Research And Application Of The Algorithm To Improve The Spatial Resolution Of Electrical Impedance Tomography

Electrical Impedance Tomography(EIT)technology,as a new type of visual inspection technology,has the advantages of non-invasive,real-time non-destructive,functional imaging and low cost.In particular,the high time resolution of EIT technology under certain conditions enables it to be applied in many important fields.However,the inherent under-determined problem and‘soft-field'effect of EIT technology cause its spatial resolution much lower than that of many existing tomography,such as Computed Tomagraphy(CT),Magnetic Resonance Imaging(MRI).The low resolution severely limits the application of EIT technology and it is hardly to apply under the conditions of complex structure and low signal-to-noise ratio of the measured object.Improving the spatial resolution of EIT technology has always been the key and difficult problem in this field.Based on the existing research results,the core of this paper focuses on overcoming key issues such as uncertainty caused by the under-determined problem and‘soft-field'effect of EIT,and applying new algorithms to improve the spatial resolution of EIT.The main research work of this paper is as follows:1.Optimization of sensitivity matrix.The sensitivity coefficient reflects the response of each pixel to boundary excitation and measurement in the measured object field,which is an important prior information in EIT.The sensitivity matrix formed by the sensitivity coefficient plays a very important role in the EIT visualization process,which directly affects the spatial resolution of the reconstructed image.However,the current EIT reconstruction algorithms mainly use the empty field sensitivity matrix when the conductivity is evenly distributed.The use of the sensitivity matrix is very inadequate,which directly increases the solution error of the EIT reconstruction image.For this reason,a sensitivity correction algorithm based on a prior information is proposed in this paper,which uses the fast fuzzy C-means algorithm and a prior information to correct the sensitivity coefficients of corresponding columns in the sensitivity matrix.In order to improve the response of sensitivity coefficient to nonlinear change,an image reconstruction algorithm based on the second order sensitivity coefficient matrix is proposed.The effectiveness of the proposed method is proved by simulation and experiment.In particular,the proposed algorithm can obtain high spatial resolution for the reconstruction of discrete distributed small targets.2.To deal with the uncertainty or fuzzy characteristics inherent in EIT process,fuzzy optimization is applied to EIT field for the first time and a set of algorithms is proposed.By revealing the fuzzy features hidden in the EIT process,namely the inaccuracy of the sensitivity coefficient,the incompleteness of the measured data,and the inconsistency of the objective function,we analyzed the rationality and interpretability of using fuzzy membership to represent the fuzzy features.Thus,fuzzy optimization is used as a new method to realize EIT visualization and target reconstruction.By analyzing different application conditions and the realization ability of EIT itself,we further divide the application of fuzzy optimization into:constrained fuzzy type,target fuzzy type and fuzzy coefficient type,and accordingly establish the corresponding EIT optimization model and implementation algorithm.Simulation and experimental results show that the proposed symmetric fuzzy linear programming has high spatial resolution and strong robustness to discrete small targets;the asymmetric fuzzy linear programming algorithm also has high spatial resolution to continuous targets;After adding appropriate constraints to the objective function,the fuzzy optimization model has stronger simulation ability,which is beneficial to the application of a priori information.3.L₁-norm optimization and artifact correction based on Bregman divergence.In the lung imaging process by EIT,the local conductivity change in the field satisfies the characteristics of spatial sparsity after resetting the empty field by using the priori structure information.Thus,L₁-norm can be introduced as a constraint penalty to obtain a clearer reconstruction target.Bregman divergence is capable to cope with more general conditions and distinguish similar target.Accordingly,an optimization method based on Bregman divergence is proposed combined with the L₁-norm,which can effectively deal with underdetermined problems in the EIT process.An unsupervised image quality evaluation index based on neighborhood information and fast fuzzy clustering is proposed.Accurate a prior information and neighborhood information are used to modify the artifact problem involved in the EIT inverse problem,enhancing the accuracy of imaging.Simulation and experimental experiments validate the proposed reconstruction algorithm and evaluation index.4.How to apply EIT technology to realize visual diagnosis and bedside monitoring of diseases in the medical field has always been a focus and difficulty in the field of EIT.To effectively identify the early functional lesions of lung cancer,an“EIT+CT”detection mode based on the priori information is proposed.The empty field can be reset by using the human lung tissue structure and the electrical characteristics,and then a more accurate matrix of sensitivity coefficients is obtained,which can improve the sensitivity of the measurement data.The simulation and static experiments show that EIT is more sensitive to the changes of local conductivity in the field after resetting the empty field.The reconstruction algorithm proposed in this paper can detect small local changes,which also indicates that the“EIT+CT”detection mode based on the priori information has a great application potential in the early diagnosis of lung diseases.The one-dimensional features extracted from the measured values can directly reflect the influence degree and sensitivity of the lesion to different electrodes.To approximate the true distribution of the human body preferably,a preliminary study was conducted on the construction of the three-dimensional lung model and the sensitivity in the three-dimensional field.The application of 3D printing technology to the production of EIT models in the medical field provides a platform for obtaining large samples of lung cancer pathology.