CN103371814A - Remote wireless electrocardiograph monitoring system and feature extraction method on basis of intelligent diagnosis - Google Patents

Abstract

A remote wireless ECG monitoring system based on intelligent diagnosis and a feature extraction method, including an ECG acquisition subsystem, a blood pressure acquisition subsystem, an intelligent diagnosis unit, a GPS positioning subsystem, a human-computer interaction subsystem, a wireless transmission subsystem, a memory unit and System processor unit; after wearing the system, the user can monitor the ECG signal in real time and display the ECG waveform in real time. When the heart is abnormal, the system processor will control the blood pressure acquisition subsystem and the GPS positioner. The system obtains the user's current blood pressure value and position coordinates, and uses the wireless transmission subsystem to send the above information to each monitoring terminal associated with the user; the feature extraction method is based on a combination of threshold and singular point detection to achieve heart The location of R wave and ST segment in the electrical signal and the extraction of other ECG features.

Description

Remote wireless ECG monitoring system and feature extraction method based on intelligent diagnosis

technical field

The invention relates to a remote wireless ECG monitoring system based on intelligent diagnosis.

Background technique

At present, the diagnosis of heart disease is mainly based on electrocardiogram and cardiac color Doppler ultrasound. Electrocardiography is mainly used in the diagnosis of functional heart disease such as arrhythmia and myocardial ischemia; it works. Since functional diseases often have the characteristics of concealment, sudden onset, high risk, and short duration, it is difficult to grasp the best diagnosis time, and it is often difficult for patients to take effective self-rescue measures when they develop. Therefore, the patient's ECG characteristics have a very important guiding role in the diagnosis and treatment of sudden heart disease.

At present, there are three types of ECG collection equipment commonly used in medical institutions: 1. ECG monitors, which can detect vital signs such as heart rate, respiration, blood pressure, pulse, and blood oxygen saturation of patients. Wards are very common. The equipment collects the above physiological indicators of the patient and displays them on the screen in real time, which is of great significance for the care of postoperative or critically ill patients. However, most of the ECG signals collected by this type of equipment are single-lead signals, which often only reflect whether the patient's heart is beating normally, and do not have much reference value for the diagnosis of heart disease. 2. Electrocardiography machine, currently the most commonly used number of leads is twelve leads. This type of ECG acquisition equipment can collect relatively complete ECG signals, and can comprehensively reflect various cardiac functional abnormalities. However, this type of equipment generally has a large structure, and can only record ECG signals for a short time each time, requiring professional operation. Therefore, only large medical institutions such as hospitals can use this equipment for detection. Many heart diseases have the characteristics of sudden onset and short duration. Patients often go to the hospital for diagnosis when they feel abnormal. It will cost a lot of human and financial resources, which is unbearable for many patients. Therefore, there are still many disadvantages in practical application of this type of ECG acquisition device. 3. Dynamic ECG acquisition equipment (Holter), this type of equipment is mainly designed to solve the problem that ordinary ECG machines are difficult to complete all-weather monitoring. This type of equipment is small in size and easy to wear, mainly including the collection and storage of ECG data Partly, the large-capacity storage device stores the 24-hour ECG signals, and then transmits these signals to the computer for preliminary screening, and sends the possible abnormal parts to the doctor for empirical judgment. This type of equipment has solved the existing problems of common electrocardiographs to a certain extent, but there are still many problems to be solved urgently. First of all, this type of equipment is very expensive, and many patients who need to wear it for a long time cannot afford its huge cost. Secondly, this kind of equipment has a low level of intelligence, and it needs to go through multiple processes to screen out the pathological signals that are finally used for diagnosis, which consumes a lot of manpower and material resources. And this equipment is also helpless for some acute and sudden fatal morbidity situations that make people lose self-rescue ability, and can not effectively shorten the time required for rescue. Therefore, it can not fully meet all the needs of cardiac patient monitoring.

Contents of the invention

The purpose of the present invention is to provide a remote wireless ECG monitoring system and feature extraction method based on intelligent diagnosis.

