Donghai Fan a, Rui Wu a, Dengke Wei a, Yingrui Li b, Tingting Tan a, Gangqiang Zha b
a State Key Laboratory of SolidifiCation Processing, and MIIT Key Laboratory of Radiation Detection Materials and Devices, Northwestern Polytechnical University, Xi'an, 710072, China
b Research & Development Institute of Northwestern Polytechnical University in Shenzhen, Shenzhen, 518063, China
## Abstract
Gamma-ray coded-aperture imaging technology has important applications in the fields of nuclear security, isolated source detection, and the decommissioning of nuclear facilities. However, artifacts can reduce the quality of reconstructed images and affect the identification of the intensity and location of radioactive sources. In this paper, a gamma-ray coded-aperture imaging method based on primitive and reversed coded functions (PRCF) was proposed to reduce imaging artifacts. Building on this, the PRCF method was improved by integrating energy spectral information collected by the detector. By selecting energy intervals corresponding to characteristic energies of different radioactive sources for data filtering, the imaging capability of the PRCF method was further enhanced for multiple radioactive sources. Through simulation, the selection range of the correction factor in the PRCF method was determined. Single-source and multi-source imaging experiments were conducted using the self-built coded-aperture imaging system based on a CdZnTe pixel detector, and the selection criteria for the energy interval ranges of different radioactive sources were established. Compared with the conventional maximum likelihood expectation maximization (MLEM) method, the improved PRCF method not only effectively reduced artifacts and enhanced the imaging quality, but also ensured the accuracy of imaging results for multiple radioactive sources. Moreover, through imaging experiments using low-activity 137Cs and high-activity 241Am and 133Ba, it was demonstrated that the PRCF method can achieve low-count imaging in complex environments, providing a solution for imaging low-activity radioactive sources.
## Introduction
Gamma-ray imaging technology plays a crucial role in the fields such as nuclear safety monitoring, isolated source search, and nuclear facility decommissioning (Caballero et al., 2018; Vetter et al., 2018; Xia et al., 2018; Yu et al., 2022). Gamma-ray coded-aperture imaging, developed from single-pinhole imaging, is a multi-aperture imaging method that enhances the detection efficiency of gamma rays while maintaining high spatial resolution (Cieślak et al., 2016). Compared to parallel-hole imaging, coded-aperture imaging can achieve a larger field of view with the same detector size (Braga, 2019; Kastis et al., 2000). In contrast to Compton imaging, coded-aperture imaging significantly improves angular resolution, allowing for more precise localization of radioactive sources (Liang et al., 2022). By combining the optical image with the radioactivity distribution map, it can provide an intuitive reference for locating radioactive sources (Paradiso et al., 2018).
However, coded-aperture imaging can be affected by artifacts. When the radioactive source is at a finite distance from the detector, non-parallel incident gamma rays cause the near-field effect, leading to near-field artifacts in the reconstructed image of the radioactive source (Li et al., 2023). A thicker coded-aperture collimator can also introduce thickness artifacts in the reconstructed image (Zhu et al., 2022). The conventional maximum likelihood expectation maximization (MLEM) and cross-correlation decoding methods are difficult to eliminate artifacts (Jeong and Kim, 2021; Mu and Liu, 2006). The dual-mode correction method uses the mask and anti-mask to image the radioactive source separately, then adds the resulting images together to reduce artifacts. This method requires rotating the collimator during the detection process, which not only increases detection time and reduces imaging efficiency, but also required additional mechanical structures (Accorsi and Lanza, 2001). Using multiple random patterns in coded-aperture imaging can help reduce artifacts to some extent. However, the process requires imaging with different collimators separately, resulting in low detection efficiency (Kasuga et al., 2020). Coded-aperture imaging based on convolutional neural networks has shown potential for reconstructing images and removing artifacts under low-count conditions. However, it has some problems such as poor performance in multi-source imaging and high computational requirements (Zhang et al., 2019).
In addition, conventional coded-aperture imaging methods usually use the full energy spectral data for imaging to ensure the detection efficiency. When gamma rays interact with the CdZnTe detector, characteristic X-rays, Compton scattering continuum, and other signals will be generated. In cases of multiple interactions between gamma rays and the detector, theoretically, only the first interaction point should be used for coded-aperture imaging. Other signals are unlikely to reflect the correct position of radioactive sources and are considered invalid events. These invalid events can lead to artifacts in the reconstructed image (Liu et al., 2020; Shy et al., 2020). When imaging involves multiple radionuclides or low-activity radioactive sources, artifacts can seriously interfere with the determination of the shape, intensity, and location of the radioactive source.
Therefore, in order to suppress artifacts and obtain low-noise reconstructed images, a gamma-ray coded-aperture imaging method based on primitive and reversed coded functions (PRCF) was proposed in this paper. Furthermore, the PRCF method was improved by integrating the energy spectral characteristics of CdZnTe pixel detectors, thereby enhancing its imaging capability for multiple radioactive sources. The signal was filtered according to the characteristic energies of the radioactive sources. The data within the corresponding energy intervals were used separately for imaging, and the images were then combined to obtain the final result. The reliability of the improved PRCF method was verified by comparing its imaging quality with that of the conventional MLEM method for multiple radionuclides and low-activity radioactive sources.
CdZnTe Association (CdZnTe.com)
https://www.cdznte.com/thesis/a-gamma-ray-imaging-method-for-multiple-radionuclides-and-low-activity-radioactive-sources.html