keywords =. doi:10.1046/j.1365-8711.2001.04937.x , abstract =

    work page doi:10.1046/j.1365-8711.2001.04937.x 2001

  2. [2]

    L., Stubbs, C

    The. ApJ , author =. 2012 , pages =. doi:10.1088/0004-637X/750/2/99 , abstract =

    work page doi:10.1088/0004-637x/750/2/99 2012

  3. [3]

    ApJ , author =

    The. ApJ , author =. 2018 , pages =. doi:10.3847/1538-4357/aae386 , abstract =

    work page doi:10.3847/1538-4357/aae386 2018

  4. [4]

    A&A , author =

    Automated detection of satellite trails in ground-based observations using. A&A , author =. 2024 , pages =. doi:10.1051/0004-6361/202451663 , abstract =

    work page doi:10.1051/0004-6361/202451663 2024

  5. [5]

    MNRAS , author =

    The. MNRAS , author =. 2015 , pages =. doi:10.1093/mnras/stv1130 , abstract =

    work page doi:10.1093/mnras/stv1130 2015

  6. [6]

    Schildknecht, Thomas and Pärli, Fiona and Vananti, Alessandro and Patole, Palash and Pessev, Peter and Flohrer, Tim and Siminski, Jan , year =. Recent. Proceedings of the 2024

    2024

  7. [7]

    AJ , author =

    Optimal and. AJ , author =. 2018 , keywords =. doi:10.3847/1538-3881/aaddff , abstract =

    work page doi:10.3847/1538-3881/aaddff 2018

  8. [8]

    AJ , author =

    Identification of. AJ , author =. 2025 , pages =. doi:10.3847/1538-3881/adb8ea , abstract =

    work page doi:10.3847/1538-3881/adb8ea 2025

  9. [9]

    and van Gemert, Jan C

    Lin, Yancong and Pintea, Silvia L. and van Gemert, Jan C. , month = jul, year =. Deep. Proceedings of the. doi:10.48550/arXiv.2007.09493 , abstract =

    work page doi:10.48550/arxiv.2007.09493 2007

  10. [10]

    Mulrooney, Mark K and Matney, Mark J and Hejduk, Matthew D and Barker, Edwin S , year =. An. Proceedings of the 2008

    2008

  11. [11]

    A&A , author =

    Impact of satellite trails on. A&A , author =. 2023 , pages =. doi:10.1051/0004-6361/202347200 , abstract =

    work page doi:10.1051/0004-6361/202347200 2023

  12. [12]

    2010 , pages =

    AJ , author =. 2010 , pages =. doi:10.1088/0004-6256/139/5/1782 , language =

    work page doi:10.1088/0004-6256/139/5/1782 2010

  13. [13]

    MNRAS , author =

    Gravitational lensing analysis of the. MNRAS , author =. 2015 , pages =. doi:10.1093/mnras/stv2140 , abstract =

    work page doi:10.1093/mnras/stv2140 2015

  14. [14]

    Nat , author =

    The impact of satellite trails on. Nat , author =. 2023 , pages =. doi:10.1038/s41550-023-01903-3 , abstract =

    work page doi:10.1038/s41550-023-01903-3 2023

  15. [15]

    A&A , author =

    Streak detection in the. A&A , author =. 2026 , note =

    2026

  16. [16]

    A&A , author =

    A method for asteroid detection using convolutional neural networks on. A&A , author =. 2025 , pages =. doi:10.1051/0004-6361/202452756 , abstract =

    work page doi:10.1051/0004-6361/202452756 2025

  17. [17]

    Acta Astron

    Space-based surveillance of. Acta Astron. , author =. 2025 , pages =. doi:10.1016/j.actaastro.2025.10.064 , abstract =

    work page doi:10.1016/j.actaastro.2025.10.064 2025

  18. [18]

    , archivePrefix = "arXiv", eprint =

    The. AJ , author =. 2015 , pages =. doi:10.1088/0004-6256/150/5/150 , abstract =

    work page doi:10.1088/0004-6256/150/5/150 2015

  19. [19]

    The Messenger , author =

    Seeing is. The Messenger , author =

  20. [20]

