CN114587307B - A non-contact blood pressure detector and method based on capacitive coupling electrode - Google Patents

Abstract

The present invention discloses a non-contact blood pressure detector and method based on capacitive coupling electrodes. Capacitive coupling excitation electrodes and capacitive coupling measurement electrodes are attached to the outside of human arm clothing to detect impedance-volume characteristics; capacitive coupling measurement electrodes and capacitive coupling reference electrodes are attached to the outside of human heart and chest clothing to detect electrocardiogram signals; an electrocardiogram and impedance-volume signal measurement module is used to synchronously collect capacitive coupling impedance-volume ratio (CCIPG) signals and capacitive coupling electrocardiogram (CCECG) signals; a data acquisition module is used to convert received chest lead CCECG and arm CCIPG signals into digital signals and transmit them to a controller, so that the controller obtains the capacitive coupling electrocardiogram of the measured person. Electrogram waveform and impedance-volume ratio waveform; feature extraction module is used to obtain pulse wave transmission time (PWTT); impedance-volume feature parameter calculation module is used to obtain human hemodynamic parameters; PTT blood pressure calculation module is used to calculate blood pressure value 1; PTT fusion impedance-volume parameter machine learning module is used to use human hemodynamic parameters and blood pressure value 1 as sample data for learning and training, and to establish the final blood pressure model; blood pressure prediction module is used to predict human blood pressure value, so as to realize real-time monitoring of blood pressure and analysis of cardiovascular health status, without directly attaching electrodes to the skin, which will not cause discomfort to the person being measured, and can be widely used in fields such as human health monitoring.

Description

A non-contact blood pressure detector and method based on capacitive coupling electrode

Technical Field

The present invention relates to a non-contact blood pressure detector and method based on capacitive coupling electrodes, which can be used for real-time monitoring and measurement of human blood pressure and belongs to the technical field of medical devices.

Background Art

Blood pressure is a common and important indicator in biomedical measurement. The heart's pumping function, the blood supply of the coronary arteries, the resistance and elasticity of peripheral blood vessels, the blood volume of the whole body and the physical state of the blood are all reflected in the parameters of blood pressure. Blood pressure can be used as an indicator of the state of the cardiovascular system. Therefore, the detection of blood pressure parameters provides an extremely important basis for clinical examinations, patient care, or physiological research.

Existing blood pressure measurement technologies are mainly divided into two methods: invasive blood pressure measurement and non-invasive blood pressure measurement. Invasive blood pressure measurement is to insert a catheter into the blood vessel or heart of the measured part, and place a pressure sensor at the end of the catheter. The pressure generated by the blood flow and impact on the catheter in the measured part is transmitted to the end of the catheter and sensed by the pressure sensor, thereby directly measuring the patient's blood pressure. This method can accurately and continuously monitor the patient's blood pressure, but it must insert the catheter into the blood vessel through the skin, which is difficult to operate and can easily cause infection of the patient's measurement site wound; the non-invasive blood pressure measurement method uses the relationship between the intravascular pressure and the blood flow changes that occur at the moment of blood blockage and opening to measure the corresponding blood pressure value from the body surface. The Korotkoff sound method and oscillometric method, which are currently widely used, both use a cuff to inflate and pressurize the artery, then slowly deflate it, detect the change of pulse or blood flow under the cuff or at the distal end of the artery, and calculate the systolic and diastolic pressures by the blood pressure algorithm. Although it can avoid the disadvantage of invasive methods causing infection to patients, it can only measure the systolic and diastolic pressures of patients during this period and cannot continuously record the blood pressure waveform for a long time.

In recent years, many foreign research institutes have begun to use photoplethysmography (PPG) to detect changes in pulse wave waveform characteristics and AgCl wet electrodes to measure electrocardiogram (ECG) signals to obtain the pulse wave conduction time (PTT), thereby achieving continuous blood pressure monitoring. However, PPG technology has high requirements for the measurer. At the same time, in order to reduce contact impedance during use, AgCl wet electrodes usually need to be used with conductive glue. There are disadvantages such as long preparation time, irritation to human skin, and easy to fall off during long-term monitoring. In addition, the diastolic pressure (DBP) extracted by relying solely on PTT parameters has large errors, and its detection accuracy still needs to be further improved.

