Inverse scattering of dielectric cylinders by a second-order Born approximation

Abstract

The inverse problem of the reconstruction of the permittivity profile of a dielectric cylinder from the knowledge of the scattered electric field is examined. The proposed method is based on a quadratic approximation of the nonlinear operator defining the scattered field. This choice overcomes the typical limitations of linear approaches and, due to the quadraticity of the operator involved, makes it possible to discuss and avoid the presence of local minima in the inversion procedure. The main features of the approach are shown for the simple case of dielectric cylinders with only angular dependence of permittivity, pointing out the higher sensitivity of the proposed approach with respect to fast varying profiles and the possibility to increase the spatial resolution when the number of available independent data are sufficiently larger than the number of unknowns.Index Terms-Dielectric profile reconstruction, electromagnetic scattering inverse problems, nonlinear inverse problems.

Tomographic diffractive microscopy: basics, techniques and perspectives

“…Methods based on the optimization of a cost function inherently converge to a minimum, but that may be a local one. To avoid such minima, a quadratic approach may be used, as suggested in [77]. Another efficient method to circumvent the problem of local minima is to apply a regularization procedure, i. e. to introduce a priori information on the object.…”

Tunnel detection and localisation via multi-monostatic radio frequency tomography using magnetic sources

“…From a mathematical viewpoint, subsurface imaging is a special case of inverse scattering problem where, at microwave and radio-frequencies, targets are reconstructed in terms of a spatial map of their electromagnetic features, usually given in terms of dielectric permittivity [10,11], electrical conductivity and, recently, magnetic permeability [12]. Inverse scattering problems are non-linear and illposed [8,9] and their solution under an exact formulation generally requires to set up some optimisation procedure (either deterministic or stochastic) [10, 13 -15].…”

“…Non-linear imaging schemes still suffer from reliability problems owing to the occurrence of false solutions (e.g. the functional to be minimised has local minima, which can trap the optimisation procedure) [10,13] and are computationally demanding in terms of both time and memory resources. Hence, they cannot be used to image electrically large spatial region when the time to compute an image is a constraint and the main aim is to obtain information about the location and geometry of the targets.…”

Fast inverse scattering solutions using the distorted Born iterative method and the multilevel fast multipole algorithm

“…Linear methods such as diffraction tomography 1-3 seek rapid solutions under the assumption that the scatterer does not strongly contrast with the background medium, allowing simple inversion of the scattering operator. Quadratic methods 4,5 seek direct inversion of a second-order approximation to the scattering operator in an attempt to avoid local minima in the solution. Iterative linearized approximations seek a compromise between these two methods, overcoming the assumptions implicit in linear inverse scattering by representing the solution as a sequence of successive linear approximations.…”