Curvelet analysis of asteroseismic data

 

Authors: P. Lambert, S. Pires, J. Ballot, R.A. Garcia, J.-L. Starck, S. Turck-Chièze
Journal: A&A
Year: 2006
Download: ADS | arXiv

 

Abstract

Context. The detection and identification of oscillation modes (in terms of their l, m, and successive n) is a great challenge for present and future asteroseismic space missions. “Peak tagging" is an important step in the analysis of these data to provide estimations of stellar oscillation mode parameters, i.e., frequencies, rotation rates, and further studies on the stellar structure.
Aims. Our goal is to increase the signal-to-noise ratio of the asteroseismic spectra computed from the time series that are representative of MOST and CoRoT observations (30- and 150-day observations).

Methods. We apply the curvelet transform – a recent image processing technique that looks for curved patterns – to echelle diagrams built using asteroseismic power spectra. In the resulting diagram, the eigenfrequencies appear as smooth continuous ridges. To test the method, we use Monte-Carlo simulations of several sun-like stars with different combinations of rotation rates, rotation-axis inclination, and signal-to-noise ratios.

Results. The filtered diagrams enhance the contrast between the ridges of the modes and the background, allowing a better tagging of the modes and a better extraction of some stellar parameters. Monte-Carlo simulations have also shown that the region where modes can be detected is enlarged at lower and higher frequencies compared to the raw spectra. In addition, the extraction of the mean rotational splitting from modes at low frequency can be done more easily using the filtered spectra rather than the raw spectra.

Weak lensing mass reconstruction using wavelets

 

Authors: J.-L. Starck, S. Pires, Alexandre Réfrégier
Journal: A&A
Year: 2006
Download: ADS | arXiv

 

Abstract

This paper presents a new method for the reconstruction of weak lensing mass maps. It uses the multiscale entropy concept, which is based on wavelets, and the False Discovery Rate which allows us to derive robust detection levels in wavelet space. We show that this new restoration approach outperforms several standard techniques currently used for weak shear mass reconstruction. This method can also be used to separate E and B modes in the shear field, and thus test for the presence of residual systematic effects. We concentrate on large blind cosmic shear surveys, and illustrate our results using simulated shear maps derived from N-Body Lambda-CDM simulations with added noise corresponding to both ground-based and space-based observations.