WindsoCC: reconstructing the wind-driven halo in MagAO-X images using wavefront sensor telemetry
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Astrophysics. The day in astrophysics: observations, models and what the data forces.
Vera: We're looking at the paper: "WindsoCC: reconstructing the wind-driven halo in MagAO-X images using wavefront sensor telemetry".
Jocelyn: Here's the gist of the paper. The windsoCC pipeline reconstructs the wind-driven halo in MagAO-X images using wavefront sensor telemetry to remove a persistent noise artifact that severely limits high-contrast imaging The Problem with the Wind-Driven Halo The wind-driven halo (WDH) is a persistent, low spatial frequency noise artifact that arises due to the servo-lag error inherent to all adaptive optics (AO) instruments.
Vera: Spatial filtering may be employed to overcome this artifact, however, filtering out the WDH while simultaneously preserving signal from an extended astrophysical object of interest is exceptionally challenging.
Jocelyn: Since wavefront sensors (WFSs) continuously track the closed-loop residual wavefront error, WFS telemetry presents the ideal resource for combating this type of noise artifact through postprocessing.
Vera: The WindsoCC Pipeline Architecture The windsoCC pipeline implements a pyramid WFS7 image processing pipeline, windsoCC, based on CC.
Title and authors: Vera: So let's talk about the title and the authors of this paper. "WindsoCC: reconstructing the wind-driven halo in MagAO-X images using wavefront sensor telemetry." It’s pretty descriptive, right?
Jocelyn: Yeah, it tells you exactly what they are doing—reconstructing that halo artifact in MagAO-X images by using the wavefront sensor telemetry. The authors are a large team from places like Arizona and Leiden.
Subrahmanyan: It’s interesting seeing how they tie together different parts of the system—the AO control loop, the image processing, and then adding this new reconstruction step based on the wavefront data. That integration is key to how they tackle dynamic noise.
Vera: And what this means for us is that instead of just trying to subtract noise after you have a final image, they are building a pipeline that fixes it at an earlier stage by using the continuous error information from the wavefront sensor.
Jocelyn: So, essentially, they’re saying the way we look at this problem needs to change from post-processing filters to real-time use of telemetry for artifact removal.
Subrahmanyan: I wonder if this approach scales well across different telescope setups; the specific details about MagAO-X are important because it sets a benchmark for how this can be applied elsewhere in high contrast imaging.
Vera: Right, the paper goes on to show that they can extract information about atmospheric wind layers and then use that to model and remove the halo components.
The paper's summary: Jocelyn: The summary of this paper explains their methodology: they use a pyramid wavefront sensor processing pipeline, which they call windsoCC, based on CC operations. It’s a multi-stage process that takes the raw WFS data and turns it into something that helps them identify these wind layers.
Vera: They start by partitioning the image data into smaller datasets over about twelve seconds of integration time, then they normalize everything, and then they apply a cross-correlation operation to each pupil quadrant folder separately using a segmented, overlapping method.
Subrahmanyan: That cross-correlation is where they get the wind vectors; by analyzing moving peaks in those response maps, they can measure the speed and direction of different atmospheric turbulence layers.
Jocelyn: And then they use something called Hierarchical Density-Based Spatial Clustering of Applications with Noise, or HDBSCAN, on that output to classify these discrete layers of atmospheric turbulence over time.
Vera: That classification step is crucial because it helps them distinguish between just random noise and actual physical layers of the atmosphere that are causing the wind-driven halo.
Subrahmanyan: It also mentions they cross-reference these extracted wind vectors with historical atmospheric properties from the Copernicus ERA5 reanalysis dataset to see if their findings align with what we already knew about atmospheric profiles.
Jocelyn: So, the summary boils down to them using structured processing on WFS data to extract wind vectors, classifying those layers, and then using a forward model to reconstruct and subtract the halo footprint.
The paper's improvements: Vera: The paper highlights a few key improvements they made over previous methods. They suggest that instead of just applying some kind of spatial filter, which we know is tricky because it might remove real smooth features, this telemetry-based approach allows them to scrub the halo components directly from the individual science frames.
