Pith. sign in

REVIEW 6 minor 48 references

This paper establishes that a lunar-orbit radio interferometer can reconstruct an all-sky map at 0.1–30 MHz, provided the beam matrix is built by pixel-averaging instead of point sampling, and that with Tikhonov regularization tuned by a tr

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-03 20:42 UTC pith:NVMH4UVT

load-bearing objection Pixel-averaging fix for lunar-array aliasing is credible; the paper's own idealized-array caveats mean 'well reconstructed' should be read narrowly.

arxiv 2511.18494 v1 pith:NVMH4UVT submitted 2025-11-23 astro-ph.IM

Synthesis imaging with a lunar orbit array: I. global sky map and its systematics

Furen Deng , Yidong Xu , Fengquan Wu , Yanping Cong , Bin Yue , Xuelei Chen This is my paper
classification astro-ph.IM
keywords lunar orbit arrayradio interferometryall-sky mappingTikhonov regularizationaliasingpixel averagingHEALPixlow-frequency radio astronomy
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper tries to show that full-sky synthesis imaging with the DSL/Hongmeng lunar-orbit array—eight small satellites flying in a line around the Moon—is feasible if a numerical artifact is handled. It claims that evaluating the beam matrix at pixel centers aliases sub-pixel sky structure into the map when baselines exceed the pixel's Nyquist length, and that averaging the beam over each pixel removes this aliasing. It further claims that a Tikhonov regularization parameter can be chosen as a trade-off between thermal noise and effective-beam error, and that with that choice a degree-scale all-sky map is well reconstructed at 3–10 MHz. A sympathetic reader would care because this is the algorithmic core of an actual mission designed to observe the sky at frequencies that are blocked by the ionosphere on Earth.

Core claim

The central claim is that the linear inversion from visibilities V=BT to sky T works for the DSL array if B is computed by pixel-averaging. Naively evaluating B at pixel centers creates strong aliasing—the correlation coefficient ρ_l between input and reconstructed maps drops well below 1—when baselines b>bp≈λ/θ_p are included. The pixel-averaging method keeps ρ_l close to 1 for l≲130 even with b up to 4bp, and the authors adopt b<2bp as sufficient. With this method, a 10 MHz map reconstructed with ε=10^-4 shows small errors outside the polar regions; at 3 MHz the reconstruction is better because more baselines satisfy b<2bp. The polar regions are systematically darkened because the shortest

What carries the argument

The beam matrix B, whose entry B_{αβ} is the response of baseline-time sample α to sky pixel β (shading, beam, fringe phase, and pixel area). The paper's key move is computing B at high resolution (Nside=1024) and downgrading to Nside=64 by averaging, which is equivalent to convolving the sky with a pixel-shape window. The inversion then uses Tikhonov regularization, adding εI to the dirty beam B^T N^{-1}B; error is quantified by the effective beam B_eff = (B^T N^{-1}B+εI)^{-1} B^T N^{-1}B, with traces giving thermal-noise and beam-imperfection estimators, and by the scale-dependent correlation coefficient ρ_l and SNR_l.

Load-bearing premise

The reconstruction assumes the beam matrix B is a known, accurate linear description of how the sky maps to visibilities; the paper explicitly defers primary-beam modeling, calibration, baseline determination, clock synchronization, and lunar reflection, so if the real instrument deviates from this model, neither the reconstruction nor the pixel-averaging correction will transfer from simulation to flight data.

What would settle it

Take a mock observation with a high-resolution sky model (Nside=1024) and reconstruct at Nside=64 using both pixel-center and pixel-averaging beam matrices. If the pixel-averaged ρ_l drops well below 1 for l between 18 and 130 when including baselines b<4bp, the claimed mitigation is false. In flight, the equivalent test is whether known point sources appear at correct positions and fluxes in the reconstructed all-sky map.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Baselines longer than twice the pixel scale add no information for a fixed Nside=64 map; b<2bp is the working limit.
  • Point sampling of the beam matrix is not a viable option beyond Nyquist baselines; pixel-averaging is the recommended default.
  • A single ε value balances thermal noise against beam error; the crossover (ε~2×10^-3 in the 10 MHz case) is a starting point but not universal.
  • The polar regions will be biased low unless a prior map supplies large-scale power; a prior with correlation 0.8 is enough to fix most of the bias.
  • Point-source sensitivity is roughly flat from 5 to 30 MHz and improves sharply below 5 MHz under sky-dominated noise.

