Automated object recognition and visualization techniques for muon tomography data analysis

In recent years an increasing interest has grown within the international community towards the prevention of terroristic threatenings. This traffic requires a broader international control and it becomes necessary to enhance systems for borders security. Currently, muon tomography is under study to be used as an experimental technique to be implemented in devices able to perform an efficient scanning to detect radioactive materials inside different kind of systems, like shipping containers or cargos, for borders control. In this paper the authors will focus on the automated identification and visualization of illicit materials inside cargo containers in the context of the Muon Portal project. The track reconstruction is a challenging task and consists in the elaboration of data from the detector planes in order to obtain information on the deflection occurred by the muons within the volume scanned. Latest results of different reconstruction algorithms, such as the well known POCA and the EM-Likelihood, are presented and discussed together with the novel application of clustering algorithms to the data analysis. The usage of these techniques allows to infer the implicit information in the data and clustering algorithms make the tracks reconstruction and the visualization of the containers content be independent from the grid and the 3D-voxels, acting as a filter for a preliminary analysis of the data.