Algebraic Reconstruction Technique for Neurofuzzy Geotomography

Abstract

Algebraic reconstruction techniques (ART) represent a set of iterative techniques that seek to obtain an estimate of material parameter distribution in which the predicted values of projection agree with the observed values. In this paper we apply a generalized ART to Neurofuzzy Geotomography to reduce the learning time or iteration times. We show the effect of ART by computer simulations. Futhermore, the proposed method is applied to the experimental data collected at a dam site by crossborehole seismic probing.