实现电阻抗层析成象的交错算法

电阻抗层析成象（ＥＩＴ）是继ＣＴ，ＭＲＩ后又一种新的图象重建技术。本文在给出ＥＩＴ的数学物理模型的基础上，提出一种新的实现算法：梯度－记忆下降方向交错算法，与已有的ＥＩＴ算法相比，它具有收敛快、内部分辨力高、不依赖于初始值等优点。模拟计算表明该算法以相当高的精度收敛于真实解，即使在测量数据带有５％的平移误差和５％的随机误差时也是如此。

An Alternate Algorithm for Realizing Electrical Impedance Tomography

Electrical Impedance Tomography(EIT),a new image reconstruction technique,is being developed after X-CT and MRI.Based on mathematical physical model,an algorithm called alternate algorithm of gradient-memory descent direction is proposed in this paper.Compared with popular algorihtms about EIT,it has advantages of high convergent speed,good inner discrimination and independence from initial value.The computer simulation result shows this algorithm converges on the theoretical solution at an exellent precision.The result is also satisfactory even when the measure data involve 5% random error and 5% shift error.