lam25 lam2 lam1 lam3 lam4 lam5
Accueil | Annuaire | English | Intranet | Webmail | Dircom |
Accueil > Recherche > Groupe R&D Optique & Instrumentation > Workshops > Past Workshops > AO data processing workshop

 AO data processing workshop

We are organizing a 2-days workshop around "data processing for AO images", at Marseille (LAM), the 26 and 27 of May. More information can be found below :



Contacts :
Benoit NEICHEL : +33 (0) 491 056 993
Thierry FUSCO : +33 (0) 491 055 923


Rationale

In order to get the best science results out of the AO-system images, and to fully optimize the return of such complex systems, dedicated and optimized reduction tools are needed. These tools must be built with a deep understanding of the system performance and limitations. These tools must be adapted to each science objectives. Toward these goals, the objective of the workshop is to gather both astronomers, AO specialist and data handling experts, in order to discuss new data reduction and analysis tools for current and future AO systems.

Those topics will be addressed in the light of the data obtained on 8m telescopes, including the newly commissioned XAO instruments (SPHERE and GPI), and WFAO instruments (GeMS, SAM, CANARY, LINC-NIRVANA, ARGOS). Lessons learned should also feed the preparation and optimization of the future instruments, for the AOF and for the ELTs.

For each topic, the following questions will be addressed :

  • What has been quantitatively achieved in terms of astrometry, photometry, kinematics, morphology, contrast ?
  • What data processing techniques are being used ?
  • Where do the data processing techniques have problems or need to be extended ?
  • What instrument characterization is needed to support data processing ?
  • Where could the community benefit from collaborations ?


Participants

Participants
NameInstituteTopic
Eric Aristidi OCA (Nice) Post-AO bispectral speckle imaging
Joana Ascenso University of Porto (Portugal) PSF reconstruction for AO photometry and astrometry pdf-link
Nicola Baccichet UCL astronomical instrumentation
Anais Bernard LAM (Marseille) Distortion correction in WFAO pdf-link
Nicolas Bouché IRAP (Toulouse) GalPak3D a recent tool for 3D spectroscopy analysis pdf-link
Amandine Caillat LAM Sparse Aperture imaging
Faustine Cantalloube ONERA / IPAG Image processing to detect and characterize exoplanet : the ANDROMEDA method pdf-link
Marcel Carbillet OCA (Nice) Post-AO bispectral speckle imaging
Valentin Christiaens Université de Liège (Belgium) / Univ. Chile High Contrast
Carlos Correia LAM Simulation-model-based PSF reconstruction for LGS-assisted wide-field Adaptive Optics pdf-link
Mirko Curti INAF (Firenze) LBT - ARGOS
Benoit Epinat LAM (Marseille) 3D spectroscopy - IFU - tools
Kacem El Hadi LAM (Marseille) Pyramid WFS
Giuliana Fiorentino INAF (Bologne) Nailing down the Main Sequence Knee method, the case of M5 pdf-link
Olivier Fauvarque LAM (Marseille) Pyramid WFS
Thierry Fusco ONERA/LAM Myopic deconvlution of AO corrected images using MISTRAL algorithm. Application to SPHERE data pdf-link
Carlos Gomez Université de Liège (Belgium) VIP, Software package for angular and spectral differential high-contrast imaging
Damien Gratadour Obs. Paris (LESIA) New approach for PSF reconstruction based on a maximum likelihood estimator
Morgan Gray LAM Adaptive Optics control
Andres Guesalaga Catolica University (Chile) & IMéRA Résident Profiling / Wide Field AO
Andrea La Camera University of Genoa, Italy Reconstruction of Fizeau images of Io volcanoes obtained with the LBTI/LMIRcam pdf-link
Nicolas Longeard OCA (Nice) Post-AO bispectral speckle imaging pdf-link
Davide Massari INAF (Bologne) Deep AO-photometry of crowded Galactic globular clusters pdf-link
Laurent Mugnier ONERA (Paris) SPHERE high contrast
Benoit Neichel LAM (Marseille) Astrometry with WFAO pdf-link
Sylvain Oberti ESO Adaptive optics Facility - EELT
Gilles Orban de Xivry MPE Commissioning progresses & first on-sky closed-loop of the GLAO system ARGOS pdf-link
Laurent Pueyo STSCI Post processing for the direct imaging : characterizing noise properties and astrophysical sources in modern ground based coronagraphs
Kalyan Radhakrishnan MPIA / LBT Adaptive Optics System of LINC­/NIRVANA
Francois Rigaut ANU (Australia) Wide Field adaptive Optics
Daniel Rouan Observatoire de Paris - LESIA General introduction to the needs in data reduction for AO images pdf-link
Gérard Rousset Observatoire de Paris - LESIA PSF reconstruction
Graeme Salter LAM Imaging and Characterisation with SPHERE in the NIR pdf-link
Jean-Francois Sauvage ONERA SPHERE
Laura Schreiber INAF (Bologne) How to handle a variable PSF in AO data pdf-link
Manash Samal LAM (Marseille) Star forming regions observed with AO
Manuel Silva University of Porto (Portugal) PSF Reconstruction pdf-link
Peter Wizinowich Keck (Hawaii) Needs in PSF reconsctruction pdf-link
Hussein Yahia INRIA Data processing tools pdf-link


