arXiv:astro-ph/0011141v1 7 Nov 2000 Reconstructing the microwave sky using a combined maximum-entropy and mexican hat wavelet analysis R. Bel´ en Barreiro 1 , Patricio Vielva 2,3 , Michael P. Hobson 1 , Enrique Mart´ ınez-Gonz´ alez 2 , Anthony N. Lasenby 1 , Jos´ e L. Sanz 2 , and Luigi Toffolatti 4 1 Astrophysics Group, Cavendish Laboratory, Madingley Road, Cambridge CB3 0HE, UK 2 Instituto de F´ ısica de Cantabria (CSIC – UC), Fac. Ciencias, Avda. de los Castros s/n, 39005 Santander, Spain 3 Departamento de F´ ısica Moderna, Universidad de Cantabria, Avda. de los Castros s/n, 39005 Santander, Spain 4 Departamento de F´ ısica, Universidad de Oviedo, c/ Calvo Sotelo s/n, 33007 Oviedo, Spain Abstract. We present a combined maximum-entropy method (MEM) and Mexi- can Hat wavelet (MHW) analysis in order to recover the different components of the microwave sky. We apply this technique to simulated observations by the ESA Planck satellite in small patches of the sky. In particular, the introduction of the MHW allows one to detect and subtract the brightest point sources present in the input data and therefore to improve the reconstructions of the CMB and foreground components achieved by MEM on its own. In addition, a point source catalogue at each Planck frequency is produced, which is more complete and accurate than those obtained by each technique independently. 1 Introduction Cosmic Microwave Background (CMB) observations carry a wealth of in- formation about the Universe. Indeed, an accurate knowledge of the CMB anisotropies can place tight constraints on fundamental parameters as well as to discriminate between competing theories of structure formation. Future CMB experiments such as the NASA MAP satellite and the Planck mission from ESA, will provide with multifrequency data at high resolution and sen- sitivity. However, these data contain not only the cosmological signal but also Galactic foregrounds, extragalactic point sources, thermal and kinetic Sunyaev-Zelodvich (SZ) emission from cluster of galaxies and instrumental noise. Therefore our capacity to recover all the valuable information encoded in the CMB will critically depend on our ability to denoise and separate the cosmological signal from the rest of components of the microwave sky. To perform such a separation, [1] has developed a Fourier MEM algorithm. This technique is particularly successful at using multifrequency data to iden- tify foreground emission from physical components whose spectral signatures