Jocelyn: That’s the big win for object recovery; they can keep those extended signals from circumstellar disks in the AO control region without having to use overly aggressive spatial filtering or PSF subtraction.
Subrahmanyan: I think that's where the real theoretical payoff is, because if you can preserve the signal while removing this dynamic noise, you get much better data for studying structure formation.
Vera: And they do this by using a custom forward-modeling framework built with a Markov Chain Monte-Carlo, or MCMC, framework. They use a 2D analytical model that mimics the shape of the wind-driven halo, including things like a Gaussian envelope along the wind direction and parameters for radial flaring <ref:2608.10302#pg2>.
Jocelyn: The paper says they fit combinations of four and five separate WDH models to match the different clusters they retrieved from their pipeline. After converging, they iterate through these individual model components to scale and subtract the artifacts frame by frame.
Subrahmanyan: That iterative fitting process sounds like a very rigorous way to handle the complexity of multi-layered turbulence, but I'm curious how computationally demanding this is for real-time application on a telescope.
Conclusion: Vera: So to wrap up, the main point of "WindsoCC: reconstructing the wind-driven halo in MagAO-X images using wavefront sensor telemetry" is that using closed-loop WFS telemetry allows them to directly scrub the wind-driven halo components from science frames before final processing.
Jocelyn: It’s a powerful technique because it keeps the extended signals, like those from disks, preserved in the AO control region without relying on overly aggressive spatial filtering or PSF subtraction.
Subrahmanyan: From a cosmological standpoint, this suggests that if we can effectively remove dynamic noise artifacts in these high-contrast images, our measurements of structure and density will be much cleaner.
Vera: And the final outputs are really useful; they produce a human-readable text file that contains information on each detected wind layer and the matched matched filter response cubes.
Jocelyn: It really emphasizes that closed-loop WFS telemetry is an incredibly valuable asset for high contrast imaging, and it’s a solid piece of postprocessing work.
Subrahmanyan: It’s important to remember that the limitations they mention are that their forward model is based on a simplified 2D analytical model, so it might not capture every single nuance of the physical interaction perfectly <ref:2608.10302#pg2>.
Vera: That's fair; they admit that their forward model is a simplification, which means it’s not a perfect physical simulation but it works well enough for this application.
Jocelyn: So this gives us a clear path forward: use that telemetry data to guide the reconstruction and removal process, and keep an eye on how they apply these ideas to other telescope systems.
Steward Observatory, University of Arizona · James C Wyant College of Optical Sciences, University of Arizona · Center for Computational Astrophysics, Flatiron Institute · Earth and Planets Laboratory, Carnegie Institution for Science · Leiden Observatory, Leiden University · Subaru Telescope National Observatory of Japan Department of Astronomy University of Michigan Northrop Grumman Corporation Draper Laboratory Northrop Grumman Starfire Optical Range
astro-ph.IM
Submitted: 2026-08-10
Updated: 2026-11-06
Comments: 9 pages, 4 figures, 1 table, published in the Proceedings of SPIE Astronomical Telescopes + Instrumentation 2026
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 83/100
The gist: The windsoCC pipeline reconstructs the wind-driven halo in MagAO-X images using wavefront sensor telemetry to remove a persistent noise artifact that severely limits high-contrast imaging The Problem…
Key concepts
- Wind-Driven Halo (WDH)
- This is a persistent, low spatial frequency noise artifact caused by servo-lag errors in adaptive optics instruments. It severely limits high-contrast imaging and makes it difficult to preserve the signal from extended objects like disks.
- Cross-Correlation (CC) Operation
- The CC operation is applied to telescope pupil images within closed-loop wavefront sensor data. This mathematical technique measures the correlation between different parts of the image, which helps extract the speed and direction of moving layers of atmospheric wind.
- Forward Modeling Framework
- This is a custom framework, including a Markov Chain Monte-Carlo (MCMC) approach, used to mimic the physical shape of the WDH. It allows researchers to fit combinations of individual WDH models to science data, enabling the iterative scaling and subtraction of artifacts.
