Erkki J. Somersalo
Abstract:In this article, we consider the problem of finding the support of an inhomogenous possibly anisotropic inclusion in a background of constant electric conductivity from the electrical impedance tomography data at the boundary of a bounded body. The article discusses the linear sampling method applied to this problem. A practical algorithm for solving this problem is suggested.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to find a support region where anisotropic inclusions may exist from Electrical Impedance Tomography (EIT) data when the background conductivity is known. Specifically, the author focuses on how to estimate the internal structure of the body through the voltage and current information measured on the boundary, especially those regions with anisotropic material properties.
### Specific description of the problem
In Electrical Impedance Tomography (EIT), the main problem faced by researchers is: applying a set of predetermined currents on the surface of an object with an unknown conductivity distribution, measuring the corresponding voltages, and then using this information to estimate the internal structure of the object. Especially in the presence of anisotropic materials, this problem becomes more complicated because the electromagnetic properties depend on the direction, which usually makes the inverse problem have no unique solution.
### Research objectives
The main objective of the paper is to study the following inverse problems:
- In an isotropic object, assume there is an unknown, possibly anisotropic inclusion.
- Given the EIT data on the surface of the object, estimate the support region of this unknown inclusion.
### Method overview
To achieve the above objectives, the author adopts the Linear Sampling Method and proposes a practical algorithm to solve this problem. The Linear Sampling Method is a technique for determining the shape of inclusions. It is based on far - field or near - field measurement data and combines multiple detection modes.
### Mathematical model
The following mathematical models are used in the paper to describe the problem:
1. **Potential equation**:
\[
\nabla \cdot (\gamma \nabla u) = 0 \quad \text{in } B
\]
where \(\gamma(x) = 1 + h(x)\chi_D(x)\), \(\chi_D\) is the characteristic function of the set \(D\), and \(h(x)\) represents the perturbation term.
2. **Boundary condition**:
\[
\frac{\partial u}{\partial n}\bigg|_{\partial B}=f
\]
3. **Neumann - to - Dirichlet mapping**:
\[
\Lambda : H^{-1/2}_0(\partial B)\to H^{1/2}_0(\partial B), \quad f\mapsto u|_{\partial B}
\]
4. **Singular solution**:
\[
\Delta \Phi=\hat{\alpha}\cdot \nabla \delta(x - y)
\]
\[
\frac{\partial \Phi}{\partial n}\bigg|_{\partial B}=0
\]
Through these mathematical tools, the author can construct a framework to estimate the support region of inclusions and propose specific algorithm steps.
### Conclusion
This paper provides an effective method, namely the Linear Sampling Method, for solving the EIT inverse problem in anisotropic materials. This method not only helps to improve the estimation accuracy of the internal structure of the body but also provides theoretical basis and technical support for practical applications.