电阻抗成像中混合罚函数正则化算法的仿真研究

该文将变差函数作为罚函数引入到电阻抗成像的正则化重构算法中，从而提出了一种新的电阻抗成像算法，文中称为混合罚函数正则化算法。与常规Tikhonov正则化算法相比，该算法的突出优点是：在确保重构解适定的同时，提高重构图像的对比度和锐度，且计算量增加不大。仿真对比实验结果显示，新算法所得重构图像目标区域与背景区域之间的边界清晰，定位更加准确，与真实医学图像更加符合，这对EIT重构成像技术早日走上实用化有积极的意义。

Mixed Regularization Algorithm in Electrical Impedance Tomography

In this paper, a variation function as a regularization penalty term has been imported into EIT image restoration, which comes to be a variation regularization algorithm. The key difference to Tikhonov regularization algorithm is that the variation regularization algorithm not only insures the inverse problem of EIT well - posed but also insures the dividing line between the goal region and the background region of restored image clearer. This made the restored image well accord with the fact of medicine. It is significative to utilize EIT technique.