The Importance of FEM model for Linearized EIT Image Reconstruction

The improvement of image quality is fundamental to further the application of Electrical Impedance Tomography (EIT) in industry and clinics. The majority of EIT image reconstruction algorithms are linearization methods based on Finite Element Method (FEM). As we know, the meshing structure of the sensing field is important to FEM performance, in this paper, we investigate the optimization of finite element mesh for linearized EIT image reconstruction. Five different scales of FEM meshes with dense elements in electrode regions and three different scales of FEM meshes with even density of elements in entire field are constructed. The FEM solution for each FEM mesh is compared with analytical solution to evaluate accuracies of FEM meshes. Our research shows that the refinement of finite elements in electrode regions can improve the accuracy of simulated boundary voltage, and achieve similar or even better simulation accuracy with less finite elements. Further, it also has better performance and less time cost in image reconstruction. In addition, the higher density of finite elements the better for both boundary voltage simulation and image reconstruction but more time cost. We propose that the density of FEM mesh should match with the performance of the EIT measurement system, so as to reconstruct the best quality image with less calculation load.