《高精度X-CT投影成像研究》-毕业论文设计（学术).doc

精品毕业论文 高精度X-CT投影成像研究 摘要 自从1971年第一台CT设备问世以来，计算机断层成像技术（CT）不断取得巨大进步。CT从理论上讲是一个从投影重建图像的反问题。卷积反投影（CBP）、直接傅立叶重建（DF）以及代数重建算法（ART）同为典型的CT重建算法。其中，DF算法在原理上简单，重建速度较快，但是由于缺乏较好的从极坐标系到直角坐标系的插值方法，使得在大多数情况下DF算法重建质量不如CBP以及ART算法重建质量。而由于CBP算法能够重建出质量足够好的图像，同时其耗费的时间也在可以接受的范围内，因此现代CT中的重建算法几乎都是用CBP算法。论文从DF算法入手，试图在投影数据足够的条件下寻找好的网格化方法，改善DF重建算法的质量。在我们的论文中，研究了DF法中应用三种常见插值器方法时，图像重建效果对比。同时，论文也指出了采用一种插值器能使DF法与CBP法等效。最后我们提出了一种新的DF重建算法。在这种算法中采用了对投影数据进行线性调频z变换近似其频谱数据的方法，使得频谱数据更密，同时在DF法的最后一步中采取线性调频z变换求重建图像，减小了插值误差。论文中把这种方法与CBP重建以及ART、SIRT进行了对比，仿真结果显示，本文中采用的方法重建图像的质量至少不比CBP重建以及ART、SIRT重建图像质量差。 关键词：计算机断层成像，直接傅立叶变换重建，卷积反投影，代数重建算法 ，同时迭代重建法 Research on High-Precision X-CT Image Reconstruction FROM PROJECTIONS ABSTRACT Since 1971 when the first CT equipment was made, CT has been making great progress all the time. In theory, CT is the inverse problem of reconstructing an image from its projection data. The convolution back projects (CBP), direct Fourier reconstructs (DF) as well as algebra reconstruction algorithm (ART) are the typical CT reconstruction algorithm. Among them，the DF method is very simple in principle, but because of lacking a good interpolation method interpolating data from the polar coordinate system to rectangular coordinate system which causes that the image reconstructed by DF method is inferior to the image reconstructed by the CBP and ART method in most cases. While the CBP method on the other side can yield to good quality reconstructions and only take a short time , so the CBP method is widely used in modern CT equipment. In this paper, we start with the DF algorithm and try to improve the reconstruction quality under the condition of enough projection data. We investigate three types of interpolators widely used in the DF method and construct the images reconstructed by using these interpolators. We also study an interpolator which can make the DF algorithm be equivalent to the CBP algorithm. At last, we provide a new way to complete the DF method. In this method, we use the C