Personnel health assessment method and device in high-risk operation environment, electronic equipment and storage medium

Abstract

The invention discloses a personnel health assessment method and device in a high-risk operation environment, electronic equipment and a storage medium. The method comprises the following steps: acquiring real-time physiological health data of an operator in the high-risk operation environment; extracting a feature vector corresponding to the real-time physiological health data, and inputting the feature vector into a preset health risk assessment model, so that the health risk assessment model assesses the probability of occurrence of the health risk of the operator according to the feature vector, and outputs the probability of occurrence of the health risk of the operator; the probability is compared with a preset health risk probability threshold value, then the health state of the operator is determined according to the comparison result, and when the health state of the operator is a dangerous state, a corresponding health early warning signal is generated. According to the invention, the overall safety and the working efficiency of a high-risk working environment can be improved.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method, device, electronic equipment and storage medium for evaluating the health of personnel in a high-risk working environment. Background Art

[0002] Safety management and control in high-risk working environments has always faced risks such as the inability of supervisors to objectively grasp the physical health status of workers before work, the difficulty in preventing and reducing muscle fatigue of workers during work, and the difficulty in reducing the risks of high-risk operations. These problems seriously affect the construction of safe enterprises.

[0003] A high-risk working environment refers to a working environment with high safety risks under specific conditions and restricted by the geographical environment, process and other harmful factors. This type of working environment usually involves dangerous factors such as high altitude, high pressure, flammable, explosive, highly toxic, and radioactive. The characteristics of a high-risk working environment include high risk, high danger, special skills, special equipment requirements, and multiple risk factors. Therefore, in a high-risk working environment, it is usually necessary to control the health status of the operators. However, the current traditional control methods often rely on regular health checks of operators or monitoring during work to ensure the safety of operators. However, such methods cannot timely understand the health status of operators and cannot respond in time, which leads to work errors and accident risks due to the health of operators, affecting the overall safety and work efficiency of the high-risk working environment. Summary of the invention

[0004] The present invention provides a personnel health assessment method, device, electronic device and storage medium in a high-risk working environment to solve the technical problem that the regular health check-up of the personnel or the monitoring method during work cannot timely understand the health status of the personnel and cannot respond in time, thus affecting the overall safety and work efficiency of the high-risk working environment.

[0005] In order to solve the above technical problems, an embodiment of the present invention provides a method for assessing personnel health in a high-risk working environment, comprising:

[0006] Collecting real-time physiological health data of workers in high-risk working environments; wherein the real-time physiological health data includes: electrocardiogram data, electrodermal data and electromyographic data;

[0007] Extracting a feature vector corresponding to the real-time physiological health data, and inputting the feature vector into a preset health risk assessment model, so that the health risk assessment model assesses the probability of occurrence of health risks for the operator according to the feature vector, and outputs the probability of occurrence of health risks for the operator;

[0008] The probability is compared with a preset health risk probability threshold, and then the health status of the operator is determined according to the comparison result, and when the health status of the operator is in a dangerous state, a corresponding health warning signal is generated;

[0009] The health risk assessment model is obtained by training a preset neural network model with the historical physiological health data of the operator as input and the probability of occurrence of health risks corresponding to the historical physiological health data as output.

[0010] As a preferred solution, the real-time physiological health data of workers in high-risk working environments are collected, including:

[0011] Obtain real-time activity intensity indicators and real-time environmental change indicators of operators;

[0012] Calculate a corresponding data collection frequency according to the real-time activity intensity index and the real-time environmental change index, and then collect real-time physiological health data of workers in a high-risk working environment according to the data collection frequency;

[0013] The calculation formula of the data collection frequency is:

[0014]

[0015] Among them, f(t) is the data collection frequency at time t; f0 is the initial collection rate; g(A(t)) is the function of activity intensity; k1 is the adjustment coefficient used to control the influence of activity intensity on the collection frequency; A(t) is the activity intensity index of the operator at time t; h(E(t)) is the function of environmental changes; k2 is the adjustment coefficient used to control the influence of environmental changes on the collection frequency; E(t) is the environmental change index of the operator at time t.

[0016] As a preferred solution, before extracting the feature vector corresponding to the real-time physiological health data, the method further includes:

[0017] Performing wavelet transform and filtering on the electrocardiogram data;

[0018] Acquire a baseline value of skin electrical data of a human body in a calm state, and then perform a baseline calibration on the skin electrical data according to the baseline value;

[0019] Decomposing the electromyographic data into a plurality of independent component signals, and performing denoising processing on each of the independent component signals to obtain denoised electromyographic data;

[0020] Data normalization was performed on the ECG data after wavelet transform and filtering, the electrodermal data after baseline calibration, and the electromyographic data after denoising.

[0021] As a preferred solution, the extracting the feature vector corresponding to the real-time physiological health data includes:

[0022] Respectively extracting electrocardiogram data features of the electrocardiogram data, electrodermal electrodermal data features of the electrodermal electrodermal data, and electromyographic electrodermal data features of the electromyographic electrodermal data;

[0023] The electrocardiogram data features, the electrodermal data features and the electromyography data features are concatenated to obtain a corresponding joint feature vector, and cross-modal features of each data feature in the joint feature vector are extracted;

[0024] The cross-modal features of each data feature in the joint feature vector are integrated to obtain a corresponding comprehensive feature vector, and then the comprehensive feature vector is used as the feature vector corresponding to the real-time physiological health data.

