CN114515151A - Electrocardiosignal acquisition system and processing method based on artificial intelligence - Google Patents

Abstract

The invention discloses an electrocardiographic signal acquisition system and processing method based on artificial intelligence, including a physiological signal acquisition module, a wireless transmission module and an intelligent terminal module. The wireless transmission module will wirelessly transmit the analog signal collected by the acquisition module after A/D conversion; the intelligent terminal will process and analyze the received digital signal, and evaluate the physiological condition of the seriously injured and sick, so as to realize the treatment of severe injuries. Wireless continuous monitoring of patients, and automatic alarm for abnormal indicators. The electrocardiographic signal acquisition system of the present invention has a simple equipment structure, can be quickly worn on the patient and continuously monitor, and the acquired signal is denoised by the internal program setting of the system, and the high-quality signal is screened for signal feature extraction and analysis. Accurately determine the origin of the patient's arrhythmia and give medical advice and treatment plans in a timely manner.

Description

ECG signal acquisition system and processing method based on artificial intelligence

technical field

The invention belongs to the technical field of signal processing, and in particular relates to an electrocardiographic signal acquisition system and a processing method based on artificial intelligence.

Background technique

An electrical excitation from the sinoatrial node in each cycle of the heart, the electrical excitation is transmitted to the atrium and ventricle in turn according to a certain route and time course, triggering the excitation of the entire heart, causing the heart to periodically contract, thereby promoting blood circulation throughout the body. In each cardiac cycle, the time, path and sequence of electrical changes in the excitation process of various parts of the heart have certain rules. The measuring electrode is placed on a certain part of the human body surface, and the recorded cardiac electrical change curve is the ECG electrocardiogram, so many cardiac physiological conditions can be obtained through the ECG signal. At the same time, in emergency medical treatment, in order to monitor the vital signs of patients in real time, various instruments need to work together to obtain various data quickly and accurately, and the existing processing methods are mostly blood pressure meters, heart rate meters, oximeters, and ECG monitoring. Instruments such as instruments are monitored separately, and then the medical staff records and analyzes whether the data is abnormal, and then can draw conclusions about the patient's life status. This method involves complicated instruments, cumbersome data, and manual results are delayed, which affects the treatment of the wounded and sick. Therefore, how to achieve efficient, accurate and long-term continuous monitoring of physiological signals such as ECG, blood pressure, electromyography, pulse and body temperature of critically wounded patients, and to carry out evaluation on the current physiological conditions of critically injured patients according to the obtained physiological data. Rapid assessment and early warning to improve the survival rate of critically wounded patients is an urgent problem.

SUMMARY OF THE INVENTION

In order to solve the above problems, the present invention designs an electrocardiographic signal acquisition system and processing method based on artificial intelligence. The equipment has a simple structure and can be quickly worn on the patient and continuously monitored. The collected signals are set by the internal program of the system. Perform noise removal, high-quality signal screening, signal feature extraction and analysis, accurately determine the origin of the patient's arrhythmia, and give medical advice and treatment plans in a timely manner.

In order to achieve the above object, the present invention adopts the following technical solutions: an electrocardiographic signal acquisition system based on artificial intelligence, including a physiological signal acquisition module, a wireless transmission module and an intelligent terminal module, the signal acquisition module includes an electrocardiographic signal acquisition module, and its characteristics are: It is as follows: the physiological signal acquisition module realizes the acquisition of various physiological signals of critically ill patients; the wireless transmission module performs A/D conversion on the analog signals collected by the acquisition module for wireless transmission; the intelligent terminal processes the received digital signals with It analyzes and evaluates the physiological condition of the seriously injured and sick, realizes wireless continuous monitoring of the critically wounded and sick, and automatically alarms the abnormal indicators.

As a further improvement of the artificial intelligence-based electrocardiographic signal acquisition system of the present invention: the signal acquisition module includes wearable equipment and sensors arranged on the wearable equipment, and the wearable equipment includes mutual leads arranged according to the human body structure The plurality of elastic installation belts are spliced movably between the plurality of elastic installation belts; the wireless transmission module adopts an anti-interference technology based on frequency agility to avoid signal interference.

As a further improvement of an artificial intelligence-based ECG signal acquisition system of the present invention: the signal acquisition module further includes a pulse, blood oxygen, and EEG data acquisition module set on the head and an electromyography set at a corresponding position on the body , Cardiopulmonary sound, ECG, body temperature data monitoring module.

