CT图像环状伪影校正方法  

Ring artifact correction for industrial CT images

作　　者：李俊江[1] 胡少兴[1] 李保磊[1] 杨民[1] 

机构地区：[1]北京航空航天大学机械工程及自动化学院,北京100083

出　　处：《北京航空航天大学学报》2007年第11期1378-1382,共5页Journal of Beijing University of Aeronautics and Astronautics

基　　金：国家自然基金资助项目(60402042);北京市科技新星资助项目(2005A14)

摘　　要：CT图像环状伪影是由探测器像元响应不理想等因素造成的,它导致CT重建图像的降质,影响了图像的后续处理和量化分析.提出了一套完整的算法完成CT伪影校正:基于探测器像元通道响应不一致性的校正,通过偏置和增益校正对硬件响应不一致性进行补偿;基于投影正弦图的校正,通过数字滤波提取正弦图上残余的线状伪影,属于CT图像重建前的一种预处理校正方法;基于重建图像的校正,通过坐标变换将直角坐标系中的环状伪影变为极坐标系中的线状伪影进行处理,属于一种图像后处理方法.这些算法已经用于基于YX-LON XRS232 9″图像增强器的CT重建,在保证图像质量的基础上环状伪影得到有效校正.Ring artifacts in computed tomography （CT） images are introduced by imperfect pixel channel in a detector. Ring artifacts degrade the CT image, hamper post processing and prohibit quantitative analysis. A set of algorithms that enable virtually complete ring artifact correction were proposed. Correction for pixel channel response nonuniformity was that offset and gain correction were done to compensate for response nonuniformity in the detector. Correction for projection sinogram data was that the vertical line artifacts in the sonogram were removed by a computationally efficient filter. Correction for already-reconstructed images was that the ring artifacts in the Cartesian coordinate were transformed into the vertical line artifacts in the polar coordinate and removed as sonogram correction. The procedures were created to deal with data acquired from an YXLON XRS232 9＂ image intensifier. Ring artifacts were reduced significantly with minimal to negligible contamination of the underlying data.

关 键 词：图像增强器 计算机层析成像 环状伪影 正弦图 响应不一致性 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]