[0025] As a preferred solution, the health risk assessment model is:

[0026]

[0027] Among them, UZZq(YCCq=1|FVUE) is the probability of health risk occurrence under given FVUE; w0 is the intercept term; w1, w2, …, w s is the weight coefficient corresponding to each dimension of FVUE; FVUE1, FVUE2, ..., FVUE s For each element of FVUE.

[0028] As a preferred solution, the training of the health risk assessment model includes:

[0029] Obtain historical physiological health data of operators;

[0030] Extracting a historical feature vector corresponding to the historical physiological health data, and training a preset neural network model based on a preset loss function with the historical feature vector as input and the probability of occurrence of the health risk corresponding to the historical physiological health data as output, to obtain the health risk assessment model;

[0031] Wherein, the loss function is:

[0032]

[0033] Among them, RT g is the number of samples of training data; ycq cap is the true health risk label of the cap-th training sample; FVUE capis the integrated feature vector corresponding to the cap-th training sample; PPI is the probability; In is the natural logarithm; LTNNQ(wujj) is the loss function; YIQ is a variable in the prediction result of the health risk assessment model.

[0034] As a preferred solution, the probability is compared with a preset health risk probability threshold, and then the health status of the operator is determined according to the comparison result. When the health status of the operator is in a dangerous state, a corresponding health warning signal is generated, including:

[0035] The probability is compared with a preset health risk probability threshold. When the probability is equal to the health risk probability threshold, the health status of the operator is determined to be normal. When the probability is not equal to the health risk probability threshold, the health status of the operator is determined to be a risky state, and a corresponding health warning signal is generated.

[0036] Based on the above embodiment, another embodiment of the present invention provides a personnel health assessment device in a high-risk working environment, including: a real-time physiological health data acquisition module, a health risk assessment module and a health warning module;

[0037] The real-time physiological health data acquisition module is used to collect real-time physiological health data of workers in high-risk working environments; wherein the real-time physiological health data includes: electrocardiogram data, electrodermal data and electromyographic data;

[0038] The health risk assessment module is used to extract the feature vector corresponding to the real-time physiological health data, and input the feature vector into a preset health risk assessment model, so that the health risk assessment model assesses the probability of the occurrence of the health risk of the operator according to the feature vector, and outputs the probability of the occurrence of the health risk of the operator; wherein the health risk assessment model is obtained by training a preset neural network model with the historical physiological health data of the operator as input and the probability of the occurrence of the health risk corresponding to the historical physiological health data as output;

[0039] The health warning module is used to compare the probability with a preset health risk probability threshold, and then determine the health status of the operator based on the comparison result. When the health status of the operator is in a dangerous state, a corresponding health warning signal is generated.

[0040] Based on the above embodiments, another embodiment of the present invention provides an electronic device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, the personnel health assessment method in a high-risk working environment described in the above invention embodiment is implemented.

[0041] Based on the above embodiments, another embodiment of the present invention provides a storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute the personnel health assessment method in a high-risk working environment described in the above invention embodiment.

[0042] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0043] The present invention provides a method for assessing the health of personnel in a high-risk working environment, which collects real-time physiological health data of personnel in the high-risk working environment; wherein the real-time physiological health data includes: electrocardiogram data, electrodermal data and electromyography data; extracts a feature vector corresponding to the real-time physiological health data, and inputs the feature vector into a preset health risk assessment model, so that the health risk assessment model assesses the probability of occurrence of health risks of the personnel according to the feature vector, and outputs the probability of occurrence of health risks of the personnel; compares the probability with a preset health risk probability threshold, and then determines the health status of the personnel according to the comparison result, and generates a corresponding health warning signal when the health status of the personnel is in a dangerous state; wherein the health risk assessment model is obtained by training a preset neural network model with the historical physiological health data of the personnel as input and the probability of occurrence of health risks corresponding to the historical physiological health data as output.

[0044] The present invention can analyze and evaluate the real-time physiological health data of the operators in real time. Through the preset health risk assessment model, the probability of the occurrence of health risks of the operators can be identified, and the potential risk of the health status of the operators can be obtained according to the probability. This real-time monitoring and evaluation capability enables the on-site operators to understand their health status in time, so as to take corresponding measures before the risk occurs to avoid accidents, achieve the effect of timely response, and ensure the life safety of operators in high-risk working environments. When the health status of the operators is in a dangerous state, the corresponding health warning signal can be generated in time. Timely health warning can effectively prevent the further deterioration of health problems, reduce the risk of work errors and accidents caused by health reasons, and improve the overall safety and work efficiency of high-risk working environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a flowchart of a method for assessing personnel health in a high-risk working environment provided by an embodiment of the present invention;

[0046] Figure 2 It is a structural schematic diagram of a personnel health assessment device in a high-risk working environment provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solutions and advantages of this application clearer, the technical solutions in this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0048] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" in the specification and claims of this application and the above-mentioned figure descriptions and any variations thereof are intended to cover non-exclusive inclusions.

[0049] In the description of the embodiments of the present application, the technical terms "first", "second", etc. are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "multiple" is more than two, unless otherwise clearly and specifically defined.

[0050] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0051] In the description of the embodiments of the present application, the term "and / or" is only a description of the association relationship of the associated objects, indicating that there may be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0052] In the description of the embodiments of the present application, the term "multiple" refers to more than two (including two). Similarly, "multiple groups" refers to more than two groups (including two groups), and "multiple pieces" refers to more than two pieces (including two pieces).

