项目名称: 显微CT图像的伪影校正方法研究
项目编号: No.61201427
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 电子学与信息系统
项目作者: 张慧滔
作者单位: 首都师范大学
项目金额: 25万元
中文摘要: X射线显微CT技术在生命科学、石油勘探等众多领域有重要应用需求。显微CT图像的伪影是影响其空间分辨率的主要因素。与常规X射线CT相比,消除显微CT图像伪影无论在理论上还是在技术上都更加困难。而系统的几何参数误差是引起伪影的主要原因。现有方法都假定系统几何参数是可重复的,即两次采样相应的几何参数是相同的。但当空间分辨率达到亚微米时,一些几何参数可能发生随机性变化,因此,"几何参数可重复"的假设不再成立。本项目将从理论与技术上研究亚微米分辨的显微CT图像伪影的有效去除方法。通过加标志物的方法,利用标志物与其投影的匹配关系,构造几何参数的自适应优化迭代算法,正确估计各角度采样下的"复合几何参数";进一步由带有"复合几何参数"的重建公式正确重建出CT图像,达到有效消除几何伪影的目的。研究成果将用于国产微纳米X射线显微CT系统的研制,并尝试进行在石油岩芯三维成像、古生物化石三维成像等方面的典型应用。
中文关键词: CT图像伪影校正;几何伪影;硬化伪影;非线性优化;约束优化
英文摘要: X-ray micro-CT technology has been used in many fields, such as life sciences, oil exploration and so on. Micro-CT image Artifact is one of the most important factors in degrading the CT image quality. Yet compared with conventional X-ray CT, reduction of artifacts in high resolution micro-CT reconstructions is more difficult, whether in theory or in technology. While the error of the geometric parameters is a major cause of micro-CT image artifacts.Traditional methods assume that the geometric parameters of the CT system are repeatable, i.e. the geometric parameters of two sampling are the same.However, some parameters of the sub-micron resolution micro-CT may be random,so the assumptions that the parameters can be repeated no longer hold.The project will study some effective methods to remove image artifacts in the sub-micron resolution micro-CT both in theory and in technology. Based on markers and their projection matching relations, we construct adaptive optimal iterative algorithm used to estimate the geometrical parameters. The iterative algorithm can correctly estimate the "complex parameter" in each projection. Further CT images are reconstructed correctly by the reconstruction formula with complex geometric parameters. This method can be efficient to eliminate the geometric artifacts. The research resu
英文关键词: CT image artifact correction;geometric artifact;beam hardening artifact;nonlinear optimization;constraint optimization