The use of artificial neural networks to distinguish naturally occurring radioactive materials from unauthorized radioactive materials using a plastic scintillation detector

Document Type: Original Article


Faculty of Nuclear Engineering‎, ‎Shahid Beheshti University‎, ‎Tehran‎, ‎Iran


‎Distinguishing naturally occurring radioactive (e.g. ceramics‎, ‎fertilizers‎, ‎etc.) from unauthorized materials (e.g. high enriched uranium‎, ‎Pu-239‎, ‎etc.) to reduce false alarms is a prominent characteristic of radiation monitoring port‎. ‎By employing the energy windowing method for the spectrum correspond to the simulation of a plastic scintillator detector using the MCNPX Monte Carlo code together with an artificial neural network‎, ‎the present work proposes a method for distinguishing naturally occurring materials and K-40 from four unauthorized sources including high enriched uranium and Pu-239 (as special nuclear materials)‎, ‎Cs-137 (as an example of dirty bombs)‎, ‎and depleted uranium‎.


  • Energy windowing is used to distinguish naturally occurring from unauthorized radioactive.
  • High enriched and depleted uranium, Pu-239, and Cs-137 are used as unauthorized radioactives.
  • Energy windows are obtained for four unauthorized radioactive materials.
  • An arti cial neural network is used for processing the energy windowing outputs.