Research On Key Technologies Of 3D Magnetic Induction Image

Magnetic Induction Tomography(MIT)is a noninvasive and contactless imaging system.It aims at the reconstruction of the electrical conductivity in objects by alternating magnetic fields.Because of the advantages of noninvasive and noncontact,MIT has great potential in the field of biomedical electrical impedance tomography.At present,the research about MIT is mainly focused on 2D method,the image is reconstructed in a cross plane of a target body.But in practice,we hope to get 3D information such as geometric shape and internal disturbance in target body,so the research of 3D MIT has great significance.Reconstruction is based on calculation of forward problem.The quality of the reconstruction image depends on the result of forward problem.In this paper,the finite element equations of magnetic induction 3D forward problem were solved by ICCG(Incomplete Cholesky Conjugate Gradient)method.A 3D finite element model was builed based on ANSYS(a kind of general finite element software),then the finite element equations were solved by ICCG method.Simulation experiment provided a optimum value of convergence tolorence in this problem.The result shows that the ICCG method has the advantage of fewer iterative times,faster convergence rate,less affected by the mesh model,better robustness.Calculated 3D sensitivity matrix distribution in target image area based on field expressions extraction.Sensitivity was pretreated before the reconstruction in order to solve the reconstruction image boundary migration and adhesion.the effectiveness of the algorithm is validated by the simulation experiment.Build a FEM MIT model of 16 coils and double-layer,the intresting area is a cylinder representing the head or chest The conductivity of background is 1 S/m when there is no perturbation,a small cylinder with conductivity 4S/m is the perturbation in intresting area.The current value in the coil was simulated,the detection signal was recorded and adding random noise.Bring in projection Landweber algorithm to improving the Landweber algorithm.let(LBP)algorithm as initial value of projection Landweber algorithm After the sensitivity pretreatment.The results shows as contour plot.The sensors of MIT system are generally air-core coils.The relative location between testing coil.exciting coil and the characteristics of incentive field are both have the influence on sensitivity and accuracy of the system.In this paper,the induced electric field in intresting area is the analysis object.Based on 3D finite element method to solve the magnetic field.A simulation analysis is done about induced electric field caused by changing coil radius and the distance.Uniform field can make MIT inverse problem simple,a 3 coils excitations 8 coils receives system is proposed which filed is uniform.A set of perturbations were carried out based on the laboratory hardware system.In this paper.First,a strong perturbation in vacuum is carry out,and filter back projection is the reconstruction method.Then a weak perturbation in vacuum is carry out and verified by simulation.A conductor perturbation in conductor test is simulationed.The results show when conductor perturbed in vacuum,the changes with the characteristics of projected bundle,filter back projection reconstruction can carry out.When conductor disturbance in conductor,the signal is not have the characteristics,so does not apply to filter back projection reconstruction.the experiment proved the feasibility of analyzing iron overload in liver with MIT system.