    Flux comparison of

    Horstmann, André and Wiedemann, Carsten and Lemmens, Stijn and Braun, Vitali and Manis, Alyssa and Matney, Mark and Gates, Drake and Seago, John and Vavrin, Andrew and Anz-Meador, Phillip , year =. Flux comparison of. 8th

  21. [21]

    AJ , author =

    Tails:. AJ , author =. 2021 , pages =. doi:10.3847/1538-3881/abea7b , abstract =

    work page doi:10.3847/1538-3881/abea7b 2021

  22. [22]

    ApJ , author =

    Impact of the. ApJ , author =. 2022 , keywords =. doi:10.3847/2041-8213/ac470a , abstract =

    work page doi:10.3847/2041-8213/ac470a 2022

  23. [23]

    2019 , keywords =

    MNRAS , author =. 2019 , keywords =. doi:10.1093/mnras/stz1096 , abstract =

    work page doi:10.1093/mnras/stz1096 2019

  24. [24]

    The Journal of the Astronautical Sciences , author =

    Exploiting. The Journal of the Astronautical Sciences , author =. 2019 , pages =. doi:10.1007/s40295-019-00155-6 , abstract =

    work page doi:10.1007/s40295-019-00155-6 2019

  25. [25]

    Artificial

    Chen, Xiaohan and Gu, Hongrui and Wang, Cunshi and Mu, Haiyang and Zheng, Jie and Du, Junju and Ren, Jing and Fan, Zhou and Li, Jing , month = sep, year =. Artificial. PASP , publisher =. doi:10.1088/1538-3873/ae051f , abstract =

    work page doi:10.1088/1538-3873/ae051f

  26. [26]

    PASP , author =

    Toward. PASP , author =. 2019 , pages =. doi:10.1088/1538-3873/ab1b18 , abstract =

    work page doi:10.1088/1538-3873/ab1b18 2019

  27. [27]

    C., Kulkarni, S

    The. PASP , author =. 2019 , keywords =. doi:10.1088/1538-3873/aaecbe , abstract =

    work page doi:10.1088/1538-3873/aaecbe 2019

  28. [28]

    J., Laher, R

    The. PASP , author =. 2019 , pages =. doi:10.1088/1538-3873/aae8ac , language =

    work page internal anchor Pith review doi:10.1088/1538-3873/aae8ac 2019

  29. [29]

    Space and

    Bloom, Amelia and Wysack, Joshua and Griesbach, Jacob D and Lawitzke, Anna , year =. Space and. Proceedings of the 2022

    2022

  30. [30]

    A&A , author =

    Unintended electromagnetic radiation from. A&A , author =. 2023 , pages =. doi:10.1051/0004-6361/202346374 , abstract =

    work page doi:10.1051/0004-6361/202346374 2023

  31. [31]

    A&A , author =

    Bright unintended electromagnetic radiation from second-generation. A&A , author =. 2024 , pages =. doi:10.1051/0004-6361/202451856 , abstract =

    work page doi:10.1051/0004-6361/202451856 2024

  32. [32]

    A&A , author =

    Analytical simulations of the effect of satellite constellations on optical and near-infrared observations , volume =. A&A , author =. 2022 , keywords =. doi:10.1051/0004-6361/202142101 , abstract =

    work page doi:10.1051/0004-6361/202142101 2022

  33. [33]

    MNRAS , author =

    The proliferation of space objects is a rapidly increasing source of artificial night sky brightness , volume =. MNRAS , author =. 2021 , keywords =. doi:10.1093/mnrasl/slab030 , abstract =

    work page doi:10.1093/mnrasl/slab030 2021

  34. [34]

    Nat , author =

    Aggregate effects of proliferating low-. Nat , author =. 2023 , pages =. doi:10.1038/s41550-023-01904-2 , language =

    work page doi:10.1038/s41550-023-01904-2 2023

  35. [35]

    The Astropy Project: Sustaining and Growing a Community-oriented Open-source Project and the Latest Major Release (v5.0) of the Core Package

    The. ApJ , author =. 2022 , keywords =. doi:10.3847/1538-4357/ac7c74 , abstract =

    work page internal anchor Pith review doi:10.3847/1538-4357/ac7c74 2022

  36. [36]

    and Green, Richard and Rawls, Meredith L

    Boley, Aaron C. and Green, Richard and Rawls, Meredith L. and Eggl, Siegfried , month = mar, year =. Research Notes of the AAS , publisher =. doi:10.3847/2515-5172/adc12f , abstract =

    work page doi:10.3847/2515-5172/adc12f

  37. [37]

