Deep learning-based focal plane wavefront sensing for classical and coronagraphic imaging

Quesnel, Maxime; Xivry, Gilles Orban de; Louppe, Gilles; Absil, Olivier

doi:10.1117/12.2562456

Cited by 4 publications

(2 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Again, a trade-off needs to be found. Finally, several authors have recently tried to use neural networks trained on instrument data to approach the instrumental models in order to produce faster algorithms [172,173].…”

Section: Focal Plane Wavefront Sensing: Modulation Of the Speckle Int...mentioning

confidence: 99%

Imaging exoplanets with coronagraphic instruments

Galicher¹,

Mazoyer²

2023

Comptes Rendus. Physique

View full text Add to dashboard Cite

Exoplanetary science is a very active field of astronomy nowadays, with questions still opened such as how planetary systems form and evolve (occurrence, process), why such a diversity of exoplanets is observed (mass, radius, orbital parameters, temperature, composition), and what are the interactions between planets, circumstellar disk and their host star. Several complementary methods are used for the detection of exoplanets. Among these, imaging aims at the direct detection of the light reflected, scattered or emitted by exoplanets and circumstellar disks. This allows their spectral and polarimetric characterization. Such imaging remains challenging because of the large luminosity ratio (10 4 -10 10 ) and the small angular separation (fraction of an arcsecond) between the star and its environment. Over the past two decades, numerous techniques, including coronagraphy, have been developed to make exoplanet imaging a reality.This paper gives a broad overview of the subsystems that make up a coronagraphic instrument for imaging exoplanetary systems. It is especially intended for non-specialists or newcomers in the field. We explain the principle of coronagraphy and propose a formalism to understand their behavior. We discuss the impact of wavefront aberrations on the performance of coronagraphs and how they induce stellar speckles in the scientific image. Finally, we present instrumental and signal processing techniques used for on-sky minimization or a posteriori calibration of these speckles in order to improve the performance of coronagraphs.Résumé. L'exoplanétologie est un domaine très actif de l'astronomie moderne avec des questions encore ouvertes : comment les systèmes planétaires se forment-ils et évoluent-ils ; pourquoi une telle diversité d'exoplanètes est-elle observée (masse, rayon, paramètres orbitaux, température, composition) ; quelles sont les interactions entre les planètes, les disques circumstellaires et leur étoile hôte ? Plusieurs méthodes complémentaires sont utilisées pour la détection d'exoplanètes. Parmi celles-ci, l'imagerie permet la détection directe de la lumière réfléchie, diffusée ou émise par les exoplanètes et les disques circumstellaires. Ceci permet une caractérisation spectrale et polarimétrique. Obtenir une image d'exoplanète n'est cependant pas simple en raison du grand rapport de luminosité (10 4 -10 10 ) et de la faible séparation angulaire (fraction de seconde d'angle) entre l'étoile et son environnement. Depuis deux décennies, de nombreuses techniques, dont la coronographie, ont été développées pour faire de l'imagerie des exoplanètes une réalité.

show abstract

Section: Focal Plane Wavefront Sensing: Modulation Of the Speckle Int...mentioning

confidence: 99%

Imaging exoplanets with coronagraphic instruments

Galicher¹,

Mazoyer²

2023

Comptes Rendus. Physique

View full text Add to dashboard Cite

show abstract

“…Deep learning with convolutional neural networks (CNN) has recently been implemented for focal-plane wavefront sensing [3][4][5][6][7][8] with good results. CNNs offer great performance and robustness as well as fast inference, making them particularly adequate for NCPA correction during observing nights.…”

Section: Introductionmentioning

confidence: 99%

A simulator-based autoencoder for focal plane wavefront sensing

Quesnel

Xivry

Absil³

et al. 2022

Adaptive Optics Systems VIII

Self Cite

View full text Add to dashboard Cite

Instrumental aberrations strongly limit high-contrast imaging of exoplanets, especially when they produce quasistatic speckles in the science images. With the help of recent advances in deep learning, we have developed in previous works an approach that applies convolutional neural networks (CNN) to estimate pupil-plane phase aberrations from point spread functions (PSF). In this work we take a step further by incorporating into the deep learning architecture the physical simulation of the optical propagation occurring inside the instrument. This is achieved with an autoencoder architecture, which uses a differentiable optical simulator as the decoder. Because this unsupervised learning approach reconstructs the PSFs, knowing the true phase is not needed to train the models, making it particularly promising for on-sky applications. We show that the performance of our method is almost identical to a standard CNN approach, and that the models are sufficiently stable in terms of training and robustness. We notably illustrate how we can benefit from the simulator-based autoencoder architecture by quickly fine-tuning the models on a single test image, achieving much better performance when the PSFs contain more noise and aberrations. These early results are very promising and future steps have been identified to apply the method on real data.

show abstract

Focal plane wavefront sensing using machine learning: performance of convolutional neural networks compared to fundamental limits

Xivry

Quesnel

Vanberg

et al. 2021

Monthly Notices of the Royal Astronomical Society

View full text Add to dashboard Cite

Focal plane wavefront sensing (FPWFS) is appealing for several reasons. Notably, it offers high sensitivity and does not suffer from non-common path aberrations (NCPA). The price to pay is a high computational burden and the need for diversity to lift any phase ambiguity. If those limitations can be overcome, FPWFS is a great solution for NCPA measurement, a key limitation for high-contrast imaging, and could be used as adaptive optics wavefront sensor. Here, we propose to use deep convolutional neural networks (CNNs) to measure NCPA based on focal plane images. Two CNN architectures are considered: ResNet-50 and U-Net which are used respectively to estimate Zernike coefficients or directly the phase. The models are trained on labelled datasets and evaluated at various flux levels and for two spatial frequency contents (20 and 100 Zernike modes). In these idealized simulations we demonstrate that the CNN-based models reach the photon noise limit in a large range of conditions. We show, for example, that the root mean squared (rms) wavefront error (WFE) can be reduced to <λ/1500 for 2 × 106 photons in one iteration when estimating 20 Zernike modes. We also show that CNN-based models are sufficiently robust to varying signal-to-noise ratio, under the presence of higher-order aberrations, and under different amplitudes of aberrations. Additionally, they display similar to superior performance compared to iterative phase retrieval algorithms. CNNs therefore represent a compelling way to implement FPWFS, which can leverage the high sensitivity of FPWFS over a broad range of conditions.

show abstract

Deep learning-based focal plane wavefront sensing for classical and coronagraphic imaging

Cited by 4 publications

References 20 publications

Imaging exoplanets with coronagraphic instruments

Imaging exoplanets with coronagraphic instruments

A simulator-based autoencoder for focal plane wavefront sensing

Focal plane wavefront sensing using machine learning: performance of convolutional neural networks compared to fundamental limits

Contact Info

Product

Resources

About