[0053] In the description of the embodiments of the present application, unless otherwise clearly specified and limited, technical terms such as "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, and it can be the internal connection of two elements or the interaction relationship between two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the embodiments of the present application can be understood according to the specific circumstances.

[0054] Embodiment 1

[0055] Please refer to Figure 1 , is a flow chart of a method for assessing personnel health in a high-risk working environment provided by an embodiment of the present invention, comprising the following specific steps:

[0056] S1. Collecting real-time physiological health data of workers in high-risk working environments; wherein the real-time physiological health data includes: electrocardiogram data, electrodermal data and electromyographic data;

[0057] Preferably, the collecting of real-time physiological health data of workers in high-risk working environments includes: obtaining real-time activity intensity index and real-time environmental change index of workers; calculating corresponding data collection frequency according to the real-time activity intensity index and real-time environmental change index, and then collecting real-time physiological health data of workers in high-risk working environments according to the data collection frequency;

[0058] The calculation formula of the data collection frequency is:

[0059] f(t)=f0×g(A(t))×h(E(t));

[0060] g(A(t))=1+k1A(t);

[0061] h(E(t))=1+k2E(t);

[0062] Among them, f(t) is the data collection frequency at time t; f0 is the initial collection rate; g(A(t)) is the function of activity intensity; k1 is the adjustment coefficient used to control the influence of activity intensity on the collection frequency; A(t) is the activity intensity index of the operator at time t; h(E(t)) is the function of environmental changes; k2 is the adjustment coefficient used to control the influence of environmental changes on the collection frequency; E(t) is the environmental change index of the operator at time t.

[0063] Preferably, before extracting the characteristic vector corresponding to the real-time physiological health data, it also includes: performing wavelet transform and filtering processing on the ECG data; obtaining the baseline value of the skin electrodermal data of the human body in a calm state, and then performing baseline calibration on the skin electrodermal data according to the baseline value; decomposing the electromyographic data into a number of independent component signals, and denoising each of the independent component signals to obtain denoised electromyographic data; and performing data normalization processing on the ECG data after wavelet transform and filtering processing, the skin electrodermal data after baseline calibration, and the electromyographic data after denoising.

[0064] Specifically, the personnel health assessment method in a high-risk working environment of the present invention comprises the following specific steps:

[0065] 1. Obtain real-time multi-source heterogeneous data (i.e., real-time physiological health data) of the smart IoT badge chip worn by workers in high-risk working environments, including ECG data, electrodermal data, and electromyographic data, and pre-process the real-time multi-source heterogeneous data. Adaptive sampling frequency technology is used to dynamically adjust the data collection frequency according to the worker's activity intensity and environmental changes, and at the same time, historical multi-source heterogeneous data of the smart IoT badge chip worn by workers in high-risk working environments is obtained.

[0066] Furthermore, the adaptive sampling frequency technology is used to dynamically adjust the data collection frequency according to the activity intensity of the operators and environmental changes. The adjustment process is as follows:

[0067] Calculate the impact of activity intensity and environmental changes on acquisition frequency:

[0068] The activity intensity impact is calculated using the following formula:

[0069] g(A(t))=1+k1A(t);

[0070] In the formula, g(A(t)) represents the function of activity intensity, k1 is the adjustment coefficient used to control the influence of activity intensity on the acquisition frequency, and A(t) represents the activity intensity index of the operator at time t;

[0071] The impact of environmental changes is calculated according to the following formula:

[0072] h(E(t))=1+k2E(t);

[0073] In the formula, h(E(t)) represents the function of environmental change, k2 is the adjustment coefficient used to control the impact of environmental change on the acquisition frequency, and E(t) represents the environmental change index of the operator at time t;

[0074] The final acquisition frequency is:

[0075] f(t)=f0×g(A(t))×h(E(t));

[0076] Where f(t) represents the data collection frequency at time t, in Hertz, and f0 represents the initial collection rate, which is the default collection frequency when the operator first starts wearing the smart IoT badge chip and does not consider activity intensity and environmental changes.

[0077] 2. Perform time series segmentation and window processing on the real-time multi-source heterogeneous data of workers in high-risk working environments, and segment the continuous real-time multi-source heterogeneous data stream into time window data segments of fixed length to facilitate subsequent parallel processing and feature extraction. For ECG data, an adaptive filtering algorithm based on wavelet transform is used to remove noise interference from ECG data. For electrodermal data, fractal dimension analysis is used to reveal the complexity of nonlinear dynamic systems and perform baseline calibration based on human physiological characteristics. For electromyographic data, electromyographic data denoising technology based on blind source separation is used to decompose the electromyographic data into multiple independent components, remove noise and interference components, and thus extract pure electromyographic data.

[0078] Furthermore, the segmentation process of segmenting the continuous real-time multi-source heterogeneous data stream into fixed-length time window data segments is as follows:

[0079] Set the time series of the original real-time multi-source heterogeneous data to x(t);

[0080] Divide real-time multi-source heterogeneous data into window data segments with a fixed length of L:

[0081]

[0082] In the formula, N represents the number of divided time windows, and T represents the total duration of the original physiological data;

[0083] For the data segment x of n windows n (t), the corresponding time range is:

[0084] t∈[(n-1)L,nL];

[0085] Wherein, n represents a variable identifying the time window sequence number, n = 1, 2, ..., N, which is used to identify different time windows, nL represents the time position at which the nth window ends, and (n-1)L represents the time position at which the nth window starts.