In order to solve the above problems, the invention discloses a non-contact blood pressure detector and method based on capacitive coupling electrodes. The coupling electrodes are designed based on the capacitive coupling principle. Without direct contact with the skin, the bioelectric signals on the human body surface are fully coupled to the detection end, and then the coupled chest electrocardiogram signal (CCECG) and coupled arm impedance volume ratio signal (CCIPG) are detected, the pulse wave transmission time is calculated, and the characteristic parameters of the impedance blood flow diagram are combined to construct a non-contact continuous blood pressure prediction model, so as to improve the accuracy of blood pressure measurement values.

Summary of the invention

The purpose of the present invention is to provide a non-contact blood pressure detector and method based on capacitive coupling electrodes. The method mainly focuses on the detection of non-contact electrocardiogram signals and impedance volume ratio signals, expands the research on non-invasive blood pressure detection, avoids direct contact of electrodes with the skin, improves the accuracy of blood pressure measurement, and establishes and optimizes a model for non-contact continuous blood pressure prediction.

To solve the above problems, the present invention provides a non-contact blood pressure detector and method based on capacitive coupling electrodes, which method consists of a capacitive coupling excitation electrode 1, a capacitive coupling measurement electrode 2, a capacitive coupling reference electrode 3, an ECG and impedance volume signal measurement module 4, a data acquisition module 5, a feature extraction module 6, an impedance volume feature parameter calculation module 7, a PTT blood pressure calculation module 8, a PTT fusion impedance volume parameter machine learning module 9 and a blood pressure prediction module 10.

The capacitive coupling excitation electrode is used to cooperate with the capacitive coupling measurement electrode to be attached to the outside of clothing on the human arm to perform impedance volume characteristic detection.

The other capacitive coupling measurement electrode is used to cooperate with the capacitive coupling reference electrode to be attached to the outside of clothing on the heart and chest of a human body to detect electrocardiogram signals.

The ECG and impedance volume signal measurement module is used to synchronously collect capacitive coupled impedance volume ratio (CCIPG) signals and capacitive coupled electrocardiogram (CCECG) signals.

The data acquisition module is used to convert the received chest lead CCECG and arm CCIPG signals into digital signals and transmit them to the controller, so that the controller can obtain the capacitive coupling electrocardiogram waveform and impedance volume ratio waveform of the measured person.

The feature extraction module is used to obtain the pulse wave transmission time (PWTT).

The impedance volume characteristic parameter calculation module is used to obtain human hemodynamic parameters.

The PTT blood pressure calculation module is used to calculate the blood pressure value 1.

The PTT fusion impedance-volume parameter machine learning module is used to use the morphological characteristic parameters of the physiological characteristics of the cardiovascular system reflected on the impedance-volume ratio waveform and the blood pressure value 1 as sample data for learning and training to establish the final blood pressure model.

The blood pressure prediction module is used to predict the blood pressure value of a human body.

Furthermore, the coupling electrode is a four-layer rectangular PCB board, and the two middle layers of the coupling electrode are both set as shielding ground layers to achieve the function of isolating spatial noise.

Furthermore, the ECG and impedance volume signal measurement module includes a coupled ECG measurement unit and a coupled impedance volume ratio measurement unit.

Among them, the coupled ECG measurement unit includes a differential amplifier circuit, a 50Hz trap circuit and a bandpass filter circuit. The input end of the differential amplifier circuit is connected to the ECG coupling measurement electrode, which is used to perform unit gain differential between the capacitively coupled ECG signal collected by the coupling measurement electrode and the ECG signal of the reference electrode, and at the same time, to induce the common mode interference of the ECG coupling electrode. The 50Hz trap circuit and the bandpass filter circuit are used to filter the differentially amplified signal to obtain the chest lead ECG signal.

