基于加权TV正则化的X射线CT系统能谱估计方法

发布时间：2018-04-18 10:01

本文选题：CT成像 + X射线能谱估计　；参考：《光谱学与光谱分析》2017年07期

【摘要】：在CT硬化伪影校正、双能CT图像重建以及CT辐射剂量计算等实际应用中,X射线的能谱信息具有重要的作用。然而,由于透射测量方程组系数矩阵的病态性、X光子的统计涨落和噪声的干扰,使得EM等能谱估计方法难以获得较精细的能谱刻画。针对该问题,提出一种基于加权TV正则化的X射线CT系统能谱估计方法。首先采用能谱能量范围内带有不同K-edge的材料作为体模以降低投影测量方程之间的相关性。然后,利用CT成像系统的几何参数来获取投影测量数据所对应的准确透射长度信息来减小投影方程的测量误差。最后,综合利用透射衰减测量数据的保真性、轫致辐射能谱部分的连续性、特征辐射能谱部分的离散性、能谱的非负性和归一性以及平均有效衰减系数等信息,使用加权TV正则化方法建立目标函数,利用正则化理论中的L曲线准则通过黄金分割参数搜索策略求得最优正则化参数及对应的能谱估计结果。分别通过不同统计波动模型下的仿真能谱和实际测量数据对算法性能进行了验证,结果表明该方法与EM等方法相比,不受初始能谱的影响,有效提高了能谱估计结果的稳定性和准确性。
[Abstract]:X-ray energy spectrum information plays an important role in the correction of CT hardening artifacts, reconstruction of dual-energy CT images and calculation of CT radiation dose.However, due to the ill-conditioned fluctuation of X photon and the interference of noise in the coefficient matrix of transmission measurement equations, it is difficult to obtain a more precise characterization of the energy spectrum by the EM equal-energy spectrum estimation method.To solve this problem, an energy spectrum estimation method for X ray CT system based on weighted TV regularization is proposed.Firstly, materials with different K-edge in the energy range of energy spectrum are used as bulk modes to reduce the correlation between the projection measurement equations.Then, the geometric parameters of the CT imaging system are used to obtain the accurate transmission length information corresponding to the projection measurement data to reduce the measurement error of the projection equation.Finally, the information of fidelity of transmission attenuation measurement data, continuity of Bremsstrahlung energy spectrum, discreteness of characteristic radiation spectrum, non-negative and normalization of energy spectrum, and average effective attenuation coefficient are synthetically used.The objective function is established by using weighted TV regularization method, and the optimal regularization parameters and the corresponding energy spectrum estimation results are obtained by using the L-curve criterion in regularization theory through the golden section parameter search strategy.The performance of the algorithm is verified by the simulation spectra and the actual measurement data under different statistical fluctuation models. The results show that the proposed method is not affected by the initial energy spectrum compared with EM and other methods.The stability and accuracy of the energy spectrum estimation results are improved effectively.
【作者单位】：解放军信息工程大学信息系统工程学院;北京大学重离子物理研究所;
【基金】：国家自然科学基金项目(61372172)资助
【分类号】：O434.1

【相似文献】