Program

Program - Day1 - Room LAVANDE (1st Floor)
SlotTitleName
09h30 - 10h Cofee
10h Introduction Neichel - Fusco
10h15 General introduction to the needs in data reduction for AO images Rouan
10h50 - 12h30 High-Contrast imaging
10h50 - 11h10 Characterizing noise properties in modern ground based coronagraphs Pueyo
11h10 - 11h30

Imaging and Characterisation with SPHERE in the NIR

Salter
11h30 - 11h50 Image processing to detect and characterize exoplanet : the ANDROMEDA method Cantalloube
11h50 - 12h10 VIP, Software package for angular and spectral differential high-contrast imaging Gomez
12h10 - 12h30 Discussions Mouillet
12h30 - 14h00 Lunch
14h00 - 17h00 Wide-Field AO
14h00 - 14h20 Nailing down the Main Sequence Knee method, the case of M5 Fiorentino
14h20 - 14h40 Deep AO-photometry of crowded Galactic globular clusters Massari
14h40 - 15h00 Distortion correction in WFAO Bernard
15h00 - 15h20 How to handle a variable PSF in AO data Schreiber
15h20 - 15h40 Break Break
15h40 - 16h00 Commissioning progresses & first on-sky closed-loop of the GLAO system ARGOS Gilles Orban de Xivry
16h00 - 16h20 Adaptive Optics System of LINC­/NIRVANA Radhakrishnan
Poster LBT - ARGOS Curti
16h20 - 17h30 Discussion Gratadour
Program - Day2 - Room LAVANDE (1st Floor)
SlotTitleName
09h30 - 10h Cofee
10h Introduction : Quantitative science with AO Wizinowich
10h30 - 10h50 Precise astrometry with AO Neichel
10h50 - 12h00 IFU - Multi Wavelength - Deconvolution
10h50 - 11h10 GalPak3D a recent tool for 3D spectroscopy analysis Bouché
11h10 - 11h30 TBD Hussein Yahia
11h30 - 11h50 Myopic deconvlution of AO corrected images using MISTRAL algorithm. Application to SPHERE data Fusco
11h50 - 12h10 LBT Fizeau image reconstruction La Camera
12h10 - 12h30 Discussion Fusco
Poster Post-AO bispectral speckle imaging Longeard
12h30 - 14h00 Lunch
14h00 - 17h00 PSF-Reconstruction
14h00 - 14h20 PSF reconstruction and application to GLAO Silva
14h20 - 14h40 PSF reconstruction for AO photometry and astrometry Ascenso
14h40 - 15h00 New approach for PSF reconstruction based on a maximum likelihood estimator Gratadour
15h00 - 15h20 Simulation-model-based PSF reconstruction for LGS-assisted wide-field Adaptive Optics Correia
15h20 - 15h40 Break
15h40 - 16h30 Concluding Remarks


