Early myocardial injury monitoring system based on medical Internet of Things

Abstract

This application relates to the field of data analysis technology and discloses an early myocardial injury monitoring system based on the medical internet of things. The system comprises a quality control module that processes peripheral blood samples and analyzes test results using immunochromatography; an extraction module that structures data and integrates it with biomarker data; an encryption module that ensures secure data transmission; a classification module that performs labeled classification and standardized processing; an evaluation module that performs risk assessment and spatiotemporal analysis; and a grading module that grades and stores early warning data. This application enables accurate early detection of myocardial injury, promptly identifying changes in the condition and providing risk warnings, thereby reducing delays in seeking medical treatment outside of the hospital.

Description

Technical Field

[0001] The present application relates to the field of data analysis, and in particular to an early myocardial injury monitoring system based on the medical Internet of Things. Background Art

[0002] Cardiovascular disease is one of the major threats to human health, and its prevalence is trending towards younger patients. Myocardial damage is a key manifestation of cardiovascular disease, with complex causes including insufficient blood or oxygen supply to cardiomyocytes, drug toxicity, and strenuous exercise. When myocardial damage occurs, the left ventricle releases specific proteins into the bloodstream, such as creatine kinase and its isoforms, and cardiac troponin. Currently, medical institutions primarily assess the extent of myocardial damage by testing for these biomarkers.

[0003] Existing myocardial injury monitoring methods have obvious shortcomings: the detection process relies on medical institutions to complete, which results in time delays; patients often seek medical treatment only after experiencing obvious symptoms, missing the best time for treatment; commonly used rapid detection methods such as the myocardial injury quadruple test card lack sensitivity and are prone to false positives; there is a lack of comprehensive analysis and real-time monitoring capabilities for multiple biomarkers, making it impossible to provide timely warnings and interventions. Summary of the Invention

[0004] This application provides an early myocardial injury monitoring system based on the medical Internet of Things, which is used to achieve early and accurate monitoring of myocardial injury, timely detect changes in the condition and provide risk warnings, thereby reducing the delay time for out-of-hospital medical treatment.

[0005] In a first aspect, the present application provides an early myocardial injury monitoring system based on the medical Internet of Things, the early myocardial injury monitoring system based on the medical Internet of Things comprising:

[0006] The quality control module is used to perform antigen-antibody reaction processing on peripheral blood samples through immunochromatography, and perform optical density analysis based on the color development of the test line and the quality control line to obtain quantitative data of myocardial injury biomarkers;

[0007] The extraction module is used to perform structured extraction of the patient's basic characteristic information and clinical history information based on a preset information collection template, and to associate and integrate it with the quantitative data of myocardial injury biomarkers to obtain the initial myocardial injury monitoring data set;

[0008] An encryption module is used to encrypt the initial myocardial injury monitoring data set based on an asymmetric encryption algorithm, generate an encrypted data packet through a data compression algorithm, transmit it through the Internet of Things protocol, and decrypt and decompress it to obtain the original myocardial injury monitoring data;

[0009] The classification module is used to label and classify the original myocardial injury monitoring data using a hierarchical clustering algorithm, and to calibrate and standardize the data by combining it with reference standard values ​​in a multi-source database to form a myocardial injury analysis data set;

[0010] An assessment module is used to perform multi-parameter risk assessment on myocardial injury analysis datasets based on a deep learning model, perform spatiotemporal correlation analysis in combination with geographic location information, and generate real-time myocardial injury risk warning data;

[0011] The grading module is used to grade the real-time myocardial injury risk warning data according to the warning level assessment rules, transmit the data through encryption, and store the assessment results in the database to form dynamic monitoring records.

[0012] In the technical solution provided in this application, the quality control module uses immunochromatography to perform antigen-antibody reaction processing, and performs optical density analysis through the color development of the test line and the quality control line, which can accurately obtain the quantitative data of myocardial injury biomarkers, ensuring the reliability and accuracy of the test results. The extraction module uses a preset information collection template to perform structured extraction of the patient's basic characteristic information and clinical history information, and associates and integrates it with the quantitative data of myocardial injury biomarkers, effectively solving the problem of data fragmentation and improving the integrity and practicality of the data. The encryption module uses an asymmetric encryption algorithm to encrypt the data and generates an encrypted data packet through a data compression algorithm, ensuring the security and efficiency of the data transmission process. The classification module uses a hierarchical clustering algorithm to label and classify the original data, and combines the reference standard values ​​in the multi-source database for data calibration and standardization, thereby improving the accuracy and scientificity of the data analysis. The evaluation module performs multi-parameter risk assessment based on a deep learning model, and combines geographic location information for spatiotemporal correlation analysis, realizing accurate prediction and timely warning of myocardial injury risks. The grading module performs graded processing based on the warning level assessment rules, feeds the assessment results back to the terminal through encrypted transmission, and stores the results in the database to form dynamic monitoring records, providing a reliable basis for medical decision-making. This modular design not only realizes the automated management of the entire process from data acquisition, processing, analysis to early warning, but also ensures patient privacy and data security through multiple data protection mechanisms. At the same time, the application of deep learning algorithms improves the accuracy of myocardial injury risk assessment. The system's real-time monitoring and early warning functions can help medical staff promptly identify potential myocardial injury risks and take appropriate preventive and intervention measures, which is of great significance to improving the diagnosis and treatment of cardiovascular diseases. In addition, the system's Internet of Things architecture design enables data to be transmitted and shared remotely, facilitating collaboration between medical institutions and improving the efficiency of medical resource utilization. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0014] Figure 1 This is a schematic diagram of an embodiment of an early myocardial injury monitoring system based on the medical Internet of Things in the embodiment of this application.