As a further improvement of the artificial intelligence-based ECG signal acquisition system of the present invention: the intelligent terminal module includes an intelligent control terminal and a central control system adapted to the network of the physiological signal acquisition module through the wireless transmission module. The module transmits the collected data information to the central control system through wireless transmission for data storage and analysis. The central control system feeds back the processed information to the intelligent control terminal. The intelligent control terminal is provided with an alarm module, and the central control system monitors the ECG. , pulse, blood oxygen, body temperature and other parameters to set thresholds, the central control system compares the processed parameters with the thresholds and feeds them back to the intelligent control terminal alarm module to achieve abnormal parameter alarms.

In order to achieve the above object, the present invention adopts the following technical solutions: the signal processing method of the ECG signal acquisition system includes:

1) ECG signal acquisition module collects ECG signals;

2) Preprocessing of ECG signal;

The ECG signal is analyzed and denoised by wavelet transform;

The low-scale wavelet coefficients d1 and d2 are processed by the soft threshold method, and the high-frequency noise is eliminated by the large-scale attenuation coefficient method;

The Scale 3 wavelet coefficients containing important input signals are processed by using the soft and hard threshold compromise algorithm;

2) Judging the quality of ECG signals, screening good quality signals;

Screen signals of good quality according to but not limited to the threshold value and the number of samples in the peak part; 3) ECG signal feature extraction

The detection of the singular point of ECG signal is completed based on the method of bi-orthogonal quadratic B-spline wavelet transform; the dynamic threshold method is used to improve the positioning accuracy of R wave. Analyze the wavelet coefficient D4, extract the extremum pairs, use threshold processing to filter to obtain the extremum pairs of accurately positioned R waves, restore the position of the extremum pairs to the reconstructed ECG signal, find the maximum value in the maximum value interval, and obtain R Wave positioning; after positioning the R wave, the Q and S waves can be located by finding the minimum point near the R wave according to the position of the R wave;

Using the positions of two adjacent R waves as benchmarks, the length of the detection interval is controlled by the mean value of the RR interval, so as to accurately detect the two heart beats, the respective T waves and the P waves between the two adjacent R waves; The respective detection areas of T wave and P wave are measured to complete the precise positioning of P, Q, R, S and T waves of ECG signals;

4) Signal analysis

The data obtained in step 3) are divided into two first-level sub-categories, namely, intraventricular block in abnormal activation origin and abnormal activation conduction, which are used to train the support vector machine SVM1;

The abnormal origin of excitation is classified into sinus rhythm and ectopic rhythm, which are used to train support vector machine SVM2;

The ectopic heart rate is divided into active ectopic heart rhythm and passive ectopic heart rhythm for training SVM3; active ectopic heart rhythm is divided into two categories, namely supraventricular premature beats and ventricular premature beats, for training SVM4;

Passive ectopic heart rhythms are divided into two categories, supraventricular escape beats and ventricular escape beats, which are used to train SVM5.

As a further improvement of the artificial intelligence-based electrocardiographic signal acquisition system of the present invention, the electrocardiographic signals are divided into seven categories by five support vector machines, so as to monitor the heart condition of the wounded and sick in real time.

The present invention collects, stores, analyzes and feeds back signals such as human electrocardiogram in real time by setting the movable installation belt of the signal acquisition sensor, and the medical staff quickly accepts the feedback results, so as to save the treatment time, help the medical staff to quickly respond to the call for help, and finally improve the injury and illness. staff rescue rate. The movable spliced installation belt is convenient for patients to wear them for vital signs monitoring at the first time. The central control system performs preset program processing on the collected ECG signals, denoises the ECG signals, and filters out low-quality signals. High-quality signals are used to extract ECG signal features, analyze and compare the signals, accurately judge the symptoms of arrhythmia, quickly trace the origin of the cause of arrhythmia, and give timely medical advice and treatment plans.

Detailed ways

The present invention will be further described below with reference to the accompanying drawings and specific embodiments. The present embodiment provides a detailed implementation manner on the premise of the technical solution of the present invention.

The artificial intelligence-based electrocardiographic signal acquisition system of the present invention includes a physiological signal acquisition module, a wireless transmission module and an intelligent terminal module, the signal acquisition module includes an electrocardiographic signal acquisition module, and the physiological signal acquisition module realizes The wireless transmission module will wirelessly transmit the analog signals collected by the acquisition module after A/D conversion; the intelligent terminal will process and analyze the received digital signals, and analyze the physiological conditions of the critically wounded and sick. Evaluation is carried out to realize wireless continuous monitoring of the seriously wounded and sick, and automatic alarm for abnormal indicators. Specifically, the signal acquisition module includes a wearable device and a sensor arranged on the wearable device. The wearable device includes a plurality of elastic installation straps that are connected to each other according to the human body structure. The plurality of elastic installation straps Active splicing, including the elastic mounting belt on the head and the elastic mounting belt on the body. The head mounting belt is mainly used for data collection such as pulse, blood oxygen, and EEG. The elastic mounting belt on the body is mainly used for For data collection such as EMG, cardiopulmonary sound, ECG, body temperature data, etc., the elastic mounting straps are connected to each other with circuit leads and can be plugged movably. It is worn on the patient's body; the wireless transmission module adopts the anti-interference technology based on frequency agility to avoid signal interference, and specifically, ZigBee/Bluetooth can be used to wirelessly transmit the physiological signal to the data intelligent terminal module.