Abstracts


Post-AO bispectral speckle imaging
pdf-link
Nicolas Longeard, Éric Aristidi, Marcel Carbillet
Adaptive Optics (AO) in the visible is more difficult because of degraded spatial and temporal coherence of the turbulent atmosphere, with respect to near-infrared. Images are more partially corrected and generally contain a large amount of speckle residuals. Dedicated image processing techniques become necessary to reach theoretical resolution. We propose to apply the "Speckle Masking" method to post-AO partially corrected images to reconstruct objects at the diffraction limit of the telescope. We present here first results based on numerical simulation of binary stars, using the "Building Block" algorithm.


PSF reconstruction for AO photometry and astrometry
pdf-link
Joana Ascenso
Existing tools to extract accurate photometry and astrometry from adaptive-optics images currently hinge on the presence of bright, isolated stars in the image from which to create a reliable model of the peculiar AO point spread function. For many science cases, and especially in the case of small fields of view, these stars are often missing, leaving the task of PSF-fitting very poorly constrained. I will review the existing methods for PSF photometry, their limitations for AO-assisted images, and present the preliminary results of a systematic study with synthetic (GL)AO data to assess the improvements of adding information about the PSF from the AO telemetry in the PSF-fitting algorithms.


Myopic deconvlution of AO corrected images using MISTRAL algorithm. Application to SPHERE data
pdf-link
T. Fusco, L. Mugnier, J-M Conan, M. Marsset, C. Dumas, B. Carry, P. Vernezza, B. Yang
Deconvolution is a necessary tool for the exploitation of a number of imaging instruments. We describe a deconvolution method developed in a Bayesian framework in the context of imaging through turbulence with adaptive optics. This method uses a noise model that accounts for both photonic and detector noises. It additionally contains a regularization term that is appropriate for objects that are a mix of sharp edges and smooth areas, meaning that it is extremely well designed for the observation of solar system bodies.
Finally, it reckons with an imperfect knowledge of the point-spread function (PSF) by estimating the PSF jointly with the object under soft constraints rather than blindly (i.e., without constraints). These constraints are designed to embody our knowledge of the PSF. The implementation of this method is called Mistral.
Mistral performance and robustness is demonstrated on simulated example and applied on various observational scenarii.
In particlar we focus obn brand new data recently obtained using the SPHERE system. This system is already provided unprecedent hihgly resolved data which are enhanced thank to MISTRAL.


Nailing down the Main Sequence Knee method, the case of M5
pdf-link
G. Fiorentino
I will present new results for the Galactic globular cluster M5 (NGC 5904) taken with near Infrared camera combined with GeMS. This system is a nearby halo globular cluster and, indeed, it is located at a distance of only 7 kpc from the Sun ; its metallicity is intermediate [Fe/H] -1.3 and it has a very low reddening E(B - V) = 0.03 (according to the updated version of the Harris catalogue, 2010). This cluster is a very good testbed to investigate the capability of the Main Sequence Knee (MSK) as age indicator. In fact for this cluster accurate distance estimation does exist (Coppola et al. 2011) thus well limiting the major uncertainty to be accounted for in the age error budget. Furthermore, together with the MSK analysis made for NGC3201 ([Fe/H] -1.5, Bono et al. 2010) and NGC7078 (M15, [Fe/H] -2.2, Monelli et al. 2015), this cluster will allow us to extend our exploration of the MSK metallicity dependence.


Deep AO-photometry of crowded Galactic globular clusters
pdf-link
D. Massari
We present the first results of our survey of Galactic globular clusters obtained through GeMS/GSAOI observations. From the analysis of the available dataset, we obtained (K, J-K) CMDs that reach several magnitudes below the Main sequence Turn Off. Cluster stars cleary show all the evolutionary sequences including to the Main Sequence knee, thus allowing us to determine cluster ages with a new and independent estimator. All the data reduction details and the calibration issues will be discussed with care.


