Abstract
The assessment of the detection performance of metal detectors (hand-worn, hand-held, and walk-through models) is based on the ability of the detectors to sound an alarm when presented with a test object. These test objects are typically actual threat items or simulated threat items. The orientation of these test objects with the magnetic field generated by the metal detectors may affect the detectability of the test objects. More importantly, small misorientations of a threat object or simulated threat object may cause the operator to incorrectly attribute to the metal detector a higher performance than it is capable of providing. Consequently, to support accurate and reproducible characterization of the performance of a metal detector, orientation effects should be minimized or eliminated. We discuss the use of spherical test objects to eliminate this potentially serious error in the assessment of the detection performance of a metal detector. In this study, we consider only hand-worn and hand-held metal detectors because of their similarity in size and operation.