用一种新滤波函数进行CT图像的局部重建(英文)  

Local Tomography with A New Filter

作　　者：徐茂林[1] 邱钧[1,2] 范惠荣[1] 张兆田[3] 李兴东[4] 

机构地区：[1]北京信息工程学院应用数学研究室 [2]中国科学院研究生院华罗庚应用数学与信息科学研究中心北京100039 [3]国家自然科学基金委员会信息科学部 [4]中国计量科学研究院

出　　处：《CT理论与应用研究（中英文）》2004年第1期54-58,共5页Computerized Tomography Theory and Applications

基　　金：This project Granted by National Natural Science Fund (No.69972047) and Project Granted by National Natural Science Fund (No.69931010)

摘　　要：本文阐述了用一种新型滤波函数进行CT图象局部重建的原理,并给出了这种重建的二次误差估计。本文选用一种新的滤波函数,它的旁瓣衰减很快,经证明在可能选取的滤波函数中,它的衰减是最快的。新滤波函数由于空域的迅速衰减性,因此具有良好的局部重建性质。同时保持该滤波函数的其它特性,能够有效地抑制噪声,补偿混迭效应,并缓解Gibbs效应。应用新滤波函数,忽略远离计算点的数据,从而直接用CBP实现图像局部重建。在计算机上对模拟和实测数据进行局部重建的实验结果表明本文建议的图像局部重建算法简便快速,重建图象Gibbs效应减小并与用全部数据的CBP重建图像比较仍具有较高的空间及密度分辨率。In this paper ,the principle of local tomography algorithm by means of a new filter is expounded and the second error estimation is presented. The new filter with its side-peaks declining rapidly is adopted. Its decline rate is also proved to be the fastest among all the possibly selected filters, which will be detailedly discussed in other papers. The new filter features itself with favorable local tomography characteristics and remains its original features unchanged. The utilization of the new filter neglects the data far from the calculating point so as to realize the local tomography simply by CBP. The numerical experiment results of local reconstruction based on the simulation and experimental projection data show that the local tomography algorithm is much simpler and faster, even compared with regular CBP reconstruction using all the projection data. The resolution and Gibbs effects of the reconstructed images get improved.

关 键 词：局部重建 旁瓣的迅速衰减 新滤波函数 时频特性 二次误差估计 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]