基于投影视角加权的直线CL重建算法

计算机分层成像对板状构件的无损检测有其独特的优势,但由于投影角度有限,导致重建图像存在有限角伪影和分层图像模糊。为提高计算机分层成像图像质量,提出了基于投影视角加权的滤波反投影算法,有效抑制了高密度特征对其它层的干扰现象。首先,根据不同投影视角分层图像间的不相似度,确定不同投影视角的加权系数;而后,对投影进行加权反投影重建;最后,分析了3种不同加权系数对重建图像质量的影响,同时引入层灵敏度曲线的调制度作为定量评价指标,未加权和3种不同加权系数层灵敏度曲线的调制度依次为0.082,0.267,0.290,0.294。实验结果表明,该算法减少了分层图像混叠,层灵敏度曲线的调制度提升了约0.2,重建图像质量显著提升。

Computed laminography and reconstruction algorithm

Computed laminography （CL） is an alternative to computed tomography if large objects are to be inspected with high resolution. This is especially true for planar objects. In this paper, we set up a new scanning geometry for CL, and study the algebraic reconstruction technique （ART） for CL imaging. We compare the results of ART with variant weighted functions by computer simulation with a digital phantom. It proves that ART algorithm is a good choice for the CL system.