- Wavefront Sensor Telemetry
- This refers to the continuous data stream from wavefront sensors that tracks residual wavefront errors in real-time. This telemetry is crucial because it provides the ideal resource for postprocessing to combat noise artifacts like the WDH.
Terminology
Summary
The windsoCC pipeline reconstructs the wind-driven halo in MagAO-X images using wavefront sensor telemetry to remove a persistent noise artifact that severely limits high-contrast imaging
The Problem with the Wind-Driven Halo
The wind-driven halo (WDH) is a persistent, low spatial frequency noise artifact that arises due to the servo-lag error inherent to all adaptive optics (AO) instruments. Spatial filtering may be employed to overcome this artifact, however, filtering out the WDH while simultaneously preserving signal from an extended astrophysical object of interest is exceptionally challenging. p. 0 ,
Tucson, AZ 85721, USA
cNational Science Foundation Graduate Research Fellow
dCenter for Computational Astrophysics, Flatiron Institute, 162 5th Ave, New York, NY
eEarth and Planets Laboratory, Carnegie Institution for Science, 5241 Broad Branch Road
NW, Washington, DC 20015-1305
fLeiden Observatory, Leiden University, PO Box 9513, 2300 RA Leiden, The Netherlands
gSubaru Telescope, National Observatory of Japan, Hilo, HI
hDepartment of Astronomy, University of Michigan, Ann Arbor, MI
iNorthrop Grumman Corporation, 600 South Hicks Rd, Rolling Meadows, IL
jDraper Laboratory, 555 Technology Square, Cambridge, MA
kNorthrop Grumman, Pasadena, CA
lStarfire Optical Range, Kirtland Air Force Base, Albuquerque, NM
ABSTRACT
The wind-driven halo (WDH) is a persistent, low spatial frequency noise artifact that arises due to the servo-lag
error inherent to all adaptive optics (AO) instruments. Spatial filtering may be employed to overcome this
artifact, however, filtering out the WDH while simultaneously preserving signal from an extended astrophysical
object of interest is exceptionally challenging. ,
Tucson, AZ 85721, USA
cNational Science Foundation Graduate Research Fellow
dCenter for Computational Astrophysics, Flatiron Institute, 162 5th Ave, New York, NY
eEarth and Planets Laboratory, Carnegie Institution for Science, 5241 Broad Branch Road
NW, Washington, DC 20015-1305
fLeiden Observatory, Leiden University, PO Box 9513, 2300 RA Leiden, The Netherlands
gSubaru Telescope, National Observatory of Japan, Hilo, HI
hDepartment of Astronomy, University of Michigan, Ann Arbor, MI
iNorthrop Grumman Corporation, 600 South Hicks Rd, Rolling Meadows, IL
jDraper Laboratory, 555 Technology Square, Cambridge, MA
kNorthrop Grumman, Pasadena, CA
lStarfire Optical Range, Kirtland Air Force Base, Albuquerque, NM
ABSTRACT
The wind-driven halo (WDH) is a persistent, low spatial frequency noise artifact that arises due to the servo-lag
error inherent to all adaptive optics (AO) instruments. Spatial filtering may be employed to overcome this
artifact, however, filtering out the WDH while simultaneously preserving signal from an extended astrophysical
object of interest is exceptionally challenging. , KLIP) that are commonly used to subtract the
starlight need to be overly-aggressive to remove both the static and dynamic noise components. Since wavefront
sensors (WFSs) continuously track the closed-loop residual wavefront error, WFS telemetry presents the ideal
resource for combating this type of noise artifact through postprocessing. , KLIP) that are commonly used to subtract the
starlight need to be overly-aggressive to remove both the static and dynamic noise components. Since wavefront
sensors (WFSs) continuously track the closed-loop residual wavefront error, WFS telemetry presents the ideal
resource for combating this type of noise artifact through postprocessing.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
The WindsoCC Pipeline Architecture
The windsoCC pipeline implements a pyramid WFS7 image processing pipeline, windsoCC, based on CC. The basic premise is that under the assumption of frozen-flow, the telescope pupil in the WFS images contain translating patterns due to layers of wind. p. 2 WINDSOCC PIPELINE DESCRIPTION
In order to gauge how different layers of atmospheric turbulence affect MagAO-X’s performance, we implemented
a pyramid WFS7
image processing pipeline, windsoCC, based on CC. The basic premise is that under the
assumption of frozen-flow, the telescope pupil in the WFS images contain translating patterns due to layers of
wind. discrete stages that we illustrate using a flowchart in Figure 1. A high-level overview of these various pipeline
stages continues below.Where the paper says it
Where the paper says it
-
Partitioning the total WFS image dataset into mini datasets spanning 12 seconds of integration time. p. 3
Where the paper says it
6kHz,8 producing a large number of image data FITS files over the course of a science observation. As a first step, windsoCC partitions the total WFS image dataset (typically encompassing hours of integration time) into mini datasets spanning 12 seconds of integration time.
