Annonce

Les commentaires sont clos.

Hyperspectral Imaging for Astrophysics

14 Octobre 2022


Catégorie : Stagiaire


Hyperspectral Imaging for Astrophysics The proposed work consists processing data from the MUSE (Very-Large-Telescope) instrument, with the aim to detect signal produced by forming planets, at very SNR. The internship is proposed to either Master 2 students or last year ingeneer students. The duration is 5-6 monts. Details are given below.

 

Integral field spectrographs are a class of instruments now deployed on most modern astronomical observatories (Very-Large-Telescope in Chile; Keck Telescope in Hawaii).The hyperspectral image cubes they produce contain a diversity of information that is beginning to be exploited to search for and characterize planets in the process of formation (proto-planets) around other stars [1].

The data are dominated by the flux halo of the star which acts as a nuisance. The signals produced by the forming planets are sparse and close to the noise level. Methods to search for weak signals in hyperspectral data cubes have been developed in similar methodological contexts [2]. These methods need to be adapted to the present problem of detecting planets in formation by searching for a specific hydrogen spectral line (Hɑ line).

The proposed work consists in taking in hand data from the MUSE (Very-Large-Telescope) instrument and to reproduce in a first step the results of the state of the art [3]. It is then necessary to highlight the possibilities and limitations of existing algorithmic solutions. A particular effort will be deployed towards the methods of subtraction of halo in the images and the analysis of detection performances on the residues obtained by subtraction of the latter. The data to be processed/analyzed are already acquired and will therefore be easily available.

The expertise developed will allow to extend and propose evolutions of these methods, possibly in the context of an extension of this work by a PhD (ANR funding acquired). The work will take place in the context of a collaboration between the GIPSA-Lab and IPAG laboratories in Saint-Martin-d'Hères.

 

Expected profile

Good knowledge in detection, estimation, Bayesian approaches. Notions on image formation. Appétences for astrophysical data processing and exchange between several scientific disciplines. Good knowledge of Matlab and Python.

 

Contacts

GIPSA-Lab : O. Michel (olivier.Michel@grenoble-inp.fr), F. Chatelain (florent.chatelain@grenoble-inp.fr)

IPAG : Bonnefoy Mickael (mickael.bonnefoy@univ-grenoble-alpes.fr)

 

Références

[1] Haffert et al. 2019, Nature Astronomy, 3, 749

[2] Meillier et al. 2016, A&A, 588, 140

[3] Xie et al. 2020, A&A, 644, 149