[0086] Furthermore, for ECG data, an adaptive filtering algorithm based on wavelet transform is used to remove noise interference from ECG data. The removal process is as follows:

[0087] Set the data segment of the ECG data in the nth window to x n，EC(t), where EC represents the ECG data; for x n，EC (t) is subjected to wavelet transform, and the transformation formula is as follows:

[0088]

[0089] Where W n，EC (a, b) represents the coefficients after wavelet transformation, a represents the scale parameter, b represents the translation parameter, represents the selected wavelet function, dt represents the integral operation symbol;

[0090] The adaptive filter is updated based on the least mean square algorithm. The process is as follows:

[0091] h n，k+1 =h n，k +2μe n，k x n，EC (tk);

[0092] In the formula, h n，k+1 Indicates h n，k The next update value, h n，k represents the coefficient of the k-th order filter of the ECG data of the n-th window, μ is the step size parameter, e n，k represents the estimation error, Represents the current filter output, x n，EC (tk) represents the value of the ECG data of the nth window at time point tk;

[0093] By continuously updating the parameters of the adaptive filter, the output of the adaptive filter As close to the original ECG data as possible n，EC (t), and perform denoising to finally obtain the denoised ECG data

[0094] Furthermore, fractal dimension analysis is used for skin electrical data to reveal the complexity of nonlinear dynamic systems and to perform baseline calibration based on human physiological characteristics. The analysis process is as follows:

[0095] For a discrete skin electrical data sequence x n，ED (i), where (i = 1, 2, ..., M), M is the number of data points in the window, and the calculation steps of the box dimension method are as follows:

[0096] First, use a box with a side length of ε to cover the spatial area where the skin electrical data sequence is located;

[0097] Set the number of boxes covering the electrodermal data sequence to M(ε);

[0098] Then, the side length ε of the box is changed to obtain different Q(ε), where Q(ε) represents the minimum number of boxes required to cover the electrodermal data sequence under a specific ε;

[0099] According to fractal theory, the fractal dimension D satisfies the following relationship:

[0100] Q(ε)~ε -D ;

[0101] By fitting Q(ε) under different ε and taking the logarithm of Q(ε) and ε, we can get logQ(ε) and Qlogε, where log represents the natural logarithm. Among them, nonlinear dynamic systems often exhibit complex and unpredictable behaviors. Through the analysis of Q(ε), this complexity can be quantified. The change of Q(ε) can reveal the behavior of the system on different time scales and help understand the dynamic characteristics of the system. By fitting the relationship between Q(ε) and ε, we can get the fractal dimension D, which is a numerical indicator used to quantify the complexity of the system.

[0102] The linear equation is obtained by linear regression fitting: logQ(ε)=-Qlogε+C; where C represents a constant;

[0103] Benchmark calibration of skin electrical data based on human physiological characteristics:

[0104] The baseline value of the skin electrical data of the human body in a calm state is set to B0, and the actual measured skin electrical data is x n，ED (t);

[0105] The calibration formula is:

[0106]

[0107] In the formula, Indicates the calibrated skin electrical data.

[0108] Electrodermal data (EDA) may produce baseline drift due to physiological or environmental factors, that is, the baseline level of the data gradually rises or falls over time. Baseline calibration can remove this drift, making the data more stable and easier to analyze. Calibration can standardize data collected from different individuals or at different time points to the same baseline level, which is helpful for comparison and analysis. Calibration can also reduce the impact of noise and improve the quality of the signal, so that meaningful physiological changes in electrodermal data will be more obvious, and calibrated data will be easier to extract features.

[0109] Furthermore, for the electromyographic data, the electromyographic data denoising technology based on blind source separation is used to decompose the electromyographic data into multiple independent components, remove the noise and interference components therein, and thus extract pure electromyographic data. The extraction process is as follows:

[0110] Set EMG data x n，EM (t) consists of multiple independent source signals S n，v (t), (v = 1, 2, ..., KU,), KU is the number of source signals;

[0111] The mixing process is represented by the following pattern:

[0112]

[0113] In the formula, a n，vj represents the elements of the mixing matrix;

[0114] The blind source separation method is used to process the electromyographic data:

[0115] For myoelectric data x n，EM (t) Perform centralization so that the mean is zero;

[0116] Estimate the separation matrix W by iterative calculation n , so that the output signal y n，v (t) = W n x n，EM (t) Be as close as possible to independent components (i.e., source signals);

[0117] The iteration formula is:

[0118]

[0119] In the formula, It is W n The updated value of g(y n ) represents a nonlinear function, E{g ACM (y n )} represents the expected operation, I represents the identity matrix, y n represents the output signal vector, represents the derivative of the output signal vector, represents g(y n )and The expected value of the product;

[0120] By continuously iteratively updating the separation matrix W n , and finally the separated independent component signal y is obtained n，v (t); Assume that the signal corresponding to the noise and interference components is z n，v (t), clean EMG data at this time It is expressed as:

[0121]

[0122] S2, extracting the feature vector corresponding to the real-time physiological health data, and inputting the feature vector into a preset health risk assessment model, so that the health risk assessment model assesses the probability of the occurrence of health risks for the operators according to the feature vector, and outputs the probability of the occurrence of health risks for the operators;

[0123] The health risk assessment model is obtained by training a preset neural network model with the historical physiological health data of the operator as input and the probability of occurrence of the health risk corresponding to the historical physiological health data as output;

[0124] Preferably, the extracting of the feature vector corresponding to the real-time physiological health data includes: respectively extracting the ECG data features of the ECG data, the electrocutaneous data features of the electrocutaneous data, and the electromyographic data features of the electromyographic data; splicing the ECG data features, the electrocutaneous data features, and the electromyographic data features to obtain a corresponding joint feature vector, and extracting the cross-modal features of each data feature in the joint feature vector; integrating the cross-modal features of each data feature in the joint feature vector to obtain a corresponding comprehensive feature vector, and then using the comprehensive feature vector as the feature vector corresponding to the real-time physiological health data.