The present invention is a remote wireless ECG monitoring system based on intelligent diagnosis and a feature extraction method. The remote wireless ECG monitoring system includes an ECG acquisition subsystem, a blood pressure acquisition subsystem, an intelligent diagnosis unit, a GPS positioning subsystem, and a human-computer interaction system. The subsystem, the wireless transmission subsystem, the memory unit and the system processor unit are characterized in that the intelligent diagnosis function and the wireless transmission function are combined and applied to the ECG monitoring equipment, wherein: the ECG signal collected by the electrode sheet is passed through A/ The D conversion module converts the digital signal to the intelligent diagnosis unit after preprocessing and feature extraction. At the same time, the blood pressure acquisition subsystem sends the collected blood pressure signal to the processor as auxiliary information for ECG characteristic diagnosis. The intelligent diagnosis unit obtains The final diagnosis result is handed over to the processor for storage, and the ECG waveform and diagnosis result are displayed on the LCD screen in real time. The GPS positioning subsystem locates the patient's current precise position, and the processor transmits the diagnosis result and position information through the wireless transmission subsystem. sent to the monitor.

The feature extraction method of the remote wireless ECG monitoring system based on intelligent diagnosis is characterized by using the R wave detection algorithm, and adopting a method based on a combination of threshold and singular point detection to realize the positioning of the R wave and S-T segment in the ECG signal and The extraction of other ECG features, its steps are:

(1) set a time window, adopt quadratic difference method in this period of time, find waveform singular point; Length is N (get N=4096) data and carry out singular point detection;

The algorithm used for singular point detection is to perform diff(sign(diff(N))) operation on the ECG signal data of one sampling period according to formulas (1) to (3):

Among them, diff is the signal difference, sign is the sign function, N is the ECG signal of a sampling period, and the point where the operation result is -2 is the maximum value point of the ECG signal;

Keep all the maximum values obtained in an array A, and record their corresponding waveform positions;

Take the threshold value Rth, compare it with the value of the above array A, keep the maximum value greater than Rth in the array B, and record its corresponding waveform position;

It is considered that the data in B is the detected R wave amplitude, and its corresponding position is the R wave position;

(2) Adaptive threshold Rth setting:

After the selected N-point ECG filtered data, after the singular point detection, find the maximum value signal amplitude range in the singular point, divide it into 15 parts according to the formula (4), and record the amplitude of each part in Th in the array;

Let the integral projection function be:

in:

$\mathrm{the\; y}\left(i\right)=\min \left(A\right)+i*\frac{\max \left(A\right)-\min \left(A\right)}{15},$ 1≤i≤15 and i∈Z (6)

That is, do the integral projection of the maximum value point on each divided amplitude segment Th(i);

Select the i value at the junction point between the zero distribution and the non-zero distribution, calculate Th(i), and determine Rth as Th(i);

After the threshold determines the R wave position and amplitude, calculate the average heart rate Rate, unit: beats/minute,

Rate＝60*Nr/(nr(end)-nr(1)/f _s ) (8)

where Nr is the number of R waves measured in a fixed time window, nr(end) is the position of the last R wave in a fixed time window, nr(1) is the position of the first R wave in a fixed time window, f _s is the data sampling rate;

According to the average heart rate, the relevant false detection and missed detection correction are added, and the correct R wave position and average heart rate Rate are finally corrected;

Then use formula (9) to calculate the HRV interval.

HRV＝RR(i+1)-RR(i) (9)

Where RR is the interval between two adjacent R waves.

Based on the above algorithm, the characteristic parameters required for subsequent diagnosis can be obtained: Rate and HRV;

(3) According to this indicator, the extracted ECG characteristics and MIT-BI H database were analyzed and summarized to obtain the diagnosis result by means of table lookup.

The invention provides an efficient and stable monitoring system which integrates intelligent diagnosis and remote monitoring, aiming at the problems existing in the existing electrocardiographic monitoring equipment in the process of diagnosing and monitoring heart disease patients. The system can monitor the patient's ECG and blood pressure changes in real time, and perform comprehensive intelligent diagnosis. When there are lesion characteristics, it will send a diagnosis report to the patient's family or doctor through the wireless network in time, and can locate the patient in real time through the GPS positioning system. The location shortens the time required for rescue and reduces the tragedy of patient death due to poor rescue.