Among them, the coupling impedance volume ratio measurement unit includes a main control circuit, a constant current source circuit, a differential amplifier circuit, a filter circuit, a full-wave rectifier circuit and a differential circuit. The main control circuit controls the constant current source circuit to output 50KHz, 0.3mA current, and adopts the four-electrode method to measure the human arm impedance. The input end of the differential amplifier circuit is connected to the coupling measurement electrode, which is used to differentially amplify the capacitive coupling impedance volume signal collected by the coupling measurement electrode. The processed signal enters the full-wave rectifier and differential circuit through the filter circuit to obtain the arm impedance volume ratio signal.

Among them, the data acquisition module is mainly composed of A/D circuits, which are used to convert the received chest lead ECG signals and arm impedance volume ratio signals into digital signals and transmit them to the controller, so that the controller can obtain the capacitive coupling ECG waveform and impedance volume ratio waveform of the measured person.

Among them, the morphological characteristic parameters of the physiological characteristics of the cardiovascular system reflected in the human CCIPG signal waveform include: the ratio of BC segment time to single cycle time ( _TRBC ), the ratio of CX segment time to single cycle time ( _TRCX ), the ratio of ejection time to single cycle ( _TR ), the slow ejection period time ( _TCX ), the ejection time ( _TBX ) and the cardiac output per stroke ( _SBX ).

The non-contact blood pressure detection method is as follows:

Step 1: locate the effective peak point of the R wave of the collected electrocardiogram waveform to obtain the effective R wave peak position of the electrocardiogram waveform; locate the rising branch end point of the impedance volume ratio signal waveform to obtain the rising branch end point position of the impedance volume ratio signal waveform.

After the jth R wave effective peak point appears on the ECG waveform, determine the position of the jth R wave effective peak point on the ECG waveform, and determine the position of the rising branch end point of the impedance volume ratio signal waveform corresponding to the jth R wave effective peak point on the ECG waveform. Calculate the jth heart rate HR _j of the measured person, HR _j = 1/T _j , where T _j is the time difference between the jth R wave effective peak point on the ECG waveform and the j+1th R wave effective peak point on the ECG waveform, and calculate the jth pulse wave conduction time PWTT _j of the measured person.

Step 2: Calculate the jth diastolic pressure DBP _j and the jth systolic pressure SBP _j according to the jth heart rate and the jth pulse wave transmission time PWTT _j .

Step 3: Introduce the morphological characteristic parameters of the impedance volume ratio signal waveform into the index data set composed of the PTT blood pressure calculation module.

Step 4: Use machine learning based on the extreme random forest model to establish a nonlinear model of blood pressure.

Furthermore, the calculation expression of the j-th diastolic pressure DBP _j is:

DBP _j ＝a*ln PWTT _j +b*HR _j +c*T _RCX +d*T _R +e*T _CX +f*S _BX +g

Furthermore, the calculation expression of the j-th systolic blood pressure SBP _j is:

SBP _j ＝α*ln PWTT _j +β*T _RBC +δ*T _RCX +ε*T _R +γ

The parameter values of a, b, c, d, e, f, g, α, β, δ, ε, γ, etc. are matched and obtained by the measurers, and the cardiopulmonary impedance volume ratio data set of m groups of measurers is obtained, m ≥ 2, and the systolic and diastolic blood pressures are measured simultaneously using a sphygmomanometer. The cardiopulmonary impedance volume ratio data set includes heart rate, pulse wave transmission time, and morphological parameters of the impedance volume ratio waveform. Substituting the m groups of cardiopulmonary impedance volume ratio data sets and the corresponding systolic and diastolic blood pressures into the training, the matching model parameter values can be obtained.