Reconstruction of Fizeau images of Io volcanoes obtained with the LBTI/LMIRcam
pdf-link
Andrea La Camera
In this talk I describe the approach developed for deconvolving the Fizeau images of Io detected with the LBTI/LMIRcam in the night of 24th Dec 2013. After a short introduction, first I present the preliminary results obtained through the standard multiple Richardson-Lucy algorithm, then I focus on the reconstructions of the Loki’s patera lava lake. Both these reconstructions have been obtained by means of the Software Package AIRY. Finally I shortly describe a multi-step technique for reconstructing the Io images.


VIP, Software package for angular and spectral differential high-contrast imaging
C. Gomez
In this poster we present our Python-based software package for post-processing of Angular Differential Imaging (ADI) / Spectral Differential Imaging (SDI) data and point-like source detection. Our goal is to provide robust, easy-to-use, well- documented, and freely available implementation of high-contrast image processing algorithms to a broader audience. It’s worth mentioning that the released version of the software is a development snapshot, many changes are being planned and any feedback will be greatly appreciated. VIP is written in Python 2.7 making use of its vast ecosystem of scientific libraries/ packages (e.g. numpy, scipy, scikit-learn, scikit-image, opencv). Python was chosen for being an open-source high level programming language, able to express a large amount of functionality per line of code. Basic image processing functionalities are available e.g. image registration, image rotation, pixel temporal and spatial subsampling, bad frame detection, among others. The Point Spread Function (PSF) subtraction and speckle noise removal, for this development version, are based on Principal Component Analysis (PCA) or Karhunen–Loeve transform. The PCA algorithm can work on the whole frames or in localized patches performing frame selection based on the amount of field rotation. Also, basic median frame subtraction is included as in the original ADI routine. Other functionalities included are : Signal-to-noise ratio (SNR) estimation, SNR map generation, algorithm throughput and fake companion injection.


Image processing to detect and characterize exoplanet : the ANDROMEDA method
pdf-link
Faustine Cantalloube
Exoplanet imaging is currently dependent on image processing techniques to disentangle quasi-statics speckles from real planetary signals in the image. Indeed, in spite of the advanced optical techniques, such as adaptive optics and sophisticated coronagraphs, there is high order optical aberrations remaining in the system that appear in the image field under the shape of speckles. The image processing method I will present, called ANDROMEDA, is an ADI-based method whose specificity is to retrieve the position and flux of companions thanks to a maximum likelihood estimation. Its principle is to make a model of the planetary signal after angular image difference and look for this pattern in the resulting differential image. In this presentation, I will explain the main steps performed by the ANDROMEDA algorithm to detect and characterize stellar companions. I will then show its performance on VLT/NaCo and VLT/SPHERE data and comparisons to other successful algorithms that are widely used today.


Commissioning progresses & first on-sky closed-loop of the GLAO system ARGOS
pdf-link
Gilles Orban de Xivry
ARGOS is the mutliple laser guide star and wavefront sensing facility for the Large Binocular Telescope.
It aims to deliver an improvement by a factor of 2 in FWHM over the 4’x4’ FoV of both LUCI instruments and this under most seeing conditions. LUCI1 and LUCI2 are two near-infrared wide field imagers and multi-object spectrographs whose capabilities and efficiency will be boosted by the increased resolution and encircled energy.
The first on-sky GLAO closure with ARGOS has been achieved in Fall 2014 producing the first observations for performance analysis since then. And while much work is still ahead, we discuss our preliminary analysis and encouraging results on those first AO data, showing the expected decrease in FWHM, together with some aspects of the past and forthcoming commissioning.


Adaptive Optics System of LINC­/NIRVANA
Kalyan Radhakrishnan
LINC/­NIRVANA (LN) is a near infrared high resolution imager for the Large Binocular
Telescope (LBT) equipped with an advanced module of Multi­Conjugate Adaptive Optics (MCAO) and capable of achieving beam combination. LN in many ways acts as a testbed for the ELTs. Atmospheric correction via layer­oriented multiple field of view MCAO using pyramid wavefront sensing and only Natural Guide Stars (NGS) make the LN special. The ground layer and a higher layer ( 7km above telescope pupil) are corrected using Ground Layer Wavefront Sensors (GWS), using NGSs in the annular 2’­6’ FoV, and High layer Wavefront Sensors (HWS), using NGSs in the inner 2’ FoV, respectively. The adaptive secondary mirrors of LBT permit exquisite ground layer correction while the Xinetics DM on the LN bench does the higher layer correction. Ground layer correction has already been achieved with the Pathfinder experiment. In the lab, we have successfully closed the loop up to 50 modes for the high layer loop. LN is going through various verification and consistency tests currently at MPIA. LN is expected to ship to LBT by September 2015.


