The hit position reconstruction algorithm for neutron and gamma imaging detectors based on artificial neural networks

Neutron and gamma imaging have continuously expanded applications in nuclear safety, national security, and materials characterization. The hit position reconstruction algorithm is a key issue that constrains the image fidelity and accuracy. This work has developed a two-dimensional planar neutron and gamma imaging system based on a monolithic lithium glass scintillator and a silicon photomultiplier array. The results from the three metrics position nonlinearity response, flood image uniformity, and useful Field-Of-View demonstrate that the proposed artificial neural network (ANN) method significantly advances over traditional reconstruction methods. Imaging results from both the ‘720’ and ‘SCU’ models confirm the ANN method's superior reconstruction quality. In addition, the ANN method achieves a systematic neutron imaging spatial resolution of approximately 0.47 mm for the ‘T’ model. These algorithms are implemented in experimental imaging system, the ANN method maintains acceptable image quality although certain noise artifacts are found, confirming its applicability in both simulated and experimental settings. This work demonstrates that the ANN method significantly enhances positioning accuracy and computational efficiency, resulting in superior neutron/gamma imaging quality.