Feasibility study of machine learning-based discrimination of , and particles from grayscale radiation images

Laffolley, H.; Tsubota, Yoichi   ; Tsuji, Tomoya   ; Honda, Fumiya

In the framework of the decommissioning of the Fukushima Daiichi Nuclear Power Station, the Japan Atomic Energy Agency analyses and classifies a variety of radioactive samples. The objective is to simplify the sample characterization process by developing multipurpose analysis tools that quickly produce results for different types of samples while reducing labor. The development of an analytical device has been started, based on the MiniPIX TPX standard detector, a hybrid semiconductor pixelated radiation detector. This detector creates grayscale images that show the interaction of ionizing particles, where brightness directly indicates energy. The final aim is to build a fast mapping device that generates 2D activity maps, distinguishing between , , and radiation, and includes simple local spectrometry for highly contaminated samples. The shape of the cluster created by the interaction between an incident particle and the semiconductor is typical of the said particle. Thus, eight supervised machine learning models have been trained on a dataset made of 9 features extracted from pure images of , and particles collected from Co, Sr, Cs and Am standard sources. The best models can distinguish the particles with nearly 80% accuracy, reaching 96% accuracy for low-energy rays exposition only, with a processing time of a few microseconds per frame. The identification of particles is 100% accurate.