Electrocardiogram monitoring system and method based on twelve-lead system

Abstract

The invention relates to an electrocardiogram monitoring system and method based on a twelve-lead system, and relates to the technical field of remote medical information processing, and the electrocardiogram monitoring system comprises an electrocardiogram signal acquisition terminal, a wireless communication transmission network, a cloud data management analysis platform and a remote monitoring interaction terminal, an analog front end and master control communication separated framework is adopted in the electrocardiosignal collecting terminal, a Wilson central electric end is synthesized through a resistance network to collect high-fidelity signals, a transmission network supports multi-mode self-adaptive switching, a partition storage and breakpoint resume strategy containing an abnormal event write-protection mechanism is started when the network is disconnected, data integrity is guaranteed, and the reliability of the electrocardiosignal collecting terminal is improved. The cloud platform generates diagnosis data by using a Demix superposition technology and an ST-T feature analysis algorithm, the remote terminal provides waveform rendering, electronic caliper gauge measurement and manual correction feedback functions, and the system solves the problems of easy signal interference, weak network data loss and low automatic diagnosis accuracy in remote monitoring.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of remote medical information processing, in particular to an electrocardiogram monitoring system and method based on a twelve-lead system. BACKGROUND

[0002] Cardiovascular disease is one of the major diseases that threaten human health worldwide. Its suddenness and concealment make long-term and dynamic electrocardiogram monitoring an important means of clinical diagnosis. Although the traditional twelve-lead electrocardiograph has high signal quality, it is mainly limited to static examination in hospitals due to its large size, complex cable and the need for professional operation, making it difficult to capture paroxysmal arrhythmia or asymptomatic myocardial ischemia events in patients' daily life.

[0003] Existing portable or wearable electrocardiogram monitoring devices attempt to solve the above problems, but most of them use single-lead or three-lead design, which cannot provide complete cardiac electrical activity information, and have limitations in locating myocardial infarction sites or diagnosing complex arrhythmias. Some portable devices that attempt to implement twelve-lead functions often face dual challenges of signal quality and transmission stability. At the hardware level, in order to pursue miniaturization, high-frequency wireless communication modules and high-sensitivity analog front-end circuits are often highly integrated, resulting in serious radio frequency interference that severely affects the accuracy of microvolt-level electrocardiogram signal acquisition, causing baseline drift or waveform distortion.

[0004] At the data transmission level, the network environment in the mobile medical scene is complex and variable. Existing remote monitoring systems usually rely on a single communication link and lack effective response mechanisms for weak or no network environments. When patients are in signal blind areas such as elevators, basements or remote areas, data transmission interruptions are likely to occur. The existing caching strategy mostly uses a simple time-sequential coverage mode. Once the network is interrupted for too long, the data collected subsequently will cover the key pathological data that has not been uploaded in the cache, resulting in the loss of abnormal event records with important diagnostic value, which seriously affects the doctor's judgment. In addition, existing automated analysis algorithms have a high false positive rate when facing dynamic interference, and lack convenient human-computer interaction tools for doctors to review and correct, making it difficult to meet the actual needs of clinical-level remote medical care. SUMMARY

[0005] The purpose of the present application is to provide an electrocardiogram monitoring system and method based on a twelve-lead system, aiming to solve the technical problems of radio frequency interference affecting signal quality caused by hardware integration in existing portable twelve-lead devices, easy loss of key pathological data in complex network environments, and lack of effective interactive verification mechanisms for remote diagnosis.

[0006] The electrocardiogram monitoring system based on a twelve-lead system provided by the first aspect of the present application adopts the following technical solution:

[0007] The application relates to a heart electrical signal acquisition terminal, a wireless communication transmission network, a cloud data management analysis platform and a remote monitoring interactive terminal.

[0008] The heart electrical signal acquisition terminal is used for acquiring twelve-lead heart electrical signals of a subject and pre-processing the heart electrical signals to generate a digitized original heart electrical data stream.

[0009] The wireless communication transmission network is connected with the heart electrical signal acquisition terminal and is used for transmitting the original heart electrical data stream according to a network signal state.

[0010] The cloud data management analysis platform is connected with the wireless communication transmission network, is used for receiving the original heart electrical data stream, and is used for performing heart beat shape classification analysis and ischemic feature analysis to generate diagnosis data.

[0011] The remote monitoring interactive terminal is communicatively connected with the cloud data management analysis platform and is used for acquiring and visually displaying the original heart electrical data stream and the diagnosis data.

[0012] The heart electrical signal acquisition terminal adopts a physically separated architecture and comprises an acquisition subunit and a sending subunit.

[0013] The acquisition subunit is used for being attached to the chest of the subject, and an analog front-end circuit is integrated in the acquisition subunit. The analog front-end circuit synthesizes a Wilson central terminal potential through a resistance network and converts a body surface potential into a digital signal through a multiplexer and an analog-to-digital converter.

[0014] The sending subunit is configured to be worn in a non-interference area and is connected with the acquisition subunit through an anti-interference flexible shielding cable. A master microprocessor and a multi-mode wireless communication module are integrated in the sending subunit and are used for receiving the digital signal and performing subsequent processing.

[0015] The heart electrical signal acquisition terminal is configured to perform local adaptive threshold early warning logic, which comprises the following steps.

