Processing and Inversion of Magnetic Gradient Tensor Data for Uxo Applications

We present a set of techniques for processing and inverting multi-component magnetic gradient<br>data for UXO detection and discrimination. A magnetic gradiometer for the UXO application, developed<br>jointly by US Geological Survey and Colorado School of Mines, measures the full gradient tensor of a<br>magnetic field. In contrast to total-field intensity measurements, gradient data have the advantage that<br>they are less affected by common-mode noise and may carry more useful information about subsurface<br>targets. Multiple tensor components also enable the extraction of a coherent common signal to further<br>increase the signal-to-noise ratio. The increased data quality opens new avenues for developing novel<br>approaches to UXO detection and discrimination. We have examined two different processing methods<br>based on the inherent relationship among the multiple tensor components, and we have developed an<br>inversion algorithm for recovering dipole and quadrupole source parameters from tensor data. We<br>outline these algorithms and illustrate the advantages of gradient data over total-field measurements.