One of the greatest impediments to the efficient cleanup of unexploded ordnance (UXO) sites is the prohibitive cost of excavating all geophysical anomalies. Several data processing techniques for geophysical survey data have been developed to discriminate between UXO and non-UXO items. The objective of this project was to evaluate the potential of the Standardized Excitation Approach (SEA) and the Surface Magnetic Charge (SMC) approach as modeling techniques for the inversion of UXO sensor data. These approaches were derived from the Method of Auxiliary Sources (MAS) approach.

Originally, the researchers planned to apply the MAS approach to invert UXO sensor data because the ability to model over such a large frequency range made it an attractive modeling technique for joint inversion of multiple data sets. However, the computational time required by the MAS to calculate sensor data prevented the researchers from applying the MAS for inversion purposes. The project was revised to evaluate the SEA and SMC approach because they are derived from the MAS approach. Although the SEA and SMC are both surface magnetic charge distribution techniques, they represent two fundamentally different approaches to UXO discrimination. The SEA is used as part of a library or hypothesis testing technique, and the SMC is part of a parameter estimation/statistical classification technique.

The SEA and SMC are relatively new modeling techniques for UXO discrimination. Therefore, the project investigated some fundamental, as well as practical, characteristics of the forward model. These included the accuracy with which the methods could model sensor data, the speed to carry out the forward modeling, and the type of discrimination algorithms amenable to each of the forward modeling methods. For the SEA, the project team researched the ease with which the sources could be derived for a particular target. For the SMC, the project team researched if the surface magnetic charge distribution was a successful discriminant, and, if so, what algorithm would be required to obtain a stable estimate of the magnetic charge.  Several data sets were collected and analyzed as part of this investigation, including Geonics EM63 data, Geophex GEM3 data, and combined data sets collected with both sensors.

The SEA and SMC models represent an increased level of sophistication relative to the dipole models that have been used in previous inversion algorithms. The successful development of a general inversion formulation utilizing the MAS has the potential to provide improved discrimination ability when processing magnetic and electromagnetic data. These improvements in discrimination could lead to fewer excavations of non-UXO geophysical anomalies, thus reducing cleanup costs.