MILCA, a modified internal linear combination algorithm to extract astrophysical emissions from multifrequency sky maps
Résumé
The present work deals with the problem of extracting a sky map of a particular emission process that has been observed in an experiment with different detectors each of them having a different spectral response. This is the arena of the so-called {{"methods of component separation"}}, especially in the field of microwave experiments, like the Cosmic Background Microwave (CMB) missions COBE, WMAP or, at present, the Planck surveyor mission. For the Internal Linear Combination methods (ILC), the difference with respect to other approaches is that it uses only the spectral behaviour of the sought component as an input. This idea has been applied to the case of CMB emission. Since this emission is itself the calibration emission in those maps, the problem is simplified. In this work, we derive the general expression for a generic spectral behaviour of the sought emission. We also apply the method to some of the common missions in the range of microwave and sub-mm emissions: Galactic dust, Sunyaev-Zel'dovich effect and as a check for consitency to CMB also. The data are simulations that resemble those performed presently by the Planck surveyor mission. Moreover, we will also show how it is possible to optimize the extraction of the chosen emission by minimizing the output noise and the bias in the extraction of the component. Therefore, we call this method MILCA: Maximum Internal Linear Component Analysis.