Research On Multi-frequency Electrical Impedance Tomography Algorithm Based On Krylov Subspac

Electrical Impedance Tomography(EIT)is a new non-invasive functional imaging technique developed in recent years.Its basic principle is to apply excitation current to the measured field through electrode array and measure the response voltage data of the boundary.Then,an image reconstruction algorithm is used to reconstruct the conductivity distribution image in the field.Because EIT technology has the advantages of non-radiation,non-invasion,fast response and low cost,it has a good development prospect in the biomedical field.Traditional EIT technology is mainly based on single-frequency excitation measurement,while multi-frequency EIT technology uses the dielectric properties of biological tissues to apply multiple current excitation with known frequencies to the field to be measured.Image reconstruction is carried out through the voltage data measured at the lower boundary of different frequencies,which makes up for the problem of insufficient measurement information of single-frequency EIT.Image reconstruction is the core of EIT technology and difficult problem,the current multifrequency EIT technology more iterative reconstruction algorithm using the Least Square Conjugate Gradient method(CGLS),convergence is not smooth,but this method for detecting unknown target,set number of iterations can only rely on experience,lead to solving the instability,poor quality of reconstruction image.In this paper,two Krylov subspace iterative methods,BICGSTAB and QMRCGSTAB,are proposed to solve the EIT image reconstruction problem,and ILUT preprocessing is used to further improve the algorithm performance and image quality.The main tasks of this project are as follows:(1)This paper makes a detailed derivation of the weighted frequency difference,and conducts simulation experiments to verify that the weighted frequency difference method can effectively suppress background noise and improve the discrimination between the detection target and the background.(2)This paper proposes to use BICGSTAB and QMRCGSTAB as the image reconstruction algorithm of multi-frequency EIT and has carried out experimental verification.Among them,the BICGSTAB algorithm has better stability and the quality of the reconstructed image is better than the CGLS method.The QMRCGSTAB algorithm has better smooth convergence and stability.As the number of iterations increases,the solution result will not deteriorate sharply.(3)It is proposed to use ILUT preprocessing to improve the performance and imaging quality of the image reconstruction algorithm.The ILUT preprocessing can effectively reduce the morbidity of the inverse problem of the multi-frequency EIT technology,improve the solution speed and the quality of image reconstruction,which is proved by simulation experiments.,After using ILUT preprocessing,the performance and image reconstruction quality of BICGSTAB algorithm and QMRCGSTAB algorithm have been significantly improved.(4)This paper develops a set of multi-frequency EIT visualization software system based on Python language and Py Qt library,and integrates BICGSTAB algorithm,QMRCGSTAB algorithm and ILUT preprocessing module in the software system.The system can communicate with EIT’s data acquisition equipment using TCP/IP network protocol,which improves the data transmission speed and reliability.