A multiscale approach to solving one dimensional inverse problems

Abstract

A multiresolution approach to solving one dimensional inverse problems is explored. Inverse problems described by that class of operators which are made sparse under the action of the wavelet transform are considered. Statistically based inversion procedures utilizing multiscale a priori stochastic models are also considered. As a concrete example, a deconvolution problem arising in wellbore induction measurement of conductivity is examined. >