[0015] Figure 2 Schematic diagram of the structure of the grading module 106 in the early myocardial injury monitoring system based on the medical Internet of Things in an embodiment of the present application. DETAILED DESCRIPTION

[0016] An embodiment of the present application provides an early myocardial injury monitoring system based on the medical Internet of Things. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0017] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 In the embodiments of the present application, an embodiment of an early myocardial injury monitoring system based on the medical Internet of Things includes:

[0018] The quality control module 101 is used to perform antigen-antibody reaction processing on the peripheral blood sample by immunochromatography, perform optical density analysis based on the color development of the test line and the quality control line, and obtain quantitative data of myocardial injury biomarkers;

[0019] Extraction module 102, configured to perform structured extraction of the patient's basic characteristic information and clinical history information according to a preset information collection template, and associate and integrate the information with the quantitative data of myocardial injury biomarkers to obtain an initial myocardial injury monitoring data set;

[0020] The encryption module 103 is used to encrypt the initial myocardial injury monitoring data set according to an asymmetric encryption algorithm, generate an encrypted data packet through a data compression algorithm, transmit it through the Internet of Things protocol, and decrypt and decompress it to obtain the original myocardial injury monitoring data;

[0021] The classification module 104 is used to classify the original myocardial injury monitoring data by labeling using a hierarchical clustering algorithm, and to perform data calibration and standardization processing based on reference standard values ​​in a multi-source database to form a myocardial injury analysis data set;

[0022] Evaluation module 105, for performing multi-parameter risk assessment on the myocardial injury analysis dataset based on a deep learning model, performing spatiotemporal correlation analysis in combination with geographic location information, and generating real-time myocardial injury risk warning data;

[0023] The grading module 106 is used to grade the real-time myocardial injury risk warning data according to the warning level assessment rules, transmit the data encrypted and fed back to the terminal, and store the assessment results in the database to form dynamic monitoring records.

[0024] It is understandable that the execution subject of this application can be an early myocardial injury monitoring system based on the medical Internet of Things, or a terminal or a server, which is not limited here. The embodiment of this application is described by taking the server as the execution subject as an example.

[0025] Specifically, the early myocardial injury monitoring system based on the medical Internet of Things achieves early monitoring of myocardial injury through the collaborative work of multiple functional modules. The quality control module 101 is responsible for the collection and testing of biological samples. A glass fiber membrane is used to make a sample pad, which is assembled to an adhesive backing plate using an adhesive. A gold-labeled mouse anti-human cMyBP-C antibody dilution solution is spray-printed on the colloidal gold-labeled conjugate pad using a film streaker, and a test line coated with cMyBP-C monoclonal antibody is streaked on a nitrocellulose membrane. The quality control line is located 5 mm above the test line and is composed of rabbit anti-mouse IgG antibody. The collected peripheral blood sample is dripped into the sample well of the test card, and chromatographic migration is carried out by capillary action. After the antigen and antibody specifically bind, color is developed at the test line and the quality control line. The color intensity is quantitatively analyzed by an optical density analyzer to obtain the concentration data of the myocardial injury biomarker. Extraction module 102 collects and structures information, mapping and normalizing fields for basic patient information (including patient number, gender, age, height, weight, and location) and clinical history information (including history of cardiovascular disease, other medical conditions, surgical trauma, and allergies). This processed structured information is then matched and associated with the quantitative biomarker data based on timestamps, and a unique identifier is added to form the initial myocardial injury monitoring dataset.

[0026] The encryption module 103 uses the RSA asymmetric encryption algorithm to securely process data. It generates a public and private key pair and uses the public key to encrypt the data in blocks. Each data block is 256 bytes in size, and a hash value is calculated for each data block as a digital signature. The encrypted data is compressed using Huffman coding, and the compressed data packet is transmitted via an IoT protocol (WiFi, NFC, or GPS). The receiving end decrypts the data packet using the private key to restore the original myocardial injury monitoring data. The classification module 104 uses a hierarchical clustering algorithm to classify the original monitoring data. A distance matrix is ​​constructed based on features such as biomarker concentration values, change trends, and time series, and clustering is performed using the Ward minimum variance method. The clustering results are compared and calibrated with standard data in the myocardial injury marker reference value database, cardiovascular disease database, and medical record database, and numerical standardization is performed to form a myocardial injury analysis data set with a unified dimension.

[0027] The assessment module 105 constructs a risk assessment model based on a deep learning framework. The input layer receives the myocardial injury analysis dataset and extracts temporal features using a multi-layer convolutional neural network. This model then combines geographic location information for spatiotemporal correlation analysis and calculates the weight coefficients for each risk factor. Assessment indicators include the degree of abnormal biomarker concentration, symptom progression rate, and geographic environmental factors. A comprehensive score is generated to generate real-time myocardial injury risk warning data. The grading module 106 sets risk level thresholds based on the ischemic time criteria of the European Society of Cardiology guidelines. Each indicator in the real-time warning data is weighted and a comprehensive score is generated based on the patient's medical history. Risk levels are categorized into mild (within 1 hour), moderate (1-2 hours), and severe (more than 2 hours) based on the total cardiac ischemic time. A corresponding warning strategy is established for each level interval, generating feedback information containing risk warnings. The assessment results are then transmitted via encrypted communication to the medical terminal and the patient's terminal, and the complete record is stored in the medical record database.

[0028] For example, a peripheral blood sample tested with an immunochromatographic test card shows normal color development on the control line, indicating a valid test. Quantitative analysis of the color intensity of the test line reveals a troponin concentration of 2.5 ng / mL. The system generates an initial monitoring data set based on the patient's basic information and prior history of coronary artery disease. This data is encrypted and compressed before being transmitted via the IoT, where it is decrypted at the receiving end to obtain the original data. The classification module clusters this data into a high-risk group, and the assessment module calculates a comprehensive risk score of 0.85. Based on the distance from the patient's current location to the nearest hospital and the duration of symptoms, the system determines the patient's risk as moderate. It immediately sends an alert to the medical institution and patient, advising them to seek medical attention promptly.