Further, the j-th pulse wave conduction time is the time difference between the j-th wave effective peak point on the PWTT _j electrocardiogram waveform and the corresponding rising branch end point of the impedance volume ratio waveform. The method for segmenting the R wave effective peak point on the electrocardiogram waveform and the rising branch end point of the impedance volume ratio waveform is as follows: if the time of the previous R wave effective peak point of the electrocardiogram waveform is earlier than the time of a rising branch end point of the impedance volume ratio waveform, and there are no other R wave effective peak points of the electrocardiogram waveform and other rising branch end points of the impedance volume ratio waveform between the two, then the two are segmented into a corresponding group.

Furthermore, the method for locating the characteristic peak point of the electrocardiogram waveform or the impedance volume ratio waveform is specifically as follows:

(1) The electrocardiogram waveform or impedance volume ratio waveform is filtered and denoised by wavelet threshold filtering.

(2) The signal curve processed by (1) is differentiated to obtain a derivative function curve. The zero point of the derivative function curve is calculated to determine the positions of the maximum and minimum points of the obtained curve.

(3) Among the extreme value points determined in (2), the points with amplitude greater than the amplitude threshold _XS are selected as the predetermined peak value, _XS = 3/4*( _XSMAX - _XSMIN ) + _XSMIN , where _XSMAX represents the value with the largest data amplitude in (1), and _XSMIN represents the value with the smallest data amplitude in (1).

(4) The n predetermined peaks are divided into i predetermined peak groups, wherein the time positions of any two adjacent predetermined peaks in the i predetermined peak groups differ by more than 0.3 to 0.5 seconds.

(5) By selecting a position point with the largest amplitude in the i determined peak groups, i valid peak points can be obtained.

Furthermore, in step 3, the method for establishing a nonlinear blood pressure model based on the machine learning method of the extreme random forest model is as follows:

First, the morphological characteristic parameters of the impedance volume ratio signal waveform are introduced into the index data set composed of the PTT blood pressure calculation module;

2. Select 6 groups of sample data from the original training set to generate 6 training sets;

3. Train 6 decision tree models for 6 training sets respectively;

4. Select the best feature for splitting a single decision tree model and repeat the above operation multiple times;

5. A random forest consisting of 6 decision trees generated;

6. Obtain the predicted value of blood pressure, and obtain the result of the validation set based on the cross-validation of the test set, so as to evaluate the correctness of the regression model results.

The present invention has the following beneficial effects:

1. The present invention is a non-contact detection of human blood pressure based on capacitive coupling electrodes, which can avoid direct contact of electrodes with the skin and solve the problems of long-term monitoring instability and electrode irritation to the skin.

2. Based on the PTT blood pressure calculation model, the present invention introduces the morphological characteristic parameters of the impedance volume waveform diagram, and further optimizes the non-contact continuous blood pressure prediction model.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

FIG1 is a schematic structural diagram of a non-contact blood pressure detector and method based on capacitive coupling electrodes;

FIG2 is a block diagram of a hardware circuit system of a non-contact blood pressure detector and method based on capacitive coupling electrodes;

FIG3 is a flow chart of the present invention using the extreme random forest method to predict blood pressure values.

DETAILED DESCRIPTION

In order to make the purpose, technical solutions and advantages of the present application more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments.

As shown in Figure 1, a non-contact blood pressure detector and method based on capacitive coupling electrodes include a capacitive coupling excitation electrode 1, a capacitive coupling measurement electrode 2, a capacitive coupling reference electrode 3, an electrocardiogram and impedance volume signal measurement module 4, a data acquisition module 5, a feature extraction module 6, an impedance volume feature parameter calculation module 7, a PTT blood pressure calculation module 8, a PTT fusion impedance volume parameter machine learning module 9 and a blood pressure prediction module 10.

