基于极坐标变换去除计算机层析图像环形伪影  

Ring Artifacts Correction of Computerized Tomography Image Based on Polar-Coordinate Transform

作　　者：张国强[1,2] 周虎[3] 和友[1] 刘慧强[1,2] 王玉丹[1,2] 薛艳玲[1] 谢红兰[1] 肖体乔[1] 

机构地区：[1]中国科学院上海应用物理研究所,上海201800 [2]中国科学院研究生院,北京100049 [3]中国科学院南京土壤研究所,江苏南京201008

出　　处：《光学学报》2012年第5期317-322,共6页Acta Optica Sinica

基　　金：国家973计划(2010CB834301);国家自然科学基金(10805071;10705020;11105213;31100680)资助课题

摘　　要：探测器像元响应的不一致、光源能不稳定等因素使得计算机层析(CT)图像中含有较多的环形伪影,严重降低了图像质量,影响了图像的三维重建和量化分析,为此提出了基于极坐标变换与傅里叶变换后低通滤波的算法去除环形伪影。通过极坐标变换将直角坐标下的环形伪影转化为极坐标下的线性伪影,然后对线性伪影图像进行傅里叶变换获得频谱图像,进而设计二维低通滤波器进行滤波处理,最后通过傅里叶逆变换与坐标逆变换获得校正后的图像。利用Matlab软件,编写程序对算法进行验证,结果表明,该算法能够有效地去除环形伪影,使图像内部细节清晰可见,并且保护了图像边缘信息,提高了图像的信噪比;另外,使用该方法处理100张切片图像只需3.5min,可满足批量处理的需求。Ring artifacts in computerized tomography（CT） images are introduced by the nonuniform response of the detector elements and the instabities of light source,which greatly influence three-dimensional reconstruction and quantitative analysis.A correction algorithm based on polar-coordinate transform and Fourier transform is proposed.Through polar coordinate transform,ring artifacts are transformed to line artifacts.Then,Fourier transform is applied to the line artifacts image.Furthermore,the frequency-domain filtering processing by a two-dimensional low-pass filter is taken.The corrected image is got by the inverse transform process of both Fourier and coordinates.Experiments are carried out by Matlab software,the result indicates that the proposed algorithm not only eliminates the ring artifacts effectively,but also keeps the image edges and increases the signal-to-noise ratio（SNR）.In addition,the correction speed proves that it can be applied for batch processing.

关 键 词：X射线光学 环形伪影 线性伪影 极坐标变换 傅里叶变换 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] O434.1[自动化与计算机技术—计算机科学与技术]