【作者】 杜强；

【导师】 白保东； 柯丽；

【作者基本信息】 沈阳工业大学 ， 电工理论与新技术， 2013， 博士

【摘要】 磁感应断层成像技术（Magnetic Induction Tomography, MIT）应用电磁感应原理，在被检测物边界测量二次磁场，结合重建算法对被检测物电导率分布进行成像，具有非接触、无创等优势，但是目前研究处于探索更好的扫描方式和成像系统的研究阶段。本文针对磁感应断层成像的关键问题，从理论建模、系统设计、磁场正逆问题及成像方法等开展了系列研究，提出了一种扇形阵列式磁感应断层成像系统，包括单激励多通道谐振检测模型，一体化数据采集系统，以及面向该系统的数据处理和新型非线性反投影图像重建算法。具体工作如下：（1）由于被测物的感应涡流磁场在空间中的真实分布和特性对MIT系统设计具有重要的指导意义，本文根据人体大脑电导率分布建立了磁感应断层成像的三维仿真模型以及系统模拟实验，并据此研究了检测区域电场分布规律，不同检测物对涡流磁场的影响，成像区域磁场变化以及边界测量值的分布规律，以及被测目标改变对边界测量值的影响等。（2）针对磁感应信号微弱、干扰严重的问题，发挥大线圈激励磁场强，小线圈检测定位准确、串扰小的优势，将检测和激励线圈功能分离。提出了双谐振非对称激励检测模式，使激励线圈和检测线圈均在谐振频率下工作，增大激励磁场同时，通过谐振选频特性提高检测线圈对磁场变化的稳定性。检测过程中将参考信号和感应信号分离，以独立通道形式提供给鉴相电路，克服鉴相的双值性问题，提高了鉴相的准确性。（3）本文提出了扇形阵列式多通道信号采集系统结构，检测线圈呈半环形排列在激励线圈对面。测量过程中通过相对旋转扫描方式，获得多角度的测量数据。原始测量数据经数据插补、校正和标准化之后，构成投影数据用于图像重建。（4）MIT系统成像区域内磁场分布非线性，导致重建过程复杂，定位精度差，本文基于反投影研究了扇形MIT系统快速图像重建理论及实现方法，根据MIT边界数据对成像区域内部电导率变化的敏感性分析结果，设计了测量数据标准化方案，进而根据测量相位差即为成像区域内电导率分布沿磁力线路径投影的集合这一结论，利用成像区域磁力线的分布构建反投影路径和区域，设计了一套适用于磁感应断层成像的反投影算法，形成了一套完整的扇形多通道磁感应断层成像系统。更多还原

【Abstract】 Magnetic Induction Tomography (MIT) detects the eddy current magnetic field on theboundary of the imaging area based on electromagnetic induction method, and gets thedetecting object tomography of the conductivity distribution by the reconstructionalgorithm. It meets requirements of a modern medical device that is contactless,non-invasive, and imaging-functional. However, it has not yet been used in the clinicalpractice due to the limitation in imaging resolution, so high precision scanningmeasurement methods and high resolution image system are still in the research phases. Inthis paper, the research on theory modeling, system design, forward and inverse problem ofmagnetic field, and imaging method was carried out, which aims to provide solutions to thekey problems inherited in MIT, and a sector array MIT system is presented in this paper,which includes single excting and multichannel harmonic resonance detecting systemstructure, integrated data acquisition system, data processing methond and a new nolinearback-projection reconstruction algorithm for the system. The main work as follows:(1) The distribution and the propertis of the eddy currents and the magnetic fieldsinduced by the imaging object may provide important message for the MIT system design.This paper simulated the3-D MIT model and mearsuring expriments according to the brainconductivity distribution, and studied on the electric field distribution of detecting erea, theeddy current magnetic field with the different object, the change in magnetic field andboundary mearsuring data.(2) For the induction signal is feeble and corrupted badly, it allows separation of thedesign of detecting coils and exciting coil, and adopts the large coil to get strong incentivemagnetic field, and small detecting coils to get the high accuracy of position and smallinterference. Meanwhile harmonic resonance detecting method is adopted to enhance theexciting magnetic fields and improve the stability due to the frequency selectivity of thedetecting coils. In the process of signal detecting, the reference signal is separated frominduction signal, and transmits to phase discriminator circuit byindependent channel, and itimproves the accuracy of phase discriminate signal. (3) A multichannel array signal acquisition system is designed to obtain measurementdata used for image reconstruction, and we design the detecting coils semicirculardistribution opposite the exciting coil. The relative rotational scanning method is used toget multiangular measurement data in the system. The raw detecting data is interpolated,adjusted, and standardizated, and got the reconstructed data.(4) The magnetic field of imaging area in MIT is nonlinear, with increase thedifficulty of image reconstruction and leads to the poor positioning accuracy. The fastimage reconstruction algorithm for sector MIT system and its implement method wasstudied in this paper based on back-projection algorithm. A new data standardizationmethod is presented according to the measureing data sensitivity with the conductivitychange in the imaging area, and then an improved back-projection image reconstructionalgorithm fit for the magnetic induction tomography is presented, based on the theory thatthe phase difference can be described as the projection of the conductivity distributionalong the magnetic field lines. The back-projection path was determined by the magneticfield lines in the imaging area, and then a complete set of sector multichannel MIT systedis enstablished.更多还原