Identification of Metallic Objects using Spectral Magnetic Polarizability Tensor Signatures: Object Classification

Abstract

The early detection of terrorist threat objects, such as guns and knives, through improved metal detection, has the potential to reduce the number of attacks and improve public safety and security. To achieve this, there is considerable potential to use the fields applied and measured by a metal detector to discriminate between different shapes and different metals since, hidden within the field perturbation, is object characterisation information. The magnetic polarizability tensor (MPT) offers an economical characterisation of metallic objects and its spectral signature provides additional object characterisation information. The MPT spectral signature can be determined from measurements of the induced voltage over a range frequencies in a metal signature for a hidden object. With classification in mind, it can also be computed in advance for different threat and non-threat objects. In the article, we evaluate the performance of probabilistic and non-probabilistic machine learning algorithms, trained using a dictionary of computed MPT spectral signatures, to classify objects for metal detection. We discuss the importances of using appropriate features and selecting an appropriate algorithm depending on the classification problem being solved and we present numerical results for a range of practically motivated metal detection classification problems.

Characterisation and Classification of Hidden Conducting Security Threats using Magnetic Polarizability Tensors

“…This chapter documents and compares probabilistic and non-probabilistic ML classifiers that are appropriate for classifying objects when the features are with MPT invariants, with the goal of developing an ML classifier that could be deployed in a walk through metal detector. The chapter has been adapted from the work presented by the author in [135]. The novelties of the chapter are as follows: Firstly, a methodology is documented to create a dictionary of threat/ non-threat objects based on principal tensor invariants.…”

“…This chapter provides a series of illustrative examples to demonstrate how the ROM approach described in Chapter 3 can be combined with an appropriate choice of eigenvalues or tensor invariants in Section 6.3 and sampling at M frequencies to form a realistic dataset for object classification. The chapter has been adapted and extended from the work presented by the author in [76,135] and the full open source dataset, MPT-Library [132].…”

“…This chapter documents results for the application of probabilistic and non-probabilistic classifiers that were discussed in Section 8.3 to the dictionary created using method described in Section 8.2.1. This chapter has been adapted from the work presented by the author in [135].…”