GalPak3D a recent tool for 3D spectroscopy analysis
pdf-link
Nicolas Bouché
We will present recent developments of 3D tools for the analysis of distant high-redshift galaxies, which may be strongly affected by the seeing. First we will discuss the parametric tool GalPaK3D, and then present a non-parametric method to access the disk kinematics.


New approach for PSF reconstruction based on a maximum likelihood estimator
D. Gratadour
We present a new PSF reconstruction method based on a maximum likelihood approach (ML) using the telemetry data of the AO system. It allows for a joint-estimation of the covariance matrix of low order modes of the residual phase, the noise variance and the Fried parameter r0 during the observations. One main challenge of such a method is the estimation of the covariance between the low order residual phase and the high order purely turbulent phase. We propose a new method to estimate this term using a temporal description of the AO loop. we present results on numerical simulations (for a SCAO system) and show that our ML method allows for an accurate estimation of the PSF in the case of a Shack-Hartmann (SH) wavefront sensor (WFS).


General introduction to the needs in data reduction for AO images
pdf-link
D. Rouan
A large variety of astrophysical fields may benefit from adaptive optics observations : from solar system objects to distant galaxies, including stellar physics, exoplanets and planetary systems, YSO, disks, AGN, stellar clusters and even interstellar medium. Paris observatory has been active in all those fields. I will make a review of the usual goals of the observing programs and of typical imaging requirements in terms of dynamics, resolution, spectral information, field, polarization, photometry, PSF, guiding star etc. The requirements of the data reduction software in each case will be underlined and the projected needs in the near-future will be tentatively guessed.


Post processing for the direct imaging : characterizing noise properties and astrophysical sources in modern ground based coronagraphs.
L. Pueyo
In this paper we present recent progress on data analysis of modern ground based coronagraph aimed at detecting and characterizing exo-planets. We first discuss the overall architecture of such instruments, that combine an Extreme Adaptive Optics system, an optimized coronagraph, low and high-order wavefront sensors and use an Integral Field Spectrograph. In this context we quantify the temporal and chromatic properties of the residual noise associated with imperfections in the optics, their correlation with wavefront sensing and control performances. We discuss post-processing strategies to subtract these speckles in order to detect exo-plantes and circumstellar structures buried under the diffracted stellar light. We then describe the intricacies associated with spectroscopy and astrometry estimation and present two recent algorithms. The first one retrieves the stellar focal plane position when the star is occulted by a coronagraphic stop. The second one yields precise astrometry and spectro-photometry estimates of faint point sources even when they are initially buried in the speckle noise. We finally highlight recent scientific results obtained using these methods.


Astrometry with WFAO.
pdf-link
B. Neichel
The Gemini multiconjugate adaptive optics system (GeMS) is a facility instrument for the Gemini South telescope. It delivers uniform, near-diffraction-limited image quality at near- infrared wavelengths over a 2 arcmin field of view. Together with the Gemini South Adaptive Optics Imager (GSAOI), a near-infrared wide-field camera, GeMS/GSAOI’s combination of high spatial resolution and a large field of view will make it a premier facility for precision astrometry. Potential astrometric science cases cover a broad range of topics including exo- planets, star formation, stellar evolution, star clusters, nearby galaxies, black holes and neutron stars, and the Galactic Centre. In this paper, we assess the astrometric performance and limita- tions of GeMS/GSAOI. In particular, we analyse deep, mono-epoch images, multi-epoch data and distortion calibration.


