Research on Direct 3D Electromagnetic Tomography Technique

Electromagnetic tomography (EMT) is an emerging imaging modality which attempts to visualize the distribution of electrically conductive or magnetically permeable materials using measurements obtained from coils distributed around the vessels or pipelines. The traditional EMT research concerns about the two-dimensional (2D) cross-sectional imaging. However, hardly any three-dimensional (3D) information of the target objects can be obtained. Effective 3D EMT imaging technique can provide further axial information and intuitional images of the target objects, which is especially suitable for processes with significant change of conductivity distribution in the axial direction. The direct 3D EMT imaging technique, which uses measurements obtained from both the same layers and the different layers and can directly acquire the 3D reconstructed images, is studied in this paper including issues such as 3D sensor layout, measurement strategy, simulation research, 3D image reconstruction algorithm. The 3D EMT finite element simulation model is established to conduct the forward problem calculations and further verify the feasibility of the sensitivity matrix. In the simulation process, a three-plane cylindrical EMT sensor with 12 coils is designed to verify the feasibility and effectiveness of the direct 3D EMT imaging technique. The images are reconstructed by three image reconstruction algorithms including Tikhonov regularization method, Projected Landweber iterative algorithm and total variation regularization method. Compared with the other two algorithms, total variation regularization method is superior in terms of the eight typical patterns due to the fact that it possesses good edge-preserving property.