[0029] In the embodiment of the present application, the quality control module uses immunochromatography to perform antigen-antibody reaction processing, and performs optical density analysis through the color development degree of the test line and the quality control line, which can accurately obtain the quantitative data of myocardial injury biomarkers, ensuring the reliability and accuracy of the test results. The extraction module uses a preset information collection template to perform structured extraction of the patient's basic characteristic information and clinical history information, and associates and integrates it with the quantitative data of myocardial injury biomarkers, effectively solving the problem of data fragmentation and improving the integrity and practicality of the data. The encryption module uses an asymmetric encryption algorithm to encrypt the data, and generates an encrypted data packet through a data compression algorithm, ensuring the security and efficiency of the data transmission process. The classification module uses a hierarchical clustering algorithm to label and classify the original data, and combines the reference standard values ​​in the multi-source database for data calibration and standardization, thereby improving the accuracy and scientificity of the data analysis. The evaluation module performs multi-parameter risk assessment based on a deep learning model, and combines geographic location information for spatiotemporal correlation analysis, realizing accurate prediction and timely warning of myocardial injury risks. The grading module performs graded processing based on the warning level assessment rules, feeds the assessment results back to the terminal through encrypted transmission, and stores the results in the database to form dynamic monitoring records, providing a reliable basis for medical decision-making. This modular design not only realizes the automated management of the entire process from data acquisition, processing, analysis to early warning, but also ensures patient privacy and data security through multiple data protection mechanisms. At the same time, the application of deep learning algorithms improves the accuracy of myocardial injury risk assessment. The system's real-time monitoring and early warning functions can help medical staff promptly identify potential myocardial injury risks and take appropriate preventive and intervention measures, which is of great significance to improving the diagnosis and treatment of cardiovascular diseases. In addition, the system's Internet of Things architecture design enables data to be transmitted and shared remotely, facilitating collaboration between medical institutions and improving the efficiency of medical resource utilization.

[0030] In a specific embodiment, the quality control module 101 is used to:

[0031] (1) The peripheral blood sample is chromatographically migrated through the capillary effect to form an immune response flow;

[0032] (2) Capture and enrich the immune reaction flow based on the specific binding of antigen and antibody to generate color reaction strips;

[0033] (3) Measure the optical density of the color reaction band based on the colorimetric method to obtain the grayscale value of the band;

[0034] (4) Perform numerical standardization on the grayscale value of the strip to obtain the standard optical density value;

[0035] (5) Calculate the concentration of the standard optical density value through the standard curve to form concentration data;

[0036] (6) Quantitative analysis of myocardial injury biomarkers is performed based on the concentration data to obtain quantitative data of myocardial injury biomarkers.

[0037] Specifically, the quality control module 101's detection process for myocardial injury biomarkers begins with sampling. When a peripheral blood sample is dripped onto the sample pad, the capillary effect causes the blood sample to migrate horizontally along the nitrocellulose membrane due to the unique capillary structure of the glass fiber membrane. In the immune reaction stream formed during this chromatographic migration, myocardial-specific proteins in the blood (including creatine kinase and its MB isoform, troponins I and T, and myosin binding protein) begin to specifically bind to the gold-labeled antibody. As the immune reaction stream continues to migrate, the gold-labeled antibody-antigen complex is captured and enriched by the immobilized monoclonal antibody at the test line. Simultaneously, excess gold-labeled antibody continues to migrate to the control line, where it binds to the immobilized rabbit anti-mouse IgG antibody. This dual-line design ensures test validation; only when the control line displays color does the test process validate. Specific immune reactions at the test and control lines cause the gold-labeled particles to aggregate there, forming a visible red color reaction band.

[0038] The quantitative analysis of the color reaction strips adopts the principle of colorimetry, and the strips are scanned by an optical density instrument. The optical density meter emits a light beam of a specific wavelength to illuminate the color strips, measures the intensity of the transmitted light, converts the light signal into an electrical signal, and then obtains a digital signal through analog-to-digital conversion to obtain the grayscale value data of the strips. The range of grayscale values ​​is usually between 0-255. The larger the value, the darker the color, indicating that the concentration of the bound marker is higher. When standardizing the grayscale values ​​of the obtained strips, background interference needs to be eliminated. By measuring the grayscale value of the nitrocellulose membrane background as the background value, the grayscale value of the detection line is subtracted from the background value to obtain the actual optical density value. In order to facilitate the comparison of results from different batches, the optical density value is normalized to obtain the standard optical density value. The calculation formula for the standard optical density value is:

[0039] ;

[0040] in, is the standard optical density value, is the gray value of the detection line, is the background gray value, is the full-scale grayscale value, is the correction factor.

[0041] When converting the standard optical density value to the actual biomarker concentration, a standard curve needs to be established. By measuring a series of standards with known concentrations, the corresponding relationship between the standard optical density value and the concentration is obtained. The calculation formula for concentration conversion is:

[0042] ;

[0043] in, is the biomarker concentration value, is the standard optical density value, is the exponential growth coefficient, is the exponential decay coefficient, is the linear growth coefficient, is the linear attenuation coefficient, is the baseline correction coefficient. These parameters are obtained by fitting the measurement results of the standard. Based on the converted concentration data, the myocardial injury biomarkers are quantitatively analyzed in combination with clinical diagnostic criteria. By analyzing the characteristic parameters such as the change trend, peak value, and duration of the marker concentration, the quantitative data of the myocardial injury biomarkers are formed. For example: when a peripheral blood sample is added to the test card, the sample migrates under the action of capillaries and binds to the gold-labeled antibody to form an immune complex. At the detection line, the biomarker labeled with the gold-labeled antibody is captured to form a red strip. The optical density meter measures the grayscale value of the detection line to be 180, the background grayscale value to be 20, and the full-scale grayscale value to be 250. After standardization, the standard optical density value of 0.696 is obtained. Substituting this optical density value into the standard curve equation, the actual concentration of the biomarker is calculated. Combined with the patient's clinical manifestations and other test indicators, a quantitative assessment of the degree of myocardial injury is completed.

[0044] In one embodiment, the extraction module 102 is configured to:

[0045] (1) Convert the basic feature information into key-value pairs through field mapping to obtain structured basic feature data;

[0046] (2) Convert clinical history information into time series records through time series coding to obtain structured clinical history data;

[0047] (3) Timestamp merging of structured basic feature data and structured clinical history data to generate time series feature data;

[0048] (4) Complement missing values ​​in time series feature data through data verification to obtain complete feature data;

[0049] (5) Align the complete feature data and the quantitative data of myocardial injury biomarkers to form associated feature data;

[0050] (6) The associated feature data are merged into fields through data identification to obtain the initial myocardial injury monitoring data set.

[0051] Specifically, the extraction module 102 performs field mapping and establishes a key-value pair relationship of basic feature information. In the specific process, basic information such as patient number, gender, age, height, weight, and location is converted according to a standardized format. Field mapping is stored in JSON format, and each field contains attribute information such as field name, field type, and value range. For example, the gender field is mapped to {field:"gender", type:"string", value:["male","female"]}, and the age field is mapped to {field:"age", type:"integer", range:[0,120]}. Through this field mapping, unstructured basic feature information is converted into a standardized key-value pair data structure.

