Study On Forward Problem And Image Reconstruction Aglorithm Of Magnetic Induction Tomography

Based on electromagnetic induction principle,magnetic induction tomography(MIT)can reconstruct conductivity distribution by measuring the induced voltages on the receiver coils.As a most promising imaging technology,it is widely used in industrial nondestructive detecting,geological exploration and so on,especially in medical functional imaging,due to its non-contact,non-invasive,safe and cheap advantages.In this dissertation the basic researches have been performed on MIT technique,including calculating the forward problem of MIT using finite element method(FEM),analyzing the influence of conductors on eddy current field,deriving three kinds of image reconstruction algorithm,achieving brain magnetic induction tomography(BMIT)imaging and obtaining the preliminary imaging results through designing a MIT detection system.The main research work and conclusions are as follows:(1)Calculation of MIT forward problem:Firstly,electromagnetic field control equation for describing the MIT forward problem is established based on Maxwell's equations.By using two-dimensional FEM,the vector magnetic potential A and induced voltage V are obtained.Furthermore,the effect of disturbance conductor on the inducted voltage is analyzed in the free space and weak conductor background.The simulated results reveal that the law of the influence of the induced voltage on conductor agrees well with the previous studies.This validates the correctness of the forward problem calculation and provides a valid theoretical route for image reconstruction.(2)Calculation and analysis of eddy current field:Magnetic induction intensity and eddy current density are calculated.And the effect of magnetic induction intensity and eddy current density on the background conductivity and disturbance conductor is analyzed.The simulated results reveal that the imaginary part of magnetic induction intensity,together with the real part of the eddy current density,roughly reflect the location of conductors.On the other hand,the real part of magnetic induction intensity and the imaginary part of the eddy current density are only related to the coil's position and structure.This significantly guides for the image reconstruction and improving the quality of image reconstruction.(3)Three kinds of image reconstruction algorithm:For overcoming poor spatial resolution in image reconstruction using Tikhonov regularization algorithm(TRA),three new regularization algorithms are derived,that is variation regularization algorithm(VRA),hybrid variation regularization algorithm(HTVA)and regularization algorithm based on Lp norm(VRAP),respectively.VRA algorithm,using variation functions as a regularization function,can distinguish the conductor region and background region effectively,which results in improving contrast ratio and quality of reconstruction image.HTVA algorithm combines the advantages of the TRA algorithm and VRA algorithm,which facilitates to improve the quality of image reconstruction through finding out the approximate location in early iterations,and realize the quick and time-saving convergence in the late iterations.VRAP algorithm,developed from VRA algorithm,improves the contrast ration and resolution,and further enhances the quality and efficiency of the reconstruction image.The simulated results reveal that these three regularization algorithms are proved to be effective as compared with the TRA algorithm.They all suppress the instability of numerical solution in MIT image reconstruction and improve the quality of reconstruction image.(4)BMIT imaging:Firstly,a simple skull-contained three-layer concentric circular head model is constructed.Based on this,the influence of the conductivity area,position of hematoma and skull conductivity on the induced voltage is analyzed.The hematoma image is reconstructed using VRA algorithm and three kinds of regularization algorithm in three different positions.The simulated results reveal that the skull with low conductivity has little effect on the induced voltage of the receiver coils,and three kinds of regularization algorithm can reconstruct the quality-improved images roughly reflected the location and size of hematoma in skull,as compared with TRA algorithm,and the reconstructed images quality of VRAP algorithm are the best.This not only reflects advantages of MIT technology in the brain impedance imaging field,but also lays a foundation for further studies on image reconstruction of BMIT.(5)MIT detection system design and the preliminary imaging results:The 8 channel MIT detection system is designed,with which the sample of plastic bottles containing saline is investigated experimentally.The experimental results reveal that the change regulation between the phase and amplitude ratio of measurement data with position and concentration of target conductor is consistent with the law obtained in the previous studies.This verifies the stability and repeatability of this detection system.Finally,the preliminary image is reconstructed for the sample in weak conductor background,and the reconstructed images quality of VRAP algorithm are much better.This provides hint for further improving MIT system.