-
Normalizing the images by subtracting off a reference and dividing by an estimate of the noise. p. 3
Where the paper says it
As a first step, windsoCC partitions the total WFS image dataset (typically encompassing hours of integration time) into mini datasets spanning 12 seconds of integration time. For each of these mini datasets, the pipeline normalizes the images by subtracting off a reference and dividing by an estimate of the noise; the former is made by the averaging and the latter is made by computing the standard deviation for each pixel using all of the images in the mini dataset.
-
Applying the CC to each pupil quadrant folder separately using a segmented, overlapping CC method. p. 3
Where the paper says it
The next step is applying the CC to each pupil quadrant folder separately. To maximize the signal-to-noise of the CC peaks due to turbulent wind, windsoCC uses a segmented, overlapping CC method, conceptually analogous to the variance-reduction techniques used in Welch’s method for power spectral density estimation.
-
Temporally averaging the response cubes for all overlapping segments within the 12-second block. p. 3
Where the paper says it
The CC was computed independently for each 2-second segment, generating a localized response cube. Finally, the response cubes for all overlapping segments within the 12-second block were temporally averaged.
-
Applying unsharp masking and matched filtering (MF) to increase the signal-to-noise ratio of the CC peaks due to the wind. p. 3
Where the paper says it
This “distill” stage then yields a single time series of length ∆tmax of CC response maps. To further increase the signal-to-noise ratio of the CC peaks due to the wind, windsoCC applies unsharp masking and matched filtering (MF).
-
Detecting and tracking peaks in the CC response maps using an annular “tripwire” region. p. 3
Where the paper says it
Finally, the last stage of the pipeline involves detecting and tracking the peaks in the CC response maps that are due to the layers of turbulence. To do this, windsoCC defines an annular “tripwire” region around the central autocorrelation peak whose width is configurable where SEP will search for peaks and ignore the rest of the image.
