基于能谱滤波分离的多谱计算机层析成像方法

发布时间：2018-06-15 07:27

本文选题：X射线光学 + X射线成像　；参考：《光学学报》2014年10期

【摘要】：随着科技的发展,适用于结构分析的传统单能X射线计算机层析(CT)成像技术,已不能满足现代工业对物质组分区分与鉴定的功能成像需求。这是由于在X射线CT系统中,现有重建算法的单能假设与CT投影的多谱性不一致,导致CT重建质量差,无法组分区分。基于光子计数探测器的能谱分离成像思想,提出了基于能谱滤波分离的多谱CT成像方法,该方法通过在X射线发射端加滤波片的方式,实现能谱滤波分离,并通过变能量成像,获得近似单能的递变能量投影序列;针对滤波后噪声水平较高问题,利用EM-TV重建算法,实现了多谱CT成像,可满足组分区分的需求。仿真实验结果表明,对于密度相近的检测对象,该方法可以满足组分区分的要求。
[Abstract]:With the development of science and technology, the traditional single-energy X-ray computed tomography (CTT) imaging technology, which is suitable for structural analysis, can no longer meet the functional imaging needs of the modern industry for the partition and identification of material groups. This is due to the fact that the single energy assumption of the existing reconstruction algorithms is not consistent with the multispectral feature of CT projection in X-ray CT system, which leads to the poor quality of CT reconstruction and the inability to group subdivision. Based on the idea of spectral separation imaging of photon counting detector, a multispectral CT imaging method based on spectral filter separation is proposed. The method realizes energy spectrum filtering separation and variable energy imaging by adding filter at the X ray emitter. In order to solve the problem of high noise level after filtering, multi-spectral CT imaging is realized by using EM-TV reconstruction algorithm, which can meet the needs of group partition. The simulation results show that the method can meet the requirements of group partition for objects with similar density.
【作者单位】：中北大学信息探测与处理山西省重点实验室;中国科学院自动化研究所中国科学院分子影像重点实验室;
【基金】：国家自然科学基金(61171179,61227003,61301259) 山西省自然科学基金(2012021011-2) 高等学校博士学科点专项科研基金(20121420110006) 山西省回国留学人员科研资助项目(2013-083) 山西省高等学校优秀创新团队支持计划
【分类号】：R814.42

【参考文献】