Estimation of radiation source distribution using machine learning with γ ray energy spectra

Abstract

. A method is proposed for estimating the original radiation source distribution by machine learning using the dose and energy spectrum of γ rays emitted from radiation sources placed at various positions. This method does not require complicated parameter set-tings and can be also applied when there is a shield liked Pb between the radiation source and the measurement point. The estimation results displayed the original radiation source distribution with high accuracy. It is expected to be used for decontamination and decommissioning by developing this method.