Bayesian signal separation applied to ground-roll removal

Accurate and adaptive noise removal is a critical part in seismic processing. Recent developments in signal separation methods have allowed a more flexible and accurate framework in which to perform ground roll and reflector separation. The use of a new Bayesian separation scheme developed at the SLIM group that contains control parameters to adjust for the uniqueness of specific problems is used. The sensitivity and variation of the control parameters is examined and this method is applied to synthetic and real data and the results are compared to previous methods.