Wind Vector Extraction and Classification
The pipeline extracts wind vectors by applying the cross-correlation (CC) operation to the telescope pupil images in the closed-loop WFS data. This process allows for measuring a given wind layer’s speed and direction by analyzing moving peaks within the resulting CC response maps. p. 1 Namely,
by applying the CC of a given WFS pupil image with itself across a range of temporal delays, turbulent layers
would manifest as moving peaks within the resulting CC response maps (e. These moving peaks
encode the turbulent layer’s speed and direction. The resulting files form a
time-resolved lookup table of exceptionally-turbulent wind layers during the observation. RETRIEVED WIND VECTORS
We utilized the Hierarchical Density-Based Spatial Clustering of Applications with Noise (HDBSCAN) algorithm
on the windsoCC output to classify discrete layers of atmospheric turbulence over the course of our observations. Cross-referencing our empirical
measurements with historical weather data from the Copernicus ERA5 reanalysis dataset10 confirmed that many
of the extracted wind vectors aligned with the multi-layer atmospheric profile above Las Campanas Observatory
at the time of observation. We show the historical wind properties at various altitudes as framed numbers
overlaid on the scatter plot area in Figure 2.Where the paper says it
Where the paper says it
Where the paper says it
Forward Modeling and Artifact Removal
To reconstruct the WDH footprint, a custom forward-modeling framework is implemented using a Markov Chain Monte-Carlo (MCMC) framework. A 2D analytical model mimics the morphology of the WDH, incorporating a Gaussian envelope along the wind direction and parameters for radial flaring. p. 5 We used the DiskFM tool11 (part of the pyKLIP package12) to handle
the forward modeling through the KLIP algorithm. We implemented a 2D analytical model to mimic the morphology of the WDH, incorporating a Gaussian
envelope along the wind direction, a pinch parameter controlling the opening angle, and a parameter to govern
radial flaring:
I(x, y) = N0 exp "
−
1
2
r
p
γ
0
2
#
exp −
1
2
x
σ+
2
; x > 0
exp −
1
2
x
σ−
2
; x < 0,
(1)
where r
2 = x
2 + y
2
, p0 is the pinch parameter which controls the opening angle of the lobes, σ+ and σ−
control the brightness of individual lobes to fit asymmetries, γ controls the radial flaring of the lobes, and N0
is a scale factor. 13 datasets, we fit combinations of 4 and 5 separate WDH models, respectively, to match the distinct clusters
retrieved by our pipeline. We determined that the walkers had converged after the number of iterations had
exceeded 50 times the autocorrelation length of the walker chains, which was achieved after testing 213k and
746k models for the UT 2023 Mar 10 and UT 2023 Mar 13 datasets, respectively. These best-fit components were compiled into a modal basis alongside a mediancollapsed PSF estimate. Utilizing a non-negative least-squares fitting routine, we evaluated this basis against
each individual coronagraphic frame to iteratively scale and subtract the WDH artifacts.Where the paper says it
Where the paper says it
Where the paper says it
Results and Conclusion
The methodology demonstrated dramatic improvements in object recovery by scrubbing the WDH components directly from the individual science frames before final processing. This approach allows extended object signals to remain preserved in the AO control region without relying on overly aggressive spatial filtering or PSF subtraction. p. 7 By scrubbing the WDH components directly from the
individual science frames before final processing, we retrieved clear images of the HR 4796A disk at r
′
-band
(615 nm) and i
′
-band (762 nm) without relying on overly aggressive spatial filtering or PSF subtraction. This
methodology allows extended object signals to remain preserved in the AO control region, establishing closed-loop
WFS telemetry as an incredibly valuable asset for high-contrast imaging. By scrubbing the WDH components directly from the
individual science frames before final processing, we retrieved clear images of the HR 4796A disk at r
′
-band
(615 nm) and i
′
-band (762 nm) without relying on overly aggressive spatial filtering or PSF subtraction. This
methodology allows extended object signals to remain preserved in the AO control region, establishing closed-loop
WFS telemetry as an incredibly valuable asset for high-contrast imaging. We use
standard ANSI flowchart shapes to denote processing steps, manual operations, data products, etc. The final outputs
include a human-readable text file that contains the information on each detected wind layer and the MF response cubes.Where the paper says it
Where the paper says it
Where the paper says it
Pruning redundant reasoning tokens