Where Pith is reading between the lines

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

  • If the real beam—antenna pattern, satellite attitude, baseline geometry—is not known to the accuracy assumed, the same aliasing or a related bias could reappear; the pixel-averaging fix does not remove the need for calibration.
  • The pixel-averaging method could extend to other wide-field interferometers or be formulated as an optimal pixel-window choice, where the window is matched to the baseline distribution rather than to the HEALPix pixel shape.
  • The polar darkening suggests a design lever: adding shorter baselines or including autocorrelation visibilities would directly recover the lost l≲18 modes, possibly making the prior map unnecessary.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

0 major / 6 minor

Summary. The paper presents an end-to-end simulation of all-sky synthesis imaging for the DSL lunar-orbit array. It constructs a mock sky with small-scale fluctuations, generates visibilities from an Nside=1024 model including Moon shading, finite bandwidth, and thermal noise, and reconstructs Nside=64 maps by Tikhonov regularization. The central result is that evaluating the beam matrix at pixel centers causes severe aliasing when baselines exceed the pixel Nyquist limit, whereas averaging the beam over the pixel area restores stable reconstruction. The authors compare pixel-center and pixel-averaging methods, study baseline cutoffs up to b<2b_p, and discuss the regularization trade-off using a scale-dependent correlation coefficient and a signal-to-noise ratio.

Significance. The paper's principal contribution is a clear, quantitative demonstration that pixel-averaged beam construction is a simple and effective anti-aliasing measure for interferometric map-making on a curved sky with finite pixelization. The simulation is unusually complete: it includes a realistic diffuse plus point-source sky with small-scale power, the orbital and breathing baseline distribution, finite bandwidth and time averaging, and thermal noise. The comparison of baseline cutoffs and the regularization discussion provide practical guidance for the DSL mission. The main limitation is that all results assume a perfectly known beam matrix; primary-beam, calibration, baseline-determination, and clock-synchronization errors are explicitly deferred. This does not invalidate the relative pixel-center versus pixel-averaging comparison, but it tempers the absolute 'well reconstructed' claim made in the abstract.

minor comments (6)
  1. [Abstract and Sec. 6] The phrase 'the sky can be well reconstructed' overstates the demonstrated result. The paper itself shows >20% relative errors in polar regions (Sec. 4.3, Figs. 9–11) and that faint structures are swamped by noise (Sec. 4.4). Moreover, the reconstruction assumes a perfectly known beam matrix, with the primary beam neglected in Sec. 3.2 and calibration/baseline/clock errors deferred in Sec. 1. Please qualify the claim, e.g., 'in the idealized simulation with a known beam,' both in the abstract and conclusion.
  2. [Appendix A, Eq. (A5)] The derivation of the map-space sub-pixel covariance uses B^T N^{-1/2} = Q W^{1/2}, which is not true in general; an orthogonal matrix from the SVD of B^T N^{-1/2} is missing. As written, Eq. (A5) is only valid if N_b is proportional to N. Since this formalism is not used in the main analysis, please correct the derivation or state explicitly that it is a heuristic approximation.
  3. [Sec. 5.2 and Figs. 16–18] The paper does not state how many thermal noise realizations are used for rho_l and SNR_l. If these curves are from a single realization, error bars or a multi-realization average should be provided, especially at low l where sample variance is large and the conclusions about the low-l behavior rely on the exact shape of the curves.
  4. [Sec. 5.1, Fig. 15] The two estimators bar{T}Tr(Sigma_s) and Tr(Sigma_n) have different units (K and K^2, respectively) but appear on the same plot. Please clarify the normalization and units, or plot dimensionless quantities, so that the claimed 'intersection' is meaningful.
  5. [Sec. 4.4] The statement that the noiseless reconstruction 'may even appear a little sharper than the original' is confusing, since the reconstruction is convolved with a 1-degree Gaussian. Please clarify whether this is an artifact of the 5-degree-smoothed input used for display, or explain the effect quantitatively.
  6. [Secs. 1–4] The term 'sub-pixel noise' is used for small-scale sky signal rather than instrumental noise. Consider defining this explicitly at first use (e.g., in Sec. 1 or Sec. 4.2) to avoid confusion with thermal noise.

