Research On Optimization Technology Of Cardiac Magnetic Imaging Based On SQUID Detection

With the aging of population and accelerating pace of life,an increasing number of people suffer from heart dysfunctions.At the same time,corresponding medical facilities are constantly improved.The diagnosis and treatment methods of heart diseases are varied.Among them,the magnetocardiography is a technology to measure weak magnetic signals generated by the human heart due to electrophysiological activities.It is noninvasive,non-contact and highly accurate,and is a research hotspot in the emerging medical technology areas.Due to the extremely weak cardiac magnetic signals,SQUID with ultra-high sensitivity is usually used for detection in the magnetic shielding chamber.However,a magnetic shielding chamber with excellent performance has problems of high cost and immobility,which limit application scenarios of SQUID.In the environment without or with simple magnetic shielding,SQUID magnetometer will be interfered by strong ambient noise,making useful signals drowned.Therefore,it is necessary to adopt signal processing algorithms to suppress the noise and extract useful signals.At present,most of the denoising algorithms are used for single-channel signals,can not process multichannel ones at the same time,and reflect the denoising effect from a whole perspective through imaging.Therefore,in this thesis,the denoising algorithms of multi-channel cardiac magnetic signals are studied in detail,and an improved method is proposed.The specific research contents are as follows:Firstly,for the preprocessing of the cardiac magnetic dataset adopted in the thesis,the method based on Ensemble Empirical Mode Decomposition with adaptive filtering is used to correct the slight baseline drift.Morphological analysis method is used to automatically locate QRS complex wave and T wave peaks,in order to prepare for the cardiac magnetic imaging.Secondly,to solve the problems of large computation and cumbersome operation of multi-channel cardiac magnetic signal processing,the thesis proposes a multi-channel cardiac magnetic signal denoising algorithm based on Independent Component Analysis and Complete Ensemble Empirical Mode Decomposition with Adaptive Noise,recorded as ICA-C.In this method,ICA is first used for decomposition,and kurtosis criterion is proposed to select the independent components generated by decomposition,and then CEEMDAN is used for denoising.In the algorithm of CEEMDAN,interval threshold method is used to avoid discontinuity of reconstructed signals.Simulation results show that the ICA-C method is effective.Finally,in order to further improve the signal-to-noise ratio of cardiac magnetic signals,a multi-channel Optimized Magnetocardiogram Imaging algorithm,recorded as ICA-C based OMI,is proposed in the thesis.In this method,the multi-channel cardiac magnetic signals are interpolated by cubic spline and magnetic field maps are drawn.Then,three typical indicators are selected and an optimization model is established by using the error minimization criteria.Further optimization of the denoised results is achieved by cardiac magnetic imaging.Through comparative experiments,the proposed method has a favorable denoising performance and a low distortion of waveform.