Combined Nsmc And Pseudo-Spectral Finite-Difference Method For Inverting A Buried Object Location

There is a wide-range of different inverse scattering methodologies (single and double dipole models, the standardized excitation approach (SEA), the normalized surface magnetic charge model (NSMC), etc) currently being used or developed for discrimination of UXO from non-UXO items. In order to utilize these approaches, first the buried object’s location and orientation have to be inverted, which usually is done by solving a time consuming, ill-posed inverse-scattering problem. In order to avoid the standard ill poised inversion procedure for location and orientation, here the NSMC is combined with a pseudo-spectral finite different (PSFD) method for inverting for the location of the buried object. In the NSMC model, over the entire magneto-quasi static regime, the scattered magnetic field outside the object is reproduced mathematically by equivalent magnetic charges. The amplitudes of the NMC, which is proportional to the primary field, are determined directly from measurement data. At the end the integrated total NSMC is employed for discriminating objects of interested from nonhazardous items. Note that the model takes into account the scatterer’s heterogeneity as well as nearand far-field effects. In addition, the NSMC that are distributed on a surface conformal to the measurement surface, can be used to extend the EMI magnetic field above the measurement surface and to generate spatially distributed monostatic as well as bi-static data. In this paper, a buried object’s location is determined by first extending the secondary measured magnetic field from xoy surface at z=h height to its parallel surface at z=h+dz/2 height. Then the PSFD method is employed in the curl and divergence free Maxwell’s equations to find the scattered magnetic field everywhere below the measurement xoy surface at z=h height, inside the observation volume. From the distribution of this calculated magnetic field the buried object’s location is estimated. To demonstrate the applicability of the combined NSMC and PSFD algorithm, several numerical tests are presented.