UXO Classification Using Characteristic Modes of the Broadband Electromagnetic Induction Response

Abstract

Electromagnetic induction methods are effective in locating unexploded ordnance (UXO). However, the induction EM instruments that are used for UXO detection generally have limited bandwidths and provide little, if any, information for UXO classification. It is well known that the broadband induction EM response from confined conductors (such as UXO) can be parameterized in the time-domain as a series of damped exponential decay curves, and in the frequency domain as a set of discrete real first order poles and their residues. Characteristic decay time or its equivalent real pole has been shown to be a function of characteristic target dimensions, target conductivity, and relative magnetic permeability. Therefore, parameterization of the broadband EM response in terms of these characteristic modes provides a basis for the classification of UXO anomalies.In this paper we have used a numerical method (Prony) to analyze TEM decay curves to obtain a set of exponential decay time-constants and their corresponding residues. Using a commercially available field data acquisition system, we have acquired fast transient TEM data from UXO. We show that these data can be analyzed and displayed in a way that is simple to understand and useful for classifying the TEM response.

Locating and Determining Dimensionality of UXOs Using Time Domain Electromagnetic Induction

“…Model-based algorithms use either an exact or approximate forward modeling algorithm to determine a set of model parameters needed to replicate the measured responses, and subsequently relating the model parameters to physical parameters (Khadr et al, 1998). One such model-based technique that has been the focus of much recent research is the determination of the time constants of the TEM response, or equivalently the poles of the frequency domain signal, to identify the buried target (Snyder et al, 1999;Baum, 1997;Collins et al, 1999). A method that represents a hybrid of the modelbased and data-based algorithm is under development at Blackhawk Geometrics (Grimm, 2000).…”

Analytical Modeling of Magnetic and Gravity Signatures of Unexploded Ordnance

“…l04.3nT -55.lnT real-time discrimination or follow-on small area discrimination of detected anomalies (Butler et al, 1998;McNeill and Bosnar, 1996;Snyder et al, 2000). If gravity surveys and in particular microgravity surveys (Butler, 1980(Butler, , 1984 can measure the gravitational anomaly produced by buried UXO, the results can be used to estimate the UXO mass excess (and the actual mass, if the soil density and density contrast is known or can be estimated).…”

Investigating the effect of data quality on time domain electromagnetic discrimination

“…The infinite sum is still needed even for an axi-symmetric target (for example Carin et al, 2001). While it is possible to truncate these representations and model the TEM response with a number of exponents (Snyder et al, 1999), the inverse problem to determine the weights and time constants is potentially difficult (Istratov and Vyvenko, 1999). Therefore, our approach is to parameterize L(t) with a simple empirical function defined by a minimum number of parameters that seems to have enough flexibility to replicate all the features of the TEM decay.…”