Circularity Check

0 steps flagged

No circular derivation: the aliasing/pixel-averaging result is an independent sampling-theory effect tested against a high-resolution forward model, and the regularization choice is heuristic with demonstrated insensitivity; acknowledged idealizations are limitations, not circularity.

full rationale

The central chain is V = BT + eta (Eq. 9) with B defined by Eq. 10 from geometry and shading; the Tikhonov solution (Eq. 15) and the effective beam B_eff (Eq. 18) are algebraic manipulations, not definitions of the target result. The aliasing effect is justified by a Nyquist argument in Sec. 4.2 and is independently demonstrated by comparing the pixel-center and pixel-averaged B rows and mock visibilities against the Nside=1024 forward model (Figs. 7 and 8); the pixel-averaging method is a standard anti-aliasing convolution, and its advantage is measured against that high-resolution reference, so it is not fitted to the low-resolution reconstruction it is meant to validate. The success metrics rho_l and SNR_l (Eqs. 23-24) do use the same simulated input map that generated the visibilities, and epsilon is chosen by inspecting those same curves, but this is in-sample simulation validation rather than a fitted parameter renamed as a prediction; moreover, Sec. 5.3 and Fig. 18 show the results are insensitive to epsilon for epsilon < 1e-2, so the 'well reconstructed' claim does not reduce to tuning epsilon. The statements in Sec. 1 that 'we neglect these systematic errors' and 'the conclusions presented here will be robust... provided the performance of the satellite array attains the design parameters', together with the Sec. 3.2 statement that 'we simply ignore the primary beam of the tripole antenna here', are limitations on transfer to flight data; they are acknowledged in Sec. 6 as an idealized model, and they do not make the internal derivation circular. Self-citations (Huang et al. 2018; Shi et al. 2022; Cong et al. 2021) supply context, simulation inputs, and the linear-inversion convention, but no uniqueness theorem or load-bearing conclusion is imported from them. No specific equation or fitted parameter can be exhibited that reduces a prediction to its own input by construction.

Axiom & Free-Parameter Ledger

5 free parameters · 6 axioms · 0 invented entities

The paper introduces no new physical entities; 'sub-pixel noise' and 'effective beam' are modeling constructs derived from standard quantities. The free parameters are simulation/processing choices (epsilon, baseline cutoff, pixel scale, array configuration, source cutoff) that the central claims depend on.

free parameters (5)
  • Tikhonov regularization parameter epsilon = 1e-4 (fiducial), 1e-6 (3 MHz), 1e-3 (sensitivity)
    Chosen by hand from the trade-off between Tr(Sigma_s) and Tr(Sigma_n); the authors state the optimal value is problem-specific and not universal (Sec 5.1, 5.3).
  • Baseline cutoff b < 2 b_p = 2 x b_p, with b_p = lambda/theta_p ~ 1.87 km at 10 MHz, Nside=64
    Data selection threshold adopted to save computation; the authors argue longer baselines add little, but this limits resolution and affects the SNR claims (Sec 4.3, 5.2).
  • Pixel resolution and post-processing smoothing = Nside=64 final map, Nside=1024/256 intermediate, Gaussian FWHM=1 deg
    Computational limit determines b_p and hence which baselines are used; the Gaussian smoothing sets the effective ~1.3 deg resolution (Sec 4.2, 4.3).
  • Array compression ratios R2=6, R=3 = R2=6, R=3 (fiducial)
    Hand-selected to reduce collision risk and produce a more uniform baseline distribution (Sec 3.1); changes uvw coverage and the polar short-baseline deficit.
  • Point-source flux cutoff = 4.36 Jy at 154 MHz
    Chosen to avoid double counting with the Haslam diffuse component; affects the small-scale power in the mock sky and therefore the sub-pixel noise level (Sec 2).
axioms (6)
  • standard math Visibility is a known linear function of sky brightness: V = BT + eta (Eq 9-10).
    Standard interferometry relation; the inversion uses Tikhonov regularization (Eq 14-15).
  • domain assumption The Moon is an opaque sphere with negligible radio emission and negligible reflection; reflected power (~7%) is incoherent and ignored (Sec 3.2, 3.4).
    If false, the shading function Eq (6) and the beam matrix B are wrong, biasing the reconstructed map.
  • domain assumption Thermal noise is Gaussian, uncorrelated between baselines, with standard deviation from Eq (7); receiver noise is subdominant to sky noise (Sec 3.2).
    The noise model drives the regularization trade-off; extra low-frequency noise components would change the conclusions.
  • domain assumption All satellites follow the same circular orbit without deviation; breathing is idealized as instantaneous velocity changes (Sec 3.1).
    The uvw coverage and the polar baseline deficit depend on this simplified orbital model.
  • ad hoc to paper The constructed mock sky (ULSA diffuse emission plus GLEAM-derived point sources with added small-scale fluctuations) is representative of the true sub-pixel sky power (Sec 2).
    The severity of aliasing and the effectiveness of pixel averaging depend on the small-scale angular power spectrum assumed.
  • ad hoc to paper After pixel averaging, beam entries for baselines b>2b_p are damped to ~0 and contribute negligibly to the final map (Sec 5.2).
    This justifies excluding longer baselines; it is a numerical observation from the simulated configuration rather than a derived theorem.

