Improved Performance of Energy-Selective Gamma-Ray Imaging System Based on Image Restoration

The application of intense current transient pulse radiation fields encompasses a wide range of fields, from scientific research to industrial application. The generation and application of this radiation have become an important direction of current research. Ray detection technology (such as neutron detection, gamma ray detection, etc.) plays a key role in understanding the working state, physical mechanism, and process of different pulse radiation devices. With a new energy-selective gamma-ray imaging technique based on a magnetic lens, the energy and spatial information of pulsed gamma rays can be obtained simultaneously. However, due to the limitations of angle and energy dispersion, the detection efficiency and spatial resolution of the system cannot be simultaneously optimized, which limits its practical application. To overcome this limitation, a method combining system optimization and imaging restoration is proposed to improve the performance of energy-selective gamma-ray imaging. In this article, the effect of beam-limiting hole size on detection efficiency and spatial resolution in an energy-selective gamma-ray imaging system is analyzed. The imaging dataset of the energy-selective gamma-ray imaging system is constructed using a Geant4 simulation, and a deblurring generative adversarial network (GAN) based on a weighted bidirectional feature pyramid is developed. The results demonstrate that by selecting Dx =60 mm and Dy =30 mm as the new beam-limiting hole size parameters, the detection efficiency of the imaging system can be improved by approximately four times, and the spatial resolution performance of 1.2 mm in the x-direction and 2 mm in the y-direction can be achieved by the proposed algorithm. Furthermore, the reliability of the proposed algorithm is substantiated by experimental verification.