Broadband Electromagnetic Detection and Discrimination of Underwater Uxo

The electromagnetic induction (EMI) response to metal objects in the marine environment,<br>including UXO and clutter, is significantly altered by the conductive seawater. Two effects can be<br>observed – propagation effects associated with the seawater skin depth that attenuate and phase rotate<br>the response, and the addition of the response associated with the perturbation of the flow of currents<br>induced in the surrounding water, referred to as the current channeling response (CCR). These effects<br>distort the multi-frequency spectral response of the target from the normal free-air eddy current response<br>(ECR, including magnetic polarization for ferrous targets).<br>If the spectral character is used to classify or identify the target, these effects must either be<br>incorporated into the model or library used as the reference, or they must be minimized so that they can<br>be ignored. Since the relative amplitude of the CCR to ECR depends on variable conditions, including<br>sensor-target geometry, seawater salinity, and the surface condition of the target (paint or corrosion can<br>result in a negative CCR by forcing background currents around the target), explicitly accounting for<br>CCR is difficult and subject to knowledge of these variables. Accounting for propagation effects also<br>depends on knowledge of sensor-target distance. Thus, restricting operational parameters to minimize<br>the effects of the seawater may be the most straightforward and reliable approach.<br>In order to quantify the CCR and propagation effects over the full range of operational<br>parameters, we have developed some analytical models for special cases and we have performed<br>controlled underwater experiments with spheres to compare with the sphere models, and UXO<br>surrogates (pipes). Another algorithm that we developed is the perfectly conducting or insulating<br>spheroid in a conductive medium, which helps quantify the effect of aspect ratio and orientation on CCR<br>and ECR at the inductive limit and for insulated targets.<br>As a preliminary test on the potential application of EMI based discrimination, we have<br>performed our simple single-point spectral matching algorithm on underwater data using a free-air<br>library. By limiting the frequency range to 12 kHz, and target-sensor geometries to ranges with high<br>signal/noise (inherently limits lateral offset distances), the ECR dominates over CCR, and propagation<br>effects are not too great (wave induced noise also is less at lower frequencies). Under these constraints,<br>the matching algorithm performance is reasonably effective at identifying the target for a small suite (5)<br>of items in the library. Identifying clutter, based on goodness-of-fit, is more difficult, but will work for<br>many clutter items.