基于基图像TV模型的CT射束硬化校正方法

Beam-hardening Correction in CT Based on Basis Image and TV Model

X射线CT中,X射束硬化导致重建图像中出现伪影,严重影响了图像质量。论文深入分析了射束硬化对原始投影数据的影响,并提出了一种基于基图像TV模型的射束硬化校正方法。该方法首先依据射束硬化的物理特征建立了带有可调参数的初步校正模型;其次,在不同的可调参数条件下,原始投影数据经该模型预处理变换得到多组预处理投影序列;再次,分别对预处理投影序列进行重建得到一系列校正基图像;最后,将基图像的线性组合作为最终重建图像。其中,加权系数的最优解是以最终重建图像的全变分(Total Variation,TV)函数作为目标函数,通过迭代法求得。为验证该算法,对真实的铝柱体模和工业检测件进行了实验,结果表明,该算法对射束硬化引起的杯状和条状伪影均有显著的抑制效果。

In X -ray computed tomography, the beam hardening leads to artifacts and reduces the image quality. It analyzes how beam hardening influences on original projection. Accordingly, it puts forward a kind of new beam- hardening correction method based on the basis images and TV model. Firstly, according to physical characteristics of the beam hardening an preliminary correction model with adjustable parameters is set up. Sec- ondly, using different parameters, original projections are operated by the correction model. Thirdly, the pro- jections are reconstructed to obtain a series of basis images. Finally, the linear combination of basis images is the final reconstruction image. Here, with total variation for the final reconstruction image as the cost function, the linear combination coefficients for the basis images are determined according to iterative method. To verify the effectiveness of the proposed method, the experiments are carried out on real phantom and industrial part. The results show that the algorithm significantly inhibits cup and strip artifacts in CT image.