pith-pipeline@v1.3.0-alltime-deepseek · 21405 in / 13756 out tokens · 131391 ms · 2026-08-03T20:42:21.084111+00:00 · methodology

0 comments
read the original abstract

Ground-based radio astronomical observation at frequencies below 30 MHz is hampered by the Ionosphere and radio frequency interference (RFI). The Discovering Sky at the Longest wavelength (DSL) mission, also known as the Hongmeng mission, employs a linear array of satellites on a circular orbit around the Moon to make interferometric observations in this band. Though vastly different from the usual ground-based arrays, the interferometric visibility data collected by such an array is linearly related to the sky map, and the reconstruction is in principle an inversion problem of linear mapping. In this paper, we investigate a number of issues in the algorithm of global map reconstruction, focusing on the impact of sub-pixel noise induced by the finite pixelization of the sky, and errors due to regularization. We find that in the reconstruction process, if one builds up the beam matrix, which relates the sky pixels to the visibilities, by naively evaluating its elements at each of the pixel centers, then the sub-pixel noise can give rise to a significant aliasing effect. However, this effect can be effectively mitigated by a simple pixel-averaging method. Based on evaluation of the image quality using the correlation coefficient between the input and reconstructed map, and the signal-to-noise ratio, we discuss the selection strategy of the regularization parameter, and show that the sky can be well reconstructed with a reasonable choice of the regularization parameter.

Figures

Figures reproduced from arXiv: 2511.18494 by Bin Yue, Fengquan Wu, Furen Deng, Xuelei Chen, Yanping Cong, Yidong Xu.