The function of the capacitive coupling excitation electrode 1, the capacitive coupling measurement electrode 2 and the capacitive coupling reference electrode 3 is to complete the synchronous acquisition of the capacitive coupling ECG signal and the capacitive coupling impedance volume signal when the human body is wearing thin clothing and the electrode sheet cannot be directly attached to the skin; the method is: according to the electrode arrangement method of the four-electrode method, the capacitive coupling excitation electrode 1 and the capacitive coupling measurement electrode 2 are fixed to the outside of the human arm clothing, and at the same time, another number of capacitive coupling measurement electrodes 2 and capacitive coupling reference electrodes 3 are also fixed to the periphery of the human chest clothing in the same way as the three-lead chest ECG method.

The ECG and impedance volume signal measurement module is used to synchronously collect capacitive coupled impedance volume ratio (CCIPG) signals and capacitive coupled electrocardiogram (CCECG) signals.

The data acquisition module is used to convert the received chest lead CCECG and arm CCIPG signals into digital signals and transmit them to the controller, so that the controller can obtain the capacitive coupling electrocardiogram waveform and impedance volume ratio waveform of the measured person.

The feature extraction module is used to obtain the pulse wave transmission time (PWTT).

The impedance-volume characteristic parameter calculation module is used to obtain the morphological characteristic parameters of the physiological characteristics of the cardiovascular system reflected on the impedance-volume ratio waveform.

The PTT blood pressure calculation module is used to calculate the blood pressure value 1.

The PTT fusion impedance-volume parameter machine learning module is used to use the morphological characteristic parameters of the physiological characteristics of the cardiovascular system reflected on the impedance-volume ratio waveform and the blood pressure value 1 as sample data for learning and training to establish the final blood pressure model.

The blood pressure prediction module is used to predict the blood pressure value of a human body.

As shown in FIG2 , the hardware system circuit of the present invention includes a coupled ECG measurement unit, a coupled impedance volumetric rate measurement unit and a data acquisition module circuit; first, the coupled ECG measurement unit includes a differential amplifier circuit, a 50Hz trap circuit and a bandpass filter circuit, the input end of the differential amplifier circuit is connected to the ECG coupling measurement electrode, and is used to perform a unit gain differential between the capacitively coupled ECG signal collected by the coupling measurement electrode and the ECG signal of the reference electrode, and at the same time, to draw out the common mode interference of the ECG coupling electrode, and the 50Hz trap circuit and the bandpass filter circuit are used to filter the differentially amplified signal to obtain the chest lead ECG signal; second, the coupled impedance volumetric rate measurement unit includes a main control circuit, a constant current source circuit, a differential amplifier circuit, and a constant current source circuit. The main control circuit controls the constant current source circuit to output 50KHz, 0.3mA current, and adopts the four-electrode method to measure the human arm impedance. The input end of the differential amplifier circuit is connected to the coupling measurement electrode, which is used to differentially amplify the capacitive coupling impedance volume signal collected by the coupling measurement electrode. The processed signal enters the full-wave rectification and differential circuit through the filtering circuit to obtain the arm impedance volume ratio signal; third, the data acquisition module is mainly composed of an A/D circuit, which is used to convert the received chest lead ECG signal and arm impedance volume ratio signal into digital signals and transmit them to the controller, so that the controller obtains the capacitive coupling ECG waveform and impedance volume ratio waveform of the measured person.

The non-contact blood pressure detection method is as follows:

Step 1: locate the effective peak point of the R wave of the collected electrocardiogram waveform to obtain the effective R wave peak position of the electrocardiogram waveform; locate the rising branch end point of the impedance volume ratio signal waveform to obtain the rising branch end point position of the impedance volume ratio signal waveform.

After the jth R wave effective peak point appears on the ECG waveform, determine the position of the jth R wave effective peak point on the ECG waveform, and determine the position of the rising branch end point of the impedance volume ratio signal waveform corresponding to the jth R wave effective peak point on the ECG waveform. Calculate the jth heart rate HR _j of the measured person, HR _j = 1/T _j , where T _j is the time difference between the jth R wave effective peak point on the ECG waveform and the j+1th R wave effective peak point on the ECG waveform, and calculate the jth pulse wave conduction time PWTT _j of the measured person.

