Linear Inversion of Band-Limited Reflection Seismograms

Abstract
We present a method for the linear inversion (deconvolution) of band-limited reflection seismograms. A large convolution problem is first broken up into a sequence of smaller problems. Each small problem is then posed as an optimization problem to resolve the ambiguity presented by the band-limited nature of the data. A new cost functional is proposed that allows robust profile reconstruction in the presence of noise. An algorithm for minimizing the cost functional is described. We present numerical experiments which simulate data interpretation using this procedure.