Maxime Quesnel | Searcharxiv

arXiv2024

ALF: an asymmetric Lyot wavefront sensor for the ELT/METIS vortex coronagraph

Non-common path quasi-static and differential aberrations are one of the big hurdles of direct imaging for current and future high-contrast imaging instruments. They increase speckle and photon noise thus reducing the achievable contrast and lead to a significant hit in HCI performance. The Mid-infrared ELT Imager and Spectrograph (METIS) will provide high-contrast imaging, including vortex coronagraphy in L, M and N bands, with the ultimate goal of directly imaging temperate rocky planets around the nearest stars. Ground-based mid-infrared observations are however also impacted by water vapor inhomogeneities in the atmosphere, which generate additional chromatic turbulence not corrected by the near-infrared adaptive optics. This additional source of wavefront error (WFE) significantly impacts HCI performance, and even dominates the WFE budget in N band. Instantaneous focal plane wavefront sensing is thus required to mitigate its impact. In this context, we propose to implement a novel wavefront sensing approach for the vortex coronagraph using an asymmetric Lyot stop and machine learning. The asymmetric pupil stop allows for the problem to become solvable, lifting the ambiguity on the sign of even Zernike modes. Choosing the Lyot plane instead of the entrance pupil for this mask is also not arbitrary: it preserves the rejection efficiency of the coronagraph and minimizes the impact of the asymmetry on the throughput. Last but not least, machine learning allows us to solve this inversion problem which is non-linear and lacks an analytical solution. In this contribution, we present our concept, our simulation framework, our results and a first laboratory demonstration of the technique.

astro-ph.IM↗

arXiv2022

A simulator-based autoencoder for focal plane wavefront sensing

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.

astro-ph.IM↗