Algorithm and Filter Selection in Geophysical Tomography

In geophysical tomography, linear system methods are used to develop a system of equations with the tomogram elements as unknowns. Usually this system of equations is inconsistent due to inaccurate measurements and the assumption that RF waves travel along straight paths in the medium to be reconstructed. In this paper two new algorithms, the distance transposed modified ART (DTM-ART) and the modified distance transposed ART (MDT-ART), are created, examined, and compared with others (such as the ART, modified ART (M-ART), and distance transposed ART (DT-ART)) for improvement and stabilization of the iterative reconstruction of geophysical tomograms. It was found that the MDT-ART algorithm exhibited the most favorable reconstruction characteristics. Six spatial smoothing filters were investigated, some of these for the first time, for their merits in accentuating geophysical image anomalies. Of these, the minimum variance partition average (MVP-AVG) filter seemed to best accentuate anomaly regions in a reconstructed tomogram.