Background

Differentiating between scrap objects that could be left in the ground and hazardous objects that must be removed is a problem in the remediation of unexploded ordnance (UXO) from former military sites. The extremely high cost of site remediation may be vastly reduced by discriminating the scrap from the UXO. Over the past few years, much progress has been made in utilizing arrays of electromagnetic induction (EMI) instruments to obtain more detailed information about anomalous conducting objects.

The MetalMapper (MM) is an advanced transient electromagnetic (TEM) system for application towards the detection and characterization of UXO. The antenna configuration includes three orthogonal transmitter loops and seven tri-axial receiver loops. The system can be deployed in mapping (or detection configuration) where it acquires data along profiles while the antenna platform is in motion. However, the most important benefit of the elaborate antenna configuration is that it permits the characterization of a buried metallic target from measurements at a single spatial point located above the target. This system is being commercialized by Geometrics, Inc.

Objective

The objective of this project was to develop and demonstrate the MM, capable of estimating detailed parameters about objects being sensed. This project leveraged its design and research from two sensor technologies; the Advanced Ordnance Locator (AOL), developed by the Naval Explosive Ordnance Disposal Technology Division; and the Lawrence Berkeley National Laboratory’s Berkeley UXO Discriminator (BUD) developed under ESTCP Project MR-200437.

Demonstration Results

In 2008, the MM was demonstrated at the Standardized UXO Technology Demonstration site located at Aberdeen Proving Ground (APG), Maryland. In 2009 and 2010, the MM participated in live site demonstrations at the former Camp San Luis Obispo (SLO), California, and the former Camp Butner, NC.

The performance improved from one demonstration to the next. This can be attributed to improvements in the ability to identify and extract the important discrimination features from the static target data and in an improved understanding of the technology of decision theory and pattern recognition. At APG, a discrimination score was achieved at the operating point of a Probability of Detection (Pd) of approximately 90 percent at a Probability of False Positive (Pfp) of approximately 10 percent where the low Pd primarily represented deep targets that were not detected by the MM.  At SLO, discrimination scores were based only on detected targets. The discrimination score at the operating point was a Pd of 98 percent with a Pfp of approximately 5 percent.  Using the same data, other demonstrators generated similar results thus showing that discrimination performance is not tied to a particular discrimination methodology.  At Camp Butner, the discrimination scores were good, while those of some other demonstrators were nearly perfect. Therefore, the data quality gathered by the instrument is high, and well suited to support further work in processing of the targets for discrimination.

Implementation Issues

The demonstrated advantage of deploying the MM lies with its ability to discriminate targets and produce a prioritized dig list. With regulatory acceptance, the use of a prioritized dig list would reduce the number of digs by 50 percent or more, depending on how conservative the dig policy was set. This would result in an overall cost savings of 30-40 percent in the cost of geophysical surveying plus digging. Therefore, the high cost for the deployment of the MM is justified on the basis that it can substantially reduce the cost of digging. However, routine deployment of the MM will require not only broad regulatory acceptance, but also that it, or other advanced systems, be specified for use in upcoming UXO remediation projects.