[0125] 3. Introduce a multimodal learning framework to integrate different types of physiological signals to extract cross-modal features;

[0126] Furthermore, a multimodal learning framework is introduced to integrate different types of physiological signals to extract cross-modal features. The extraction process is as follows:

[0127] Normalize the real-time data:

[0128]

[0129] In the formula, Xn represents the normalized data, Xnq represents the original data, and Xnq min Indicates the minimum value in the data set, Xnq max Indicates the maximum value in the data set;

[0130] Construct a convolutional neural network to extract features of ECG data:

[0131] Convolutional Layer 1:

[0132] Set the number of input channels to 1, the number of output channels to c1, the convolution kernel size to EJ1×EJ1, and the step size to Scx1;

[0133] The calculation formula for the convolution operation is:

[0134]

[0135] In the formula, represents the output feature map of convolutional layer 1, Represents the normalized ECG data matrix, which is used as the input of convolutional layer 1. Conv1 represents the first convolutional layer in the ECG data feature extraction network.

[0136] Activate layer 1:

[0137] Perform a ReLU activation operation on the output of convolutional layer 1 to introduce nonlinear features:

[0138]

[0139] In the formula, Represents the output feature map after being processed by the ReLU activation function. Represents the output feature map of convolutional layer 1;

[0140] Pooling layer 1:

[0141] The maximum pooling operation is adopted, the pooling kernel size is Ej1×Ej1, and the step size is Scx2;

[0142]

[0143] In the formula, represents the output feature map of pooling layer 1, MaxPool represents the maximum pooling layer, Represents the output feature map of activation layer 1;

[0144] The output of pooling layer 1 is flattened and connected to a fully connected layer with the number of neurons fvb1;

[0145]

[0146] Where FE EC represents the output of the fully connected layer, FC represents the fully connected layer itself, Represents the output feature map of pooling layer 1, and Flatten represents the flattening operation;

[0147] Among them, the extracted ECG data features are local features, nonlinear features, global features, etc. These features are used for classification, anomaly detection or other analysis tasks of ECG data. Useful features can be learned from the original ECG data through CNN.

[0148] The process of extracting features from skin electrical data is as follows:

[0149] Construct a multi-layer perceptron structure to extract the features of skin electrical data:

[0150] Input Layer:

[0151] The input dimension is m2, which is the feature dimension of the electrical skin data after preprocessing;

[0152] Hidden Layer 1:

[0153] The number of neurons is mr1, and the ReLU activation function is used:

[0154]

[0155] In the formula, represents the normalized skin electrical data matrix, Represents the output after being processed by hidden layer 1 and activated by the ReLU function;

[0156] Hidden Layer 2:

[0157] The number of neurons is mr2, and the ReLU activation function is also used:

[0158]

[0159] Where APT2 ED represents the output of hidden layer 2;

[0160] Output layer:

[0161] The number of neurons is fvb2;

[0162]

[0163] Where, FPF ED Represents the output of the output layer, OL represents the output layer;

[0164] Among them, the multi-layer perceptron is usually used to extract the time domain features, frequency domain features, time-frequency domain features and nonlinear features of the DEA signal.

[0165] The process of extracting features from electromyographic data is as follows:

[0166] Build a long short-term memory network to extract features from EMG data:

[0167] LSTM Layer:

[0168] The output dimension is mmjz3, and the number of hidden units is lerc1;

[0169]

[0170] In the formula, represents the output of the LSTM layer, Represents the normalized EMG data matrix;

[0171] Fully connected layer (FC):

[0172] Connect the output of the last time step of the LSTM layer to a fully connected layer with fvb3 neurons:

[0173]

[0174] Where, FDD EM is the output of the fully connected layer;

[0175] Among them, the extracted EMG data features are time series features, long-term dependency features, nonlinear features and context features. These features can be used for tasks such as EMG signal classification, recognition, and anomaly detection. Through LSTM, useful features can be learned from the original EMG data.

[0176] The features extracted from three different modal data are spliced ​​together to form a joint feature vector using the splicing fusion method:

[0177] FGG fusion =Contee(FE EC , FPF ED , FDD EM );

[0178] In the formula, FGG fusion represents the joint feature vector after concatenation, and Contee represents the concatenation operation;

[0179] Construct a new fully connected neural network to further extract cross-modal features:

[0180] Input Layer:

[0181] The input dimension is fvb1+fvb2+fvb3, which is the dimension of the joint feature vector after concatenation;

[0182] Hidden Layer 1 (HL1):

[0183] The number of neurons is cso3, and the ReLU activation function is used:

[0184]

[0185] In the formula, represents the output of hidden layer 1, FQQ fusion Represents the joint feature vector after multimodal fusion;

[0186] Hidden Layer 2 (HL2):

[0187] The number of neurons is cso4, and the ReLU activation function is used:

[0188]

[0189] In the formula, represents the output of hidden layer 2;

[0190] Output Layer (OL):

[0191] The number of neurons is kxc:

[0192]

[0193] In the formula, FQyy caross Represents the output of the output layer.