The intelligent terminal module includes an intelligent control terminal and a central control system adapted to the network of the physiological signal acquisition module through the wireless transmission module, and the physiological signal acquisition module transmits the collected data information to the central control system through wireless transmission for data storage. , analysis, the central control system feeds back the processed information to the intelligent control terminal, the intelligent control terminal is set with an alarm module, the central control system sets thresholds for parameters such as ECG, pulse, blood oxygen, body temperature, etc., and the central control system will The processed parameters are compared with the threshold value and fed back to the intelligent control terminal alarm module to realize abnormal parameter alarm. For example, the central control system sets the blood oxygen saturation threshold to 90%. When the parameter information collected by the blood oxygen signal acquisition module is processed, the data is compared with the set threshold. When the blood oxygen saturation is low At 90%, the central control system feeds back the comparison result to the intelligent control terminal, and triggers the alarm module to give an alarm; set the minimum diastolic blood pressure threshold Bm to 60, the maximum diastolic blood pressure threshold Bx to 90, and the minimum systolic blood pressure threshold Dm It is set to 90, and the maximum systolic blood pressure threshold Dx is set to 140. When the collected data is processed and compared with the threshold, an alarm is triggered after abnormal data results are obtained.

The ECG signal acquisition system of the present invention can automatically screen and analyze the collected ECG signals, specifically, the collected signals can be filtered, the noise of the ECG signal can be screened out, the quality of the ECG signals can be judged, and the screening After the high-quality ECG signal is extracted, the characteristics of the ECG signal are extracted and displayed on the display in the form of graphics. At the same time, the above data can be used to accurately judge the symptoms of arrhythmia, quickly trace the origin of the cause of arrhythmia, and give timely medical advice. and treatment options. The signal processing method of the ECG signal acquisition system includes: 1) ECG signal acquisition module to acquire ECG signals;

2) Preprocessing of ECG signal;

After the original ECG signal data is obtained, it needs to be preprocessed first to remove the noise contained in the ECG signal. In the present invention, the method of wavelet transform will be used to analyze and denoise the signal. Choose a biorthogonal quadratic B-spline wavelet, which is a first-order smooth function. ECG signal noise mainly includes baseline drift, power frequency interference and EMG noise. The frequency distribution of each noise signal included in the ECG signal is also different: baseline drift 0-0.5Hz, power frequency interference 50~60Hz; EMG noise 50~2KHz , a wider range, similar to Gaussian sound. In addition, the center frequency of the QRS complex, the main component of the ECG signal, is about 17 Hz. The energy of the QRS complex in the present invention is mainly distributed on the 3rd and 4th scales after wavelet transformation; the power frequency interference energy is mainly distributed on the 2nd scale after wavelet transformation; the EMG noise energy is mainly distributed on the 1st, 2nd and 3rd scales after wavelet transformation superior. The filtering method adopted here is: processing the low-scale wavelet coefficients d1 and d2 by the soft threshold method, and eliminating the high-frequency noise by the large-scale attenuation coefficient method. For the scale 3 wavelet coefficients including important input signals, the present invention adopts a compromise algorithm between soft and hard thresholds, which not only eliminates EMG noise, but also retains the original input signal as much as possible, without affecting the accurate extraction of subsequent ECG characteristic parameters. At the same time, the baseline drift is suppressed by an all-pass low-pass filter.

2) Judging the quality of ECG signals, screening good quality signals;

Screen signals of good quality according to, but not limited to, the threshold value and the number of samples in the peak part; the specific evaluation criteria are:

A1: The part with the threshold value greater than 3mv exceeds 40%, and the standard is considered to be satisfied.

A2: This criterion is met when the first derivative is greater than 0.3 (that is, the number of samples in the peak part) is greater than 40%.

A3: When the lead-off part is greater than 80%, this criterion is met.

A4: The part that meets the judgment conditions in A1, A2, and A3 is a potential failure point. When the potential failure part is greater than 68.5%, this standard is satisfied.

