A Novel Markov Random Field-Based Clustering Algorithm to Detect High-Z Objects With Cosmic Rays

We have developed a novel algorithm based on Markov random fields that uses cosmic ray muons to detect high-Z material, such as special nuclear material, in large-scale volumes, such as cargo containers. Since the amount of muon scattering is approximately dependent on the Z and the density of the material traversed, strong scattering in a localized area is indicative of high-Z material being present. For scanning purposes in freight harbors and similar, a decision should be made in ~ 1 minute. The performance of our algorithm has been evaluated on a variety of scenarios reflecting the composition of real-life cargo, using simulations tuned with our detector performance; we show that the algorithm can clear 64% of these containers using 60 seconds of cosmic muon exposure, and 88% using 90 seconds, with a run-time of the algorithm between 1 and 5 seconds.