Research And Application Of ECG Abnormal State Monitoring Technology

Portable ECG monitor has great application value.It enables people to grasp their own heart health status at any time without leaving home.It can alarm in time when abnormal conditions occur and fight for seconds to save patients’ lives.In addition,for special user groups such as athletes and firefighters,portable ECG monitor can provide more comprehensive life or work security for these special user groups by combining other vital signs such as EEG and human posture.In this paper,a solution of ECG monitoring based on "ECG acquisition front-end + user App + ECG health management platform" is proposed.Firstly,ECG acquisition circuit is built to realize real-time acquisition and data upload of human ECG signals.Secondly,Android App application program is developed in smart phone terminal to realize user information input,ECG drawing and abnormal alarm.Finally,ECG health management platform is developed based on LabVIEW on PC,which realizes the functions of ECG data preservation,ECG preprocessing,ECG feature extraction and ECG abnormal diagnosis.In order to keep the ECG signal undistorted,the Mallat fast decomposition and reconstruction algorithm based on wavelet transform is used to denoise the ECG signal.The denoising effect of commonly used wavelet is evaluated by peak signal-to-noise ratio(PSNR)and root mean square error(MSE).Based on the principle of singularity detection of signal by wavelet transform,the peak value of QRS wave is located,and the starting and ending point of QRS wave is detected near zero baseline.In order to improve the accuracy of peak detection of R wave,a retrospective comparison mechanism is added to the R wave detection algorithm to eliminate false detection and missed detection.It is verified by MIT-BIH Arrhythmia database samples.The sensitivity(Se)and positive detection rate(+P)are used to evaluate the reliability of the QRS wave detection algorithm.On the basis of locating R-wave,we segment ECG records and further reduce the dimension of ECG features by principal component analysis(PCA).25 principal components are fed into the ECG detection model.Four typical heartbeat classifications of MIT-BIH Arrhythmia database are validated by improved BP neural network,support vector machine(SVM)and convolutional neural network(CNN),and the results are compared.This paper establishes the ECG acquisition system,establishes the ECG abnormal diagnosis model,realizes the functions of human ECG signal acquisition,upload,analysis,early warning,and can meet the requirements of users for ECG monitoring system.The research results have certain reference value and practical significance for the design and application of portable ECG monitoring instrument.