Multi-sensor explosive detection system using an articifical neural

Abstract

A nuclear detection system and method uses an artificial neural system to efficiently detect explosives in checked airline baggage or other parcels with a high probability of detection (PD) and a low probability of false alarms (PFA). The detection system detects the presence of nitrogen and other elements and their rough density distribution within the object under investigation by performing a nuclear-based analysis of the object. The detection system includes a source of neutrons; an array of gamma ray detectors; at least one neutron detector; means for irradiating the object being examined with neutrons from the neutron source, which neutrons interact with the atomic nuclei of one or more specific elements, e.g., nitrogen, chlorine, or hydrogen present within the object so as to cause elemental-specific gamma rays to be emitted; means for capturing and counting the elemental-specific gamma rays and determining their approximate origin within the object, thereby providing a rough measure of the density distribution of these specific elements within the object; and means for detecting neutrons that pass through the object without substantially interacting with atomic nuclei, thereby providing an image of the density distribution of elements in the object. The invention further includes an X-radiography system for generating a high-resolution two-dimensional density image of the object. An artificial neural system (ANS) provides an efficient means for analyzing the recorded gamma rays, image, and other data so as to accurately discriminate between objects carrying explosives and objects not carrying explosives.