Characterise the stellar populations using seismic indices ; Caractériser les populations stellaires à l'aide d'indices sismiques

Like terrestrial seismology, asteroseismology is the study of star quakes. These internal vibrations are detected by observing the luminosity (or velocity) variations at the stellar surfaces. From these time series, one computes power spectra which contain a wealth of information. In particular, for solar-like pulsators, we are able to observe the signal of granulation as well as the eigenmodes of stellar oscillations, both of which are a direct consequence of the convection in the star's envelope. Asteroseismology allows us to probe the interior of stars much like an ultrasound. Furthermore, with the large spatial photometric surveys CoRoT and Kepler, a new scope for seismology appeared. Indeed, it is possible to characterise to first order the oscillation spectra of solar-like pulsators with few indices or parameters, called seismic indices. Using simple relations, they allow us to estimate fundamental parameters of these stars. Asteroseismology is by consequence a very powerful tool for the study of stellar populations.In the perspective of the development of the Stellar Seismic Indices (SSI - The opening of the SSI database is planned for the summer of 2016, see http://ssi.lesia.obspm.fr/).) database, the purpose of my thesis was to develop an automatic method able to extract simultaneously the seismic indices and the parameters characterizing the granulation signature of solar-like pulsators. This method, called MLEUP, was optimized for red giants because for the few hundred main-sequence solar-like pulsators observed by CoRoT and Kepler, several tens of thousands of red giants have been observed by these same missions. MLEUP has a major advantage over most existing methods: it relies on the use red-giant stellar oscillation universal pattern (UP) to fit the oscillation spectra. This allows us to analyse the unsmoothed spectrum and fit simultaneously both granulation and oscillations with the maximum likelihood estimate (MLE).As a first step, MLEUP was tested on Monte Carlo simulations in order to quantify ...