[0052] To convert clinical history data into time series records, time series encoding is used to process medical information such as history of cardiovascular disease, other medical conditions, surgical trauma, and allergies. Each medical history record contains information such as onset time, disease type, severity, and treatment plan. Time series encoding converts this information into structured data with timestamps in the form [timestamp, event_type, severity, treatment]. This encoding facilitates subsequent time series analysis and association mining of medical history data. When merging the structured basic feature data and clinical history data with timestamps, a unified time base is required. Using the patient's first visit as the reference point, all time information is converted to relative time. Timestamps are used to align data from different sources along the temporal dimension, generating feature data containing complete time series information. Each time point in the time series feature data is associated with the basic feature state and clinical history events at that moment.

[0053] When performing data verification on time series feature data, the focus is on dealing with missing values. For missing numerical features, the average value of the adjacent time points is used to fill in the missing values; for missing categorical features, the nearest neighbor interpolation method is used to supplement them. Data verification also includes outlier detection, marking and correcting data that exceeds the normal range. These processes ensure the integrity and validity of the feature data. The complete feature data needs to be aligned with the quantitative data of myocardial injury biomarkers. The quantitative data of biomarkers contains the detection time and the corresponding concentration value. These data are matched with the feature data along the time axis. In the case where the detection time point does not completely overlap with the feature data time point, the time window matching method is used to associate the data at the closest time point.

[0054] Data identifiers are used to merge fields of associated feature data. Each data record is assigned a unique identifier that includes information such as patient ID, recording time, and data type. All relevant fields are organized according to a predefined format to form a unified initial myocardial injury monitoring dataset. This dataset contains a complete spectrum of patient information.

[0055] For example, when a patient undergoes myocardial injury monitoring, basic feature information is collected, such as patient number P20240122001, male, and 45 years old, and converted into a standardized key-value pair format through field mapping. At the same time, medical history information is collected, such as the patient's diagnosis of coronary heart disease five years ago and stent implantation surgery three years ago. This information is converted into structured clinical history data through time series encoding. The system establishes a timeline based on the time of admission and aligns the basic feature data with the medical history data. During the processing, missing blood pressure data was found and supplemented with data from a nearby time point. The myocardial injury biomarker test performed that day showed an elevated troponin concentration, and this test result was associated with other feature data. All data is organized by unique identifiers to form a monitoring data set.

[0056] In a specific embodiment, the encryption module 103 is configured to:

[0057] (1) Obtain quantitative data of myocardial injury biomarkers and patient clinical information data from the initial myocardial injury monitoring dataset;

[0058] (2) Inputting the quantitative data of myocardial injury biomarkers and the patient's clinical information data into a data encryption management model to generate a segmented encryption scheme that matches the myocardial injury monitoring, wherein the segmented encryption scheme includes a biomarker data encryption segment, a clinical information data encryption segment, and an overall encryption path consisting of multiple encryption segments, and setting corresponding data verification parameters for each data encryption segment;

[0059] (3) Perform RSA encryption on each data encryption segment and verify the encryption validity based on the data verification parameters;

[0060] (4) performing block compression on the encrypted data to generate an encrypted data packet with a check bit, wherein the encrypted data packet includes encryption position information of the data encryption segment, encryption parameter information, and data check information;

[0061] (5) Transmit the encrypted data packet in blocks through the network protocol and perform data integrity verification at the receiving end;

[0062] (6) Decrypt the encrypted data packet according to the encrypted location information and encrypted parameter information to obtain the original myocardial injury monitoring data.

[0063] Specifically, the encryption module 103 processes data for secure transmission by separating two key data types from the initial myocardial injury monitoring dataset: quantitative myocardial injury biomarker data (including troponin concentration, test time, and test sequence number) and patient clinical information data (including basic characteristic information, medical history, and treatment records). The separated data is categorized and labeled according to data type and sensitivity to facilitate subsequent differentiated encryption. The categorized and labeled data is input into the data encryption management model, which automatically generates a segmented encryption scheme based on data characteristics. The biomarker data encryption segment specifically processes quantitative test data, packaging the test value, timestamp, and sequence number into fixed-length data blocks. The clinical information data encryption segment processes patient information, encoding and packaging text and numerical data separately. The overall encryption path specifies the order and hierarchy of data encryption, prioritizing encryption of highly sensitive biomarker data, followed by clinical information data. Verification parameters are set for each data encryption segment, including a data length check code, a cyclic redundancy check (CRC), and a hash check value.

[0064] RSA encryption uses a 1024-bit key pair, performing block encryption on each encrypted data segment. A random session key is generated and encrypted with the recipient's public key. The data is then symmetrically encrypted using the session key, with each encrypted block size being 117 bytes (to avoid exceeding the RSA encryption limit of 128 bytes after padding). Digital signatures are calculated for the encrypted data block and session key ciphertext, respectively, and together with the checksum, form an integrity verification chain. The encrypted data undergoes block compression using the DEFLATE algorithm. This compression process represents adjacent repeated data using a shorter encoding while preserving the necessary decompression dictionary. The compressed data consists of three main components: a header (recording the encryption location and parameter information), a data body (the compressed encrypted data), and checksum information (including a CRC code and digital signature). The size of each compressed block is kept below 1KB to ensure efficient transmission.

[0065] Data packets are transmitted via the IoT communication protocol, supporting multiple transmission methods such as WiFi, NFC, or GPS. Transmission adopts a block strategy, with each data block carrying a sequence number and checksum. The receiving end reassembles the data blocks by sequence number and verifies data integrity through checksums. For data blocks lost or damaged during transmission, an automatic repeat request (ARQ) mechanism is used for retransmission. The receiving end verifies the integrity of the data packet, including checking the CRC code and verifying the digital signature. After verification, decryption is performed based on the encrypted position and parameters in the header information. The session key is first decrypted with the recipient's private key, and then the decrypted session key is used to decrypt the data body. The decrypted data is decompressed and restored to the original myocardial injury monitoring data.