The windsoCC pipeline reconstructs the wind-driven halo in MagAO-X images using wavefront sensor telemetry to remove a persistent noise artifact that severely limits high-contrast imaging. The WDH is a dynamic and high-contrast noise artifact, which makes it difficult to remove through postprocessing. p. 1 The wind-driven halo (WDH) is a pervasive source of starlight leakage within the AO control region and
can severely limit contrast. It is a dynamic and high-contrast noise artifact,2 which makes it difficult to remove
through post-processing. Using archival WFS telemetry from
MagAO-X, which is the “extreme” AO instrument for the 6. 5-meter Magellan-Clay telescope, we demonstrate a
novel workflow for WDH reconstruction and removal in individual coronagraphic science images.Where the paper says it
Where the paper says it
How the training signal is built
The cross-correlation (CC) operation on closed-loop WFS data is used to measure the wind vectors of several turbulent layers of the atmosphere above Las Campanas Observatory. Applying the CC of a given WFS pupil image with itself across a range of temporal delays can extract wind vectors for each turbulent layer. p. 1 In this scenario, and specifically for a pyramid WFS, applying the cross-correlation (CC) operation to the
telescope pupil images in the closed-loop WFS data can extract wind vectors for each turbulent layer. Namely,
by applying the CC of a given WFS pupil image with itself across a range of temporal delays, turbulent layers
would manifest as moving peaks within the resulting CC response maps (e. Namely,
by applying the CC of a given WFS pupil image with itself across a range of temporal delays, turbulent layers
would manifest as moving peaks within the resulting CC response maps (e. These moving peaks
encode the turbulent layer’s speed and direction.Where the paper says it
Where the paper says it
Where the gains come from
The dramatic improvement in object recovery is achieved by scrubbing the WDH components directly from individual science frames before final processing. This technique allows extended object signals to remain preserved in the AO control region, which establishes closed-loop WFS telemetry as an incredibly valuable asset for high-contrast imaging. p. 7 By scrubbing the WDH components directly from the
individual science frames before final processing, we retrieved clear images of the HR 4796A disk at r
′
-band
(615 nm) and i
′
-band (762 nm) without relying on overly aggressive spatial filtering or PSF subtraction. This
methodology allows extended object signals to remain preserved in the AO control region, establishing closed-loop
WFS telemetry as an incredibly valuable asset for high-contrast imaging. Our results demonstrated dramatic improvements. By scrubbing the WDH components directly from the
individual science frames before final processing, we retrieved clear images of the HR 4796A disk at r
′
-band
(615 nm) and i
′
-band (762 nm) without relying on overly aggressive spatial filtering or PSF subtraction.Where the paper says it
Where the paper says it
Summary
In summary, we built the automated windsoCC pipeline leveraging the CC operation on closed-loop pyramid WFS telemetry to empirically map the atmospheric wind vectors throughout MagAO-X operations. By bridging this telemetry with a custom forward-modeling framework (using DiskFM), we successfully parameterized and removed complex, multi-layered WDH artifacts from MagAO-X focal plane images. p. 7 SUMMARY
In summary, we built the automated windsoCC pipeline leveraging the CC operation on closed-loop pyramid
WFS telemetry to empirically map the atmospheric wind vectors throughout MagAO-X operations. By bridging
this telemetry with a custom forward-modeling framework (using DiskFM), we successfully parameterized and
removed complex, multi-layered WDH artifacts from MagAO-X focal plane images. By bridging
this telemetry with a custom forward-modeling framework (using DiskFM), we successfully parameterized and
removed complex, multi-layered WDH artifacts from MagAO-X focal plane images. Ultimately, this allowed for
cleaner recovery of the HR 4796A debris disk and highlights the importance of retaining telemetry for advanced
post-processing techniques. The wind-driven halo (WDH) is a pervasive source of starlight leakage within the AO control region and
can severely limit contrast. It is a dynamic and high-contrast noise artifact,2 which makes it difficult to remove
through post-processing. RETRIEVED WIND VECTORS
We utilized the Hierarchical Density-Based Spatial Clustering of Applications with Noise (HDBSCAN) algorithm
on the windsoCC output to classify discrete layers of atmospheric turbulence over the course of our observations. We show the results for the dataset collected on UT Mar 12 2023 in Figure 2. Cross-referencing our empirical
measurements with historical weather data from the Copernicus ERA5 reanalysis dataset10 confirmed that many