Description of drawings

Fig. 1 overall system block diagram of the present invention, Fig. 2 (a) framework diagram of the ECG acquisition subsystem of the present invention, Fig. 2 (b) flow chart of the intelligent diagnosis part of the ECG acquisition subsystem of the present invention, Fig. 3 (a) blood pressure acquisition of the present invention Subsystem architecture diagram, Fig. 3 (b) flow chart of blood pressure acquisition subsystem of the present invention, Fig. 4 (a) architecture diagram of GPS positioning subsystem of the present invention, Fig. 4 (b) flow chart of GPS positioning subsystem of the present invention, Fig. 5 ( a) Architecture diagram of the human-computer interaction subsystem of the present invention, Fig. 5 (b) flow chart of the human-computer interaction subsystem of the present invention, Fig. 6 (a) architecture diagram of the wireless transmission subsystem of the present invention, Fig. 6 (b) wireless transmission of the present invention Subsystem flow chart, symbol description in the figure: 11-ECG characteristic signal collected by the ECG collection subsystem, 12-blood pressure characteristic signal collected by the blood pressure collection subsystem, 13-processor's control of the ECG collection subsystem Signal, 14-the control signal of the processor to the blood pressure acquisition subsystem, 15-the result of intelligent diagnosis of physiological parameters such as comprehensive ECG and blood pressure, 16-the orientation information returned by the GPS positioning subsystem to the processor, 17-the processor sends to Screen refresh information of the LCD screen, 18-the touch screen key position returned to the processor by the touch screen driver, and 19-instructions and diagnostic information sent by the processor to the wireless module.

Detailed ways

As shown in Figure 1, the remote wireless ECG monitoring system based on intelligent diagnosis of the present invention includes: ECG acquisition subsystem, blood pressure acquisition subsystem, intelligent diagnosis unit, GPS positioning subsystem, human-computer interaction subsystem, wireless transmission subsystem System, memory unit, system processor unit. The 24-hour intelligent monitoring of patients is completed through the above eight subsystems. The processor realizes the switching of leads and the realization of the A/D conversion function through the control signal 13 of the ECG acquisition subsystem; the processor realizes the inflation and deflation of the air pump and the reading of the blood pressure value through the control signal 14 of the blood pressure acquisition subsystem. The electrocardiographic characteristic signal 11 extracted by the electrocardiographic acquisition subsystem is sent to the intelligent diagnosis unit in the processor for further analysis; the blood pressure value signal 12 collected by the blood pressure acquisition subsystem is sent to the processor as an auxiliary electrocardiographic characteristic diagnosis Information, the final diagnosis result 15 obtained by the intelligent diagnosis unit is stored by the processor; the GPS positioning subsystem sends time, coordinates, speed and other information 16 to the processor in the form of an interrupt through the serial port, and the processor extracts from it according to the format matching method The required coordinate information; the human-computer interaction subsystem is composed of a touch screen and a display screen, and various interaction requirements and function displays between the user and the processor are realized through signals 17 and 18; the communication between the wireless transmission subsystem and the processor Communicate with each other through serial ports, and the diagnostic information and position coordinates 19 obtained by the processor are sent to the monitoring terminal through the wireless module.

As shown in Figure 2(a), the ECG acquisition subsystem mainly includes ECG electrodes, a lead controller and a signal preprocessing part; the ECG electrodes are connected to the lead controller, and the output of the lead controller is amplified, A/D conversion is performed after the filter circuit, the output end of the A/D conversion module is connected to the wavelet filter circuit, and the output result of the wavelet filter circuit is sent to the intelligent diagnosis unit.

In this embodiment, the lead controller is implemented by the analog switch MAX397, the amplifier circuit and the filter circuit are built by the operational amplifier AD620 and TLV2254, the A/D conversion chip adopts TLC1549, the processor adopts Altera Cyclonell 2C70FPGA, and the coif5 wavelet is implemented as a hardware circuit on the FPGA. way of design and implementation.

The ECG acquisition subsystem collects ECG signals from the surface of the human body through electrode sheets, and the lead controller connects different leads in time divisions in a time-division multiplexing manner, realizing the simultaneous acquisition of twelve-lead ECG signals. The collected ECG signal is very weak, usually between 0.5 and 2mv. The differential signal is superimposed on the DC voltage component of about 300mv generated by the contact between the electrode and the skin, which also contains the electric potential between the electrode and the ground. The generated common-mode voltage of 1.5v, so the present invention adopts the instrument amplifier AD620 to build an amplifying circuit, the operational amplifier is close to 1kHz with a common-mode rejection ratio up to 100dB, a maximum input offset voltage of 50uV, and a maximum input bias current of 1nA , the performance can meet the requirements of developed ECG signal here. Since the frequency of the ECG signal is between 0.05 and 100Hz, a low-pass filter is needed to eliminate high-frequency noise above 100Hz, a high-pass filter is used to filter out low-frequency noise below 0.05Hz, and a The notch circuit is used to eliminate 50Hz power frequency interference. The cut-off frequency of the filter is f=1/2πRC, and 0.05Hz, 50Hz, and 100Hz are brought into the above formula respectively, and the parameters of each filter can be obtained. The filter is realized by using the operational amplifier TLV2254. This design uses TLC 1549 as the analog-to-digital converter. This chip is a serial 10-bit A/D conversion chip with few data ports and can meet the requirements of the present invention for data accuracy. The ECG signal collected from the body surface is delivered to the wavelet filter circuit in the FPGA in the form of a digital signal after the above preprocessing. The present invention adopts the coif5 filter, and through three layers of decomposition and reconstruction, the filtering process is carried out in the FIFO in the FPGA. The (FIFO) unit is completed, and 512 data are filtered each time.

