北京航空航天大学学报 ›› 2009, Vol. 35 ›› Issue (2): 170-173.

• 论文 • 上一篇    下一篇

单圆轨道扫描的Grangeat改进算法

魏东波, 张立凯, 傅 健   

  1. 北京航空航天大学 机械工程及自动化学院, 北京 100191
  • 收稿日期:2008-07-20 出版日期:2009-02-28 发布日期:2010-09-16
  • 作者简介:魏东波(1952-),男,湖北赤壁人,教授,weidongbo1019@yahoo.com.cn.
  • 基金资助:

    国家自然科学基金资助项目(50575015, 50875013); 航空科学基金资助项目(20070951023);北京市教育委员会共建项目

Modified Grangeat algorithm as apply to a circular scan orbit

Wei Dongbo, Zhang Likai, Fu Jian   

  1. School of Mechanical Engineering and Automation, Beijing University of Aeronautics and Astronautics, Beijing 100191, China
  • Received:2008-07-20 Online:2009-02-28 Published:2010-09-16

摘要: 为减少ICT(Industrial Computed Tomography)重建图像的伪影,研究了Grangeat精确重建算法的原理.基于平板探测器依次通过对投影数据沿行、列一阶偏微分,求Radon导以及两步反投影等步骤仿真实现了Grangeat算法.针对单圆轨道锥束扫描投影数据的不完备性,分析了Radon空间的数据缺失情况,提出了通过近邻插值、线性插值、二次三点插值和一元线性回归插值的方法对Radon空间缺失数据进行插值改进该算法的几种途径.计算机仿真结果证明了改进算法的正确性和可行性.

Abstract: To compensate for artifacts in reconstructed images of industrial computed tomography (ICT), Grangeat type exact cone beam reconstruction algorithm was studied. Grangeat algorithm based on flat detector was implemented by the steps of computing first partial derivatives of cone-beam projection data along rows and lines respectively, first partial derivatives of Radon data, back-projection and so on. Due to the missed data of a circular scan orbit, the incomplete Radon space was analyzed. The nearest-neighbor interpolation, linear interpolation, 3-points quadratic interpolation and one-dimensional linear regression interpolation methods were used to estimate missed data respectively, and the effects of different methods on the image quality were examined. The compute simulation results prove the validity of proposed methods.


版权所有 © 《北京航空航天大学学报》编辑部
通讯地址:北京市海淀区学院路37号 北京航空航天大学学报编辑部 邮编:100191 E-mail:jbuaa@buaa.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发