Through the above four evaluation criteria, the signals of poor contact, lead shedding and poor quality can be screened out and not processed, so that the signals of good quality can be used for disease analysis.

3) ECG signal feature extraction

The detection of singular points of ECG signal is completed by the method based on bi-orthogonal quadratic B-spline wavelet transform. The R wave of the ECG signal exhibits the form of extreme value pairs with the largest amplitude at scale 4, and the frequency band range of scale 4 (11.3~22.5Hz) is the closest to the center frequency of the QRS wave (17Hz), which effectively avoids unfiltered outside the bandwidth. In addition to the influence of clean noise, the detection accuracy of the R wave is guaranteed. The dynamic threshold method is used to improve the positioning accuracy of R wave. Analyze the wavelet coefficient D4, extract the extremum pairs, use threshold processing filtering to obtain the extremum pairs that accurately locate the R wave, restore the position of the extremum pairs to the reconstructed ECG signal (time shift), and find the maximum value in the maximum value interval. value to get the R-wave localization. After the R wave is located, the Q and S waves can be located by finding the minimum point near the R wave according to the position of the R wave.

The detection of P wave and T wave can be summed up by using the positions of two adjacent R waves as benchmarks, and using the average value of the RR interval to control the length of the detection interval, so as to accurately detect the two adjacent R waves. Heart beat. The respective T waves and P waves. The respective detection areas of the T wave and the P wave are measured. So far, the P, Q, R, S, and T waves of the ECG signal have been accurately located, providing a data source for the subsequent model building.

4) Signal analysis

The incidence of arrhythmia in patients with heart disease is as high as 80-100%. Therefore, to accurately judge the symptoms of arrhythmia, quickly trace the origin of the cause of arrhythmia, and give timely medical advice and treatment plans, specifically:

The data obtained in step 3) are divided into two first-level sub-categories, namely, intraventricular block in abnormal activation origin and abnormal activation conduction, which are used to train the support vector machine SVM1;

The abnormal origin of excitation is classified into sinus rhythm and ectopic rhythm, which are used to train support vector machine SVM2;

The ectopic heart rate is divided into active ectopic heart rhythm and passive ectopic heart rhythm for training SVM3; active ectopic heart rhythm is divided into two categories, namely supraventricular premature beats and ventricular premature beats, for training SVM4;

Passive ectopic heart rhythms are divided into two categories, supraventricular escape beats and ventricular escape beats, which are used to train SVM5;

In the present invention, seven types of heart rhythms, namely, sinus rhythm, supraventricular premature beat, ventricular premature beat, supraventricular escape beat, ventricular escape beat, ventricular block and ventricular flutter wave are selected as classification targets. The processed data is classified by the Support Vector Machine (SVM) method. This module requires the heart rhythm to be divided into 7 categories, so the binary tree classification method is used. First divide all categories into two sub-categories, then sub-categories are further divided into two sub-categories, and so on, until a single category is obtained. The data were first divided into two first-level sub-categories, namely, intraventricular block in abnormal activation origin and abnormal conduction conduction, which were used to train the support vector machine SVM1. Abnormal origins of activation were then classified as sinus rhythm and ectopic rhythm, respectively, for training the support vector machine SVM2. The ectopic heart rate was divided into active ectopic heart rhythm and passive ectopic heart rhythm, and the training SVM3 was used. Active ectopic heart rhythms are divided into two categories, supraventricular premature beats and ventricular premature beats, respectively, for training SVM4. Passive ectopic heart rhythms are divided into two categories, supraventricular escape beats and ventricular escape beats, which are used to train SVM5. The ECG signals are divided into seven categories by five support vector machines, so as to monitor the heart condition of the wounded and sick in real time.

The above are only preferred embodiments of the present invention, and do not limit the present invention in any form. Although the present invention has been disclosed above with preferred embodiments, it is not intended to limit the present invention. The technical personnel, within the scope of the technical solution of the present invention, can make some changes or modifications by using the technical content disclosed above to be equivalent embodiments of equivalent changes, provided that they do not depart from the technical solution content of the present invention, according to the technical solution of the present invention. Any simple modifications, equivalent changes and modifications made to the above embodiments still fall within the scope of the technical solutions of the present invention.

Claims (7)

1. Electrocardiosignal collection system based on artificial intelligence, including physiological signal collection module, wireless transmission module and intelligent terminal module, signal collection module includes electrocardiosignal collection module, its characterized in that: the physiological signal acquisition module acquires various physiological signals of the critically ill patients; the wireless transmission module performs wireless transmission after A/D conversion on the analog signals acquired by the acquisition module; the intelligent terminal processes and analyzes the received digital signals, evaluates the physiological condition of the serious sick and wounded, realizes wireless continuous monitoring of the serious sick and wounded, and automatically alarms abnormal indexes.