Implementation of the blood pressure collection subsystem: many heart disease patients often suffer from abnormal blood pressure, so the present invention adds a blood pressure collection function. As shown in Figure 3(a), the blood pressure acquisition subsystem mainly includes an air pump controller, an air pump, a pressure sensor, and an A/D conversion module. The processor is connected to the air pump controller, and the air pump controller is connected to the air pump. The air pump controller can control the inflation and deflation of the air pump, and can read the air pressure value in the cuff through the pressure sensor. The pressure sensor is installed in the cuff, and the processor reads the current pressure value measured by the pressure sensor through the TLC1549 analog-to-digital converter, and the pulse condition can be determined through the change of pressure.

As shown in Figure 3(b), when the processor needs to collect the current blood pressure, the air pump controller controls the air pump to inflate the cuff, and the pressure sensor detects the current pulse at all times. When the arterial pulse disappears, the air pump controller controls the air pump to deflate slowly , when the pressure sensor detects the pulse again, write down the air pressure value P1 in the cuff, and the air pump continues to deflate, when the pulse disappears again, record the air pressure value P2 in the cuff, and the two air pressure values obtained after data correction The final blood pressure value including systolic blood pressure and diastolic blood pressure can be obtained.

As shown in Figure 2(b), the intelligent diagnosis unit is the core of the present invention, and this unit includes two main functions of feature extraction and intelligent diagnosis, and this unit is realized by a hardware circuit designed in FPGA. A standard electrocardiogram mainly includes P wave, QRS wave group, T wave, PR interval, ST segment, QT interval, HRV interval, etc. Heart disease often causes changes in the amplitude and interval of these waves, which is The basis of intelligent diagnosis of the present invention.

The feature extraction function unit in the intelligent diagnosis unit mainly detects R wave, average heart rate Rate and HRV interval. The QRS complex is the most significant part of an electrocardiographic cycle feature. It has high amplitude and high slope, time-domain waveform features with a certain width, and frequency-domain features (peak frequency at 10-20Hz).

As shown in Figure 4(a), the GPS module uses RS232 level logic, so it needs to go through the serial interface chip for level conversion before connecting with the processor. Refer to Figure 4(b) for the workflow of the GPS positioning subsystem. The GPS module sends information including time, coordinates, altitude, speed, etc. to the processor in the form of a serial port interrupt. To locate the complete data package, and extract the coordinate information from it to determine the current location of the patient.

As shown in Figure 5(a), the human-computer interaction subsystem mainly includes liquid crystal display and touch screen input functions. The liquid crystal display part of the present invention adopts Y70-4024-65K color liquid crystal module, which is a color TFT liquid crystal display widely used in industrial control and other equipment. It has a 7-inch super large display area and a resolution of 400*240 , which greatly reduces the pressure of system data transmission and storage, adopts 16-bit standard 8080 bus interface mode, supports 65536 colors to make the image more delicate, unique 2-page video memory, and operating one page alone does not affect other pages, which is convenient to meet the requirements of this project Requires real-time partial screen refresh. The invention develops a set of friendly human-computer interaction interface based on the screen, which can conveniently display various detection information and function menus. The display part of the touch screen mainly includes two parts: the touch screen and the touch screen driver. The present invention adopts a 7-inch resistive touch screen. After the touch screen is powered on, pressing different positions of the touch screen can cause corresponding changes in the output potential. The situation can accurately locate the user's touch position and function options. The present invention uses ADS7843 as a touch screen controller. ADS7843 is a 4-wire resistive touch screen conversion interface chip produced by TI. It is a 12-bit synchronous serial interface. The sampling analog-to-digital converter can locate the abscissa and ordinate of the touch position of the touch screen connected to it through two A/D conversions. This subsystem can complete the display of equipment operation and user command input, that is, the human-computer interaction function.