[0194] Among them, cross-modal features refer to features extracted and fused from multiple different types of physiological signals. These features can provide more comprehensive information than a single modality. Cross-modal features include: electrocardiogram features, electrodermal features, electromyography features, time series features, nonlinear features, contextual features, and fusion features. These cross-modal features can be used to monitor and analyze multiple physiological states, such as emotion recognition, stress level assessment, and motion monitoring. By fusing data from different modalities, the accuracy and robustness of the model can be improved, because data from different modalities may show different sensitivities under different physiological states.

[0195] Preferably, the health risk assessment model is:

[0196]

[0197] Among them, UZZq(YCCq=1|FVUE) is the probability of health risk occurrence under given FVUE; w0 is the intercept term; w1, w2, …, w s is the weight coefficient corresponding to each dimension of FVUE; FVUE1, FVUE2, ..., FVUE s For each element of FVUE.

[0198] Preferably, the training of the health risk assessment model includes: obtaining historical physiological health data of the operator; extracting historical feature vectors corresponding to the historical physiological health data, and training a preset neural network model according to a preset loss function, taking the historical feature vectors as input and the probability of occurrence of health risks corresponding to the historical physiological health data as output, to obtain the health risk assessment model; wherein the loss function is:

[0199]

[0200] Among them, RT g is the number of samples of training data; ycq cap is the true health risk label of the cap-th training sample; FVUE capis the integrated feature vector corresponding to the cap-th training sample; PPI is the probability; In is the natural logarithm; LTNNQ(wujj) is the loss function; YIQ is a variable in the prediction result of the health risk assessment model.

[0201] 4. Use big data and machine learning algorithms to build a health risk assessment model to evaluate the probability of health risks occurring to workers.

[0202] Furthermore, the construction process of building a health risk assessment model using big data and machine learning algorithms is as follows:

[0203] Integrate the concatenated joint feature vectors into a comprehensive feature vector FVUE:

[0204] FVUE=[FQyy caross ];

[0205] The health risk assessment model was constructed using logistic regression algorithm:

[0206]

[0207] Where UZZq(YCCq=1|FVUE) represents the probability of health risk occurrence under a given FVUE, w0 represents the intercept term, w1, w2, …, w s Represents the weight coefficients corresponding to each dimension of FVUE, FVUE1, FVUE2, ..., FVUE s Represents the various elements of FVUE;

[0208] Training health risk assessment models through historical multi-source heterogeneous data:

[0209]

[0210] Where RT g Indicates the number of samples of training data, ycq cap represents the true health risk label of the cap-th training sample, FVUE cap represents the integrated feature vector corresponding to the cap-th training sample, PPI represents probability, In represents natural logarithm, LTNNQ(wujj) represents loss function, and YIQ represents a variable of the prediction result of the health risk assessment model;

[0211] The gradient descent method is used to update the weight parameters so that the loss function LTNNQ(wujj) reaches the minimum value. The formula is as follows:

[0212]

[0213] In the formula, wujj jsxis the jsx-th weight parameter, Represents the learning rate, which is used to control the step size of each update. Represents the weight parameter wujj of the loss function LTNNQ(wujj) jsx The partial derivative of

[0214] After the health risk assessment model is trained, the real-time multi-source heterogeneous data is extracted through the above steps to obtain the feature vector FVUE new ;

[0215] FVUE new Bring it into the trained health risk assessment model and calculate the probability of health risk of the operator as UZZq(YCCq=1|FVUE new ).

[0216] S3. Compare the probability with a preset health risk probability threshold, and then determine the health status of the operator based on the comparison result. When the health status of the operator is in a dangerous state, generate a corresponding health warning signal.

[0217] Preferably, the probability is compared with a preset health risk probability threshold, and then the health status of the operator is determined based on the comparison result, and when the health status of the operator is in a dangerous state, a corresponding health warning signal is generated, including: comparing the probability with a preset health risk probability threshold, when the probability is equal to the health risk probability threshold, the health status of the operator is determined to be a normal state, when the probability is not equal to the health risk probability threshold, the health status of the operator is determined to be a risky state, and a corresponding health warning signal is generated.

[0218] 5. When the health risk of the operators reaches a certain threshold, a health warning is triggered.

[0219] Set the health risk probability threshold to

[0220] if The health of the operators is judged to be at high risk and a Level 1 warning is issued;

[0221] if The health of the operator is judged to be normal and no warning is issued;

[0222] if The health of the operators is judged to be at low risk and a Level 2 warning is issued;

[0223] When the first-level warning is triggered, the smart IoT badge chip will sound an alarm, and the alarm will sound continuously to remind the operator;

[0224] During the second-level warning, the smart IoT badge chip will sound an alarm, and the alarm will intermittently remind the operating personnel.

[0225] It can be seen that the present invention provides a method for assessing the health of personnel in a high-risk working environment, and the present invention can achieve the following beneficial effects:

[0226] Through adaptive sampling frequency technology, the data collection frequency can be flexibly adjusted according to the actual activity intensity of the operators and environmental changes. When the operators perform high-intensity physical labor or the environmental parameters change dramatically, the frequency can be automatically increased to ensure more accurate real-time data, providing a reliable basis for subsequent analysis and evaluation. This helps to promptly detect changes in the physiological indicators of operators under special circumstances, improve the ability to monitor potential risks, and appropriately reduce the sampling frequency under relatively stable conditions, which not only saves energy, but also reasonably utilizes data storage space, while also extending the service life of the smart IoT badge chip.