Figure 1
Figure 1. Figure 1: The input sky model. (a) A patch of the sky of the diffuse component at 10 MHz, shown in the ecliptic coordinates (λec, βec). The location of this patch is marked by the lowest red rectangle in [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: The full-sky map with both diffuse component and point sources in ecliptic coordinates at 10 MHz. The map is smoothed by a Gaussian filter with Full-Width-Half￾Maximum (FWHM)≈ 1.3 ◦ to match the resolution of our reconstructed map. The red boxes mark small regions where we take detailed study. 3. SIMULATION SETUP 3.1. Satellite array configuration and orbital motion The array of satellites rely on the orbi… view at source ↗
Figure 3
Figure 3. Figure 3: The geometric factor q0 as a function of the com￾pression ratio R for the satellites configuration. The blue, orange, and green lines correspond to the compression ratio between satellites 1 and 2 of R2 =3, 6, and 9, respectively [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: The integration time spent per unit baseline length as a function of baseline length b for different com￾pression ratios R2 and R of the ‘breathing’ of the satellite array, after one cycle of precession (1.3 years). The integration time spent per unit baseline length as a function of baseline length b after one cycle of preces￾sion (1.3 years) for different compression ratios R2 and R is shown in [PITH_FU… view at source ↗
Figure 6
Figure 6. Figure 6: The directions of incipient wave and reflected wave for two satellites. In the case where the height and distance between satellites are much smaller than the radius of the Moon, the images seen by the two satellites are nearly in the same direction, and the reflected wave geometric delay between the two satellites are close to that of the incipient wave. non-uniform reflections. These are beyond the scope… view at source ↗
Figure 7
Figure 7. Figure 7: The comparison of a row of the B matrix, computed with the pixel-center method (a) and that with the pixel￾averaging method (b). In each panel, the real part of fringe term exp −2πi ν c nˆβ · rα  is shown with Nside = 64. The baseline is assumed to be aligned with y-axis, with a length of ≈ 250.2λ, corresponding to b ≈ 4bp. The shading of Moon is not included in this plot. For a sky with a total pixel num… view at source ↗
Figure 8
Figure 8. Figure 8: Comparison of the real part of the mock visibility ℜ (BT) computed with a high resolution of Nside = 1024 (left panel), with that evaluated at a low resolution of Nside = 64 using the pixel-averaging method (middle panel) and the pixel-center method (right panel) method, respectively. The baseline length is b ≈ bp for Nside = 64. Nside = 64. Compared with the visibility generated in the high resolution cas… view at source ↗
Figure 9
Figure 9. Figure 9: The input and reconstructed maps at 10 MHz, with the regularization parameter ϵ = 10−4 and baseline combination b < 2bp. The upper panel shows the input sky map. The middle-left panel shows the reconstructed map without thermal noise, and the middle-right panel shows the reconstructed map with thermal noise. The lower-left panel shows the percentage relative error between the input and reconstructed map wi… view at source ↗
Figure 10
Figure 10. Figure 10: Same as [PITH_FULL_IMAGE:figures/full_fig_p013_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: The 10 MHz maps reconstructed with a non-zero prior map (upper panels), and the fractional errors (lower panels). The input map is the same as before. In each row, the left panel shows the result without thermal noise, and the right panel shows the result with thermal noise. these regions. In polar regions, the loss of large-scale power is alleviated as shown by the reconstructed map, but the relative err… view at source ↗
Figure 12
Figure 12. Figure 12: The reconstructed maps (upper panels), and the fractional error maps (lower panels) using a non-zero but incorrect prior map at 10 MHz. The input map is the same as before. In each row, the left panel shows the result without thermal noise, and the right panel shows the result with thermal noise. finite pixel size ∼ 1 ◦ of the reconstructed map, though this can be suppressed in post-processing. However, t… view at source ↗
Figure 13
Figure 13. Figure 13: Three 20 deg × 20 deg sky patches of the input and reconstructed sky maps at 10 MHz. The input map, reconstructed maps without thermal noise and with thermal noise are shown in left, middle and right panels, respectively. [PITH_FULL_IMAGE:figures/full_fig_p016_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Same as [PITH_FULL_IMAGE:figures/full_fig_p016_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: The variation of the estimators of the error from the imperfect beam T¯Tr(Σs) and that from the thermal noise Tr(Σn) for different ϵ at 10 MHz with baseline combination b < 2bp. of ϵ at 10 MHz with the baseline combination b < 2bp. Note that Tr(Σs) does not have the same dimension as Tr(Σn). Now T¯Tr(Σs) increases monotonically with in￾creasing ϵ, while Tr(Σn) is flat for small ϵ, and decrease significant… view at source ↗
Figure 16
Figure 16. Figure 16: Comparison of the correlation coefficient ρl at 10 MHz for the pixel-center (left) and pixel-averaging (right) methods, for different baseline combinations. The dashed vertical line marks l = 18, below which ρl drops [PITH_FULL_IMAGE:figures/full_fig_p018_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: The SNRl for the pixel-averaging method at 10 MHz with different baseline combinations. Here we set ϵ = 10−4 and the scale corresponding to l = 18 is indicated by the dashed vertical line. array configuration, the shortest projected baseline at the polar region is rproj,min = 100 m × cos 30◦ ≈ 2.9λ at 10 MHz, and the corresponding angular scale is l ∼ 2πrproj,min/λ ≈ 18, which corresponds exactly to the s… view at source ↗
Figure 18
Figure 18. Figure 18: The correlation coefficient ρl and SNRl as a function of l for different regularization parameters ϵ at 10 MHz. including longer baselines, the SNRl improves signifi￾cantly, especially at the high-l end. The results presented above show that the pixel￾averaging method is robust against sub-pixel variation [PITH_FULL_IMAGE:figures/full_fig_p018_18.png] view at source ↗
Figure 19
Figure 19. Figure 19: The 5σ point source sensitivity (left axis, blue line) estimated from the noise-only map and the standard deviation of the same map (right axis, orange line) for base￾line combination b < 2bp and ϵ = 10−3 after post-processing and removal of large-scale component. the standard deviation of the noise-only map. We con￾vert the result to the unit of Jy/beam using the esti￾mated beam size, and plot this by th… view at source ↗
Figure 20
Figure 20. Figure 20: Noise covariance induced by the sub-pixel noise Nb estimated using Eq. (A3) for the same baseline and time points as in [PITH_FULL_IMAGE:figures/full_fig_p021_20.png] view at source ↗

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

48 extracted references · 12 canonical work pages · 1 internal anchor

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter doi edition editor eprint howpublished institution journal key month number organization pages publisher school series title misctitle type volume year version url label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts ...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION format.url url empty "" new.block "" url * "" * if FUNCTION format.eprint eprint empty "" archivePrefix empty "" archivePrefix "arXiv" = new.block " " eprint * " " * new.block " " eprint * " " * if if if FUNCTION format.doi doi empty "" " " doi * " " * if FUNCTION format.pid doi empty eprint empty ur...