As shown in Figure 6(a), the present invention mounts the electrocardiographic diagnosis equipment on the communication network in the form of a mobile terminal, so that the communication between the terminal and the monitoring terminal can be realized very conveniently by means of a public wireless network. The present invention adopts MC323CDMA module, which adopts RS232 level logic, and is connected with the processor through a serial port conversion chip, and the processor sends instructions to the wireless module through the serial port in the form of AT commands. The design uses short messages to send patient diagnosis information to the monitoring terminal. The short message sending process is shown in Figure 6(b). After the processor completes the initialization of the baud rate for the wireless module, it enters the short message sending process. After the above operations are completed, the wireless module can send diagnostic information to the monitoring terminal through the telecom CDMA network.

The feature extraction method of the remote wireless ECG monitoring system based on intelligent diagnosis, using the R wave detection algorithm, adopts the method based on the combination of threshold and singular point detection to realize the positioning of the R wave and S-T segment in the ECG signal and other ECG signals. Feature extraction, the steps are:

(1) set a time window, adopt quadratic difference method in this period of time, find waveform singular point; Length is N (get N=4096) data and carry out singular point detection;

The algorithm used for singular point detection is to perform diff(sign(diff(N))) operation on the ECG signal data of one sampling period according to formulas (1) to (3):

Among them, diff is the signal difference, sign is the sign function, N is the ECG signal of a sampling period, and the point where the operation result is -2 is the maximum value point of the ECG signal;

Keep all the maximum values obtained in an array A, and record their corresponding waveform positions;

Take the threshold value Rth, compare it with the value of the above array A, keep the maximum value greater than Rth in the array B, and record its corresponding waveform position;

It is considered that the data in B is the detected R wave amplitude, and its corresponding position is the R wave position;

(2) Adaptive threshold Rth setting:

After the selected N-point ECG filtered data, after the singular point detection, find the maximum value signal amplitude range in the singular point, divide it into 15 parts according to the formula (4), and record the amplitude of each part in Th in the array;

$\mathrm{Th}\left(i\right)=\min \left(A\right)+i*\frac{\max \left(A\right)-\min \left(A\right)}{15},$ 1≤i≤15 and i∈Z (4)

Let the integral projection function be:

$S\left(i\right)=\underset{\mathrm{the\; y}(i-1)}{\overset{\mathrm{the\; y}\left(i\right)}{\mathrm{\Σ}}}I\left(x\right),$ 1≤i≤15 and i∈Z (5)

in.

$\mathrm{the\; y}\left(i\right)=\min \left(A\right)+i*\frac{\max \left(A\right)-\min \left(A\right)}{15},$ 1≤i≤15 and i∈Z (6)

That is, do the integral projection of the maximum value point on each divided amplitude segment Th(i);

Select the i value at the junction point between the zero distribution and the non-zero distribution, calculate Th(i), and determine Rth as Th(i);

After the threshold determines the R wave position and amplitude, calculate the average heart rate Rate, unit: beats/minute,

Rate＝60*Nr/(nr(end)-nr(1)/f _s ) (8)

where Nr is the number of R waves measured in a fixed time window, nr(end) is the position of the last R wave in a fixed time window, nr(1) is the position of the first R wave in a fixed time window, f _s is the data sampling rate;

According to the average heart rate, the relevant false detection and missed detection correction are added, and the correct R wave position and average heart rate Rate are finally corrected;

Then use formula (9) to calculate the HRV interval.

HRV＝RR(i+1)-RR(i) (9)

Where RR is the interval between two adjacent R waves.

Based on the above algorithm, the characteristic parameters required for subsequent diagnosis can be obtained: Rate and HRV;

(3) According to this indicator, the extracted ECG characteristics and MIT-BI H database were analyzed and summarized to obtain the diagnosis result by means of table lookup.

The heart disease that can be diagnosed by the present invention includes atrial premature beat, ventricular premature beat, ventricular tachycardia, asystole, atrial fibrillation, ventricular fibrillation, atrial flutter, ventricular fibrillation, supraventricular tachycardia and handover escape.