Step 2: Calculate the jth diastolic pressure DBP _j and the jth systolic pressure SBP _j according to the jth heart rate and the jth pulse wave transmission time PWTT _j ; obtain the calculation expression of the jth diastolic pressure DBP _j as: DBP _j = a*ln PWTT _j + b*HR _j + c* _TRCX + d* _TR + e* _TCX + f* _SBX + g, and the calculation expression of the jth systolic pressure SBP _j as: SBP _j = α*ln PWTT _j + β* _TRBC + δ* _TRCX + ε* _TR + γ. Among them, the values of parameters such as a, b, c, d, e, f, g, α, β, δ, ε, γ, etc. are matched and obtained by the measurers, and the cardiac impedance volume ratio data groups of m groups of measurers are obtained, m≥2, and the systolic pressure and diastolic pressure are measured synchronously using a sphygmomanometer. The cardiotonic impedance volume ratio data set includes heart rate, pulse wave transmission time and morphological parameters of impedance volume ratio waveform. By substituting m groups of cardiotonic impedance volume ratio data sets and corresponding systolic pressure and diastolic pressure into training, matching model parameter values can be obtained.

Step three, introduce the morphological characteristic parameters of the impedance-volume ratio signal waveform, such as the ratio of BC segment time to single cycle time ( _TRBC ), the ratio of CX segment time to single cycle time ( _TRCX ), the ratio of ejection time to single cycle ( _TR ), the slow ejection period time ( _TCX ), the ejection time ( _TBX ) and the cardiac output per stroke ( _SBX ), into the indicator data set composed of the PTT blood pressure calculation module.

Step 4. Use machine learning based on extreme random forest model to establish a nonlinear model of blood pressure. Random forest is a supervised learning algorithm and an ensemble learning algorithm based on decision tree. Its construction process is as follows: 1. Introduce the morphological characteristic parameters of the impedance volume ratio signal waveform into the index data set composed of the PTT blood pressure calculation module; 2. Select 6 groups of sample data from the original training set to generate 6 training sets; 3. Train 6 decision tree models for the 6 training sets respectively; 4. Select the best features for a single decision tree model to split, and repeat the above operation many times; 5. Create a random forest composed of the 6 generated decision trees; 6. Obtain the predicted value of blood pressure, and obtain the result of the validation set based on the cross-validation of the test set, so as to evaluate the correctness of the regression model results.

What is disclosed above is only a preferred embodiment of the present invention, and it certainly cannot be used to limit the scope of rights of the present invention. Ordinary technicians in this field can understand that all or part of the processes of the above embodiment and equivalent changes made according to the claims of the present invention still fall within the scope of the invention.

Claims (4)

1. The non-contact blood pressure detector based on the capacitive coupling electrode is characterized by comprising a capacitive coupling excitation electrode, a capacitive coupling measurement electrode, a capacitive coupling reference electrode, an electrocardio and impedance volume signal measurement module, a data acquisition module, a characteristic extraction module, an impedance volume characteristic parameter calculation module, a PTT blood pressure calculation module, a machine learning module of PTT fusion impedance volume parameters and a blood pressure prediction module, wherein the capacitive coupling excitation electrode, the capacitive coupling measurement electrode, the capacitive coupling reference electrode, the electrocardio and impedance volume signal measurement module, the data acquisition module, the characteristic extraction module, the impedance volume characteristic parameter calculation module, the PTT blood pressure calculation module, the machine learning module of PTT fusion impedance volume parameters and the blood pressure prediction module are sequentially connected;

The capacitive coupling excitation electrode is used for detecting impedance volume characteristics by being matched with the capacitive coupling measurement electrode and attached to the outer side of the clothes of the human arm;

The capacitive coupling measuring electrode is used for carrying out electrocardiographic signal detection by being matched with the capacitive coupling reference electrode and being attached to the outer side of clothes of the chest of a human body;

the electrocardio and impedance volume signal measuring module is used for synchronously acquiring a capacitive coupling impedance volume rate CCIPG signal and a capacitive coupling electrocardiogram CCECG signal;