[0016] In a baseline learning stage, the peak amplitude distribution of QRS complexes is counted to initialize a signal threshold and a noise threshold.

[0017] In a dynamic monitoring stage, a signal that has been corrected by baseline drift is subjected to difference and integral operations to generate an energy feature sequence, and local maximum points in the energy feature sequence are detected.

[0018] If the local maximum point is greater than the current signal threshold, it is determined that a valid QRS complex is detected, and the signal threshold and the noise threshold are updated through a weighted average algorithm according to the local maximum point.

[0019] The heart electrical signal acquisition terminal is further configured to identify time interval abnormalities of adjacent QRS complexes and mark a data segment at a corresponding moment as abnormal event data.

[0020] The wireless communication transmission network is configured to perform multi-mode network state detection and adaptive switching strategy:

[0021] Periodically detecting the received signal strength indication values of the wireless local area network and the cellular mobile communication network;

[0022] When the wireless local area network is available, preferentially transmitting the full amount of the original electrocardio data stream through the wireless local area network;

[0023] When the wireless local area network is unavailable and the received signal strength indication value of the cellular mobile communication network is greater than a first signal strength threshold value, switching to a high-speed cellular transmission mode;

[0024] When the received signal strength indication value of the cellular mobile communication network is between the first signal strength threshold value and a second signal strength threshold value, switching to a low-speed transmission mode and enabling a data compression algorithm.

[0025] The wireless communication transmission network is further configured to perform breakpoint resume logic in flight mode:

[0026] When the received signal strength indication values of all available networks are lower than the second signal strength threshold value, entering flight mode and redirecting the original electrocardio data stream to a local non-volatile memory;

[0027] The local non-volatile memory is divided into an event data area and a normal data area, and the segments marked as abnormal event data and their context data are written to the event data area and are in a write protection state, and the conventional monitoring data is written to the normal data area and follows the first-in-first-out coverage principle;

[0028] When the network signal is restored, the data in the event data area is preferentially uploaded, and the write protection state is released after receiving a cloud confirmation instruction.

[0029] The cloud data management and analysis platform is configured to perform heart beat classification analysis based on Demix superposition technology:

[0030] The original electrocardio data stream is segmented to extract heart beat waveform vectors, and the Pearson correlation coefficient of the current heart beat waveform vector and each template vector in the pre-stored heart beat template library is calculated;

[0031] If the maximum Pearson correlation coefficient is greater than a preset classification matching threshold value, the current heart beat is classified as the heart rhythm type of the corresponding template and the template feature is updated;

[0032] If the maximum Pearson correlation coefficient is less than the classification matching threshold value, the current heart beat waveform vector is added to the heart beat template library as a new template.

[0033] The cloud data management and analysis platform is further configured to perform ischemic ST-T segment analysis:

[0034] Based on the QRS complex start and end point positioning result, determine the isopotential line reference level, position the junction point of the QRS complex end point and the ST segment, and measure the actual level of the ST segment at a preset time offset after the junction point;

[0035] Calculate the offset of the actual level relative to the isopotential line reference level, and when the offset of a plurality of consecutive heartbeats exceeds the myocardial ischemia alarm threshold and appears in the relevant lead group, determine a myocardial ischemia suspected event.

[0036] The remote monitoring interactive terminal is configured to perform graphical rendering of twelve-lead electrocardiogram data:

[0037] Receive the original electrocardiogram data stream and parse it into sample point values of each lead, and construct a coordinate system according to a preset background grid pixel density;

[0038] According to the gain control parameter and the paper speed control parameter set by the user, calculate the vertical coordinate and horizontal coordinate of each sample point on the screen, draw the coordinate point sequence on the background grid, and support multi-view layout switching.

[0039] The remote monitoring interactive terminal integrates an electronic caliper measurement tool:

[0040] In response to the start cursor point and the end cursor point selected by the user in the waveform display area, obtain the data index values and vertical pixel coordinates corresponding to the two points;

[0041] Calculate the time difference according to the data index value difference and the sampling rate of the two points, and calculate the potential difference according to the vertical pixel coordinate difference of the two points, the gain control parameter and the background grid pixel density.

[0042] The remote monitoring interactive terminal is further configured to provide a diagnostic feedback mechanism:

[0043] Display an abnormal event list generated by the cloud data management and analysis platform in the interface side bar, and position the corresponding waveform segment in response to the user's triggering operation on the list item;

[0044] Receive the user's correction instruction for the classification label of the abnormal event list, and send the feedback instruction containing the event unique identifier and the corrected label to the cloud data management and analysis platform to update the record.

