MURA编码孔成像中γ源分布的数值重建  

Digital Reconstruction of γ-Source Distribution in MURA Coded Aperture Imaging

作　　者：肖洒[1] 兰明聪[1] 党晓军[1] 韦孟伏[1] 张松柏[2] 

机构地区：[1]中国工程物理研究院,四川绵阳621900 [2]北京应用物理与计算数学研究所,北京100088

出　　处：《核电子学与探测技术》2013年第8期913-918,共6页Nuclear Electronics & Detection Technology

基　　金：中国工程物理研究院基金(2011B0302054)

摘　　要：采用Geant4模拟软件模拟不同形状γ源的MURA(19×19)嵌套型编码成像,分别利用两种相关解码算法(精细采样平衡相关和δ解码)和最大似然-期望最大化迭代算法对各个编码像进行数值重建,用对比度-噪声比(CNR)对各种算法的重建图像质量进行比较。结果表明,三种图像重建算法均能为各种不同形状的平面γ源给出清晰、准确的图像,其中MLEM算法重建图像的分辨率和CNR比传统的相关解码要高,该算法可以较好地抑制统计噪声,图像质量高。因此,编码孔成像技术用于γ相机成像是可行的。Coded aperture imaging of various shaped γsources was simulated by Geant4 toolkit, with a mosa-icked MURA（19 ×19）mask.Two correlation decoding algorithms （Finely Sampled Balanced and δdecoding）, as well as the iterative Maximum Likelihood -Expectation Maximization （MLEM） algorithm, were applied for digital reconstruction of the coded images .The qualities of reconstructed images were compared by calculating the contract-to-noise ratio （CNR）.The results showed that each of the algorithms represented a clear , exact image for every single source and that high -quality images with a better resolution and CNR were obtained by MLEM algorithm, as the statistic noise was effectively inhibited .Accordingly, coded aperture imaging was ap-plicable in γcamera.

关 键 词：改进型均匀冗余阵列 编码孔 相关解码 最大似然一期望最大化法 

分 类 号：TL99[核科学技术—核技术及应用]