[0227] Through targeted processing of different types of real-time multi-source heterogeneous data, such as using an adaptive filtering algorithm based on wavelet transform to remove noise interference from ECG data, the accuracy and readability of ECG data can be improved, providing purer data for subsequent heart health analysis. Fractal dimension analysis and baseline calibration of skin electrical data can help reveal the complexity of its nonlinear dynamic system and make it more consistent with the actual physiological situation of the human body, better reflecting the changes in the physiological state of operators. The electromyography data denoising technology based on blind source separation can extract pure electromyography data, providing an accurate basis for the assessment of muscle fatigue, injury, etc. These processing methods make various types of physiological data more valuable and lay a solid foundation for subsequent comprehensive analysis and risk assessment.

[0228] A multimodal learning framework is introduced to integrate different types of physiological signals such as electrocardiogram, electrodermal signal, and electromyography to extract cross-modal features. This integration can comprehensively and comprehensively reflect the physical state of the workers, overcoming the limitations of single physiological signal analysis. Through the extraction of cross-modal features, the correlation and synergy between different physiological signals can be captured, and a deeper understanding of the physical adaptation and potential health problems of workers in high-risk working environments can be achieved, providing richer and more accurate information for health risk assessment.

[0229] The health risk assessment model constructed by using big data and machine learning algorithms can analyze and evaluate the physiological data of workers in real time. By learning and analyzing a large amount of multi-source heterogeneous data, the model can accurately identify the changing trends and potential risks of the workers' health status. This real-time monitoring and evaluation capability enables on-site workers to understand their health status in a timely manner so that they can take appropriate measures before risks occur to avoid accidents, thereby achieving the effect of timely response and ensuring the life safety of workers in high-risk working environments. When the health risk of workers reaches a certain threshold, a health warning can be triggered in time. Timely health warnings can effectively prevent further deterioration of health problems, reduce work errors and accident risks due to health reasons, and improve the overall safety and work efficiency of high-risk working environments.

[0230] Embodiment 2

[0231] Please refer to Figure 2 , is a structural schematic diagram of a personnel health assessment device in a high-risk working environment provided by an embodiment of the present invention, the device comprising: a real-time physiological health data acquisition module, a health risk assessment module and a health warning module;

[0232] The real-time physiological health data acquisition module is used to collect real-time physiological health data of workers in high-risk working environments; wherein the real-time physiological health data includes: electrocardiogram data, electrodermal data and electromyographic data;

[0233] The health risk assessment module is used to extract the feature vector corresponding to the real-time physiological health data, and input the feature vector into a preset health risk assessment model, so that the health risk assessment model assesses the probability of the occurrence of the health risk of the operator according to the feature vector, and outputs the probability of the occurrence of the health risk of the operator; wherein the health risk assessment model is obtained by training a preset neural network model with the historical physiological health data of the operator as input and the probability of the occurrence of the health risk corresponding to the historical physiological health data as output;

[0234] The health warning module is used to compare the probability with a preset health risk probability threshold, and then determine the health status of the operator based on the comparison result. When the health status of the operator is in a dangerous state, a corresponding health warning signal is generated.

[0235] It should be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art may understand and implement it without paying any creative effort.

[0236] Those skilled in the art can clearly understand that for the sake of convenience and brevity, the specific working process of the device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0237] Embodiment 3

[0238] Accordingly, an embodiment of the present invention provides an electronic device, comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the method for assessing personnel health in a high-risk working environment as described in the above-mentioned embodiment of the invention is implemented.

[0239] The electronic device may be a computing device such as a desktop computer, a notebook, a palm computer, a cloud server, etc. The device may include, but is not limited to, a processor and a memory.

[0240] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the device, and various interfaces and lines are used to connect various parts of the entire device.

[0241] Embodiment 4

[0242] Accordingly, an embodiment of the present invention provides a storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute the personnel health assessment method in a high-risk working environment described in the above-mentioned embodiment of the invention.

[0243] The memory can be used to store the computer program, and the processor realizes various functions of the device by running or executing the computer program stored in the memory and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function, etc.; the data storage area can store data created according to the use of the mobile phone, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0244] The storage medium is a computer-readable storage medium, and the computer program is stored in the computer-readable storage medium. When the computer program is executed by the processor, the steps of each method embodiment described above can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0245] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for assessing personnel health in a high-risk working environment, characterized in that: include: Collecting real-time physiological health data of workers in high-risk working environments; wherein the real-time physiological health data includes: electrocardiogram data, electrodermal data and electromyographic data; Extracting a feature vector corresponding to the real-time physiological health data, and inputting the feature vector into a preset health risk assessment model, so that the health risk assessment model assesses the probability of occurrence of health risks for the operator according to the feature vector, and outputs the probability of occurrence of health risks for the operator; The probability is compared with a preset health risk probability threshold, and then the health status of the operator is determined according to the comparison result, and when the health status of the operator is in a dangerous state, a corresponding health warning signal is generated; The health risk assessment model is obtained by training a preset neural network model with the historical physiological health data of the operator as input and the probability of occurrence of health risks corresponding to the historical physiological health data as output.