[0045] The second aspect of the present application provides a twelve-lead system-based electrocardiogram monitoring method, comprising the following steps:

[0046] S10, the electrocardiosignal acquisition terminal acquires the body surface analog potential signal of the measured target through a twelve-lead electrode array, synthesizes a reference potential based on a Wilson central terminal algorithm, and generates a twelve-channel electrocardiosignal vector containing standard limb leads, pressurized monopolar limb leads, and chest leads;

[0047] S20, the signal processing unit in the electrocardiosignal acquisition terminal performs analog-to-digital conversion on the collected analog potential signal, and uses a digital filter to filter out power frequency interference, electromyographic noise, and baseline drift, generating a digitized raw electrocardio data stream;

[0048] S30, the electrocardiosignal acquisition terminal performs local real-time analysis on the preprocessed electrocardio data stream based on an adaptive threshold algorithm, extracts QRS complex features, calculates instantaneous heart rate, and identifies suspected arrhythmia events according to a preset alarm logic, generating a data packet containing raw data and event markers;

[0049] S40, the electrocardiosignal acquisition terminal detects the signal strength parameter of the current wireless communication transmission network, switches the transmission strategy between the wireless local area network mode, the cellular mobile network mode and the flight mode according to the signal strength parameter, when the network connection is available, the data packet is encrypted and sent to the cloud data management analysis platform, when the network connection is unavailable, the flight mode is entered, the data packet is written into the local ring memory, and the breakpoint resume is executed after the network is restored;

[0050] S50, the cloud data management analysis platform receives and decrypts the data packet, uses the Demix superposition analysis algorithm to perform template matching classification on the heart beat morphology, performs ST-T segment level measurement and heart rate variability analysis, generates comprehensive diagnostic data and stores it in a distributed database;

[0051] S60, the remote monitoring interactive terminal responds to the user's review request, obtains and renders twelve-lead electrocardio waveforms and comprehensive diagnostic data from the cloud data management analysis platform, and displays electrocardiogram spectrum, heart rate trend and abnormal event list.

[0052] In summary, the present application includes at least one of the following beneficial technical effects:

[0053] 1. The present application adopts a physical split architecture design, which separates the analog front-end circuit and the main control communication module in space, and connects them through an anti-interference flexible shielding cable. This design effectively isolates the interference of high-frequency radio frequency signals of the wireless communication module on the microvolt-level electrocardio analog signal, reduces the baseline drift and power frequency noise by combining the Wilson central terminal synthesis technology, and ensures the clinical-level acquisition quality of twelve-lead electrocardiosignal in dynamic and long-term monitoring scenarios;

[0054] 2. The application switches WiFi, 4G and low-speed transmission modes adaptively by monitoring network signal strength in real time, and enables a dual-zone storage mechanism in a flight mode without network coverage, writes protection locks abnormal event data, and implements cyclic coverage on ordinary data, which solves the technical problem that key pathological data is easy to lose in a weak network and network interruption environment, and realizes the integrity of data and the reliability of breakpoint continuation;

[0055] 3. The application constructs a cloud intelligent diagnosis process based on Demix superposition classification and ST-T segment quantitative analysis, cooperates with electronic caliper measurement of a remote terminal and artificial feedback correction mechanism, the system can not only automatically identify complex arrhythmia and myocardial ischemia events, but also allows doctors to review and correct the automatic diagnosis results through a visual tool, forming a closed-loop diagnosis and treatment mode of intelligent preliminary screening, artificial diagnosis and feedback optimization, and improving the diagnosis accuracy and clinical practical value of remote electrocardio monitoring. BRIEF DESCRIPTION OF DRAWINGS

[0056] Figure 1 is a system architecture diagram of the application;

[0057] Figure 2 is a method flowchart of the application;

[0058] Figure 3 is a hardware structure block diagram of the electrocardio signal acquisition terminal of the application;

[0059] BRIEF DESCRIPTION OF DRAWINGS: 10, electrocardio signal acquisition terminal; 20, wireless communication transmission network; 30, cloud data management and analysis platform; 40, remote monitoring and interaction terminal. DETAILED DESCRIPTION

[0060] The following will be combined with the Figure 1 -Appendix Figure 3 , the present application is further described in detail.

[0061] The electrocardio monitoring system based on twelve-lead system, referring to Figure 1 , includes: an electrocardio signal acquisition terminal 10, a wireless communication transmission network 20, a cloud data management and analysis platform 30, and a remote monitoring and interaction terminal 40.

[0062] The electrocardio signal acquisition terminal 10 is connected with the body surface of the measured target through electrodes, and is used for acquiring the electrocardio physiological potential signal of the human body. The electrocardio signal acquisition terminal 10 is configured to perform amplification, filtering and analog-to-digital conversion operations of analog signals, convert continuous analog body surface potential into digitized twelve-lead electrocardio data stream. The electrocardio signal acquisition terminal 10 internally integrates a data processing unit, which corrects baseline drift of the collected raw data and preliminarily extracts features, and packs the data according to a preset communication protocol.

[0063] The wireless communication transmission network 20 establishes a data transmission channel between the electrocardiogram signal acquisition terminal 10 and the cloud data management and analysis platform 30. The wireless communication transmission network 20 supports multiple wireless communication standards and is configured to carry the interaction of electrocardiogram data packets and control instructions according to the current network signal quality parameters. The wireless communication transmission network 20 routes the encrypted data packets sent by the electrocardiogram signal acquisition terminal 10 to the cloud data management and analysis platform 30.

[0064] The cloud data management and analysis platform 30 is configured to receive data streams from the wireless communication transmission network 20. The cloud data management and analysis platform 30 performs integrity checking, decryption, and analysis on the received data packets, restores twelve-lead electrocardiogram waveform data, integrates a high-performance computing cluster for executing arrhythmia recognition algorithms, ST-T segment analysis algorithms, and heart rate variability analysis algorithms, and stores the analysis results and original waveform data in a distributed database.

