Soutenance de thèse: Arseniy Kuznetsov

Soutenance de thèse : “Towards the next generation of AO-assisted instruments: Self-learning techniques for system optimization & science exploitation” par Arseniy Kuznetsov.

Dans l’amphithéâtre du LAM, à 3h00, en anglais.

Abstract:

Key astronomical fields, such as galaxy studies, the Galactic Center, exoplanet detection, imaging of the bodies within the solar system, and more, rely on adaptive optics (AO) in ground-based observations to achieve near-diffraction-limited resolution by correcting atmospheric turbulence. As such, the quality of scientific results depends on imaging quality, which is directly tied to the shape of the point spread function (PSF). In addition, accurate PSF knowledge is crucial for reliable data interpretation and advanced post-processing, including deconvolution, object subtraction, and precise astrometry and photometry. However, the complex structure of AO-assisted PSFs makes realistic modeling challenging, particularly when the PSF cannot be directly extracted from the field of view. This thesis introduces a fast and realistic method for predicting AO-corrected science PSFs using a data-augmented analytical model. This approach accurately infers PSF shape from a limited set of data inputs that can be easily associated with scientific observations. The proposed method was validated with on-sky data from two ESO Very Large Telescope (VLT) instruments: SPHERE and MUSE. Additionally, the second part of this work addresses focal plane wavefront sensing, focusing on reliably extracting quasi-static aberrations from the PSF shape during the on-sky operations.

Jury:

Laurent Jolissaint (HES-SO) – Reviewer

Jean-Pierre Véran (UVic) – Reviewer

Annie Zavagno (Aix Marseille Université) – President

David Mary (Université Côte d’Azur) – Examiner

Jessica Lu (UC Berkley) – Examiner

Joël Vernet (ESO) – Examiner

Supervisors:

Benoît Neichel (LAM)

Sylvain Oberti (ESO)

Thierry Fusco (ONERA)