Blanc SIMI 5 - Sciences de l'information, de la matière et de l'ingénierie : Sciences de l’univers

Processing of Polychromatic Interferometric Data for Astrophysics – POLCA

POLCA: Processing of pOLychromatic interferometriC data for Astrophysics

From multispectral interferometric data to astrophysical breakthroughs<br /><br />The project POLCA brought together people from optical stellar interferometry and from signal processing to make breakthroughs in the potential of stellar interferometry by fully using the polychromatic information of interferometric data, and by demonstrating them on significant science cases with real data.<br />

Joining together experts in stellar interferometry and signal processing to make the most of optical interferometers

By combining the light from several telescopes, Optical Interferometry (OI) provides the highest angular resolution (milliarcsecond resolution) in many fields such as stellar physics, star and planet formation, and environment of extra galactic black holes. In this field, the pioneering work in France has put Europe at the forefront with the European interferometer, the VLTI (over 50% of the publications in 2010-2014). Still very active through the JMMC, French laboratories provide tools to prepare, reduce, and interpret the observations. The first image reconstructions came out. But current and future instruments (GRAVITY, MATISSE) provide a rich multi-spectral information not fully exploited yet. By joining experts in signal processing and OI, the POLCA project has innovated new methods to jointly process spatial and spectral information, and to improve both the performances (limiting magnitude, precision) and the observational capabilities (physics of the objects). The ability to reconstruct spatio-spectral images will increase the value and impact of the OI.

The project POLCA was based on actual data prompted by specific astrophysical goals. We selected both existing data and new ones that would benefit from new polychromatic processing. The areas eventually addressed were young stars (stars and environment), Mira stars, and stars of the Galactic Center. We have combined a thorough knowledge of the instruments to the latest achievements in signal processing in the study of the interferometric observables, particularly the differential measurements (differences in wavelength), the noise statistics and correlations, and develop new algorithms suited to a spatio-spectral global approach. Work on astrophysics interpretation and search for new approaches were carried out in conjunction. Thus short-term pragmatic solutions were carried out in parallel to exploratory fundamental research.

Among the results of POLCA, we can notice achievements on the morphology of the Herbig Ae/Be star environment with PIONIER/VLTI, the first multispectral reconstructions of Mira stars (R For and R Car), and the demonstration of the extraction of the positions and spectra of stars in the Galactic Center, as the instrument GRAVITY will observe them. Concerning multispectral image reconstruction, SPARCO algorithm, efficient with HAe/Be polychromatic data, was followed by MiRA-3D and PAINTER, two algorithms based on the alternating direction method of multipliers (ADMM) with various regularizations and constraints, including sparse approaches that have also inspired the best blind deconvolution algorithm in 3D microscopy. The study of interferometric data themselves yielded an estimator that can increase the accuracy of stellar diameter measurements by a factor of 2 to 10, and a robust method for cophasing fringes with 10 times more sensitivity. The PAINTER software is publicly available.

The main results achieved are summarized on POLCA web site:

The project POLCA produced 28 publications including 9 peer-reviewed papers and participated in the development of the new version of the OIFITS standard format for interferometric data, that now conveys spectral information (spectra, spectro-differential measurements, noise correlations). The Jean-Marie Mariotti Center (JMMC) is now working at building on POLCA results to make new tools available to the community.

The full list of the publications is available on the POLCA web site:

Publications in peer reviewed journals:

Thibaut Paumard, Oliver Pfuhl, Fabrice Martins, et al., 2014, «GCIRS 7, a pulsating M1 supergiant at the Galactic centre. Physical properties and age«, A&A Vol 568, id.A85, 10 pp.

Antony Schutz, André Ferrari, David Mary, et al., 2014, PAINTER: a spatio-spectral image reconstruction algorithm for optical interferometry, JOSA A, Vol. 31, pp. 2334-2345.

Antony Schutz, Martin Vannier, David Mary, et al., Statistical characterization of polychromatic absolute and differential squared visibilities obtained from AMBER/VLTI instrument, Astronomy and Astrophysics, Vol. 565, pp A88, 2014

Jacques Kluska, Fabien Malbet, Jean-Philippe Berger, et al., 2014, SPARCO : a semi-parametric approach for image reconstruction of chromatic objects. Application to young stellar objects, Astronomy & Astrophysics, Vol 564, id.A80, 11 pp.