[0066] For example, an initial monitoring data set contains a patient's troponin quantitative test data (test value, test time, serial number) and clinical information (medical history, medication records). The encryption management model divides the test data into a primary encryption segment and the clinical information into a secondary encryption segment. After generating the encryption scheme, the test data is encrypted using the RSA algorithm, and a digital signature and CRC checksum are calculated. The encrypted data is compressed to form a data packet containing header information, compressed data, and checksum information. The data packet is transmitted in blocks to the receiver via the WiFi protocol. The receiver verifies the data integrity and then decrypts and decompresses it to restore the original monitoring data.

[0067] In one embodiment, the classification module 104 is configured to:

[0068] (1) Obtaining biomarker quantitative data and medical history characteristic data from original myocardial injury monitoring data;

[0069] (2) Clustering and stratifying the biomarker quantitative data according to the concentration value range to generate multi-level concentration distribution data, where the multi-level concentration distribution data includes the baseline concentration value, change trend value, peak point and the corresponding time series information of the biomarker;

[0070] (3) Compare and analyze the multi-level concentration distribution data with the reference values ​​of myocardial injury markers, cardiovascular disease data, and medical record data in a multi-source database to establish a grading standard for the severity of myocardial injury;

[0071] (4) Labeling the concentration distribution data based on the myocardial injury severity grading standard to generate graded data containing risk level labels;

[0072] (5) Calibrate the abnormal numerical points in the graded data to eliminate the data deviation caused by interference factors;

[0073] (6) The calibrated classification data were correlated with the medical history characteristic data to form a myocardial injury analysis data set.

[0074] Specifically, classification module 104 extracts two core types of information from the raw myocardial injury monitoring data: quantitative biomarker data and patient history data. Quantitative biomarker data includes concentration values ​​for various markers, including creatine kinase and its MB isoform, cardiac troponin I and T, and myosin binding protein. Each data set is timestamped. Patient history data includes information such as previous cardiovascular disease history, surgical history, and medication records. When clustering and stratifying the quantitative biomarker data, a hierarchical clustering algorithm is used to process the concentration values. First, the concentration values ​​of each marker are arranged in a time series, and the concentration differences between adjacent time points are calculated to generate a trend sequence. Peak times are determined by finding the maximum concentration value, and the concentration level corresponding to each peak is recorded. Based on these characteristic values, multi-level concentration distribution data is constructed, including the baseline concentration (initial detection value) of each marker, the trend (rise / fall rate), the peak point (maximum concentration and time of occurrence), and the time series data for the entire monitoring process.

[0075] When comparing and analyzing multi-level concentration distribution data with multi-source databases, the first step is to query the myocardial injury marker reference value database to obtain the standard concentration ranges for each marker at different severity levels. The cardiovascular disease database is then searched to obtain characteristic patterns of marker changes across different cardiovascular disease types. Simultaneously, the medical record database is used to extract marker change patterns from similar cases. By comparing and analyzing data from these three dimensions, a grading system for myocardial injury severity is established. The labeling process involves assigning risk levels to the concentration distribution data. Based on the established grading system, the marker data is mapped to corresponding risk levels. Specifically, multiple risk indicators are defined: the degree of single concentration exceedance (the ratio of the current concentration to the reference value), the rate of concentration increase (the rate of increase per unit time), and the duration (the duration of the exceedance). Each indicator is assigned a specific threshold range. For example, a troponin I concentration exceeding 5 times the normal value for more than 2 hours is considered high risk, while a rate of increase exceeding 100% per hour is considered an acute exacerbation. By comprehensively evaluating these indicators, the data at each time point is labeled with a risk level: mild, moderate, or severe.

[0076] When calibrating outliers in the graded data, first identify data points that do not conform to physiological patterns. A tolerance range is set, and data points outside this range are marked as suspected anomalies. These anomalies are traced back to identify any interfering factors (such as instrument failure, sample contamination, medication interference, etc.). Abnormal data confirmed to be caused by interference will be eliminated and replaced with the average value of the adjacent valid data points. Finally, the calibrated graded data is correlated with the patient's medical history data. Based on the timeline, the graded data is matched with the patient's medical history, treatment records, complications, and other information to analyze the correlation between marker changes and clinical events. Through this correlation analysis, a myocardial injury analysis dataset containing complete monitoring data and clinical background is formed.

[0077] For example, a patient's monitoring data includes 72 consecutive hours of troponin I concentration test values. After arranging these data in chronological order and calculating the concentration change trend, it was found that the first peak appeared at the 24th hour. Comparing with the reference value database, it was found that the peak value exceeded the normal range by 8 times, and the rate of increase in concentration reached 150% per hour within the first 4 hours, which is consistent with the characteristics of acute myocardial injury. The system automatically labeled the data for this period as "severe risk". At the same time, it was found that an abnormal concentration drop occurred at the 48th hour. After verification, it was found that a certain drug that interfered with the detection was used during this period. The abnormal value was then eliminated and replaced with the average value of the previous and next time points. These processed data were associated with the patient's coronary heart disease history, stent surgery records and other information to form an analysis data set.

[0078] In one embodiment, the evaluation module 105 is configured to:

[0079] (1) Calculate the difference between the biomarker concentration values ​​in the myocardial injury analysis data set and the reference standard range to obtain concentration deviation data;

[0080] (2) Correlate the concentration deviation data with the medical history characteristic data in time series and generate a risk trend curve according to the time nodes;

[0081] (3) Calculate the slope and analyze the inflection point of the risk trend curve to identify the time point when myocardial damage worsens;

[0082] (4) Map the time point of myocardial injury aggravation with geographic location information to establish a correlation map between the patient's activity trajectory and the degree of myocardial injury;

[0083] (5) Conduct risk factor attribution analysis based on the association map and calculate the contribution weight of each risk factor;

[0084] (6) Perform weighted calculation on the contribution weight and current monitoring data to generate real-time myocardial injury risk warning data.

[0085] Specifically, the assessment module 105 first performs a difference calculation on the myocardial injury analysis dataset. The concentration values ​​of key myocardial injury biomarkers (including troponin, myoglobin, and creatine kinase) are compared with the standard reference ranges. Each marker has a corresponding upper limit of normal value, and the deviation between the actual measured value and the upper limit is calculated. The deviation value includes not only the absolute difference but also the multiple exceeding the normal range, thereby generating standardized concentration deviation data. When temporally correlating the concentration deviation data with patient medical history characteristics, a unified timeline is required. Medical history data includes key information such as the time of onset of previous cardiovascular disease and the timing of treatment interventions. By arranging the deviation data in chronological order and aligning them with historical events on the timeline, a risk trend curve is plotted. This curve reflects the temporal evolution of myocardial injury severity.

