Impedance imaging and electrode models (Q4356946)

We consider the impedance imaging problem of estimating the unknown resistivity distribution in a body from a finite number of current-voltage measurements on the surface of the body. In practice, the current is injected into the body through electrodes attached on the surface, and usually the same electrodes are used for voltage measurements. The focus of this article is on the effect of the electrodes on the resistivity estimation. If the model is furnished with a properly chosen contact impedance between the electrodes and the body, the computed voltages agree with measured voltages with a reasonably high precision. To reduce the electrode noise in the inverse solution, one should therefore use algorithms that take the contact impedance into account in one way or another. In this article, we discuss impedance imaging algorithms that take the electrodes into account. The first algorithm is a Newton type algorithm based on a finite element direct solver using the contact impedance electrode model. The second algorithm is a simple linearized algorithm based on the invariant imbedding equation used in the layer stripping algorithms. In the latter algorithm, we introduce an idea that can be applied to basically any scheme to reduce electrode noise. This idea is called here quasistatic imaging. For the entire collection see [Zbl 0871.00041].