  3. [3]

    r o[eٲu! dND|<d lG 躮# B !

    thebibliography [1] 20pt to REFERENCES 6pt =0pt -12pt 10pt plus 3pt =0pt =0pt =1pt plus 1pt =0pt =0pt -12pt =13pt plus 1pt =20pt =13pt plus 1pt \@M =10000 =-1.0em =0pt =0pt 0pt =0pt =1.0em @enumiv\@empty 10000 10000 `\.\@m \@noitemerr \@latex@warning Empty `thebibliography' environment \@ifnextchar \@reference \@latexerr Missing key on reference command E...

    2021

  4. [4]

    2025, RAS Techniques and Instruments, 4, rzae061, 10.1093/rasti/rzae061

    Artuc , K., & de Lera Acedo , E. 2025, RAS Techniques and Instruments, 4, rzae061, 10.1093/rasti/rzae061

    doi:10.1093/rasti/rzae061 2025

  5. [5]

    D., Bassett , N., Burns , J

    Bale , S. D., Bassett , N., Burns , J. O., et al. 2023, arXiv e-prints, arXiv:2301.10345, 10.48550/arXiv.2301.10345

    Pith/arXiv arXiv doi:10.48550/arxiv.2301.10345 2023

  6. [6]

    2021, in 2021 IEEE Aerospace Conference (50100), 1--25, 10.1109/AERO50100.2021.9438165

    Bandyopadhyay, S., Mcgarey, P., Goel, A., et al. 2021, in 2021 IEEE Aerospace Conference (50100), 1--25, 10.1109/AERO50100.2021.9438165

    arXiv 2021

  7. [7]

    Burns , J. O. 2021, Philosophical Transactions of the Royal Society of London Series A, 379, 20190564, 10.1098/rsta.2019.0564

    arXiv 2021

  8. [8]

    I., et al

    Chen, L., Aminaei, A., Gurvits, L. I., et al. 2018, Experimental Astronomy, 45, 231

    2018

  9. [9]

    2020, Phil

    Chen, X., Yan, J., Deng, L., et al. 2020, Phil. Trans. Roy. Soc. Lond. A, 379, 20190566, 10.1098/rsta.2019.0566

    arXiv 2020

  10. [10]

    2019, arXiv e-prints, arXiv:1907.10853

    Chen , X., Burns , J., Koopmans , L., et al. 2019, arXiv e-prints, arXiv:1907.10853. 1907.10853

    Pith/arXiv arXiv 2019

  11. [11]

    2023, Chinese Journal of Space Science, 43, 43

    Chen, X., Yan, J., Xu, Y., et al. 2023, Chinese Journal of Space Science, 43, 43

    2023

  12. [12]

    2024, Philosophical Transactions of the Royal Society A, 382, 20230094

    Chen, X., Gao, F., Wu, F., et al. 2024, Philosophical Transactions of the Royal Society A, 382, 20230094

    2024

  13. [13]

    2024, arXiv preprint arXiv:2501.00431

    Cong, Y., Yue, B., Xu, Y., et al. 2024, arXiv preprint arXiv:2501.00431

    Pith/arXiv arXiv 2024

  14. [14]

    2021, , 914, 128, 10.3847/1538-4357/abf55c

    Cong , Y., Yue , B., Xu , Y., et al. 2021, , 914, 128, 10.3847/1538-4357/abf55c

    work page doi:10.3847/1538-4357/abf55c 2021

  15. [15]

    2022, , 940, 180, 10.3847/1538-4357/ac9df7

    Cong , Y., Yue , B., Xu , Y., Shi , Y., & Chen , X. 2022, , 940, 180, 10.3847/1538-4357/ac9df7

    work page doi:10.3847/1538-4357/ac9df7 2022

  16. [16]