Claims (6)

1. Remote wireless ECG monitoring system based on intelligent diagnosis, including: ECG acquisition subsystem, blood pressure acquisition subsystem, intelligent diagnosis unit, GPS positioning subsystem, human-computer interaction subsystem, wireless transmission subsystem, memory unit and system The processor unit is characterized in that the intelligent diagnosis function and the wireless transmission function are combined and applied to the ECG monitoring equipment, wherein: the ECG signal collected by the electrode piece is converted into a digital signal by the A/D conversion module and then preprocessed and processed. After the feature is extracted, it is handed over to the intelligent diagnosis unit. At the same time, the blood pressure acquisition subsystem sends the collected blood pressure signal to the processor as auxiliary information for ECG feature diagnosis. The final diagnosis result obtained by the intelligent diagnosis unit is handed over to the processor for storage and stored. The ECG waveform and diagnosis results are displayed on the LCD screen in real time, the GPS positioning subsystem locates the patient's current precise location, and the processor sends the diagnosis results and location information to the monitoring terminal through the wireless transmission subsystem. 2. The remote wireless ECG monitoring system based on intelligent diagnosis according to claim 1, wherein the ECG acquisition subsystem is composed of ECG signal conditioning circuit, wavelet filter and ECG feature extraction module, wherein wavelet filter Using coif5 wavelet filtering. 3. The remote wireless ECG monitoring system based on intelligent diagnosis according to claim 1, characterized in that the GPS positioning subsystem is composed of a serial port interface chip, a GPS module, and an external GPS antenna. 4. The remote wireless ECG monitoring system based on intelligent diagnosis according to claim 1, characterized in that the feature extraction function unit in the intelligent diagnosis unit is used to detect R wave, average heart rate Rate and HRV interval. 5. The remote wireless ECG monitoring system based on intelligent diagnosis according to claim 1, characterized in that the wireless transmission subsystem includes a serial port conversion chip, a GSM/CDMA modem, and an external antenna, and the wireless transmission subsystem can be used to transfer the device Mount it on the public network in the form of a mobile terminal. 6. The feature extraction method of the remote wireless ECG monitoring system based on intelligent diagnosis, which is characterized in that the R wave detection algorithm is used, and the method based on the combination of threshold and singular point detection is used to realize the R wave and S-T segment in the ECG signal Location and extraction of other ECG features, the steps are: (1) set a time window, adopt quadratic difference method in this period of time, find waveform singular point; Length is N (get N=4096) data and carry out singular point detection; The algorithm used for singular point detection is to perform diff(sign(diff(N))) operation on the ECG signal data of one sampling period according to formulas (1) to (3):

Among them, diff is the signal difference, sign is the sign function, N is the ECG signal of a sampling period, and the point where the operation result is -2 is the maximum value point of the ECG signal; Keep all the maximum values obtained in an array A, and record their corresponding waveform positions; Take the threshold value Rth, compare it with the value of the above array A, keep the maximum value greater than Rth in the array B, and record its corresponding waveform position; It is considered that the data in B is the detected R wave amplitude, and its corresponding position is the R wave position; (2) Adaptive threshold Rth setting: After the selected N-point ECG filtered data, after the singular point detection, find the maximum value signal amplitude range in the singular point, divide it into 15 parts according to the formula (4), and record the amplitude of each part in Th in the array;

Let the integral projection function be: in: $\mathrm{the\; y}\left(i\right)=\min \left(A\right)+i*\frac{\max \left(A\right)-\min \left(A\right)}{15},$ I≤i≤15 and i∈Z (6)

That is, do the integral projection of the maximum value point on each divided amplitude segment Th(i); Select the i value at the junction point between the zero distribution and the non-zero distribution, calculate Th(i), and determine Rth as Th(i); After the threshold determines the R wave position and amplitude, calculate the average heart rate Rate, unit: beats/minute, Rate＝60*Nr/(nr(end)-nr(1)/f _s ) (8) where Nr is the number of R waves measured in a fixed time window, nr(end) is the position of the last R wave in a fixed time window, nr(1) is the position of the first R wave in a fixed time window, f _s is the data sampling rate; According to the average heart rate, the relevant false detection and missed detection correction are added, and the correct R wave position and average heart rate Rate are finally corrected; Then use formula (9) to calculate the HRV interval. HRV＝RR(i+1)-RR(i)            (9) Where RR is the interval between two adjacent R waves. Based on the above algorithm, the characteristic parameters required for subsequent diagnosis can be obtained: Rate and HRV; (3) According to this indicator, the extracted ECG characteristics and MIT-BI H database were analyzed and summarized to obtain the diagnosis result by means of table lookup.

Images