Search

The container scanning system based on the elemental analysis of suspect cargo by use of fast 14 MeV neutrons with detection of associated alpha particle has been recently developed under the EU-FP6 “EURITRACK” project. The system is currently set-up in port of Rijeka, Croatia, to be tested and upgraded under the EU-JLS ERITR@C project. Here we present results of analysis performed on 152 screened containers selected by the Croatian Custom and detail descriptive statistical analysis of their shipping manifests. Statistics such as distribution functions of container types, type of packages, type of cargos and average cargo densities were obtained in order to establish the properties of standard cargo traffic. The analyses by using multivariate statistical approach that included principal component analysis (PCA) and between group analysis (BGA) have shown that cargo matrices can be classified as metallic (Fe, Al, Cu, Zn, Ni…), ceramic–glass (Si, O) and organic (C, O, N). In this paper the organic matrices have been analysed in more detail since they are identified as most difficult because of their potential interference with main elements contained in threat materials such as explosives or drugs. Density as a discriminating factor has been included for better recognition of different types of goods.