    Aerospace , author =

    Distribution and. Aerospace , author =. 2026 , note =. doi:10.3390/aerospace13040303 , abstract =

    work page doi:10.3390/aerospace13040303 2026

  38. [38]

    SPHEREx confirms predictions for artificial satellite trail pollution in Low Earth Orbit

    Borlaff, Alejandro S. and Marcum, Pamela M. and Howell, Steve B. and Sánchez-Alarcón, Pablo M. and Dubois, David and McDowell, Jonathan , month = may, year =. doi:10.48550/arXiv.2605.27501 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2605.27501

  39. [39]

    DebrisWatch II: Digging deeper for geosynchronous debris

    Blake, James A. and Cooke, Benjamin F. and Paragini, Cristina and Feline, William and Onken, Christopher A. and Pollacco, Don and Privett, Grant and Yanagisawa, Toshifumi and Airey, Robert and Apergis, Ioannis and Armellin, Roberto and Beesley, Lily and Chote, Paul and Cutolo, Anna-Maria and Eves, Stuart and Fujiwara, Tomoko and Kuroda, Daisuke and Lockle...

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2605.27516

  40. [40]

    2024 , keywords =

    Proceedings of the Conference on Robots and Vision , author =. 2024 , keywords =. doi:10.21428/d82e957c.8ef38e6b , abstract =

    work page doi:10.21428/d82e957c.8ef38e6b 2024

  41. [41]

    2026 , note =

    Astronomy & Astrophysics , author =. 2026 , note =. doi:10.1051/0004-6361/202558754 , abstract =

    work page doi:10.1051/0004-6361/202558754 2026

  42. [42]

    Earth's Future , author =

    Radiative. Earth's Future , author =. 2026 , keywords =. doi:10.1029/2025EF007229 , abstract =

    work page doi:10.1029/2025ef007229 2026

  43. [43]

    Monthly Notices of the Royal Astronomical Society , author =

    Linear feature detection algorithm for astronomical surveys -. Monthly Notices of the Royal Astronomical Society , author =. 2018 , note =. doi:10.1093/mnras/stx3085 , abstract =

    work page doi:10.1093/mnras/stx3085 2018

  44. [44]

    How long can you trust a Starlink TLE? An empirical comparison of SGP4 and high-fidelity propagation against operator-updated truth across a megaconstellation

    Jankovic, Dimitrije , month = may, year =. How long can you trust a. doi:10.48550/arXiv.2605.19850 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2605.19850

  45. [45]

    Impact of Satellite Constellations on Observations with the 80-cm Telescope and the Mini-SiTian at the Xinglong Observatory, NAOC

    Ren, Jing and Fan, Zhou and Gu, Hong-rui and Zhang, Qi-qian and Xu, Yun-fei and Du, Jun-ju and Chen, Xiao-han and Mi, Lin-ying and Wu, Hong , month = may, year =. Impact of. doi:10.48550/arXiv.2605.17948 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2605.17948

  46. [46]

    Mallama, Anthony , month = dec, year =. The. doi:10.48550/arXiv.2012.05100 , abstract =

    work page doi:10.48550/arxiv.2012.05100 2012

  47. [47]

    Sarazin, Marc and Tokovinin, Andrei , year =. The. European

  48. [48]

    Bulletin of the AAS , author =

    Impact of. Bulletin of the AAS , author =. doi:10.3847/25c2cfeb.346793b8 , language =

    work page doi:10.3847/25c2cfeb.346793b8

  49. [49]

    and Rangwala, Naqiya , month = feb, year =

    Lemonwala, Murtuza and Taylor, Kayla D. and Rangwala, Naqiya , month = feb, year =. Classification of. doi:10.1109/SoutheastCon63549.2026.11476213 , abstract =

    work page doi:10.1109/southeastcon63549.2026.11476213 2026

  50. [50]

