【作者】 龚绍润；

【导师】 高峰；

【作者基本信息】 天津大学 ， 生物医学工程， 2010， 博士

【摘要】 与基于强度衰减的传统X光成像技术相比,X光相衬成像技术在对软组织进行成像时具有低辐射量,高可见度等优点,因而在过去十年间引起了研究者的广泛重视并得到了迅速发展。本研究中讨论的微焦点源同轴相衬成像方法由于其对于光源限制较小且无需在光路中附加光学器件,因此被视为最具临床应用前景的相衬成像方法。本论文包括以下三方面内容,首先,采用刀口法定量评估相衬成像系统的成像质量;其次,建立数值模拟模型来指导实验室成像系统下各物理参数的最优化配置;最后,采用维纳滤波法来提高相衬成像中干涉条纹对的可见度。实验室构建了基于微焦点源的同轴相衬成像系统,为定量评估该成像系统下成像质量,本文中采用刀口法测量了该系统的调制传递函数。为修正因刀口加工精度和探测器上各像素点响应差异引入的测量误差,文中在对边缘响应曲线采用了分段指数拟合后,又对线扩散函数采用了拖尾截断和指数沿拓。刀口法所获系统调制传递函数与线对卡的离散结果非常接近,说明使用文中刀口法能够有效地测量系统的调制传递函数。为获得最佳的相衬成像结果,需要建立一套准确可靠的数值模拟工具来指导相衬成像系统下各物理参数的最优化配置。本文中数值模拟工具基于菲涅耳-基尔霍夫衍射积分,综合考虑了点源到样本距离、样本到探测器距离、微焦点源尺寸、复色光频谱分布、荧光屏的点扩散函数及探测器上像素点大小对相衬成像结果的影响。数值模拟模型下获得的强度剖面图与相同物理参数配置下实验结果较吻合,验证了该数值模拟模型的正确性和有效性。实际相衬成像结果因受到系统噪声、总的点扩散函数和空间采样率的影响,导致相衬成像下干涉条纹对可见度低。该章中首先通过数值模拟的方法验证了采用维纳滤波器来提高相衬图像可见度的可行性,随后采用维纳滤波器处理了一维情况下的实验数据,结果显示该滤波方法有效地提高了干涉条纹对的信噪比,增强了相衬成像结果的可见度。更多还原

【Abstract】 Compared with the conventional X-ray imaging technique based only onattenuation, X-ray phase contrast imaging, in the field of soft-tissue imaging has suchvirtues as low radiation, high visibility and so on, thus leading to widespread respectand rapid development in the past decade.Without beam restriction or optical components along the beam path, the simplestmethod mentioned in this study, referred to as in-line X-ray phase contrast imaging(XPCi), is considered as the most straightforward XPCi implementation from the labto bedside. Three parts compose the main content of this paper. Firstly, edge methodwas used to quantitatively evaluate image quality from the XPCi system. Secondly, anumerical simulation tool was developed to guide the optimization of parametersunder the XPCi system. Finally, Wiener filtering was used to improve the visibility ofinterference patterns from the XPCi system.An in-line phase contrast imaging system using micro focus X-ray source wasbuilt in the laboratory. For quantitative evaluation of the image quality, edge methodwas used to measure the modulation transfer function (MTF) of the imaging system.In order to correct the measuring errors produced by both flaws in the edge andresponse difference in the pixels, curve-fitting on the different segmentation of edgeresponse function were used, as well as tail-truncating and exponent-stretching on theline spread function. Consistency was obtained between the MTF from the edge andthe discrete value from the test card, which verified the effectiveness of the edgemethod in the MTF measurement.To obtain optimal phase-contrast results, an accurate and reliable numericalsimulation tool is needed for the optimization of parameters under the lab-equippedimaging system. The simulation tool in this study is based on the Fresnel/Kirchhoffdiffraction integral with comprehensive considerations on the source-to-sampledistance, the sample-to-detector distance, the source size, the polychromaticdistribution, the point spread function of fluorescent screen and the pixel dimension ofCMOS detector. Consistency between the profile from the numerical simulation andthat from the experiment under the same parameters has verified the validity and theeffectiveness of the simulation tool.Because of the negative effects such as noise, total point spread function and spatial resolution under the lab-equipped imaging system, the visibility of interferencepatterns was low. Firstly the Wiener filter was used to validate its feasibility inimproving the image visibility under the simulation results, and then this filter wastested under one-dimensional experiment result. The experiment profile after filteringshowed a raise in signal-to-noise ratio, in other words, the visibility of phase contrastimage was improved.更多还原