Research On Electrical Impedance Tomography For Lung Disease Diagnosis

Electrical impedance tomography(EIT)is a non-invasive visualization technique.The electrical responses are measured by applying the electrical excitation to the boundary of measured field.Then the electrical conductivity distributions and the image changes of the measured field are reconstructed by using the reconstruction algorithm.Compared with the mainstream techniques for diagnosing lung diseases,EIT has the advantages of non-radiation,non-invasive,fast response,low cost and portability.Therefore,EIT has the potential to avhieve the long-term continuous bedside monitoring,and has an attractive development prospect in the field of medical monitoring.This paper investigated the EIT for lung disease diagnosis in depth,including the analysis and design of EIT sensor system,signal processing in data acquisition system and EIT imaging algorithm based on pulmonary prior information.The main research work of this paper is as follows:1.Research on EIT sensor system.Through the study of two/multi-layer EIT sensor,it was found that there was an optimal layer spacing to achive the best quality for the EIT images.The use of two/multi-layer EIT sensor can effectively improve the “3D effects” of EIT sensitive field,and a compensation relationship of measured voltages for the two-layer EIT sensor was given.In addition,to solve the problem of dynamic change of measurement points,a data acquisition device for thoracic contour,a displacement measurement sensor and the related measurement methods for human thoracic EIT were proposed and designed.2.The traditional phase-sensitive demodulation(PSD)of EIT required the reference signals and integer cycle sampling.Aiming to the limitations of PSD,a singular value decomposition(SVD)based amplitude demodulation method was proposed.This method gave the mathematical relation between the signal amplitude and the effective singular values.It overcame the disadvantages of traditional PSD,and its validity and feasibility were verified by the numerical simulation and experiment.3.Research on impulsive noise reduction in multi-frequency PSD.In actual EIT measurement systems,due to the factors of cardiac fibrillation,circuitry/switch switching and external vibrations,the measured signal was often contaminated by impulsive noise.In order to effectively reduce the impact of impulsvie noise on multi-frequency PSD,two methods were proposed,i.e.2D median filter based multi-frequency PSD and SVD based multi-frequency PSD.Numerical simulation and experimental results indicated that the two methods can effectively reduce the impact of impulsvie noise on multi-frequency PSD,and provide the demodulation results with higher SNR and stronger robustness.4.A Tikhonov regularization algorithm based on pulmonary prior information was proposed to monitor the lung cancer and its metastasis.The simulations and experiments validated the predictability of lung cancer disease through EIT.The results demonstrated that the algorithm can effectively improve the spatial resolution of EIT images,and can be used to identify the cancerous tissues and monitor their metastasises.