2. The personnel health assessment method in a high-risk working environment as claimed in claim 1, characterized in that: The real-time physiological health data of workers in high-risk working environments are collected, including: Obtain real-time activity intensity indicators and real-time environmental change indicators of operators; Calculate a corresponding data collection frequency according to the real-time activity intensity index and the real-time environmental change index, and then collect real-time physiological health data of workers in a high-risk working environment according to the data collection frequency; The calculation formula of the data collection frequency is: f(t)=f0×g(A(t))×h(E(t)); g(A(t))=1+k1A(t); h(E(t))=1+k2E(t); Among them, f(t) is the data collection frequency at time t; f0 is the initial collection rate; g(A(t)) is the function of activity intensity; k1 is the adjustment coefficient used to control the influence of activity intensity on the collection frequency; A(t) is the activity intensity index of the operator at time t; h(E(t)) is the function of environmental changes; k2 is the adjustment coefficient used to control the influence of environmental changes on the collection frequency; E(t) is the environmental change index of the operator at time t.

3. The personnel health assessment method in a high-risk working environment as claimed in claim 1, characterized in that: Before extracting the feature vector corresponding to the real-time physiological health data, the method further includes: Performing wavelet transform and filtering on the electrocardiogram data; Acquire a baseline value of skin electrical data of a human body in a calm state, and then perform a baseline calibration on the skin electrical data according to the baseline value; Decomposing the electromyographic data into a plurality of independent component signals, and performing denoising processing on each of the independent component signals to obtain denoised electromyographic data; Data normalization was performed on the ECG data after wavelet transform and filtering, the electrodermal data after baseline calibration, and the electromyographic data after denoising.

4. The personnel health assessment method in a high-risk working environment as claimed in claim 1, characterized in that: The extracting the feature vector corresponding to the real-time physiological health data includes: Respectively extracting electrocardiogram data features of the electrocardiogram data, electrodermal electrodermal data features of the electrodermal electrodermal data, and electromyographic electrodermal data features of the electromyographic electrodermal data; The electrocardiogram data features, the electrodermal data features and the electromyography data features are concatenated to obtain a corresponding joint feature vector, and cross-modal features of each data feature in the joint feature vector are extracted; The cross-modal features of each data feature in the joint feature vector are integrated to obtain a corresponding comprehensive feature vector, and then the comprehensive feature vector is used as the feature vector corresponding to the real-time physiological health data.

5. The personnel health assessment method in a high-risk working environment as claimed in claim 1, characterized in that: The health risk assessment model is: Among them, UZZq(YCCq=1|FVUE) is the probability of health risk occurrence under given FVUE; w0 is the intercept term; w1, w2, …, w s is the weight coefficient corresponding to each dimension of FVUE; FVUE1, FVUE2, ..., FVUE s For each element of FVUE.

6. The method for assessing personnel health in a high-risk working environment as claimed in claim 5, characterized in that: The training of the health risk assessment model includes: Obtain historical physiological health data of operators; Extracting a historical feature vector corresponding to the historical physiological health data, and training a preset neural network model based on a preset loss function with the historical feature vector as input and the probability of occurrence of the health risk corresponding to the historical physiological health data as output, to obtain the health risk assessment model; Wherein, the loss function is: Among them, RT g is the number of samples of training data; ycq cap is the true health risk label of the cap-th training sample; FVUE cap is the integrated feature vector corresponding to the cap-th training sample; PPI is the probability; In is the natural logarithm; LTNNQ(wujj) is the loss function; YIQ is a variable in the prediction result of the health risk assessment model.

7. The personnel health assessment method in a high-risk working environment as claimed in claim 1, characterized in that: The probability is compared with a preset health risk probability threshold, and then the health status of the operator is determined according to the comparison result. When the health status of the operator is in a dangerous state, a corresponding health warning signal is generated, including: The probability is compared with a preset health risk probability threshold. When the probability is equal to the health risk probability threshold, the health status of the operator is determined to be normal. When the probability is not equal to the health risk probability threshold, the health status of the operator is determined to be a risky state, and a corresponding health warning signal is generated.

8. A device for assessing personnel health in a high-risk working environment, characterized in that: include: Real-time physiological health data acquisition module, health risk assessment module and health warning module; The real-time physiological health data acquisition module is used to collect real-time physiological health data of workers in high-risk working environments; wherein the real-time physiological health data includes: electrocardiogram data, electrodermal data and electromyographic data; The health risk assessment module is used to extract the feature vector corresponding to the real-time physiological health data, and input the feature vector into a preset health risk assessment model, so that the health risk assessment model assesses the probability of the occurrence of the health risk of the operator according to the feature vector, and outputs the probability of the occurrence of the health risk of the operator; wherein the health risk assessment model is obtained by training a preset neural network model with the historical physiological health data of the operator as input and the probability of the occurrence of the health risk corresponding to the historical physiological health data as output; The health warning module is used to compare the probability with a preset health risk probability threshold, and then determine the health status of the operator based on the comparison result. When the health status of the operator is in a dangerous state, a corresponding health warning signal is generated.

9. An electronic device, characterized in that: It comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the personnel health assessment method in a high-risk working environment as described in any one of claims 1 to 7 is implemented.

10. A storage medium, characterized in that: The storage medium includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute the personnel health assessment method in a high-risk working environment as described in any one of claims 1 to 7.

Images