Artificial Lossy Backgrounds to Improve Linear Electromagnetic Imaging Inside PEC Enclosures

Abstract

Herein, we propose the use of artificial (or numerical) backgrounds for electromagnetic inversion inside conducting enclosures. Such enclosed imaging systems are used both due to external constraints (e.g., grain bin imaging), or in order to provide external shielding and easy-to-model boundary conditions (e.g., in stroke or breast imaging systems). We use the orthogonality sampling method (OSM), although the use of such artificial backgrounds is independent of the imaging method. Using a set of 2-D simulated examples, we show that the performance of the OSM is generally poor inside such perfect electric conductor (PEC) enclosures if the background material surrounding the target is lossless. However, we show that the performance of the OSM can be greatly improved through the use of artificial lossy backgrounds. With such artificial backgrounds, the imaging experiment can take place in a lossless PEC enclosure, but one can compute the scattered fields relative to an arbitrary lossy background, then perform the OSM inside that artificial lossy enclosure. The introduction of loss in the background problem can improve the performance of OSM when the average target permittivity is approximately known.