Deep learning for wavefront sensing and control in view of exoplanet imaging

20/06/2022 – 21/06/2022 all-day

Offre en lien avec l’Action/le Réseau : – — –/Doctorants

Laboratoire/Entreprise : Observatoire de Paris (LESIA) / ONERA (DOTA)
Durée : 3 ans
Contact :
Date limite de publication : 2022-06-20

Contexte :
Direct imaging of circumstellar environments is aspiring to detect and analyze faint exoplanets (planet orbiting another star). The flux ratio between an Earth-like planet and a sun-type star is typically 10 billion in the visible, at a fraction of arcsecond. The observation and spectral characterization of these objects therefore require a combination of high angular resolution and of high contrast imaging techniques, with instruments called coronagraphs.

Sujet :
State of the art coronagraph designs (space or ground-based) are currently limited by residual aberrations, due to the atmosphere or directly to the optics surface quality. These aberrations create stellar leakage (speckles) in the focal plane that masks the fainter planets or disks. The goal of active instrumentation for coronagraphy is to correct these aberrations to detect exoplanets. Recent improvements in deep neural network architectures, as well as increasing computational capabilities, offer new solutions to control these aberrations. We propose to combine previous model-based control techniques with machine learning to produce fast and efficient correction of optical aberrations.

Profil du candidat :
We are looking for a candidate with a dual set of interests both in physics / Astronomy and in Computer Science / Applied math. The candidate can be a specialist in only one of those fields but should be willing to quickly learn about the other.
The mentoring team is composed of experts in Machine Learning / Astronomie / Applied Math to help the candidate lear efficiently in those subjects.

Formation et compétences requises :
Master degree in Signal/Image processing or Applied Math, with an interest in astronomy / experimental physics
Master degree in Physics or Astronomy or with an strong interest for AI and experimental physics

Level of French required: None

Level of English required: Advanced: You can speak the language more complexly, spontaneously and on a variety of topics.

Adresse d’emploi :
Observatoire de Paris, site de Meudon
5 Pl. Jules Janssen, 92190 Meudon