Galaxy decomposition in multispectral images using markov chain monte carlo algorithms

Abstract

Astronomers still lack a multiwavelength analysis scheme for galaxy classification. In this paper we propose a way of analysing multispectral observations aiming at refining existing classifications with spectral information. We propose a global approach which consists of decomposing the galaxy into a parametric model using physically meaningful structures. Physical interpretation of the results will be straightforward even if the method is limited to regular galaxies. The proposed approach is fully automatic and performed using Markov Chain Monte Carlo (MCMC) algorithms. Evaluation on simulated and real 5-band images shows that this new method is robust and accurate. © 2009 Springer Berlin Heidelberg.