少投影钢管截面CT重建研究  

Research on Reconstruction of Steel Tube Section from few Projections

作　　者：彭帅军[1] 邬海峰[1] 王凯[1] 

机构地区：[1]清华大学核能与新能源技术研究院,北京100084

出　　处：《核电子学与探测技术》2007年第5期882-885,共4页Nuclear Electronics & Detection Technology

摘　　要：通过钢管截面CT图像,可以获取其大部分的参数指标,全面检测钢管质量。但是CT图像重建往往需要大量投影数据,这些数据的采集和计算非常耗时。为了加快钢管检测的速度,本文研究了少投影条件下钢管截面的图像重建算法并通过模拟数据验证了其重建的效果。我们分别用ART、MAP和OSEM这三种迭代重建算法对仿真模型进行了重建试验,并根据钢管本身的先验知识分布对算法做了改进,使重建图像质量有了明显的改善。模拟试验的结果表明,ART、MAP和OSEM算法可以实现少于20组投影数据下钢管截面较为精确的重建,并且可以做到10组投影下的近似截面重建。Most parameters of steel tube can be acquired from CT image of the section so as to evaluate its quality.But large numbers of projections are needed in order to reconstruct the section image,so the collection and calculation of the projections consume lots of time.In order to solve the problem,reconstruction algorithms of steel tube from few projections are researched and the results are validated with simulation data in the paper.Three iterative algorithms,ART、MAP and OSEM,are attempted to reconstruct the section of steel tube by using the simulation model.Considering the prior information distributing of steel tube,we improve the algorithms and get better reconstruction images.The results of simulation experiment indicate that ART、MAP and OSEM can reconstruct accurate section images of steel tube from less than 20 projections and approximate images from 10 projections.

关 键 词：钢管检测 CT成像 少投影 重建算法 

分 类 号：TL99[核科学技术—核技术及应用]