MULTICHANNEL SEISMIC DECONVOLUTION

This paper deals with Bayesian estimation of two-dimensional (2-D) str atified structures from echosounding signals. This problem is of inter est in seismic exploration, but also for nondestructive testing or med ical imaging. The proposed approach consists of a multichannel Bayesia n deconvolution method of the 2-D reflectivity based upon a theoretica lly sound prior stochastic model. The Markov-Bernoulli random field re presentation introduced in [1] is used to model the geometric properti es of the reflectivity, and the emphasis is placed on representation o f the amplitudes and on deconvolution algorithms. It is shown that the algorithmic structure and computational complexity of the proposed mu ltichannel methods are similar to those of single channel B-G deconvol ution procedures, but that explicit modeling of the stratified structu re results in significantly better performances. Simulation results an d examples of real-data processing illustrate the performances and the practicality of the multichannel approach.