[0065] The remote monitoring interactive terminal 40 is network-connected with the cloud data management and analysis platform 30 for visual display and interactive operation of data. The remote monitoring interactive terminal 40 is configured to provide a graphical user interface to authorized doctors or users, which real-time renders twelve-lead electrocardiogram waveforms, displays heart rate trend graphs, and shows abnormal event warning information generated by the cloud data management and analysis platform 30.

[0066] The remote monitoring interactive terminal 40 retrieves historical data or real-time data in the cloud data management and analysis platform 30 according to user instructions. The remote monitoring interactive terminal 40 supports gain adjustment and paper speed adjustment of waveform data, and allows doctors to manually review and correct the diagnostic conclusions generated by automatic analysis. The corrected diagnostic data is fed back to the cloud data management and analysis platform 30 for updating the electronic health records of the patient.

[0067] The electrocardiogram monitoring method based on the twelve-lead system refers to Figure 2 and includes the following steps:

[0068] S10, the electrocardiogram signal acquisition terminal 10 acquires the body surface analog potential signal of the measured target through a twelve-lead electrode array, synthesizes a reference potential based on the Wilson central terminal algorithm, and generates a twelve-channel electrocardiogram signal vector containing standard limb leads, pressurized monopolar limb leads, and chest leads;

[0069] S20, the signal processing unit in the electrocardiogram signal acquisition terminal 10 performs analog-to-digital conversion on the collected analog potential signal, and uses a digital filter to filter out power frequency interference, electromyographic noise, and baseline drift, generating digitized raw electrocardiogram data streams;

[0070] S30, the electrocardiosignal acquisition terminal 10 performs local real-time analysis on the preprocessed electrocardio data stream based on an adaptive threshold algorithm, extracts QRS complex features, calculates instantaneous heart rate, and identifies suspected arrhythmia events according to a preset alarm logic to generate a data packet containing original data and event markers;

[0071] S40, the electrocardiosignal acquisition terminal 10 detects the signal strength parameter of the current wireless communication transmission network 20, switches the transmission strategy between the wireless local area network mode, the cellular mobile network mode and the flight mode according to the signal strength parameter, encrypts and sends the data packet to the cloud data management and analysis platform 30 when the network connection is available, and enters the flight mode when the network connection is unavailable, writes the data packet into the local ring memory, and performs breakpoint resume transmission after the network is restored;

[0072] S50, the cloud data management and analysis platform 30 receives and decrypts the data packet, uses the Demix superposition analysis algorithm to perform template matching classification on the heart beat morphology, performs ST-T segment level measurement and heart rate variability analysis, generates comprehensive diagnostic data and stores it into a distributed database;

[0073] S60, the remote monitoring and interaction terminal 40 acquires and renders twelve-lead electrocardio waveforms and comprehensive diagnostic data from the cloud data management and analysis platform 30 in response to the user's review request, displays electrocardiogram spectrum, heart rate trend and abnormal event list.

[0074] The electrocardiosignal acquisition terminal 10 adopts a physically separated architecture, including an acquisition subunit and a sending subunit, and the acquisition subunit and the sending subunit are connected by an anti-interference flexible shielding cable to realize transmission of analog signals or digital signals. The acquisition subunit is used to adhere to the skin of the subject's chest to obtain weak bioelectric signals, and the sending subunit is used to be worn on the subject's waist or other non-interference area to perform data processing and wireless communication. This spatially separated structure design can effectively reduce the interference of electromagnetic radiation generated by the work of the wireless radio frequency module on the front-end high-sensitivity analog acquisition circuit, and at the same time, it reduces the weight of the device attached to the chest and reduces the contact artifacts generated during movement.

[0075] In the specific circuit implementation, the acquisition subunit internally integrates an analog front-end circuit, which includes an input protection circuit, a multiplexer, a low-noise instrument amplifier and a programmable gain amplifier. The input protection circuit is connected to ten standard electrocardio electrode interfaces, corresponding to right arm, left arm, left leg, right leg and chest V1 to V6 electrodes. The multiplexer switches the conduction channel according to the preset time sequence to transmit the body surface potential picked up by each electrode to the instrument amplifier.

[0076] The ECG signal acquisition terminal 10 synthesizes the Wilson central terminal potential through specific circuit logic. Specifically, the resistance network in the acquisition subunit is connected to the input terminals of the right arm, left arm, and left leg electrodes, and the common connection point of the three resistors constitutes the Wilson central terminal potential reference point. This reference point serves as the differential negative input terminal for the chest leads V1 to V6, thereby constructing a standard twelve-lead system. For limb leads, the system directly calculates the potential difference between each limb electrode through logical operation circuit.

[0077] The analog-to-digital converter in the ECG signal acquisition terminal 10 samples and quantizes the amplified analog signal. Considering the frequency band characteristics of ECG signals and the need for high-precision analysis, the analog-to-digital converter is used to have at least 24-bit resolution and a sampling rate of 1000 Hz or higher. The sampled digital signal is transmitted to the main microprocessor in the transmission subunit through a serial peripheral interface.

[0078] The main microprocessor in the transmission subunit performs preprocessing operations on the received raw digital signal. The preprocessing process includes a baseline drift correction step, and the specific correction logic is as follows:

[0079] A time sliding window with a length of is set, and the median of the sample point sequence is calculated within the window to obtain the baseline component . Then, the original sample point is subtracted by the baseline component to obtain the corrected signal . This step eliminates low-frequency baseline fluctuations caused by respiratory motion or body position changes.