    Communications Earth & Environment , author =

    Measurement of a lithium plume from the uncontrolled re-entry of a. Communications Earth & Environment , author =. 2026 , keywords =. doi:10.1038/s43247-025-03154-8 , abstract =

    work page doi:10.1038/s43247-025-03154-8 2026

  51. [51]

    Advances in Space Research , author =

    Space waste:. Advances in Space Research , author =. 2026 , keywords =. doi:10.1016/j.asr.2026.03.026 , abstract =

    work page doi:10.1016/j.asr.2026.03.026 2026

  52. [52]

    and Malagón, Andrés Alejandro Plazas , month = mar, year =

    Mendoza, Alexandra Serrano and Rawls, Meredith L. and Malagón, Andrés Alejandro Plazas , month = mar, year =. Identifying and. doi:10.48550/arXiv.2603.10790 , abstract =

    work page doi:10.48550/arxiv.2603.10790

  53. [53]

    Diffuse and specular brightness models applied to LEO satellites. Case study: The ONEWEB constellation

    Romero-Colmenares, María and Vieira, Katherine and Tregloan-Reed, Jeremy and Kim, Yonggi and Yoon, Joh-Na and Kim, Ha-eun and Jeon, Hyo-ri and Kim, Chae-rin and Adam, Christian and Hinse, Tobías C. and Soto, Mario and Unda-Sanzana, Eduardo and Longa-Peña, Penélope and Otarola, Ángel , year =. Diffuse and specular brightness models applied to. doi:10.48550...

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2603.21859

  54. [54]

    Mallama, Anthony , month = apr, year =. Secular. doi:10.48550/arXiv.2604.13145 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2604.13145

  55. [55]

    Large or bright satellite constellations: Effects on observations, including on the background sky brightness

    Hainaut, Olivier R. , month = apr, year =. Large or bright satellite constellations:. doi:10.48550/arXiv.2604.09427 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2604.09427

  56. [56]

    doi:10.48550/arXiv.2504.10550 , abstract =

    Kyselica, Daniel and Hrobár, Tomáš and Šilha, Jiří and Ďurikovič, Roman and Šuppa, Marek , month = apr, year =. doi:10.48550/arXiv.2504.10550 , abstract =

    work page doi:10.48550/arxiv.2504.10550

  57. [57]

    Flohrer, T and Lemmens, S and Virgili, B Bastida and Krag, H and Klinkrad, H and Parrilla, E and Sanchez, N and Oliveira, J and Pina, F , year =. 6th

  58. [58]

    Drolshagen, Gerhard , year =. In-situ. Advanced

  59. [59]

    Concerns about ground based astronomical observations: quantifying satellites' constellations damages , doi =

    Gallozzi, Stefano and Paris, Diego and Scardia, Marco and Dubois, David , year =. Concerns about ground based astronomical observations: quantifying satellites' constellations damages , doi =

  60. [60]

    M., Brasseur, C

    astroquery:. The Astronomical Journal , author =. 2019 , keywords =. doi:10.3847/1538-3881/aafc33 , abstract =

    work page doi:10.3847/1538-3881/aafc33 2019

  61. [61]

    Experimental Astronomy , author =

    The. Experimental Astronomy , author =. 2013 , keywords =. doi:10.1007/s10686-012-9306-1 , abstract =

    work page doi:10.1007/s10686-012-9306-1 2013

  62. [62]

    Collision frequency of artificial satellites: The creation of a debris belt,

    Collision frequency of artificial satellites:. Journal of Geophysical Research: Space Physics , author =. 1978 , pages =. doi:10.1029/JA083iA06p02637 , abstract =

    work page doi:10.1029/ja083ia06p02637 1978

  63. [63]

    Adam: A Method for Stochastic Optimization

    Kingma, Diederik P. and Ba, Jimmy Lei , year =. Adam:. Third. doi:10.48550/arXiv.1412.6980 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1412.6980

  64. [64]

    and Block, Adam , year =

    Krantz, Harrison and Pearce, Eric C. and Block, Adam , year =. Characterizing the. Advanced. doi:10.48550/arXiv.2110.10578 , abstract =

    work page doi:10.48550/arxiv.2110.10578

  65. [65]

    Starlink

    Mallama, Anthony , month = nov, year =. Starlink. doi:10.48550/arXiv.2111.09735 , abstract =

    work page doi:10.48550/arxiv.2111.09735

  66. [66]

