ISSN 1004-4140
CN 11-3017/P
刘萍, 黎淼, 张彦琦, 牟晓霜, 张尤峰, 席守智, 张保强. 基于高紧支性移动最小二乘法的骨密度能谱探测技术研究[J]. CT理论与应用研究, 2021, 30(5): 555-565. DOI: 10.15953/j.1004-4140.2021.30.05.03
引用本文: 刘萍, 黎淼, 张彦琦, 牟晓霜, 张尤峰, 席守智, 张保强. 基于高紧支性移动最小二乘法的骨密度能谱探测技术研究[J]. CT理论与应用研究, 2021, 30(5): 555-565. DOI: 10.15953/j.1004-4140.2021.30.05.03
LIU Ping, LI Miao, ZHANG Yanqi, MOU Xiaoshuang, ZHANG Youfeng, XI Shouzhi, ZHANG Baoqiang. Research of Bone Mineral Density Energy Spectrum Detection Technology Based on High-compact Support Moving Least Square Method[J]. CT Theory and Applications, 2021, 30(5): 555-565. DOI: 10.15953/j.1004-4140.2021.30.05.03
Citation: LIU Ping, LI Miao, ZHANG Yanqi, MOU Xiaoshuang, ZHANG Youfeng, XI Shouzhi, ZHANG Baoqiang. Research of Bone Mineral Density Energy Spectrum Detection Technology Based on High-compact Support Moving Least Square Method[J]. CT Theory and Applications, 2021, 30(5): 555-565. DOI: 10.15953/j.1004-4140.2021.30.05.03

基于高紧支性移动最小二乘法的骨密度能谱探测技术研究

Research of Bone Mineral Density Energy Spectrum Detection Technology Based on High-compact Support Moving Least Square Method

  • 摘要: 双能X射线骨密度仪系统具有测量精度高、时间短、剂量低等独特优势,是目前X射线骨密度测定技术的“金标准”。本文采用半导体CdZnTe探测器模块搭建半导体光子计数骨密度能谱测量平台,针对CdZnTe半导体探测器信号特点,使用移动最小二乘算法进行不同材料高低能拟合校正。移动最小二乘算法通过权函数改变目标数据点周围节点对其影响程度,可以让数据点的拟合方向更具灵活性。研究分析移动最小二乘算法对双能X射线骨密度测量数值的数据处理流程,完成基于移动最小二乘高低能拟合校正算法的骨密度测量实验。实验结果表明,本文所设计实现的基于移动最小二乘算法的骨密度双能拟合技术在实际应用中能达到较好的拟合误差精度,其中探测器像素单元在高能条件下拟合平均误差为0.032%,低能条件下拟合平均误差为0.036%。进一步数据分析表明,边缘像素单元与中心像素单元在高低能条件下的拟合误差差异仅为0.012%和0.011%。该算法能够有效提高半导体光子计数探测器的骨密度测量精度,降低探测器边缘像素单元信号不均匀性差异带来的误差影响,对目前线阵或面阵光子计数半导体探测器像素差异对骨密度诊断的影响具有良好的改善作用。

     

    Abstract: The dual-energy X-ray bone densitometer system is the "gold standard" for current X-ray bone density measurement technology, which holds unique advantages such as high measurement accuracy, little time cost and low dosage. In this paper, semiconductor CdZnTe detector module was used to build a bone mineral density energy spectrum measurement platform operated by semiconductor photon. Based on the characteristics of CdZnTe semiconductor detector, we performed the moving least square algorithm on the high-low energy fitting and correction of different materials. The moving least-squares algorithm changes the degree influenced by the surrounding nodes of the target data point through the weight function, which can make the fitting direction of the data point more flexible. The research analyzed the data processing flow of the moving least- squares algorithm performed on the dual-energy X-ray bone density measurement value, and completed the bone density measurement experiment operated by the high-low energy fitting and correction algorithm based on the moving least-squares. According to the experimental results, in this paper we designed and intended to achieve that the dual-energy fitting technology of bone density based on the moving least-squares algorithm can achieve better fitting error accuracy in practical applications, the average fitting error of the detector pixel unit under high-energy conditions is 0.032% while the average fitting error under low-energy conditions is 0.036%. Further data analysis showed that the difference of fitting error between the edge pixels and the center pixels under high and low energy conditions was only 0.012% and 0.011% respectively. This algorithm can effectively improve the bone mineral density measurement accuracy of semiconductor detectors, and reduce the influence of errors caused by the the pixels at the edge of the detector. The photon-counting detectors have an improving effect on the bone density mineral diagnosis influenced by pixel differnce of linear array or plane array photon counting semiconductor detectors.

     

/

返回文章
返回