of the extracted wind vectors aligned with the multi-layer atmospheric profile above Las Campanas Observatory
at the time of observation. We show the historical wind properties at various altitudes as framed numbers
overlaid on the scatter plot area in Figure 2. WDH forward modeling results using the KLIP-reduced image of HR 4796A from the observations collected
on UT 2023 Mar 12. Top left: The best-fitting model consisting of a sum of 4 individual WDH models. Our results demonstrated dramatic improvements. By scrubbing the WDH components directly from the
individual science frames before final processing, we retrieved clear images of the HR 4796A disk at r
′
-band
(615 nm) and i
′
-band (762 nm) without relying on overly aggressive spatial filtering or PSF subtraction. By bridging
this telemetry with a custom forward-modeling framework (using DiskFM), we successfully parameterized and
removed complex, multi-layered WDH artifacts from MagAO-X focal plane images. Ultimately, this allowed for
cleaner recovery of the HR 4796A debris disk and highlights the importance of retaining telemetry for advanced
post-processing techniques.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
The wind-driven halo (WDH) is a pervasive source of starlight leakage within the AO control region and can severely limit contrast. It is a dynamic and high-contrast noise artifact, which makes it difficult to remove through postprocessing. p. 1 The wind-driven halo (WDH) is a pervasive source of starlight leakage within the AO control region and
can severely limit contrast. It is a dynamic and high-contrast noise artifact,2 which makes it difficult to remove
through post-processing. It is a dynamic and high-contrast noise artifact,2 which makes it difficult to remove
through post-processing. The WDH is often the most limiting source of noise in high-contrast images, especially images of extended objects like circumstellar disks.Where the paper says it
Where the paper says it
Under Taylor’s “frozen-flow” hypothesis, layers of wind within the telescope’s line-of-sight will appear as 2D phase screens undergoing translation across the pupil. Specifically for a pyramid WFS, applying the cross-correlation (CC) operation to the telescope pupil images in the closed-loop WFS data can extract wind vectors for each turbulent layer. p. 1 Under Taylor’s “frozen-flow” hypothesis,3
layers
of wind within the telescope’s line-of-sight will appear as 2D phase screens undergoing translation across the
pupil. In this scenario, and specifically for a pyramid WFS, applying the cross-correlation (CC) operation to the
telescope pupil images in the closed-loop WFS data can extract wind vectors for each turbulent layer.Where the paper says it
The authors utilize on-sky MagAO-X images of the disk around HR 4796A at visible wavelengths to demonstrate a dramatic improvement in object recovery. The final methodology allows extended object signals to remain preserved in the AO control region, establishing closed-loop WFS telemetry as an incredibly valuable asset for high-contrast imaging. p. 7 By scrubbing the WDH components directly from the
individual science frames before final processing, we retrieved clear images of the HR 4796A disk at r
′
-band
(615 nm) and i
′
-band (762 nm) without relying on overly aggressive spatial filtering or PSF subtraction. This
methodology allows extended object signals to remain preserved in the AO control region, establishing closed-loop
WFS telemetry as an incredibly valuable asset for high-contrast imaging.Where the paper says it
The WDH is a dynamic and high-contrast noise artifact, which makes it difficult to remove through postprocessing. Spatial filtering may be employed to overcome this artifact, however, filtering out the WDH while simultaneously preserving signal from an extended astrophysical object of interest is exceptionally challenging. The WDH changes in intensity and position angle through an observation, data-driven algorithms (e. p. 0 Spatial filtering may be employed to overcome this
artifact, however, filtering out the WDH while simultaneously preserving signal from an extended astrophysical
object of interest is exceptionally challenging. Additionally, since the WDH changes in intensity and position
angle through an observation, data-driven algorithms (e. Additionally, since the WDH changes in intensity and position
angle through an observation, data-driven algorithms (e. , KLIP) that are commonly used to subtract the
starlight need to be overly-aggressive to remove both the static and dynamic noise components.Where the paper says it
Where the paper says it
The WFS telemetry provides the ideal resource for combating this type of noise artifact through postprocessing. Using WFS data processing, windsoCC cross-correlates the recorded closed-loop wavefront to measure the wind vectors of several turbulent layers of the atmosphere. p. 0 MagAO-X is
