Two-dimensional Born-type velocity inversion using multiple-frequency information

The aim of inverse-profiling problems is to reconstruct the spatial distribution of one or more material parameters of an object from scattered-field data. A well-known method for solving such problems is the distorted-wave Born iterative technique. Since its introduction, this technique been applied by several authors, mainly to one- and two-dimensional geometries. Numerical experiments have shown that the convergence of the procedure depends on the magnitude of the contrast between the configuration and the surrounding medium, and on the frequency content of the incident field. As a consequence, the reconstruction obtained may suffer from band limitations caused entirely by the method of solution. The approach is feasible only when an efficient algorithm is available for the repeated determination of the fields that would be present in a geometry with an estimated distribution of the unknown material parameter. The present authors address both these issues for the two-dimensional tomography problem.