The data acquisition module is used for converting the received chest lead CCECG and arm CCIPG signals into digital signals and transmitting the digital signals to the controller, so that the controller obtains a capacitive coupling electrocardiogram waveform and an impedance volume rate waveform diagram of a measurer;

The feature extraction module is used for acquiring pulse wave transmission time PWTT;

The impedance volume characteristic parameter calculation module is used for obtaining morphological characteristic parameters of the physiological characteristics of the cardiovascular system reflected on the impedance volume rate waveform diagram;

The PTT blood pressure calculation module is used for calculating a blood pressure value 1;

the machine learning module of PTT fusion impedance volume parameter is used for learning and training morphological characteristic parameters of cardiovascular system physiological characteristics reflected on the impedance volume rate waveform chart and the blood pressure value 1 as sample data and establishing a final blood pressure model;

the blood pressure prediction module is used for predicting the blood pressure value of the human body.

2. The non-contact blood pressure monitor based on capacitive coupling electrode according to claim 1, wherein the morphological characteristic parameters of the physiological characteristics of the cardiovascular system of the human body reflected on the impedance volume ratio waveform chart comprise: the ratio of time in the BC segment TRBC to the monocycle time ratio TRCX to the CX segment, the monocycle ratio TR to the ejection time, the slow ejection period time TCX, the ejection time TBX and the stroke volume SBX.

3. The capacitive coupling electrode-based noncontact blood pressure monitor as claimed in claim 1, wherein:

the blood pressure value 1 includes a jth diastolic blood pressure DBP _j and a jth systolic blood pressure SBP _j;

The calculation expression of the j-th diastolic pressure DBP _j is:

DBP_j＝a*ln PWTT_j+b*HR_j+c*T_RCX+d*T_R+e*T_CX+f*S_BX+g;

The calculation expression of the jth systolic pressure SBP _j is:

SBP_j＝α*ln PWTT_j+β*T_RBC+δ*T_RCX+ε*T_R+γ；

Wherein HR _j＝1/T_j,T_j represents the time difference between the jth R wave effective peak point on the capacitively coupled electrocardiogram waveform and the (j+1) th R wave effective peak point on the capacitively coupled electrocardiogram waveform; PWTT _j represents the time difference between the jth R wave effective peak point on the capacitively coupled electrocardiogram waveform and the rising branch end point of the corresponding impedance volume rate waveform diagram; a. b, c, d, e, f, g, alpha, beta, delta, epsilon and gamma parameter values are matched and valued by a measurer to obtain electrocardiographic impedance volume rate data sets of m groups of measurer, wherein m is more than or equal to 2, a sphygmomanometer is synchronously used for measuring systolic pressure and diastolic pressure, morphological parameters including heart rate, pulse wave conduction time and impedance volume rate oscillograms in the electrocardiographic and impedance volume rate data sets, and the m groups of electrocardiographic impedance volume rate data sets and corresponding systolic pressure and diastolic pressure are substituted into training to obtain model parameter values matched with the electrocardiographic and diastolic pressure data sets.

4. The capacitive coupling electrode-based noncontact blood pressure monitor as claimed in claim 1, wherein: the method for establishing the nonlinear blood pressure model based on the machine learning mode of the extremely random forest model comprises the following steps of firstly, introducing morphological characteristic parameters of an impedance volume rate signal oscillogram into an index data set formed by a PTT blood pressure calculation module, secondly, selecting 6 groups of sample data from an original training set, and generating 6 training sets; 3. respectively training 6 decision tree models for the 6 training sets; 4. selecting the best feature for splitting aiming at a single decision tree model, and repeating the operation for a plurality of times; 5. a random forest consisting of 6 generated decision trees is formed; 6. and obtaining a predicted value of the blood pressure, and obtaining a result of the verification set according to the cross verification of the test set, thereby evaluating the correctness of the regression model result.