[0086] The following mathematical model is used to analyze the risk trend curve:

[0087] ;

[0088] in, is the risk change rate at time t, is the measured value of the i-th marker at time t, is the corresponding threshold, is the weight coefficient of each marker, is the weight of the second-order derivative term, and n is the number of marker types. By calculating the maximum value and inflection point of R(t), the critical time point of myocardial injury aggravation is identified.

[0089] The identified time points of worsening symptoms are spatiotemporally matched with GPS positioning data, and the geographic location information corresponding to each key time point is recorded. By connecting these locations, a patient's activity trajectory is constructed. Myocardial injury severity data is simultaneously overlaid onto this trajectory to construct a three-dimensional correlation map encompassing time, space, and disease severity. When conducting attribution analysis on this correlation map, risk factors from multiple dimensions are considered: environmental factors (such as air quality and altitude), activity factors (such as strenuous exercise and emotional state), and underlying medical conditions (such as hypertension and diabetes). Multivariate regression analysis is used to determine the contribution of each factor to the worsening of myocardial injury, and the corresponding weight coefficients are calculated. Each weight coefficient is then weighted and calculated with real-time monitoring data to determine the current risk warning value. The warning value is directly related to the warning level, which in turn generates real-time myocardial injury risk warning data.

[0090] For example, a patient's troponin I test value consistently exceeded the upper limit of normal by three times and showed a continuous upward trend. Combined with the patient's history of coronary artery disease, the test data points were plotted onto a risk trend curve. Calculating the slope of the curve revealed a sudden increase at a certain point in time, with a clear inflection point, indicating worsening myocardial damage. The GPS location at that point in time indicated that the patient was engaging in moderate-intensity exercise. This information was integrated into a correlation map, taking into account factors such as the patient's underlying medical conditions (hypertension) and local environmental conditions (high temperatures). After weighting, exercise intensity and underlying medical conditions were identified as major risk factors. Based on the current test value and a weighted analysis of these factors, a high-risk warning message was generated.

[0091] In one embodiment, the classification module 106 is configured to:

[0092] (1) Analyze the troponin and creatine kinase indicators in the real-time myocardial injury risk warning data, compare them with the standard thresholds in the myocardial injury marker reference value database, and obtain the risk score of each indicator;

[0093] (2) Based on the risk score and the patient's previous cardiovascular disease history and medical history time series, a multidimensional correlation calculation is performed to generate a comprehensive risk score that includes time series characteristics;

[0094] (3) Divide the comprehensive risk score into intervals according to the total cardiac ischemia time threshold, and establish a risk level interval map including mild, moderate, and severe;

[0095] (4) Perform spatiotemporal matching of the risk level interval map and the patient's geographic location information, and generate a feedback data packet using an asymmetric encryption algorithm;

[0096] (5) Transmit the feedback data packet to the medical terminal and the patient terminal through the Internet of Things communication protocol to generate graded early warning prompt information;

[0097] (6) Write the graded warning information and the corresponding assessment basis data into the medical record database to form a dynamic monitoring record.

[0098] Specifically, the grading module 106 processes key indicators in the real-time myocardial injury risk warning data. For the troponin marker, the current test value is read and compared with the standard threshold in the reference value database, and the excess multiple and duration are calculated. For the creatine kinase marker, the proportion of MB subtypes and total activity are analyzed to obtain the corresponding risk score. The marker reference value database stores standard threshold ranges for different pathological conditions. Through comparative analysis, the specific risk score for each indicator is derived. The correlation analysis between the risk score and the patient's medical history data uses a multidimensional data fusion method. The patient's previous cardiovascular disease history, including the number of episodes, severity, and treatment plan, is extracted from the medical record database. This information is arranged chronologically to form a medical history time series. A multidimensional correlation calculation matches the current risk score with key events in the medical history time series and calculates the weight of each risk factor. The weight calculation takes into account multiple dimensions such as disease recurrence frequency, treatment efficacy, and complications, ultimately generating a comprehensive risk score that incorporates complete time series characteristics.

[0099] The European Society of Cardiology guidelines were used to categorize the intervals for the comprehensive risk score. Risk levels were divided into three levels based on the total cardiac ischemia time: mild risk if the ischemia time was less than one hour, moderate risk if it was one to two hours, and severe risk if it was more than two hours. Within each risk level, multiple sublevels were further subdivided based on the changing trends in biomarker levels. This multi-level grading approach established a refined risk level interval map. The risk level interval map was matched to the patient's geographic location information using spatiotemporal data association technology. GPS positioning was used to obtain the patient's real-time location coordinates, and the patient's activity patterns were analyzed in combination with movement trajectory data. The location information and risk level data were superimposed in the spatiotemporal dimensions to form a risk distribution map that included geographic attributes. The integrated data was encrypted using the RSA asymmetric encryption algorithm to generate a feedback data packet containing encrypted data and verification information.

[0100] Feedback data packets are transmitted via IoT communication protocols, supporting multiple communication methods such as WiFi, NFC, or GPS. At the medical terminal, the data packet is decrypted and displays detailed risk assessment results, including the specific values ​​of each indicator, the basis for determining the risk level, and recommended measures. At the patient terminal, this information is converted into easy-to-understand early warning information, such as risk level prompts and medical advice. Complete early warning information and its assessment basis are written to the medical record database. The assessment basis includes biomarker test data, clinical history records, geographic location information, etc. This data is stored in a standard format to form a dynamic monitoring record for the patient. Each record contains information such as timestamp, data source, and processing flow to facilitate subsequent disease tracking and analysis.

[0101] For example, real-time monitoring data for a patient showed a persistently elevated troponin concentration, exceeding the upper limit of normal by six times, and the proportion of the MB isoform of creatine kinase reached 8%. A review of the patient's medical history revealed two episodes of angina within a year and a history of hypertension. A comprehensive analysis of these indicators generated a risk score, which, compared to the ischemic time threshold, was determined to be moderate risk. The patient was currently within 5 kilometers of their residence. This information, along with the risk level, was encrypted and transmitted to the medical terminal. Upon receiving the warning, the doctor advised the patient to seek medical attention immediately. Data from the entire monitoring and warning process was stored in the medical record database for subsequent follow-up analysis.