    J., Golap , K., & Bhatnagar , S

    Cornwell , T. J., Golap , K., & Bhatnagar , S. 2008, IEEE Journal of Selected Topics in Signal Processing, 2, 647, 10.1109/JSTSP.2008.2005290

    arXiv 2008

  17. [17]

    J., & Perley , R

    Cornwell , T. J., & Perley , R. A. 1992, , 261, 353

    1992

  18. [18]

    J., Voronkov , M

    Cornwell , T. J., Voronkov , M. A., & Humphreys , B. 2012, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, Vol. 8500, Image Reconstruction from Incomplete Data VII, ed. P. J. Bones , M. A. Fiddy , & R. P. Millane , 85000L, 10.1117/12.929336

    work page doi:10.1117/12.929336 2012

  19. [19]

    M., et al

    de Oliveira-Costa, A., Tegmark, M., Gaensler, B. M., et al. 2008, Mon. Not. Roy. Astron. Soc., 388, 247, 10.1111/j.1365-2966.2008.13376.x

    arXiv 2008

  20. [20]

    W., Anderson , M

    Eastwood , M. W., Anderson , M. M., Monroe , R. M., et al. 2018, , 156, 32, 10.3847/1538-3881/aac721

    doi:10.3847/1538-3881/aac721 2018

  21. [21]

    Evans , J. V. 1969, , 7, 201, 10.1146/annurev.aa.07.090169.001221

    arXiv 1969

  22. [22]

    Franzen , T. M. O., Vernstrom , T., Jackson , C. A., et al. 2019, , 36, e004, 10.1017/pasa.2018.52

    doi:10.1017/pasa.2018.52 2019

  23. [23]

    H., Heath, M., & Wahba, G

    Golub, G. H., Heath, M., & Wahba, G. 1979, Technometrics, 21, 215. http://www.jstor.org/stable/1268518

    arXiv 1979

  24. [24]

    M., Hivon , E., Banday , A

    G \'o rski , K. M., Hivon , E., Banday , A. J., et al. 2005, , 622, 759, 10.1086/427976

    doi:10.1086/427976 2005

  25. [25]

    J., Burns , J

    Hibbard , J. J., Burns , J. O., MacDowall , R., et al. 2025, arXiv e-prints, arXiv:2503.09842, 10.48550/arXiv.2503.09842

    Pith/arXiv arXiv doi:10.48550/arxiv.2503.09842 2025

  26. [26]

    2018, , 156, 43, 10.3847/1538-3881/aac6c6

    Huang , Q., Sun , S., Zuo , S., et al. 2018, , 156, 43, 10.3847/1538-3881/aac6c6

    doi:10.3847/1538-3881/aac6c6 2018

  27. [27]

    2019, Science China Physics, Mechanics, and Astronomy, 62, 989511, 10.1007/s11433-018-9333-1

    Huang , Q., Wu , F., & Chen , X. 2019, Science China Physics, Mechanics, and Astronomy, 62, 989511, 10.1007/s11433-018-9333-1

    work page doi:10.1007/s11433-018-9333-1 2019

  28. [28]

    2024, in American Astronomical Society Meeting Abstracts, Vol

    Klein Wolt , M., Falcke , H., & Koopmans , L. 2024, in American Astronomical Society Meeting Abstracts, Vol. 243, American Astronomical Society Meeting Abstracts \#243, 264.01

    2024

  29. [29]

    A., Wayth , R

    Kriele , M. A., Wayth , R. B., Bentum , M. J., Juswardy , B., & Trott , C. M. 2022, , 39, e017, 10.1017/pasa.2022.2

    work page doi:10.1017/pasa.2022.2 2022

  30. [30]

    2023, Nature Astronomy, 7, 1025, 10.1038/s41550-023-02057-y

    Mondal , R., & Barkana , R. 2023, Nature Astronomy, 7, 1025, 10.1038/s41550-023-02057-y

    doi:10.1038/s41550-023-02057-y 2023

  31. [31]

    Plice , L., Galal , K., & Burns , J. O. 2017, arXiv e-prints, arXiv:1702.00286, 10.48550/arXiv.1702.00286

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1702.00286 2017

  32. [32]

    J., Bigot-Sazy , M

    Remazeilles , M., Dickinson , C., Banday , A. J., Bigot-Sazy , M. A., & Ghosh , T. 2015, , 451, 4311, 10.1093/mnras/stv1274

    doi:10.1093/mnras/stv1274 2015

  33. [33]

