Convex backscattering support in electric impedance tomography (Q623331)

The inverse obstacle problem in electric impedance tomography consists of the reconstruction of an unknown inclusion in some object of known homogeneous conductivity from electrostatic measurements at the boundary of the object. Appropriate measurements are, for instance, the electrostatic potential at the boundary of the object when a certain current is injected through the boundary into the object. In this article, the authors investigate the two dimensional inverse obstacle problem with a nonstandard set of data that has been introduced and termed backscattered data: a pair of electrodes is used to derive a given current and record the resulting voltage while being rotated around the object. The notion of ``backscattering'' has been adopted for this setup because of the relation to the ``usual backscattering'' of electromagnetic waves with very low frequencies. This paper reinvestigates a recently introduced notion of backscattering for the inverse obstacle problem in impedance tomography. Under mild restrictions on the topological properties of the obstacles, it is shown that the corresponding backscatter data are the boundary values of a function that is holomorphic in the exterior of the obstacle(s), which allows to reformulate the obstacle problem as an inverse source problem for the Laplace equation. For general obstacles, the convex backscattering support is then defined to be the smallest convex set that carries an admissible source, i.e., a source that yields the given (backscatter) data as the trace of the associated potential. The convex backscattering support can then be computed numerically; numerical reconstructions are included in the article in order to illustrate the viability of the method.