Maximum-likelihood blind deconvolution: non-white Bernoulli-Gaussian case

The authors present a maximum-likelihood deconvolution (MLD) algorithm for estimating nonwhite Bernoulli-Gaussian signals μ(k ), which were distorted by a linear time-invariant system υ( k) taking into account the measured spectrum of μ(k) such as that obtained from sonic logs. The proposed MLD algorithm can recover both the phase of a minimum-phase coloring filter υ₁(k) and that of υ(k) as long as the spectrum of μ(k) is known in advance. The authors also present some simulation results which support the proposed MLD algorithm