Hybrid Full MAS and Combined MAS/TSA Algorithm for Electromagnetic Induction Sensing

摘要

Electromagnetic induction (EMI) sensing, in both frequency and time domains, is emerging as one of the mostpromising remote sensing technologies for detection and discrimination of buried metallic objects, particularly unexploded ordinance (UXO). UXO sites are highly contaminated with metallic clutter so that the major problem is discrimination not detection. This requires high fidelity forward modeling for successful inversion and classification. Recently, the method of auxiliary sources (MAS) has been applied for solving a large range of ultra-wideband (1 Hz-300kHz) electromagnetic induction problems 1 - 6. For a highly conducting and permeable metallic object, when the skin depth becomes small (at high frequency, i.e. induction number >100) the efficiency of the MAS is reduced significantly 6. Other methods are stressed in this region as well. At the same time the Thin Skin Approximation (TSA) 7 - 9, which is based on the divergence free Maxwell's equation in a thin layer, infinitely close to the boundary interior, has shown very accurate results at high frequency/induction number. In this paper a hybrid algorithm, with standard MAS and also the MAS with TSA, is introduced and applied for solving the electromagnetic induction forward problem. Once the broadband frequency domain (FD) electromagnetic response is found it is translated into time domain (TD) using an inverse Fourier transform specialized for the characteristic TD input form. Numerical experiments are performed for highly conducting and permeable canonical objects, illuminated by a magnetic dipole or a loop antenna. These tests indicate that an algorithm using either the full MAS or MAS-TSA formulation, where appropriate, should provide a simulator that is applicable and efficient enough for fast 3-D solutions on a PC, under all conditions across the EMI band in both frequency and time domains.