为提高多相流CT系统重建图像质量,提出基于Delaunay三角剖分的射线层析成像方法。根据模型构造点集,对管道截面进行多尺度Delaunay三角剖分,根据三角形重心在探测阵列上的投影到其最近探测阵列元的归一化距离,确定各剖分单元的投影系数,并采用不同的重建算法进行图像重建。仿真与实验结果表明,该方法可利用少量投影数据重建图像,改善了工业多相流检测中投影数据不完整造成的图像失真;与均匀剖分相比,多尺度剖分明显改进了图像精度和实时性。
In order to improve the image quality of multiphase flow CT system,Delaunay triangulation based CT technique is developed.A point set is constructed based on the model of the imaging region,and multiple dimensioned triangulation for the conduit section is carried out using Delaunay method,then according to the normalized distance between the projection of the triangle gravity center and its nearest detector,the projection coefficient of every triangular cell can be calculated.Thereafter various reconstruction algorithms are utilized to reconstruct the image.Simulation and experiment results demonstrate that the proposed method can reconstruct image from limited projection data,thus can improve the poor image quality caused by the incompleteness of the projection data in industrial multiphase flow measurement,which proves that multiple dimensioned triangulation is superior to uniform tessellation both in image accuracy and real time performance.