Research And Implementation Of Magnetic Induction Tomography System

Since the new century,with the rapid development of China's economy and science technology,industry,medical treatment,food and pharmaceuticals have developed vigorously and have huge volume.However,more sophisticated automated testing methods are needed in their production and transportation.Magnetic induction tomography(MIT),also called as electromagnetic tomography,is a new detection technology which is based on electromagnetic induction theory.This technology has great research value,because it can be applied to multi-phase flow detection in industrial transportation,tissue detection in biomedical field,process detection in food processing and pharmaceuticals,etc.The project aims to design a MIT hardware system which works at 1MHz,and to rebuild the two-dimensional cross-section image of detection area on this hardware platform.The main research contents include:1.In principle,the sensitive field equation of MIT system described by vector magnetic potential is derived.The relationship between the physical parameters of the detection area and the magnetic vector potential is established.The magnetic dipole model of single-turn coil and the physical simulation model of MIT system with 8-coil antenna are constructed by electromagnetic simulation software.The influence of magnetic vector potential at the coil is simulated when the iron ball moves transversely.2.Existing MIT detection systems often use an integrated acquisition card to sample signals.In this project,a complete hardware test platform based on FPGA is designed.The hardware system uses analog switch structure to make each coil act as both excitation coil and detection coil at different moments.The system uploads the detection data to the computer through serial port.The detection data is processed by MATLAB.3.Using the Digital Correlation Method to calculate the phase difference of the detected information.In this paper,the average measurement value of the empty detection area is used as local signal,and correlating with the signal which is detected when the detection object exists.The phase change between when the empty field and when the detection object are present is obtained.The traditional method generally samples the excitation signal and the detection signal at the same time,and uses some phase detection method to calculate the phase difference between the two channels.This way requires at least two detection channels,and the normalization problem caused by the different gain should be considered.4.Based on the design of the system,the two-dimensional image of 40×40 pixels in the field is rebuilt using back projection method and regularization algorithm.The imaging effect of the two algorithms is analyzed,and the influence of regularization factor on the regularization algorithm is analyzed.