[0102] In a specific embodiment, if Figure 2 FIG. 1 is a schematic diagram of the structure of the grading module 106 in the early myocardial injury monitoring system based on the medical Internet of Things in an embodiment of the present application. The grading module 106 includes:

[0103] The sampling unit 1061 is used to perform time series sampling on the comprehensive risk score to obtain risk score data at different time points;

[0104] The calculation unit 1062 is used to perform cumulative ischemia time calculation on the risk score data to generate ischemia duration data;

[0105] The grading unit 1063 is configured to set a risk grading threshold based on the ischemia duration data, wherein an ischemia duration of less than one hour is considered mild risk, one to two hours is considered moderate risk, and more than two hours is considered severe risk;

[0106] Extraction unit 1064, used to extract features for different risk level intervals to form risk level feature vectors;

[0107] Comparison unit 1065, used to compare and analyze the risk level feature vector with the current monitoring data in real time to determine the interval to which the current risk belongs;

[0108] The updating unit 1066 is used to dynamically update the determined risk intervals and generate a danger level interval map.

[0109] Specifically, the sampling unit 1061 in the grading module 106 first obtains the time series data of the comprehensive risk score. Sampling is performed at a fixed time interval, and the risk score is recorded every 5 minutes to form discrete time series data points. Each sampling point contains a timestamp, a risk score value, and the corresponding biomarker data. Through this timed sampling method, continuous monitoring of the patient's status is ensured. The calculation unit 1062 calculates the duration of the biomarker level exceeding the standard based on the risk score data obtained by sampling. The cumulative timing starts from the time point when the normal value is first exceeded, and the trend of the marker concentration change in each time period is recorded. The calculation of the duration of ischemia takes into account the fluctuation of the marker concentration. When the concentration falls back to the normal range, the timing is paused, and when it exceeds the standard again, the accumulation continues, thereby accurately reflecting the actual duration of myocardial ischemia.

[0110] Grading unit 1063 sets risk grading thresholds according to the European Society of Cardiology guidelines. Using the duration of ischemia as the primary grading basis and combined with marker concentration levels, the risk level is divided into three intervals: within one hour, mild risk, indicating the need for close observation; one to two hours, moderate risk, with prompt medical attention recommended; and more than two hours, severe risk, requiring emergency treatment. Each risk level is further subdivided into multiple sublevels for more refined risk classification. Extraction unit 1064 extracts data features from different risk level intervals. Four main features are extracted: time features (duration, attack frequency), concentration features (peak value, rate of change), symptom features (accompanied manifestations), and medical history features (previous attacks). These features are organized into vector form, with each risk level corresponding to a feature vector describing the typical characteristic pattern of that level.

[0111] The comparison unit 1065 compares the currently monitored data with the feature vector in real time. A pattern matching algorithm is used to calculate the similarity between the current data and the feature vectors of each risk level. The similarity calculation comprehensively considers the weights of each feature and classifies the current state into the most matching risk interval. The comparison process is carried out in real time to ensure the timeliness of risk level determination. The update unit 1066 is responsible for the dynamic maintenance of the hazard level interval map. As new monitoring data is continuously generated, the feature vectors of each risk level are regularly updated. The update process includes recalculation of feature values, dynamic adjustment of interval boundaries, and identification and processing of abnormal patterns. Through this dynamic update mechanism, the accuracy of the risk assessment standard is maintained.

[0112] For example: When monitoring a patient at high risk of myocardial infarction, the sampling unit records the levels of troponin and creatine kinase every 5 minutes. When the troponin begins to rise, the calculation unit starts the timer to record the duration of the exceedance. After 30 minutes, the concentration continues to rise and chest pain symptoms appear, and the grading unit classifies it as a mild risk. The extraction unit analyzes the data features and finds that the attack pattern is similar to that recorded in the previous medical history. The comparison unit compares the current status with the feature library to confirm the accuracy of the risk level judgment. At 1 hour and 15 minutes, due to the worsening symptoms and the continued increase in markers, the update unit raises the risk level to moderate and updates the patient's risk feature model.

[0113] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An early myocardial injury monitoring system based on medical Internet of Things, characterized by: The early myocardial injury monitoring system based on the medical Internet of Things includes: The quality control module is used to perform antigen-antibody reaction processing on peripheral blood samples through immunochromatography, and perform optical density analysis based on the color development of the test line and the quality control line to obtain quantitative data of myocardial injury biomarkers; The extraction module is used to perform structured extraction of the patient's basic characteristic information and clinical history information based on a preset information collection template, and to associate and integrate it with the quantitative data of myocardial injury biomarkers to obtain the initial myocardial injury monitoring data set; An encryption module is used to encrypt the initial myocardial injury monitoring data set based on an asymmetric encryption algorithm, generate an encrypted data packet through a data compression algorithm, transmit it through the Internet of Things protocol, and decrypt and decompress it to obtain the original myocardial injury monitoring data; The classification module is used to label and classify the original myocardial injury monitoring data through a hierarchical clustering algorithm, and perform data calibration and standardization processing in combination with the reference standard values ​​in the multi-source database to form a myocardial injury analysis data set. It is specifically used to: obtain the biomarker quantitative data and medical history characteristic data in the original myocardial injury monitoring data; cluster and stratify the biomarker quantitative data according to the concentration value range to generate multi-level concentration distribution data, wherein the multi-level concentration distribution data includes the baseline concentration value, change trend value, peak point and corresponding time series information of the biomarker; compare and analyze the multi-level concentration distribution data with the myocardial injury marker reference value, cardiovascular disease data, and medical record data in the multi-source database to establish a myocardial injury severity grading standard; label the concentration distribution data based on the myocardial injury severity grading standard to generate graded data containing risk level labels; perform data calibration on abnormal numerical points in the graded data to eliminate data deviations caused by interference factors; and perform correlation analysis on the calibrated graded data with the medical history characteristic data to form a myocardial injury analysis data set; The evaluation module is used to perform multi-parameter risk assessment on the myocardial injury analysis data set based on the deep learning model, conduct spatiotemporal correlation analysis in combination with geographic location information, and generate real-time myocardial injury risk warning data. Specifically, it is used to: calculate the difference between the biomarker concentration value in the myocardial injury analysis data set and the reference standard range to obtain concentration deviation degree data; perform time-series correlation between the concentration deviation degree data and the medical history feature data, and generate a risk trend curve according to the time node; perform slope calculation and inflection point analysis on the risk trend curve to identify the time point of myocardial injury aggravation; map the time point of myocardial injury aggravation with geographic location information to establish a correlation map between the patient's activity trajectory and the degree of myocardial injury; perform risk factor attribution analysis on the correlation map to calculate the contribution weight of each risk factor; perform weighted calculation on the contribution weight and current monitoring data to generate real-time myocardial injury risk warning data; The grading module is used to grade the real-time myocardial injury risk warning data according to the warning level assessment rules, transmit the data through encryption, and store the assessment results in the database to form dynamic monitoring records.