Éric Thiébaut, Ferréol Soulez and Loïc Denis, 2013, Exploiting spatial sparsity for multi-wavelength imaging in optical interferometry, JOSAA, Vol 30, 160-170

Jacques Kluska, Fabien Malbet, Jean-Philippe Berger, et al., 2013, High Angular Resolution and Young Stellar Objects: Imaging the Surroundings of MWC 158 by Optical Interferometry, EAS Publications Series Vol 59 pp 141-154

Ferréol Soulez, Éric Thiébaut and Loïc Denis, 2013, Restoration of Hyperspectral Astronomical Data with Spectrally Varying Blur, EAS Publications Series Vol 59 pp 403-416

Gilles Duvert, John Young and Christian Hummel, 2015, OIFITS 2: the 2nd version of the Data Exchange Standard for Optical (Visible/IR) Interferometry, submitted to A&A, sept. 2015

Alexis Matter, Lucas Labadie, Jean-Charles Augereau, et al., 2015, «Inner disk clearing around the Herbig Ae star HD 139614: Evidence for a planet-induced gap ?«, A&A, Vol 561, A26

The aim of the POLCA project is to bring together people from optical stellar interferometry (OSI) and from signal processing to make a breakthrough in the potential of OSI by fully using the polychromatic information of interferometric data, and to demonstrate it on significant science cases with real data.

Europe and in particular France is a leading actor in OSI. Thanks to the pioneering work made in France, the VLTI European facility is the most productive in the world (40% of the publications for years 2002-2007). This success allows OSI to enter into mainstream astronomy providing milli-arcsecond spatial resolution on a wide range of domain (stellar physics, star and planet formation, environment of extra galactic black holes,...).

As of today, French OSI laboratories, through JMMC, distribute software for preparing observations, for finding calibrators, for data reduction of the VLTI instruments, and for model fitting. Efforts are ongoing for image reconstruction and first images are obtained.

But a new step forward needs to be accomplished in order to make the most of the somehow complex interferometric data. Indeed, the current interferometric data carries rich spectral information that is not fully exploited because of the difficulty to process it at the same time as the spatial information.

Most usually, the data are processed for spectral channels separately, and wavelength dependency is considered afterward. Impact of stellar interferometry would be significantly enhanced with efficient means to process (x,y,lambda) as a whole. But the heterogeneity of the (x,y,lambda) space needs specific progress and significant advances in signal processing. Further both the size of the parameter space for image reconstruction and the numerous possible representations of the spectrum call for innovative R&D in the fields of compressed sensing and regularization methods.

As another example, the observation of complex objects shows that model fitting often captures only a part of the object features, and efficient means are needed to explore the residuals of a fit, for instance with an image reconstruction approach, with semi-parametric modeling and spatio-spectral regularizations.

Other specific difficulties come from the correlations seen between the measurements in different spectral channels, or from the different methods used by the instruments for measuring the so-called differential visibilities (differential along wavelengths). These problems call for bringing together experts from stellar interferometry and from signal processing with the aim to clarify the correct modeling of data, including noises and biases.

The outcome should be to improve the current measurements or to derive more suitable tools for data processing.

In the context of the existing instruments (e.g. MIDI, AMBER, VEGA, IOTA) and the ones currently in development (PIONIER, GRAVITY, MATISSE), several scientific cases can be considered as promising but currently limited by the lack of suitable inversion tool. For instance, the study of chemical composition and fine spatial structure of dusty disk (protoplanetary or evolved ones), the study of faint companion to stars up to exoplanets, or the study of the kinematics of young or evolved circumstellar environments.

We propose to select existing data on specific objects that really need new polychromatic processing to solve some open questions, and to add new data acquired with the same intention. Around these data, we intend to formalize physical models of object, model of data, study noise statistics and correlations, and develop new algorithms suitable for a 3D approach.

This project would allow French astronomers to reap the benefits of long investments in OSI by getting the most of available and soon to come OSI facilities, and also to keep a leading position in this field.

Project coordinator


The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.



Help of the ANR 400,000 euros
Beginning and duration of the scientific project: - 48 Months

Useful links

Explorez notre base de projets financés



ANR makes available its datasets on funded projects, click here to find more.

Sign up for the latest news:
Subscribe to our newsletter