有限角逆向螺旋锥束CT扫描与图像重建

项目名称： 有限角逆向螺旋锥束CT扫描与图像重建

项目编号： No.61271313

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 无线电电子学、电信技术

项目作者： 曾理

作者单位： 重庆大学

项目金额： 80万元

中文摘要： 本课题研究改进移动式C型臂锥束CT的扫描方式和图像重建算法，可应用于在役管道的无损检测。由于检测现场条件限制，C型臂（包括射线源和探测器）的旋转角度有限，往往不能形成通常的螺旋扫描。针对长管道检测问题，提出通过C型臂的平移和往返圆弧扫描运动，得到管壁的有限角逆向螺旋锥束扫描的投影数据。对对应的离散线性系统用奇异值分解方法分析解的稳定性，并分析相关迭代方法的收敛性。针对初步重建图像亮度不均、噪声较大且存在伪影等问题，提出鲁棒的几何主动轮廓线图像分割模型，用以改进全变分最小化图像重建算法。针对管壁腐蚀检测的需求，结合改进的局部滤波器，直接从投影数据重建管壁轮廓。本课题提出的扫描方式及重建算法，能够适应在役管道现场检测条件，成像质量好，具有算法理论创新和重要的实际意义。

中文关键词： 计算机断层成像；有限角；逆向螺旋；图像重建；图像分割

英文摘要： The scan way and image reconstruction algorithm of C-arm cone-beam CT are researched and updated, in order to apply in the nondestructive testing of service pipeline. Because of the condition limit in testing field, the rotation angle of C-arm (include ray source and detector) is limited, the usual helical scan may not be formed. Aim at the testing of long pipeline, the projection of limited-angle reverse helical cone-beam scan is gotten by the translation and back-forth limited arc scan of C-arm. The solution stability of the discrete linear system is analysed by singular value decomposition. And the convergence of the iterative algorithm is analysed. Aim at the problem of preliminary reconstruction image with inhomogeneities, noise and artifact, the robust image segmentation model with region-based active contour is presented, in order to update the image reconstruction algorithm with total variation minimization. According to the requirement of the nondestructive testing of service pipeline corrosion, the contour of shell of pipe can be reconstructed directly from projections combining with an updated local filter. Proposed scan way and reconstruction algorithm are adapt to the condition in testing field for service pipeline. The quality of image is good. Thus, there are algorithm update and important practi

英文关键词： Computed Tomography；limited angle；reverse helical；image reconstruction；image segmentation