Data processing method for team accounting safety management system

Abstract

The invention relates to the field of data processing, in particular to a data processing method for a team accounting safety management system, which comprises the following steps: acquiring original data of employees from the team accounting safety management system; calculating a vigilance index, a facial feature index, a health index and a safety index of the employee according to the original data of the employee; weight parameters are attached to the vigilance index, the facial feature index, the health index and the safety index, and weighted summation is carried out to calculate the state value of the employee; inputting the state value into a correlation model of the state value and the total output for learning to obtain a function relationship between the state value and the total output; and predicting the overall state of the employee according to the function relationship between the state value and the total output, taking optimization measures, and adjusting the method for calculating the state value. According to the method, each index is converted into the index to calculate the employee state value, and the association model of the state value and the total yield is established, so that the safety risk is effectively reduced, and the production efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to a data processing method for a team accounting safety management system. Background Art

[0002] In the past, the way of data processing was often to directly perform simple calculations on the collected data and then make various intuitive charts, while the relationship and prediction between the data could only rely on experience and guesswork.

[0003] Nowadays, with the popularization of big data technology, enterprises' processing and use of data have also moved away from traditional manual calculations and turned to big data analysis.

[0004] Nowadays, many companies have used big data analysis to establish management systems, which has greatly reduced the workload of employees. However, they are still limited to the integration and intuitive display of data, and have not further processed the data. Management decisions also rely more on intuition and experience, resulting in the management system still having problems such as inadequate management, inaccurate assessment, and untimely rectification during actual operation, making it difficult to discover and rectify some safety hazards in a timely manner.

[0005] The patent document with Chinese patent application number CN113869706A discloses a production performance index accounting method based on MES, including: maintaining the basic working time data of products measured by the production line to the standard working time maintenance database of the production information management system, calculating the standard working time according to the basic working time data of the products, and saving the standard working time of the products to the standard working time database; counting the product scanning records of the production information management system, obtaining the number of output products, and calculating the product output working time according to the number of products and the standard working time of the products; monitoring the time difference of the adjacent product scanning records of the production information management system, if the time difference exceeds the set time threshold, generating a line stop maintenance task and calculating the lost working time according to the time difference and time threshold in the line stop state; connecting to the attendance system, setting up employee teams and binding the employee teams with the production lines, counting the employee information and attendance time of each team, and calculating the attendance working time of the team according to the employee information and attendance time; and calculating the overall factory efficiency, unit man-hour capacity and comprehensive production efficiency according to the output working time, product quantity, lost working time and attendance working time.

[0006] This shows the following problem: In the existing technology, it is impossible to predict future production based on employee information, which makes it difficult to adjust output efficiency. Summary of the invention

[0007] To this end, the present invention provides a data processing method for a team accounting safety management system, which is used to overcome the problem in the prior art that future production cannot be predicted based on employee information, resulting in difficulty in adjusting output efficiency, based on feedback calculation of an association model.

[0008] To achieve the above object, the present invention provides a data processing method for a team accounting safety management system, comprising:

[0009] Step S1, obtaining the original data of employees from the team accounting safety management system, which includes questionnaire scores, facial images, lung capacity, body temperature, blood alcohol content, and the number of safety items carried by employees;

[0010] Step S2, calculating the employee's alertness index, facial feature index, health index and safety index based on the employee's original data;

[0011] Step S3, adding weight parameters to the alertness index, the facial feature index, the health index and the safety index, and performing weighted summation to calculate the employee's status value;

[0012] Step S4, inputting the state value into a correlation model between the state value and the total output for learning, and obtaining a functional relationship between the state value and the total output;

[0013] Step S5, predicting the overall status of the employees and taking optimization measures according to the functional relationship between the status value and the total output, and adjusting the method for calculating the status value.

[0014] Furthermore, in the step S2, it includes:

[0015] Step S21, calculating the employee's vigilance index according to the questionnaire score;

[0016] Step S22, extracting features from the facial image to calculate the employee's facial feature index;

[0017] Step S23, determining the employee's health index according to the vital capacity, body temperature and blood alcohol content;

[0018] Step S24, determining the employee's safety index according to the number of safety items carried.