    Sathyanarayana Rao , M., Singh , S., K. S. , S., et al. 2023, Experimental Astronomy, 56, 741, 10.1007/s10686-023-09909-5

    doi:10.1007/s10686-023-09909-5 2023

  34. [34]

    R., Sigurdson , K., Pen , U.-L., Stebbins , A., & Sitwell , M

    Shaw , J. R., Sigurdson , K., Pen , U.-L., Stebbins , A., & Sitwell , M. 2014, , 781, 57, 10.1088/0004-637X/781/2/57

    doi:10.1088/0004-637x/781/2/57 2014

  35. [35]

    2022, Astrophys

    Shi, Y., Deng, F., Xu, Y., et al. 2022, Astrophys. J., 929, 32, 10.3847/1538-4357/ac5965

    work page doi:10.3847/1538-4357/ac5965 2022

  36. [36]

    2022, , 510, 3046, 10.1093/mnras/stab3623

    Shi , Y., Xu , Y., Deng , L., et al. 2022, , 510, 3046, 10.1093/mnras/stab3623

    doi:10.1093/mnras/stab3623 2022

  37. [37]

    M., Callingham , J

    Slob , M. M., Callingham , J. R., R \"o ttgering , H. J. A., et al. 2022, , 668, A186, 10.1051/0004-6361/202244651

    work page doi:10.1051/0004-6361/202244651 2022

  38. [38]

    A., & Datta, A

    Tanti, H. A., & Datta, A. 2021, in 2021 IEEE Indian Conference on Antennas and Propagation (InCAP), 312--315, 10.1109/InCAP52216.2021.9726500

    arXiv 2021

  39. [39]

    2021, in 43rd COSPAR Scientific Assembly

    Vecchio , A., Bentum , M., Falcke , H., et al. 2021, in 43rd COSPAR Scientific Assembly. Held 28 January - 4 February, Vol. 43, 1525

    2021

  40. [40]

    2024, Chinese journal of radio science, 39, 589

    Wu, F., Xu, j., Zhu, J., et al. 2024, Chinese journal of radio science, 39, 589

    2024

  41. [41]

    I., et al

    Yan , J., Wu , J., Gurvits , L. I., et al. 2023, Experimental Astronomy, 56, 333, 10.1007/s10686-022-09887-0

    work page doi:10.1007/s10686-022-09887-0 2023

  42. [42]

    2024, Res

    Yu, K., Zuo, S., Wu, F., Wang, Y., & Chen, X. 2024, Res. Astron. Astrophys., 24, 025002, 10.1088/1674-4527/ad1223

    work page doi:10.1088/1674-4527/ad1223 2024

  43. [43]

    2016, Mon

    Zhang, J., Ansari, R., Chen, X., et al. 2016, Mon. Not. Roy. Astron. Soc., 461, 1950, 10.1093/mnras/stw1458

    doi:10.1093/mnras/stw1458 2016

  44. [44]

    2012, Research in Astronomy and Astrophysics, 12, 1297, 10.1088/1674-4527/12/9/010

    Zhang , X.-Z., Gray , A., Su , Y., et al. 2012, Research in Astronomy and Astrophysics, 12, 1297, 10.1088/1674-4527/12/9/010

    work page doi:10.1088/1674-4527/12/9/010 2012

  45. [46]

    2017 b , , 465, 2901, 10.1093/mnras/stw2910

    ---. 2017 b , , 465, 2901, 10.1093/mnras/stw2910

    work page doi:10.1093/mnras/stw2910 2017

  46. [47]

    Zhu, J., Acedo, E. d. L., Artuc, K., & Chen, X. 2025, RAS Techniques and Instruments, 4, rzae064

    2025

  47. [48]

    2021, Research in Astronomy and Astrophysics, 21, 116, 10.1088/1674-4527/21/5/116

    Zhu , X.-Y., Su , Y., Ji , Y.-C., et al. 2021, Research in Astronomy and Astrophysics, 21, 116, 10.1088/1674-4527/21/5/116

    doi:10.1088/1674-4527/21/5/116 2021

  48. [49]

    2019, Journal of Open Source Software, 4, 1298, 10.21105/joss.01298

    Zonca, A., Singer, L., Lenz, D., et al. 2019, Journal of Open Source Software, 4, 1298, 10.21105/joss.01298

    doi:10.21105/joss.01298 2019