同步辐射X射线同轴相衬成像算法研究及生物医学应用
发布时间：2018-07-28 16:41
【摘要】：传统的X射线吸收成像是利用样品不同部分对X射线的吸收不同而成像的。但对于生物软组织等以轻元素为主的样品，由于其对X射线吸收差别很小，吸收衬度的成像方法很难获得它的结构信息。近年来同步辐射光源的迅速发展使得利用X射线的相位衬度成像成为研究的一个焦点。由于X射线能量在10KeV或更高时，生物组织对X射线的相移是吸收的一千倍以上，所以利用X射线的相位信息成像，将会大大提高成像质量。常见的相衬成像方法中，同轴相衬成像易于实现，应用广泛，而相位恢复算法是同轴相衬成像中重要的一环。本文主要研究了同轴相衬成像原理和相位恢复算法，及同轴相衬成像在生物医学中的应用。 本文首先建立了X射线和物质相互作用的关系模型，然后从光的自由空间传播的菲涅尔-基尔霍夫衍射公式出发，导出了基于维纳格分布的同轴相衬成像模型，介绍了几种相位恢复算法：CTF算法，一阶Born近似算法，TIE算法，Bronnikov算法和MBA算法。并通过模拟实验比较和分析了距离，吸收率，噪声三种因素的变化对不同算法相位恢复结果的影响，得出结论：TIE算法和CTF算法都不耐噪声，，MBA算法可以抑制噪声但是只能用于弱吸收样品。模拟的结果对实验中选取最适的算法有重要的指导意义。然后本文提出了一种基于迭代的改进相位恢复算法，这种改进算法可以改善MBA算法只适用于弱吸收样品的情况，并且可以利用MBA算法较好的抑制噪声的性质，随后通过模拟实验证实了这种算法可以适用于更广的、非弱吸收样品的情形，而且计算复杂度不会显著增加。由于实际的成像样品是复杂多样的，很多情况下吸收性未知，所以这种改进算法有重要的现实意义。 本文在上海光源做了标准物理样品和小鼠脑样品的同轴相衬成像实验，然后利用本文提出的相位恢复算法做相位恢复重构。实验结果表明，同轴相衬成像可以更好的区分软组织不同成分，而且不同组织的重构值之间有较大的区分度，可以获得比吸收成像更多的组织结构信息，实验结果显示出同轴相衬成像在对生物软组织成像上的巨大潜力。
[Abstract]:Traditional X-ray absorption imaging is based on the different absorption of X-ray in different parts of the sample. However, for samples with light elements, such as biological soft tissue, the absorption contrast imaging method is difficult to obtain the structure information because the difference of X-ray absorption is very small. In recent years, the rapid development of synchrotron radiation light sources makes the use of X-ray phase contrast imaging become a focus of research. Because the phase shift of biological tissue to X-ray is more than one thousand times that of absorption when the energy of X-ray is higher than 10KeV, the imaging quality will be greatly improved by using the phase information of X-ray. Among the common phase contrast imaging methods, the coaxial phase contrast imaging is easy to realize and widely used, and the phase recovery algorithm is an important part of the coaxial phase contrast imaging. In this paper, the principle and phase recovery algorithm of coaxial phase contrast imaging and the application of coaxial phase contrast imaging in biomedicine are studied. In this paper, a model of the interaction between X-ray and matter is established, and then a coaxial phase contrast imaging model based on the Wiener distribution is derived from the Fresnel Kirchhoff diffraction formula for the free space propagation of light. This paper introduces several phase recovery algorithms: the first order Born approximation algorithm, the Bronnikov algorithm and the MBA algorithm. The effects of distance, absorptivity and noise on the phase recovery results of different algorithms are compared and analyzed by simulation experiments. It is concluded that both the CTF algorithm and the CTF algorithm can suppress noise but can only be used in weak absorption samples. The results of the simulation are of great significance to the selection of the optimal algorithm in the experiment. Then, an improved phase recovery algorithm based on iteration is proposed. This improved algorithm can improve the MBA algorithm which is only suitable for weak absorption samples, and can make use of the MBA algorithm to suppress noise. The simulation results show that the proposed algorithm can be used in the case of a wide range of non-weak absorption samples, and the computational complexity is not significantly increased. Because the actual imaging samples are complex and diverse, and the absorbability is unknown in many cases, this improved algorithm has important practical significance. In this paper, the coaxial phase contrast imaging experiments of standard physical samples and mouse brain samples are carried out at Shanghai Light Source, and then the phase recovery reconstruction is done by using the phase recovery algorithm proposed in this paper. The experimental results show that the coaxial phase contrast imaging can better distinguish different components of soft tissue, and there is a greater degree of differentiation between the reconstruction values of different tissues, so more information of tissue structure can be obtained than that of absorption imaging. The experimental results show the great potential of coaxial phase contrast imaging in the imaging of biological soft tissue.