[0080] After completing signal preprocessing, the ECG signal acquisition terminal 10 executes a local adaptive threshold early warning logic. This logic aims to solve the problem that traditional fixed threshold algorithms cannot adapt to the amplitude differences of ECG signals of different patients. The logic includes a baseline learning phase and a dynamic monitoring phase.

[0081] In the baseline learning phase, the ECG signal acquisition terminal 10 continuously calculates the peak amplitude distribution of the QRS complex within a certain period of time at the initial wearing stage, calculates the mean and variance of the signal strength, and initializes the signal threshold and the noise threshold .

[0082] In the dynamic monitoring phase, the ECG signal acquisition terminal 10 performs the following detection steps:

[0083] First-order difference operation and squaring operation are performed on the corrected signal to enhance high-frequency components and suppress the interference of low-frequency P waves and T waves, generating an energy feature sequence. The energy feature sequence is integrated by a sliding window to generate an integrated waveform , and the presence of an arrhythmia is detected Local maxima in ,Will With the current signal threshold and noise threshold Perform a comparison, if Greater than If a valid QRS complex is detected, its occurrence time and morphological characteristics are recorded, and the detected data are used to... The value is updated using a weighted average algorithm to determine the signal threshold. and noise threshold The specific update formula is: the new threshold equals the old threshold multiplied by the forgetting factor. Add the current peak value multiplied by This dynamic update mechanism allows the threshold to be automatically adjusted as the signal amplitude changes slowly, ensuring a stable detection rate even when the signal fluctuates.

[0084] The ECG signal acquisition terminal 10 is also configured to identify and mark abnormal heart rhythm events. The main control microprocessor calculates the time interval (RR interval) between two adjacent QRS complexes. If the RR interval is detected to be less than a preset tachycardia threshold, or greater than a preset arrest threshold, or the rate of change of the RR interval exceeds the fluctuation range of normal sinus rhythm, the main control microprocessor marks the ECG data segment at the current moment as abnormal event data. This type of data will be given high priority processing permissions in the subsequent transmission and storage process.

[0085] The wireless communication transmission network 20 adopts a multi-mode converged communication architecture. The ECG signal acquisition terminal 10 integrates a wireless local area network communication module and a cellular mobile communication 4G / 3G / GPRS module. The system monitors the availability and signal quality of each communication link in real time through the built-in network status management unit, and dynamically adjusts the data transmission strategy accordingly to optimize power consumption and transmission delay while ensuring data integrity.

[0086] The network status management unit periodically reads the received signal strength indication parameters of each communication module and executes the following multi-mode network status detection and adaptive switching logic:

[0087] The system sets the first signal strength threshold. Second signal strength threshold ,in Greater than .

[0088] The connection state of the WiFi module and the received signal strength indication parameter value are detected in priority. If the WiFi connection is established and the received signal strength indication parameter value is greater than or equal to a preset WiFi available threshold, the system selects the WiFi network as the current main transmission channel and performs a real-time full data transmission mode. In this mode, the original waveform data of the twelve leads is directly uploaded without lossy compression.

[0089] If the WiFi connection is unavailable or weak, the system activates the cellular mobile communication module and detects the received signal strength indication parameter value. If the detected cellular network received signal strength indication parameter value is greater than or equal to a first signal strength threshold , the system switches to a 4G / 3G high-speed cellular transmission mode and continues to maintain real-time data upload.

[0090] If the cellular network received signal strength indication parameter value is between a second signal strength threshold and the first signal strength threshold , the system determines that it is currently in a weak network environment and switches to a GPRS or low-speed transmission mode. In this mode, the system enables a data compression algorithm to losslessly or slightly lossily compress the electrocardiogram data to reduce the transmission bandwidth requirement and ensure the connectivity of the core monitoring data. If the received signal strength indication parameter values of all available networks are detected to be lower than the second signal strength threshold , the system determines that the current network environment cannot meet the minimum transmission requirement and automatically enters a flight mode.

[0091] In the flight mode, the electrocardiogram signal acquisition terminal 10 starts a local storage protection mechanism and performs a breakpoint resume transmission logic, which specifically includes:

[0092] The main control microprocessor cuts off the wireless transmission link to reduce power consumption and redirects the acquired electrocardiogram data stream to a local large-capacity non-volatile memory. The local storage space is divided into a normal data area and an event data area, and a ring buffer management method is used.

[0093] When abnormal event data marked in the foregoing steps is detected during data writing, the system writes the data segment and the context data of a preset time length before and after it into the event data area. The data in the event data area is set to a write protection state and will not be overwritten by new data unless it is successfully uploaded to the cloud and an acknowledgement is received.

[0094] For non-abnormal regular electrocardiogram monitoring data, the system writes it into the normal data area. When the storage space is full, the new regular data will overwrite the earliest regular data, following the first-in-first-out principle, but always avoiding the locked event data area.

[0095] During flight mode operation, the network status management unit wakes up the communication module at low frequency to perform network scans. Once the network signal strength is detected to have recovered to the second signal strength threshold, the unit will initiate a network scan. The system exits flight mode and re-establishes the data connection with the cloud data management and analysis platform 30. After the connection is re-established, the system performs a data retransmission operation. The priority logic of the retransmission queue is as follows:

[0096] Prioritize uploading locked abnormal event data packets from the event data area, followed by uploading historical monitoring data from the ordinary data area, and finally restore the pass-through of real-time data. Whenever the cloud server successfully receives and verifies a data packet, it will return a confirmation command to the terminal. After receiving the confirmation command, the terminal will release the write protection or locking status of the corresponding data in the local storage and free up storage space for subsequent use.

