| Cardiovascular disease is always a threat to people’s physical and mental health,and daily monitoring of cardiovascular function parameters has become an urgent need in people’s hearts.Impedance rheogram extracts human’s hemodynamic information by detecting the impedance volume change of the corresponding part,and realizes the non-invasive detection and evaluation of cardiovascular function parameters.However,people are inevitably subjected to noise interference from measurement jitter and limb movement when performing impedance rheogram detection,which affects the accuracy of feature point extraction and causes measurement errors of cardiovascular function parameters.According to the waveform characteristics of impedance rheogram,this paper proposes a sample division method combining data segmentation and interval search,which realizes the division of impedance rheogram’s periodic samples and the effective division of feature point positions.according to the divided effective interval and the physiological significance of the impedance differential signal,a sample selection method of waveform amplitude constraint and feature point position constraint is proposed,this method constrain the amplitudes of A and O waves in the impedance differential graph and limit the position of the feature point B in the QC interval,so as to eliminate invalid samples and retain the commonality between samples.Affected by noise,the QC interval band is prone to multiple inversions and rapid slope changes.This paper mainly considers the inflection point in the QC interval,and designs an automatic identification method for the characteristic points of impedance blood flow signals,which improves the accuracy and anti-interference of feature point recognition.Blood pressure is an important index of cardiovascular function parameters.Photoplethysmography is used to continuously measure blood pressure at home and abroad.There is a certain error in blood pressure values,and the detection accuracy still needs to be improved.In this paper,according to the feature point information of the impedance differential graph,the amplitude,interval time,slope of the feature points C,B,and X points are extracted to reflect the impedance rheogram’s characteristic parameters of the heart pumping process;and according to the correlation between blood pressure and the characteristic parameters,the multiple linear regression method was used to establish the blood pressure prediction model.In addition,according to the mathematical relationship between the impedance volume model and the characteristic parameters,this paper realizes the acquisition of cardiovascular function parameters such as stroke volume,cardiac output,and ejection fraction;and designs the impedance rheogram detection system software to intuitively present the detection results of cardiovascular function parameters.Finally,the error analysis and evaluation of the blood pressure prediction results of the blood pressure regression model are carried out.The mean absolute error between the predicted value of systolic blood pressure and the results of the electronic sphygmomanometer was 4.1mm Hg,and the mean relative error was 3.81%;the mean absolute error of the predicted value of diastolic blood pressure was 2.85 mm Hg,and the mean relative error was 3.85%.the deviation of the predicted results of systolic and diastolic blood pressure from the calibration value is less than 5mm Hg,and has good consistency,the measurement system has good performance with the regression model. |
PDF Full Text Request