偏置探测器CT重建算法研究及实现
发布时间:2018-07-23 17:38
【摘要】:计算机断层成像技术(Computed Tomography,CT),利用测量不同视角下X射线通过被成像物体的衰减值,重建物体内部二维或三维结构图像。它能够以无损、精确、直观的方式表现物体的内部结构信息,现已在工业检测、医学诊断等领域得到了广泛的应用,是最佳的现代无损检测技术之一。在众多扫描方式中,圆轨迹锥形束扫描,由于其机械结构简单,具有X射线利用率高、扫描速度快、水平分辨率和轴向分辨率一致等优点,是目前最为流行的CT扫描结构之一。 本文介绍了一种偏置探测器CT系统,,与传统的锥形束圆轨迹扫描方式不同,通过将面阵探测器水平平移,半覆盖扫描将视场区域扩大了近一倍。不仅未增加机械复杂度,还大大减少了重建对象受辐射的剂量。 针对这种特殊的扫描结构,本文从CT基础理论和重建基础算法出发,对插值重排偏置重建算法、预加权偏置重建算法进行了详细的分析以及严格的数学推导。比较了两者在偏置重建时的不同特点,插值重排偏置重建计算复杂度高,对于倾斜非中心平面的插值重排会导致空间分辨率的损失,使图像质量下降。预加权偏置重建充分利用了数据的冗余特性,在滤波之前对投影数据进行了加权,避免了计算复杂度的增加。此外,本文分别对两者进行了GPU加速,改善了算法性能,使其更有实际应用价值。最后从扇形束二维重建和锥形束三维重建两方面出发,分别设计了仿真实验和实际系统实验对上述各方法的可靠性以及准确性进行了验证。
[Abstract]:Computed Tomography CT (Computed Tomography CT) is used to reconstruct 2D or 3D structure images of objects by measuring the attenuation value of X-ray through the object under different angles of view. It has been widely used in the field of industrial detection, medical diagnosis and so on. It is one of the best modern nondestructive testing techniques. Among the many scanning methods, circular track conical beam scanning is one of the most popular CT scanning structures because of its simple mechanical structure, high X-ray utilization rate, fast scanning speed and uniform horizontal and axial resolution. In this paper, a bias detector CT system is introduced, which is different from the traditional conical beam circle path scanning method. By translating the plane array detector horizontally, the semi-coverage scan expands the field of view nearly twice. Not only does the mechanical complexity not increase, but also greatly reduce the radiation dose of the reconstructed object. In view of this special scanning structure, the interpolation rearrangement bias reconstruction algorithm and the preweighted offset reconstruction algorithm are analyzed in detail and the strict mathematical derivation is given based on the basic theory of CT and the basic reconstruction algorithm. The difference between them in bias reconstruction is compared. The computational complexity of interpolation rearrangement is high. The interpolation rearrangement of tilted non-central plane will result in the loss of spatial resolution and decrease the image quality. Pre-weighted bias reconstruction takes full advantage of the redundancy of the data and weights the projected data before filtering to avoid the increase of computational complexity. In addition, this paper accelerates the two algorithms by GPU, which improves the performance of the algorithm and makes it more practical. Finally, the reliability and accuracy of the above methods are verified by simulation experiments and practical system experiments from two aspects of sector beam 2D reconstruction and conical beam 3D reconstruction.
【学位授予单位】:西安电子科技大学
【学位级别】:硕士
【学位授予年份】:2014
【分类号】:TP391.41;R814
本文编号:2140144
[Abstract]:Computed Tomography CT (Computed Tomography CT) is used to reconstruct 2D or 3D structure images of objects by measuring the attenuation value of X-ray through the object under different angles of view. It has been widely used in the field of industrial detection, medical diagnosis and so on. It is one of the best modern nondestructive testing techniques. Among the many scanning methods, circular track conical beam scanning is one of the most popular CT scanning structures because of its simple mechanical structure, high X-ray utilization rate, fast scanning speed and uniform horizontal and axial resolution. In this paper, a bias detector CT system is introduced, which is different from the traditional conical beam circle path scanning method. By translating the plane array detector horizontally, the semi-coverage scan expands the field of view nearly twice. Not only does the mechanical complexity not increase, but also greatly reduce the radiation dose of the reconstructed object. In view of this special scanning structure, the interpolation rearrangement bias reconstruction algorithm and the preweighted offset reconstruction algorithm are analyzed in detail and the strict mathematical derivation is given based on the basic theory of CT and the basic reconstruction algorithm. The difference between them in bias reconstruction is compared. The computational complexity of interpolation rearrangement is high. The interpolation rearrangement of tilted non-central plane will result in the loss of spatial resolution and decrease the image quality. Pre-weighted bias reconstruction takes full advantage of the redundancy of the data and weights the projected data before filtering to avoid the increase of computational complexity. In addition, this paper accelerates the two algorithms by GPU, which improves the performance of the algorithm and makes it more practical. Finally, the reliability and accuracy of the above methods are verified by simulation experiments and practical system experiments from two aspects of sector beam 2D reconstruction and conical beam 3D reconstruction.
【学位授予单位】:西安电子科技大学
【学位级别】:硕士
【学位授予年份】:2014
【分类号】:TP391.41;R814
【参考文献】
相关期刊论文 前5条
1 李亮;陈志强;张丽;邢宇翔;康克军;;一种用于小体积偏置探测器锥束CT系统的反投影滤波重建算法[J];CT理论与应用研究;2007年01期
2 傅健,路宏年;扇束X射线ICT偏置扫描方式及其重构算法[J];光学技术;2003年01期
3 傅健;路宏年;龚磊;;锥束射线三维大视场工业CT成像方法研究[J];光学技术;2006年02期
4 傅健 ,路宏年;工业CT大视场扫描成像方法研究(英文)[J];Chinese Journal of Aeronautics;2003年01期
5 ;Micro-computed tomography for small animal imaging: Technological details[J];Progress in Natural Science;2008年05期
相关博士学位论文 前1条
1 郭吉强;锥束工业CT截断投影数据重建和轮廓特征重建算法研究[D];重庆大学;2011年
本文编号:2140144
本文链接:https://www.wllwen.com/yixuelunwen/fangshe/2140144.html