Automated neutron dosimetry with bubble detectors using CNN-based object detection

Bubble detectors are widely used for measuring neutron yields in nuclear fusion and various neutron sources, as well as in monitoring personal neutron radiation dose. Manual counting of high-density bubbles in neutron detectors is labor-intensive and prone to error, while conventional algorithms struggle with bubble occlusion. We developed an automated dosimetry system integrating CNN-based object detection with a multi-view tracking algorithm to address these challenges. The system achieved over 95% relative accuracy compared to expert manual counts. By leveraging multi-angle imagery, it effectively resolves bubble overlapping and occlusion in three-dimensional volumes. Successful deployment at the XingGuang III laser facility validated the system’s robustness under high-yield conditions, reducing processing time from several minutes to approximately 1 s per image. This automated framework significantly improves the precision and efficiency of neutron diagnostics, providing a reliable, bias-free solution for large-scale data analysis in nuclear fusion and radiation protection experiments.