Search

A method based on the combination of the tagged neutron method and the support vector machine (SVM) algorithm was proposed for the fast and effective identification of explosives behind concrete walls. Geant4 was used for modeling the experimental setup. The high degree of agreement in the experimental and simulated gamma spectra from selected time window validated the developed methodology. The difference between the experimental and simulated spectra could be inferred that the difference between the experimental and simulated spectra mainly comes from the difference in the composition of the cement. Geant4 was used for modeling the situation where explosives were hidden behind walls with thicknesses of 5 cm, 10 cm, and 15 cm, and the minimum detectable mass was calculated. The minimum detection limit (MDL) for C element in explosives hidden behind concrete walls is used to determine the minimum mass of explosives that can be detected and identified. The types of explosives hidden inside the wall were identified using SVM. The mass of explosives with 95% identification accuracy was considered to be the minimum measurable mass that can identify the type of explosives hidden within the wall. Cph/N and Cse/N were selected as input vectors to the SVM where Cph is the peak area of the inelastic scattering peak of element C, Cse is the peak area of the first escape peak of element C, and N is the peak area of the inelastic scattering peak of element N. Calculations of the minimum detectable mass of the type of explosive (TNT, AN, RDX) that could be effectively identified behind walls of 5, 10, and 15 cm resulted in 452 g, 539 g, and 893 g, respectively. [ABSTRACT FROM AUTHOR]

The de-broadening algorithm based on RBFNN is proposed to solve the problems of overlapping peaks and ROI selection in the traditional peak area ratio method. This algorithm is combined with the tagged neutron method for the detection of hidden explosives inside packages and Geant4 is applied for modeling the experimental setup. The broadened energy spectra from 3.5 MeV to 6.8 MeV are chosen as input for the prediction of the unbroadened energy spectra from 3.5 MeV to 6.5 MeV. The result shows that the de-broadening algorithm improves the degree of separation of the samples compared to the peak area ratio method. In the reliability validation of the model, the trained RBFNN can effectively predict the samples within the database and has 99 % identification accuracy in a blind experiment. Meanwhile, the established model predicts other kinds of samples, different masses of samples, and untrained experimental spectra, the good prediction results verify the universality of the model. [ABSTRACT FROM AUTHOR]