Material Classification from Imprecise Chemical Composition : Probabilistic vs Possibilistic Approach

Abstract

In this paper we propose a method of explainable material classification from imprecise chemical compositions. The problem of classification from imprecise data is addressed with a fuzzy decision tree whose terms are learned by a clustering algorithm. We deduce fuzzy rules from the tree, which will provide a justification of the result of the classification. Two opposed approaches are compared : the probabilistic approach and the possibilistic approach.

Non-intrusive inspection of cargo containers using the C-BORD Rapidly Relocatable Tagged Neutron Inspection System

“…Further refinements of the data processing are foreseen to improve material identification, as a function of depth in the cargo container, in particular with organic cargo materials. Moreover, the artificial intelligence techniques employed in the material classification [6] will be improved.…”

“…If anomalies cannot be cleared in the first-line, the second-line technologies (RRTNIS or photofission) are selected and employed, depending on the density and atomic number of the anomaly. The RRTNIS is used to detect contraband of explosives, illicit drugs and chemical agents in specific voxels within the container [1,[3][4][5][6]8]. Voxels are elementary volume units, i.e.…”

“…The European project C-BORD (effective Container inspection at BORDer control points) was aimed at developing and testing a comprehensive cost-effective solution for the generalized inspection of large-volume freight containers [1][2][3][4][5][6][7][8][9][10]. In order to contribute to the protection of EU borders, the C-BORD collaboration developed a framework of Non-Intrusive Inspection (NII) technologies.…”

Sea container inspection with tagged neutrons

“…Further improvements of data processing are now foreseen in the frame of a new H2020 project called ENTRANCE, which will start in October 2020. In particular, artificial intelligence with fuzzy logic trees will be used to improve material classification [67]. In addition, recent studies showed that, although no hydrogen peak can be measured using tagged neutrons, the use of artificial neural networks allows assessing hydrogen concentration using the APT, thus improving greatly the capability to identify organic materials [68].…”