电阻抗层析成像的电导网络解法

阻抗ＣＴ技术是通过在物体的边界进行一些电测量来重建内部阻抗分布。本文提出利用网络理论来解决阻抗ＣＴ的一种新的重建方法。首先将非均匀电导分布区域离散化为具有一些节点的电导网络，然后在边界节点上施加一组正交的电流激励，测量感生出的电压。根据网络论，这样就可利用网络的节点方程来进行求解．为求得最符合观测值的节点电导矩阵Ｇ的估计，我们从Ｇ的性质中提取了两类约束，根据广义投影理论，采用施加约束的迭代方法来重建这一反映电导分布的节点电导矩阵Ｇ．计算机模拟表明，迭代是收敛的并且能得到较好的重建结果。

A Solution to Resistivily Tomography by Means of Conductive Network

The inner electrical conductivity distributinn of an inhomogeneous objectcan be modeled by an n-node conductance-network.The node conductance matrix ofthis network can be fully described by its conductance matrix G.Our algorithm seeks toget the most reasonable estimation of G which equally represents the conductivity distribution We find that G must satisfy two kinds of constraints,i.e.Y -constraint and Z-constraint. These constraints are employed in an iteration procedure for modification of Y andZ by using projection theory.Y and Z are modifyed altnatively until the voltages calculatedare very close to the voltages measured at the eletrodes.The algorithm yields satisfatory results from computer simulation. The reconstruc tion examples for an 8 * 8 square region at several phantoms are obtained.The iterationconverges pretty well with only a few steps,even for some samples with large conductivitycontrast.