【学位授予单位】：上海交通大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：R811;O434.1
[Abstract]:Traditional X-ray absorption imaging is based on the different absorption of X-ray in different parts of the sample. However, for samples with light elements, such as biological soft tissue, the absorption contrast imaging method is difficult to obtain the structure information because the difference of X-ray absorption is very small. In recent years, the rapid development of synchrotron radiation light sources makes the use of X-ray phase contrast imaging become a focus of research. Because the phase shift of biological tissue to X-ray is more than one thousand times that of absorption when the energy of X-ray is higher than 10KeV, the imaging quality will be greatly improved by using the phase information of X-ray. Among the common phase contrast imaging methods, the coaxial phase contrast imaging is easy to realize and widely used, and the phase recovery algorithm is an important part of the coaxial phase contrast imaging. In this paper, the principle and phase recovery algorithm of coaxial phase contrast imaging and the application of coaxial phase contrast imaging in biomedicine are studied. In this paper, a model of the interaction between X-ray and matter is established, and then a coaxial phase contrast imaging model based on the Wiener distribution is derived from the Fresnel Kirchhoff diffraction formula for the free space propagation of light. This paper introduces several phase recovery algorithms: the first order Born approximation algorithm, the Bronnikov algorithm and the MBA algorithm. The effects of distance, absorptivity and noise on the phase recovery results of different algorithms are compared and analyzed by simulation experiments. It is concluded that both the CTF algorithm and the CTF algorithm can suppress noise but can only be used in weak absorption samples. The results of the simulation are of great significance to the selection of the optimal algorithm in the experiment. Then, an improved phase recovery algorithm based on iteration is proposed. This improved algorithm can improve the MBA algorithm which is only suitable for weak absorption samples, and can make use of the MBA algorithm to suppress noise. The simulation results show that the proposed algorithm can be used in the case of a wide range of non-weak absorption samples, and the computational complexity is not significantly increased. Because the actual imaging samples are complex and diverse, and the absorbability is unknown in many cases, this improved algorithm has important practical significance. In this paper, the coaxial phase contrast imaging experiments of standard physical samples and mouse brain samples are carried out at Shanghai Light Source, and then the phase recovery reconstruction is done by using the phase recovery algorithm proposed in this paper. The experimental results show that the coaxial phase contrast imaging can better distinguish different components of soft tissue, and there is a greater degree of differentiation between the reconstruction values of different tissues, so more information of tissue structure can be obtained than that of absorption imaging. The experimental results show the great potential of coaxial phase contrast imaging in the imaging of biological soft tissue.
【学位授予单位】：上海交通大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：R811;O434.1
【参考文献】
相关期刊论文 前6条
1 高鸿奕,陈建文,谢红兰,朱化风,李儒新,徐至展,朱佩平,袁清习,田玉莲,黄万霞,王[镌
本文编号：2150916
本文链接：https://www.wllwen.com/yixuelunwen/yundongyixue/2150916.html
最近更新
教材专著
