PASEO Group / Project / CubeDeconvolution

We consider the problem of deconvolution of astronomical data cubes. This project is part of ANR Dahlia (2009-2013) which aims at developping new methods for image analysis applied to the new integral field unit MUSE. Because in this application the data are huge (300x300x4096) and the PSF varies spatially and spectrally, no existing method could succesfully be applied. Therefore, we propose to develop a new framework in which we model the PSF variations. We also need to propose new priors and models to take into account the complexity of the spectral information in the cube. In a second hand, we will tune the method to the particular use for MUSE. This model will be developed in partnership with LATT (Toulouse, France).

Description

Nowadays, we are thinking about the modeling of the PSF which contains the PSF of the atmosphere, the telescope (and the adaptive optics if needed), as well as the PSF of the instrument MUSE. This step is essential for the choice of the method to enhance the spatial resolution of the cubes.

The modelization of the PSF is difficult due to many problems. Especially, the PSF is not necessary separable ; that is, it could not be written as the product of a «spatial PSF» (or field spread function) and a «spectral PSF» (or line spread function). Besides, the PSF may vary spatially and spectrally, both in size and shape. These variations can be very important and non negligeable. At last, some PSF are explicitly known (lucky case) : an expression is available to model the PSF, but most of the time its parameters have to be measured. On the contrary, some PSF are measured and, thus, no explicit expression exists. The other PSF are unknown and are to be simulated using numerical models. The huge quantity of data for measured and simulated PSF is also one of the difficulty of the project.

Bibliography

A. Jarno, Développement d'un modèle numérique de l'instrument MUSE/VLT. PhD of the Institut National des Sciences Appliquées de Lyon, 2008. PDF (41 Mb)