[0097] This breakpoint resume mechanism ensures zero loss of critical pathological data in mobile healthcare scenarios with unstable networks by differentiating data priorities and protecting storage areas. Even during prolonged network outages, it can completely preserve abnormal electrocardiogram evidence from patients.

[0098] The cloud-based data management and analysis platform 30 receives ECG data packets from the wireless communication transmission network 20 and performs unpacking, verification, and deep analysis tasks. To meet the needs of automatic diagnosis of massive ECG data, the platform adopts a morphological classification algorithm based on Demix overlay technology, combined with ST-T segment analysis and heart rate variability analysis, to achieve accurate identification of complex arrhythmias and myocardial ischemia events.

[0099] The cloud-based data management and analysis platform 30 executes the following Demix overlay classification analysis logic:

[0100] The unpacked ECG data stream is segmented using QRS complexes to extract the vector of each individual heartbeat waveform. Let the extracted QRS complex vector be... ,in This represents the number of sampling points.

[0101] The system maintains a dynamically updated cardiac rhythm template library. ,in Representing the A known morphological template for heartbeats, such as a normal sinus heartbeat template or a premature ventricular contraction template.

[0102] Calculate the current heartbeat vector With each template in the template library Pearson correlation coefficient between The calculation formula is:

[0103] ;

[0104] wherein, is the mean of the vector , is the mean of the template vector , and are the i-th element of the vector ,

[0105] Set a classification matching threshold If there exists a template such that , the current beat is classified as the rhythm type of the template, and the morphology feature of the template is updated by using a weighted average method to fuse the new waveform. If the correlation coefficient of the current beat with all known templates is less than , it is determined that the beat is a new morphology class. The system adds the beat as a new template to the template library and marks it as a pending manual confirmation state, prompting the doctor to review. Based on the beat classification results, the occurrence frequency and distribution law of abnormal beats such as premature ventricular contractions and premature atrial contractions are calculated, and a histogram and a scatter plot are generated.

[0106] The cloud data management and analysis platform 30 simultaneously performs ischemic ST-T segment analysis logic:

[0107] Based on the location of the start and end points of the QRS complex, the isoelectric line reference level is determined , which is usually the average voltage value of a predetermined interval before the start of the QRS wave.

[0108] The J point, the junction point of the QRS complex end and the ST segment, is located, and the actual level of the ST segment is measured at a predetermined time offset after the J point .

[0109] The relative offset of the ST segment is calculated .

[0110] The system sets a myocardial ischemia alarm threshold , if consecutive beats exceed , and such offset occurs in two or more related leads such as the II, III, and aVF lead groups, the system determines a suspected event of myocardial ischemia and records the duration and maximum offset of the event.

[0111] The cloud data management and analysis platform 30 further performs heart rate variability analysis, extracts the RR interval sequence of consecutive normal sinus beats, calculates time domain indicators including the standard deviation of all sinus RR intervals and the root mean square of adjacent RR interval differences, which reflect the regulation function of the autonomic nervous system to the heart, and constructs a Lorraine scatter plot with the first a RR interval a RR interval a RR interval scatter plot is drawn as the ordinate, and the system analyzes the geometric shape of the scatter plot such as comet shape, fan shape, and torpedo shape by using an image recognition algorithm to assist in identifying arrhythmia modes such as atrial fibrillation and premature beat.

[0112] All diagnostic data, abnormal event segments, and full-quantity original waveforms generated by the above algorithm are stored in a distributed database system. The database uses a multiple-copy redundancy mechanism to ensure data recoverability in the event of a single storage node failure. Meanwhile, the platform logically isolates sensitive patient identity information and pathological data, and only terminals with authorized keys can perform associated queries.

[0113] The remote monitoring interactive terminal 40 runs on a computing device with a graphical display capability and a network communication capability, including but not limited to a desktop computer, a tablet computer, or a smart mobile terminal. The remote monitoring interactive terminal 40 performs encrypted communication with the cloud data management and analysis platform 30 through an application program interface, obtains electrocardiogram waveform data and analysis results, and provides a visual diagnostic assistance tool.

[0114] The remote monitoring interactive terminal 40 runs on a computing device with a graphical display unit, a user input interface, and a network communication interface. The computing device includes a central processing unit and a memory. The memory stores computer program instructions that, when executed, implement the following interactive logic. The remote monitoring interactive terminal 40 establishes a communication channel with the cloud data management and analysis platform 30 through a WebSocket long connection or an HTTPS polling mechanism, and obtains electrocardiogram waveform data packets and diagnostic metadata.

[0115] The graphical rendering engine in the remote monitoring interactive terminal 40 performs real-time rendering and historical playback of twelve-lead electrocardiogram data. The specific rendering and display logic is as follows:

[0116] The terminal unpacks the received binary electrocardiogram data stream and extracts the original sampling point values of each lead.

[0117] A standard electrocardiogram background grid is constructed in the memory of the display buffer. The grid is drawn in the form of a pixel array, and defines that each 1 mm corresponds to a pixel point.

