基于改进卷积反投影模型的CT成像伪影消除方法

Artifact elimination method for CT imaging based on improved convolution back projection model

针对传统卷积反投影法的CT重建图像容易产生伪影的问题,提出了一种基于改进卷积反投影算法的CT成像伪影消除方法.该方法通过分析给定模板的测量参数,标定CT系统的旋转中心、探测器间距及各探测方向等参数,使用Radon和傅里叶变换得到反投影方程,并利用卷积反投影技术得到每一点的吸收率与投影值的积分关系,使用三次样条插值算法得到较清晰的重建图像.仿真结果表明,所提出的算法具有简明易懂、操作方便和精确度高等诸多优点,相较于传统方法拥有更高的运算效率.

Focusing on the fact that the CT reconstruction image obtained with the traditional convolution back projection method is easy to produce artifacts,an artifact elimination method for the CT imaging based on improved convolution back projection algorithm was proposed.Through analyzing the measurement parameters of a given template,such parameters as the rotation center,detector spacing and detection direction of CT system were calibrated,and the back projection equation was obtained through Radon and Fourier transformation.In addition,the integral relationship between the absorption rate and projection value of each point was obtained with the convolution back projection technique,and the clear reconstructed image was obtained with the cubic spline interpolation algorithm.The simulation results showthat the proposed algorithm has many advantages,such as concise and easy understanding,easy operate and high precision,and has a higher computational efficiency than traditional methods.