equipped with a pyramid WFS capable of recording wavefront telemetry at a high-cadence which is saved during
data acquisition. Given this, we detail how our WFS data processing pipeline, windsoCC, cross-correlates the
recorded closed-loop wavefront to measure the wind vectors of several turbulent layers of the atmosphere above
Las Campanas Observatory.Where the paper says it
The pipeline uses a segmented, overlapping CC method to maximize the signal-to-noise of the CC peaks due to turbulent wind. The final detection stage rejects noise by defining an annular “tripwire” region where all peaks due to a wind layer will traverse linearly, radially-outward, and at a constant velocity. p. 3 To do this, windsoCC defines an annular “tripwire” region around the
central autocorrelation peak whose width is configurable where SEP will search for peaks and ignore the rest of
the image. The concept of the tripwire search region was motivated by the idea that all peaks due to a wind
layer will traverse this region going linearly, radially-outward, and at a constant velocity. For every matched-filter response cube, the pipeline reports the wind layer track
identifier, wind speed, propagation direction, and the peak’s average area in pixels. The resulting files form a
time-resolved lookup table of exceptionally-turbulent wind layers during the observation.Where the paper says it
Where the paper says it
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Automated wind reconstruction via windsoCC pipeline allows AI systems to empirically map atmospheric wind vectors throughout MagAO-X operations by applying cross-correlation (CC) to closed-loop wavefront sensor telemetry. This enables the system to measure
the wind vectors of several turbulent layers of the atmosphere
and provides this information for reconstructing the WDH footprint. -
Integration of windsoCC output with a custom forward modeling framework like DiskFM enables AI systems to parameterize and remove complex, multi-layered WDH artifacts from science images without relying on
overly aggressive spatial filtering or PSF subtraction.
This allows extended object signals to remain preserved in the AO control region while achieving dramatic improvements in object recovery. -
The system can utilize the extracted wind vectors to inform parameter priors for forward modeling, such as the
number of simultaneous WDH models and their expected position angles.
This allows for a more accurate simulation of the WDH, leading to better performance when fitting combinations of individual WDH models to match observed clusters in KLIP-reduced data.
Abstract
The wind-driven halo (WDH) is a persistent, low spatial frequency noise artifact that arises due to the servo-lag error inherent to all adaptive optics (AO) instruments. Spatial filtering may be employed to overcome this artifact, however, filtering out the WDH while simultaneously preserving signal from an extended astrophysical object of interest is exceptionally challenging. Additionally, since the WDH changes in intensity and position angle through an observation, data-driven algorithms (e.g., KLIP) that are commonly used to subtract the starlight need to be overly-aggressive to remove both the static and dynamic noise components. Since wavefront sensors (WFSs) continuously track the closed-loop residual wavefront error, WFS telemetry presents the ideal resource for combating this type of noise artifact through postprocessing. Using archival WFS telemetry from MagAO-X, which is the ``extreme" AO instrument for the 6.5-meter Magellan-Clay telescope, we demonstrate a novel workflow for WDH reconstruction and removal in individual coronagraphic science images. MagAO-X is equipped with a pyramid WFS capable of recording wavefront telemetry at a high-cadence which is saved during data acquisition. Given this, we detail how our WFS data processing pipeline, windsoCC, cross-correlates the recorded closed-loop wavefront to measure the wind vectors of several turbulent layers of the atmosphere above Las Campanas Observatory. We then make use of the wind parameters learned through windsoCC to reconstruct the WDH footprint by leveraging a parametric model. Notably, we demonstrate a dramatic improvement in object recovery using on-sky MagAO-X images of the disk around HR 4796A at visible wavelengths.
Sources
Related papers
- On-sky demonstration of a vector Zernike wavefront sensor in a cascaded adaptive optics system
- Euclid Data Release 1 (DR1): Calibration strategy and observations
- The General Antiparticle Spectrometer (GAPS) Antarctic Balloon Payload
- Detection and excision of radio-frequency interference in radio arrays
- Unified matched-filter and excess-power pipeline for gravitational-wave detector characterization
- Ushering in sub-arcsecond angular resolution observatories in the far-infrared: The case for a double-Fourier interferometer