2. The artificial intelligence based cardiac electrical signal acquisition system of claim 1 wherein: the signal acquisition module comprises wearable equipment and a sensor arranged on the wearable equipment, the wearable equipment comprises a plurality of elastic mounting belts which are arranged in a mutual lead mode according to the structure of a human body, and the elastic mounting belts are movably spliced; the wireless transmission module adopts an anti-interference technology based on frequency agility, and avoids signal interference.

3. The artificial intelligence based cardiac electrical signal acquisition system of claim 2 wherein: the signal acquisition module also comprises a pulse, blood oxygen and electroencephalogram data acquisition module arranged on the head and a myoelectricity, cardiopulmonary sound, electrocardio and body temperature data monitoring module arranged at the corresponding position on the body.

4. The artificial intelligence based cardiac electrical signal acquisition system of claim 3 wherein: the intelligent terminal module comprises an intelligent control terminal and a central control system, the intelligent control terminal is matched with the physiological signal acquisition module through a wireless transmission module in a network mode, the physiological signal acquisition module transmits acquired data information to the central control system through wireless transmission to be stored and analyzed, the central control system feeds back the processed information to the intelligent control terminal, the intelligent control terminal is provided with an alarm module, the central control system sets threshold values for parameters such as electrocardio, pulse, blood oxygen and body temperature, and the central control system compares the processed parameters with the threshold values and feeds back the processed parameters to the alarm module of the intelligent control terminal to achieve abnormal parameter alarm.

5. The signal processing method of an artificial intelligence based cardiac signal acquisition system as claimed in claim 1, comprising:

1) an ECG signal acquisition module acquires electrocardiosignals;

2) preprocessing of ECG signals;

analyzing and denoising the ECG signal by adopting a wavelet transform method;

processing the low-scale wavelet coefficients d1 and d2 by a soft threshold value method, and eliminating high-frequency noise by a large-amplitude attenuation coefficient method;

filtering the scale 3 wavelet coefficient containing the important input signal by adopting a soft and hard threshold compromise algorithm;

3) judging the quality of the ECG signal, and screening signals with good quality;

screening a signal with good quality according to characteristics including but not limited to a threshold value and a peak part sample number; 4) electrocardiosignal feature extraction

The detection of the singular points of the ECG signal is finished based on a biorthogonal quadratic B-spline wavelet transform method; improving the positioning precision of the R wave by adopting a dynamic valve value method, analyzing a wavelet coefficient D4, extracting an extreme value pair, obtaining the extreme value pair for accurately positioning the R wave by adopting threshold processing and filtering, restoring the position of the extreme value pair to a reconstructed ECG signal, finding the maximum value in the maximum value interval, and obtaining the R wave positioning; after the R wave is positioned, Q, S waves can be positioned by searching a minimum value point near the R wave according to the position of the R wave;

The length of a detection interval is controlled by taking the positions of two adjacent R waves as a marker post and the mean value of RR intervals, so that two heart beats between the two adjacent R waves, respective T waves and P waves are accurately detected; measuring respective detection areas of the T wave and the P wave to finish accurate positioning of P, Q, R, S, T waves of the ECG signal;

5) signal analysis

Dividing the data obtained in the step 3) into two first-level subclasses which are indoor blocks in excitation origin abnormality and excitation conduction abnormality respectively and are used for training a support vector machine (SVM 1);

classifying the excitation origin abnormality into sinus rhythm and ectopic rhythm respectively for training a support vector machine (SVM 2);

dividing the ectopic heart rate into an active ectopic heart rate and a passive ectopic heart rate for training the SVM 3; dividing the active ectopic heart rhythm into two categories, namely supraventricular premature beat and ventricular premature beat, and using the two categories to train the SVM 4;

passive ectopic rhythms are classified into two categories, supraventricular and ventricular escapes, respectively, for training SVM 5.

6. The signal processing method of the artificial intelligence based electrocardiographic signal acquisition system according to claim 5, wherein: the electrocardiosignals are divided into seven classes by five support vector machines, so that the heart condition of the sick and wounded is monitored in real time.

7. The signal processing method of an artificial intelligence based cardiac signal acquisition system as recited in any of claims 1-6, further comprising: the electrocardiosignal acquisition system based on artificial intelligence is applied to wearable whole-course vital sign wireless monitoring equipment.