2. The early myocardial injury monitoring system based on medical Internet of Things according to claim 1 is characterized in that: The quality control module is used to: The peripheral blood sample is subjected to chromatographic migration through the capillary effect to form an immune response flow; The immune reaction flow is captured and enriched according to the specific binding of antigen and antibody to generate color reaction bands; The optical density of the color reaction band is measured based on the colorimetric method to obtain the gray value of the band; The grayscale value of the strip is numerically normalized to obtain the standard optical density value; The standard optical density value is converted into concentration through standard curve calculation to form concentration data; The myocardial injury biomarkers are quantitatively analyzed based on the concentration data to obtain quantitative data of the myocardial injury biomarkers.

3. The early myocardial injury monitoring system based on medical Internet of Things according to claim 1 is characterized in that: The extraction module is used to: Convert the basic feature information into key-value pairs through field mapping to obtain structured basic feature data; The clinical history information is converted into time series records through time series coding to obtain structured clinical history data; Merge the structured basic feature data and structured clinical history data with timestamps to generate time series feature data; Through data verification, missing values ​​in time series feature data are supplemented to obtain complete feature data; Align the complete feature data and the quantitative data of myocardial injury biomarkers to form associated feature data; The associated feature data were merged into fields through data identification to obtain the initial myocardial injury monitoring data set.

4. The early myocardial injury monitoring system based on medical Internet of Things according to claim 1 is characterized in that: The encryption module is used to: Obtain quantitative data of myocardial injury biomarkers and patient clinical information data from the initial myocardial injury monitoring dataset; Inputting quantitative data of myocardial injury biomarkers and patient clinical information data into a data encryption management model to generate a segmented encryption scheme that matches the myocardial injury monitoring, wherein the segmented encryption scheme includes a biomarker data encryption segment, a clinical information data encryption segment, and an overall encryption path consisting of multiple encryption segments, and setting corresponding data verification parameters for each data encryption segment; Perform RSA encryption on each data encryption segment and verify the encryption validity based on data verification parameters; Performing block compression on the encrypted data to generate an encrypted data packet with a check bit, wherein the encrypted data packet includes encryption position information of the data encryption segment, encryption parameter information, and data check information; The encrypted data packet is transmitted in blocks through the network protocol, and the data integrity is checked at the receiving end; the encrypted data packet is decrypted according to the encrypted position information and encryption parameter information to obtain the original myocardial injury monitoring data.

5. The early myocardial injury monitoring system based on medical Internet of Things according to claim 1 is characterized in that: The evaluation module uses the following mathematical model to analyze the risk trend curve: ； in, is the risk change rate at time t, is the measured value of the i-th marker at time t, is the corresponding threshold, is the weight coefficient of each marker, is the weight of the second-order derivative term, and n is the number of marker types. By calculating the maximum value and inflection point of R(t), the critical time point of myocardial injury aggravation is identified; The identified aggravation time points are matched with the GPS positioning data in time and space, and the geographical location information corresponding to each key time point is recorded. By connecting the location points, the patient's activity trajectory is formed, and the myocardial injury degree data is superimposed on the trajectory to construct a three-dimensional correlation map that includes time, space and disease severity. When performing attribution analysis on the correlation map, environmental factors, activity factors and underlying disease factors are considered. The contribution of each factor to the aggravation of myocardial injury is determined through multivariate regression analysis, and the corresponding weight coefficient is calculated. The weighted calculation of each weight coefficient and the real-time monitoring data is performed to obtain the current risk warning value, and then form real-time myocardial injury risk warning data.

6. The early myocardial injury monitoring system based on medical Internet of Things according to claim 1 is characterized in that: The grading module is used to: Analyze the troponin and creatine kinase indicators in the real-time myocardial injury risk warning data and compare them with the standard thresholds in the myocardial injury marker reference value database to obtain the risk level score for each indicator; A comprehensive risk score that includes time series characteristics is generated by performing multidimensional correlation calculations based on the risk score and the patient's previous cardiovascular disease history and medical history time series; The comprehensive risk score is divided into intervals according to the total cardiac ischemia time threshold, and a risk level interval map including mild, moderate and severe ischemia is established; Perform spatiotemporal matching of the risk level interval map and the patient's geographic location information, and generate a feedback data packet using an asymmetric encryption algorithm; Transmit feedback data packets to medical terminals and patient terminals through the Internet of Things communication protocol to generate graded early warning prompt information; The graded warning prompt information and the corresponding assessment basis data are written into the medical record database to form a dynamic monitoring record.

7. The early myocardial injury monitoring system based on medical Internet of Things according to claim 6 is characterized in that: The grading module includes: The sampling unit is used to perform time series sampling on the comprehensive risk score to obtain risk score data at different time points; a calculation unit, configured to perform cumulative ischemia time calculation on the risk score data to generate ischemia duration data; A grading unit is used to set risk grading threshold points based on ischemia duration data, where within one hour is mild risk, one to two hours is moderate risk, and more than two hours is severe risk; An extraction unit, configured to extract features from different risk level intervals to form a risk level feature vector; The comparison unit is used to compare and analyze the risk level feature vector with the current monitoring data in real time to determine the interval to which the current risk belongs; The updating unit is used to dynamically update the determined risk interval and generate a hazard level interval map.

Images