Distortion correction in WFAO.
pdf-link
A. Bernard
Static and quasi-static distortion are limiting the performance of Wide Field AO images, as when one wants to combine multiple frames, the resulting resolution is severely affected. We are exploring an optimal way of combining images, following the approach developed by D. Gratadour and L. Mugnier, but generalized to several distortion modes. The formalism as well as first simulation results will be presented.


Simulation-model-based PSF reconstruction for LGS-assisted wide-field Adaptive Optics
pdf-link
C. Correia
We cover a simulation-based approach for long exposure point spread function reconstruction (PSFR) for laser guide star (LGS) multi conjugate adaptive optics (MCAO). The approach is based on (i) processing the on-axis high- order LGS wavefront sensor (WFS) second-order statistics to compute via simulation a tip/tilt removed (TTR) structure function (SF) for the science target at infinite range, and (ii) processing similarly the multiple low-order natural guide star (NGS) WFSs controlling tip/tilt (TT) and tilt anisoplanatism (TA) over an extended field of view (FoV) to compute via simulation a tip/tilt (TT) SF for the science target. Such an approach to PSFR is general enough that is fully applicable to any type of tomographic and non- tromographic AO system, employing either laser or exclusively natural guide stars.


How to handle a variable PSF in AO data
pdf-link
L. Schreiber
Adaptive Optics offer the opportunity to work with diffraction limited images.
The extraction of high-precision quantitative information from the astronomical data seems still to be one of the hard points of this technology. In fact, AO data are characterized by a structured and complex PSF, sometimes significantly varying in space and time. Depending on the AO flavour and on the necessary degree
of PSF stability imposed by science requirements, space variation of the PSF could be a crucial issue to be addressed by image processing methods. In the literature one can find many ’exotic’ methods that are difficult to reproduce and to standardise.
A possible approach is to decompose the image in sub-domains where the PSF variation is compatible with the science requirements. Starfinder, a code specifically designed to deal with structured PSFs, offers this possibility, taking properly into account for the sub-domains boundary effects. A new code, called PATCH, also can work on ’fragmented’ images, deconvolving each sub-domaing for the proper PSF. The PSF relative to each sub-domain has to be estimated separately.
Work is in progress to release also a new version of Starfinder that can deal with an analitical model of the PSF, whose parameters are variable across the FoV.


Imaging and Characterisation with SPHERE in the NIR
G. Salter
pdf-link
In this talk we shall take a quick look at how data from SPHERE’s dual band imager (IRDIS) and integral field spectrograph (IFS) are analysed. We shall also take a look at some of the recent results with SPHERE.


Local Information

Take the Metro Line number 1 until the terminus "La Rose".
Go out of the metro station, turn right at the bottom of the electric stairs, and go to the bus station at 20m in front.
Take the bus B3B, it is a long bus, and leave it after 2 stations, at the "Technopole Centrale Marseille" LAM is just in front, a little bit toward your left.

  • More informations can be found here
  • Lunch and cofee are provided.
  • A dinner (not covered by the workshop) will be organized on Tuesday night


Organizers

Organizers
NameInstituteTopic
Carlos Correia LAM (Marseille) Wide Field AO
Thierry Fusco ONERA/LAM (Marseille) High Contrast
Andres Guesalaga Catolica University (Chile) & IMéRA Résident Profiling / Wide Field AO
Damien Gratadour Observatoire de Paris (LESIA) Wide Field AO
David Mouillet IPAG (Grenoble) High Contrast
Benoit Neichel LAM (Marseille) Wide Field AO
Laurent Pueyo STSCI (USA) High Contrast
Peter Wizinowich Keck (Hawaii) LGS AO


Sponsors

The workshop is supported by the OPTICON-FP7 program, the IMéRA Foundation, the ASHRA, and the French ANR program WASABI.



ccorreia15_psfreconstru...
(PDF 768.8 ko)

© LAM - Laboratoire d’Astrophysique de Marseille

Pôle de l’Étoile Site de Château-Gombert
38, rue Frédéric Joliot-Curie 13388 Marseille cedex 13 FRANCE

Tél : +33 4 91 05 59 00
Fax : +33 4 91 62 11 90