The use of muon scattering tomography for the non-invasive characterisation of nuclear waste is well established. We report here on the application of a combination of feature discriminators and multivariate analysis techniques to locate and identify materials in nuclear waste drums. After successful training and optimisation of the algorithms they are then tested on a range of material configurations to assess the system's performance and limitations. The system is able to correctly identify uranium, iron and lead objects on a ~few \text{ cm} scale. The system's sensitivity to small uranium objects is also established as $0.90^{ +0.07}_{ -0.12}$, with a false positive rate of $0.12^{ +0.12}_{ -0.07}$.
Categories: physics.ins-det hep-ex