[0118] Coordinate mapping operations are performed. The graphical rendering engine reads the current gain setting and paper speed , and calculates the ordinate and abscissa of each sampling point on the screen according to the following formula:

[0119] ;

[0120] ;

[0121] wherein, is the baseline pixel coordinate of the lead, is the sample point value, is the voltage conversion coefficient determined by the ADC reference voltage and bit width, is the horizontal coordinate of the previous sample point, is the sampling rate, the calculated coordinate point sequence is drawn on the background grid through the graphics processing unit or the graphics interface, the system supports multi-view layout switching, including 12-channel longitudinal cascade view, and 6x2 or 3x4 array view, in the array view mode, the system automatically adjusts the position parameters of to adapt to different window segmentation areas.

[0122] The remote monitoring interactive terminal 40 integrates an electronic caliper measurement tool, when the user selects the start cursor point and the end cursor point in the waveform area through the input interface, the system obtains the data index values , and the vertical pixel coordinates , corresponding to the two points, the system calculates the physical quantities and displays according to the following logic:

[0123] time difference milliseconds;

[0124] potential difference millivolts, the tool is used for the doctor to accurately review the PR interval, QT interval or ST segment offset.

[0125] The remote monitoring interactive terminal 40 also provides an artificial correction and confirmation mechanism for the diagnosis conclusion, the interface side column loads the abnormal event list generated by the cloud algorithm, in response to the user's triggering operation on an event in the list, the main view area recalculates the X-axis offset, and positions the waveform window center to the timestamp of the event occurrence, the user modifies the classification label of the event through the interactive control, and the terminal encapsulates the feedback instruction containing the unique identifier of the event, the original label and the modified label, and sends it to the cloud data management and analysis platform 30, to trigger the field update of the corresponding record in the database.

[0126] In the report generation link, the remote monitoring interactive terminal 40 reads the final confirmed electrocardio statistical data, typical abnormal waveform screenshot data and doctor's text diagnosis opinion, the system renders the above data into portable document format according to the preset layout template, the system calls the digital signature module, encrypts the hash digest of the report file by using the doctor's private key, and generates a tamper-proof diagnosis report with an electronic seal.

[0127] In addition, the remote monitoring interactive terminal 40 performs function control according to the permission identification code of the login account. When the permission of the primary medical institution is identified, the system disables the report review function. When the permission of the superior expert is identified, the system opens the data access interface and the consultation opinion input interface across institutions, and realizes the hierarchical diagnosis and treatment cooperation.

Claims

1. An electrocardiogram monitoring system based on a twelve-lead system, characterized in that, include: An electrocardiogram (ECG) signal acquisition terminal is used to acquire twelve-lead ECG signals from the subject and preprocess the ECG signals to generate a digital raw ECG data stream. A wireless communication transmission network is connected to the ECG signal acquisition terminal and is used to transmit the raw ECG data stream according to the network signal status. A cloud-based data management and analysis platform, connected to the wireless communication transmission network, is used to receive the raw electrocardiogram data stream, perform cardiac morphology classification analysis and ischemic characteristic analysis, and generate diagnostic data. The remote monitoring and interactive terminal is connected to the cloud-based data management and analysis platform to acquire and visualize the raw electrocardiogram data stream and the diagnostic data.

2. The electrocardiogram monitoring system based on a twelve-lead system according to claim 1, characterized in that, The ECG signal acquisition terminal adopts a physically split architecture, including an acquisition subunit and a transmission subunit; The acquisition subunit is used to attach to the chest of the subject. It integrates an analog front-end circuit. The analog front-end circuit synthesizes the Wilson center terminal potential through a resistor network and converts the body surface potential into a digital signal through a multiplexer and an analog-to-digital converter. The transmitting subunit is configured to be worn in a non-interference area and is connected to the acquisition subunit via an anti-interference flexible shielded cable. The transmitting subunit integrates a main control microprocessor and a multi-mode wireless communication module for receiving the digital signal and performing subsequent processing.

3. The electrocardiogram monitoring system based on a twelve-lead system according to claim 1, characterized in that, The ECG signal acquisition terminal is configured to execute local adaptive threshold early warning logic, including: During the baseline learning phase, the peak amplitude distribution of the QRS group is statistically analyzed to initialize the signal threshold and noise threshold; During the dynamic monitoring phase, differential and integral operations are performed on the signal after baseline drift correction to generate an energy feature sequence, and local maxima points in the energy feature sequence are detected. If the local maximum point is greater than the current signal threshold, it is determined that a valid QRS group has been detected, and the signal threshold and the noise threshold are updated according to the local maximum point using a weighted average algorithm. The ECG signal acquisition terminal is also configured to identify abnormal time intervals between adjacent QRS complexes and mark the data segments at the corresponding times as abnormal event data.

4. The electrocardiogram monitoring system based on a twelve-lead system according to claim 1, characterized in that, The wireless communication transmission network is configured to implement a multi-mode network state detection and adaptive handover strategy: Periodically detect the received signal strength indication values ​​of wireless local area networks and cellular mobile communication networks; When the wireless local area network is available, the full amount of the original ECG data stream shall be transmitted via the wireless local area network first. When the wireless local area network is unavailable and the received signal strength indication value of the cellular mobile communication network is greater than the first signal strength threshold, switch to high-speed cellular transmission mode; When the received signal strength indication value of the cellular mobile communication network is between the first signal strength threshold and the second signal strength threshold, it switches to low-speed transmission mode and enables data compression algorithm.