    Mallama, Anthony , month = jan, year =. The. doi:10.48550/arXiv.2101.00374 , abstract =

    work page doi:10.48550/arxiv.2101.00374

  67. [67]

    On the Convergence of Adam and Beyond

    Reddi, Sashank J. and Kale, Satyen and Kumar, Sanjiv , year =. On the. The. doi:10.48550/arXiv.1904.09237 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1904.09237 1904

  68. [68]

    Photometric monitoring of non-resolved space debris and databases of optical light curves , abstract =

    Schildknecht, Thomas and Linder, Esther and Silha, Jiri and Hager, Monika and Koshkin, Nikolay and Korobeinikova, Elen and Melikiants, Seda and Shakun, Leonid and Strakhov, Svetlana , year =. Photometric monitoring of non-resolved space debris and databases of optical light curves , abstract =. Advanced

  69. [69]

    Wijnen, T. P. G. and Stuik, R. and Rodenhuis, M. and Langbroek, M. and Wijnja, P. , year =. Using all-sky optical observations for automated orbit determination and prediction for satellites in. 1st. doi:10.48550/arXiv.2004.08137 , abstract =

    work page doi:10.48550/arxiv.2004.08137 2004

  70. [70]

    Kessler, Donald J and Johnson, Nicholas L , pages =. The

  71. [71]

    Matney, Mark and Vavrin, Andrew and Poehlmann, Michael and King, Austen and Seago, John H and Anz-Meador, Phillip and Liou, Jer-Chyi and Cruz, Corbin and Manis, Alyssa , year =. 9th

  72. [72]

    Advances in Space Research , author =

    Optical observations of space debris with the. Advances in Space Research , author =. 1997 , pages =. doi:10.1016/S0273-1177(97)00004-5 , language =

    work page doi:10.1016/s0273-1177(97)00004-5 1997

  73. [73]

    Monthly Notices of the Royal Astronomical Society , author =

    The. Monthly Notices of the Royal Astronomical Society , author =

  74. [74]

    Karpov, S and Beskin, G and Biryukov, A and Bondar, S and Ivanov, E and Katkova, E and Perkov, A and Sasyuk, V , year =. Mini-. Proceedings of the

  75. [75]

    Estimating the impact to astronomy from the

    Adam, Christian and Tregloan-Reed, Jeremy and Unda-Sanzana, Eduardo , month = dec, year =. Estimating the impact to astronomy from the. Proceedings of the. doi:10.1017/S1743921324000620 , abstract =

    work page doi:10.1017/s1743921324000620

  76. [76]

    Advances in Space Research , author =

    Space debris spectroscopy:. Advances in Space Research , author =. 2023 , pages =. doi:10.1016/j.asr.2022.12.001 , abstract =

    work page doi:10.1016/j.asr.2022.12.001 2023

  77. [77]

    Advances in Space Research , author =

    Modelling of spacecraft apparent brightness. Advances in Space Research , author =. 2024 , pages =. doi:10.1016/j.asr.2024.04.010 , abstract =

    work page doi:10.1016/j.asr.2024.04.010 2024

  78. [78]

    Recent developments in space debris environment modelling, verification and validation with

    Braun, V and Horstmann, A and Lemmens, S and Wiedemann, C and Bottcher, L , year =. Recent developments in space debris environment modelling, verification and validation with. 8th

  79. [79]

    and Busko, Ivo and Donath, Axel and Günther, Hans Moritz and Cara, Mihai and Lim, P

    Bradley, Larry and Sipőcz, Brigitta and Robitaille, Thomas and Tollerud, Erik and Vinícius, Zé and Deil, Christoph and Barbary, Kyle and Wilson, Tom J. and Busko, Ivo and Donath, Axel and Günther, Hans Moritz and Cara, Mihai and Lim, P. L. and Meßlinger, Sebastian and Burnett, Zach and Conseil, Simon and Droettboom, Michael and Bostroem, Azalee and Bray, ...

    work page doi:10.5281/zenodo.14889440

  80. [80]

    2011 , pages =

    The Messenger , author =. 2011 , pages =

    2011

Showing first 80 references.