Electromagnetic imaging inside metallic enclosures using the normal boundary field components

Summary form only given. In recent years, electromagnetic imaging inside chambers with metallic walls has been investigated both theoretically and experimentally. Potential advantages in utilizing these chambers include: (i) shielding the inside of the imaging chamber from outside noise, (ii) better signal-to-noise ratio that may improve the resolution of the imaging modality, (iii) easier system modeling in comparison to open-boundary problems, (iv) the ability to use a lossless matching medium which means more energy is delivered to the target.First, an overview of the research conducted in the area of electromagnetic imaging inside metallic enclosures at the University of Manitoba is presented. This includes two-dimensional transverse magnetic (TM) and transverse electric (TE) microwave tomography (MWT) inside metallic chambers with different boundary shapes (A. Zakaria et al., IEEE TAP, 59, 2012); MWT inside rotating conductive chambers (P. Mojabi and J. LoVetri, IEEE TAP, 59, 2011); MWT using different metallic enclosures simultaneously; and methods and algorithms used to ease studying and performing electromagnetic imaging inside metallic enclosures (A. Zakaria et al., Inverse Problems, 26, 2010). Next, a novel approach for electromagnetic imaging in metallic enclosures is introduced and investigated. The new method utilizes normal field component measurements near the metallic chamber walls to perform imaging; near the chamber boundary the normal electric field components are dominant while the the tangential components vanish. The study is performed both synthetically and experimentally. Using an in-house parallelized full-vectorial electromagnetic finite-element solver, various chamber configurations are modeled and used to collect synthetic datasets (A. Zakaria et al., PIER, 147, 2013). The data are inverted using the finite-element contrast source inversion (FEMCSI) method. The goal of the synthetic study is to understand the effects of frequency-selecti- n, number of observation points, and transceiver modeling on the imaging results. Experimentally, normal electric field data are collected from two configurations. The first setup consists of an air-filled circular metallic chamber with an open-top; within the chamber 24 antennas distributed in three layers are used to measure the normal electric field component near the chamber walls. In the second experiment, electromagnetic imaging inside a grain-bin storage facility is performed. The bin is an enclosed metallic chamber of 4.7 m radius and 7.5 m height. The data are collected using 12 monopole antennas normal to the bin walls.