5. The electrocardiogram monitoring system based on a twelve-lead system according to claim 4, characterized in that, The wireless communication transmission network is also configured to execute breakpoint resume logic in flight mode: When the received signal strength indication values ​​of all available networks are lower than the second signal strength threshold, enter flight mode and redirect the raw ECG data stream to local non-volatile memory; The local non-volatile memory is divided into an event data area and a normal data area. The fragments marked as abnormal event data and their context data are written to the event data area and are in a write-protected state. Regular monitoring data is written to the normal data area and follows the first-in-first-out overwrite principle. Once the network signal is restored, the data in the event data area will be uploaded first, and the write protection status will be lifted after receiving a confirmation instruction from the cloud.

6. The electrocardiogram monitoring system based on a twelve-lead system according to claim 1, characterized in that, The cloud-based data management and analysis platform is configured to perform cardiac classification analysis based on Demix overlay technology: The original ECG data stream is segmented to extract the heartbeat waveform vector, and the Pearson correlation coefficient between the current heartbeat waveform vector and each template vector in the pre-stored heartbeat template library is calculated. If the maximum Pearson correlation coefficient is greater than the preset classification matching threshold, the current heartbeat is classified into the heart rhythm type of the corresponding template and the template features are updated. If the maximum Pearson correlation coefficient is less than the classification matching threshold, the current heartbeat waveform vector is added to the heartbeat template library as a new template.

7. The electrocardiogram monitoring system based on a twelve-lead system according to claim 6, characterized in that, The cloud-based data management and analysis platform is also configured to perform ischemic ST-T segment analysis: The reference level of the equipotential line is determined based on the location results of the start and end points of the QRS group. The intersection point between the end point of the QRS group and the ST segment is located, and the actual level of the ST segment is measured at a preset time offset after the intersection point. The offset of the actual level relative to the isoelectric line reference level is calculated. When the offset of multiple consecutive heartbeats exceeds the myocardial ischemia alarm threshold and appears in the relevant lead group, it is determined to be a suspected myocardial ischemia event.

8. The electrocardiogram monitoring system based on a twelve-lead system according to claim 1, characterized in that, The remote monitoring interactive terminal is configured to perform graphical rendering of twelve-lead electrocardiogram data: The system receives the raw ECG data stream and parses it into sampling point values ​​for each lead, then constructs a coordinate system based on a preset background grid pixel density. Based on the user-defined gain control parameters and paper feed speed control parameters, the vertical and horizontal coordinates of each sampling point on the screen are calculated, and the coordinate point sequence is drawn on the background grid, supporting multi-view layout switching.

9. The electrocardiogram monitoring system based on a twelve-lead system according to claim 8, characterized in that, The remote monitoring and interactive terminal integrates electronic caliper measurement tools: In response to the start and end cursor points selected by the user in the waveform display area, the data index values ​​and vertical pixel coordinates of the two points are obtained. The time difference is calculated based on the difference in data index values ​​between the two points and the sampling rate, and the potential difference is calculated based on the difference in vertical pixel coordinates between the two points, the gain control parameters, and the background grid pixel density.

10. A method for electrocardiogram (ECG) monitoring based on a 12-lead system, and an ECG monitoring system based on a 12-lead system according to any one of claims 1-9, characterized in that, Includes the following steps: S10. The ECG signal acquisition terminal acquires the simulated potential signal of the body surface of the target through a twelve-lead electrode array, synthesizes the reference potential based on the Wilson central electrical terminal algorithm, and generates a twelve-channel ECG signal vector including standard limb leads, pressurized unipolar limb leads and chest leads. S20. The signal processing unit in the ECG signal acquisition terminal performs analog-to-digital conversion on the acquired analog potential signal and uses a digital filter to filter out power frequency interference, electromyographic noise and baseline drift, generating a digital raw ECG data stream. S30: The ECG signal acquisition terminal performs local real-time analysis on the preprocessed ECG data stream based on an adaptive threshold algorithm, extracts QRS complex features, calculates instantaneous heart rate, identifies suspected arrhythmia events according to preset alarm logic, and generates a data packet containing raw data and event markers. S40: The ECG signal acquisition terminal detects the signal strength parameters of the current wireless communication transmission network and switches the transmission strategy between wireless LAN mode, cellular mobile network mode and flight mode according to the signal strength parameters. When the network connection is available, the data packet is encrypted and sent to the cloud data management and analysis platform. When the network connection is unavailable, the terminal enters flight mode, writes the data packet to the local ring memory, and performs breakpoint resume transmission after the network is restored. The S50 cloud-based data management and analysis platform receives and decrypts data packets, uses the Demix overlay analysis algorithm to perform template matching and classification of heartbeat morphology, performs ST-T segment level measurement and heart rate variability analysis, generates comprehensive diagnostic data and stores it in a distributed database. The S60 remote monitoring interactive terminal responds to the user's access request by obtaining and rendering twelve-lead electrocardiogram waveforms and comprehensive diagnostic data from the cloud data management and analysis platform, and displays electrocardiograms, heart rate trends and lists of abnormal events.

Images