[0019] Furthermore, in step S21, the vigilance index is calculated by the scores of different categories of questions in the questionnaire scores, the questions in the questionnaire include psychological problems and safety knowledge problems, and the scores of employees on these two types of problems are respectively called psychological problem scores and safety knowledge problem scores. The vigilance index is calculated based on the psychological problem scores and safety knowledge problem scores.

[0020] Furthermore, in step S22, the facial feature index is calculated as follows:

[0021] Extracting feature values ​​and feature vectors from the facial image, comparing the feature values ​​and feature vectors with the facial image of the corresponding employee stored in the security management system to obtain facial expression parameters of the employee, and then obtaining facial features corresponding to the employee according to the facial expression parameters of the employee;

[0022] Searching for a facial expression reference that is highly similar to the eigenvalue, the eigenvector, and the corresponding facial image in the security management system, and obtaining facial feature parameters corresponding to the facial expression reference;

[0023] The facial feature index is obtained according to the facial feature parameters, wherein the facial feature index is positively correlated with the corresponding facial feature parameters.

[0024] Furthermore, in step S23, the value of the health index is calculated by lung capacity, body temperature, and blood alcohol content. The health index is positively correlated with lung capacity, and negatively correlated with the square difference between body temperature and standard human temperature, and blood alcohol content.

[0025] Furthermore, in step S24, the method for determining the safety index is:

[0026] The number of safety items carried by employees is determined based on data provided by the safety management system, which is set with a rated carrying quantity. The safety index is positively correlated with the number of safety items. When the number of safety items exceeds the rated carrying quantity, the safety index does not change even if the number of safety items continues to increase.

[0027] Furthermore, in step S3, the state value is obtained by weighted summation of the alertness index, the facial feature index, the health index and the safety index.

[0028] Furthermore, in the step S4, it includes:

[0029] S41, inputting the calculated state value into the association model;

[0030] S42, using reinforcement learning to obtain a functional relationship between the state value and the total output, wherein:

[0031] Reinforcement learning is that the safety management system continuously receives raw data and processes the raw data. As the data processing proceeds, the total output and the state value are continuously updated according to the actual situation. The total output after reaching the new state is used as a standard, and the learning rate in the function is adjusted according to the standard. The optimal learning rate is obtained through continuous iteration, and then the state value of the new state is introduced into the function of the state value and the total output for calculation, and the discount factor is continuously adjusted to obtain the functional relationship between the state value and the total output that predicts the trend of total output changes;

[0032] S43, predicting the change trend of the total output according to the state value.

[0033] Furthermore, in the step S5, it includes:

[0034] S51, taking corresponding optimization measures according to the values ​​of the alertness index, the facial feature index, the health index and the safety index of the calculated state value to improve the state value;

[0035] S52, adjusting the weight parameter corresponding to each index when calculating the state value according to the degree of change of different indexes and state values ​​after the optimization measures are taken.

[0036] Furthermore, in step S52, when the degree of change of a certain index after taking the optimization measures is less than a preset degree of change threshold, the weight parameter corresponding to it when calculating the state value is increased, and the weight parameter of the index with the largest degree of change is reduced.

[0037] Compared with the prior art, the beneficial effect of the present invention lies in that, by converting various indicators into indexes to calculate employee status values, a quantitative evaluation of employee status is achieved, the adaptability and accuracy of the evaluation are improved, and management personnel can intuitively grasp the employee status. By establishing a correlation model between status values ​​and total output, not only can the employee status be evaluated, but also its potential impact on production efficiency can be predicted, providing a scientific basis for management decisions. By predicting the overall status of employees and taking optimization measures in a timely manner, a management mode transformation from passive response to active prevention is achieved, which effectively reduces safety risks and improves production efficiency.

[0038] Furthermore, by subdividing employee status into four indexes, a comprehensive assessment of the overall status of employees is achieved, avoiding the one-sidedness that may be caused by a single indicator, and greatly improving the objectivity and accuracy of the assessment.

[0039] Furthermore, by calculating the psychological problem score and the safety knowledge problem score separately and combining these two scores to derive the vigilance index, the assessment results are more accurate and reliable, and can better reflect the actual status of employees, so as to track the changing trends of employee vigilance, timely discover and solve potential problems, and thus formulate more effective management strategies.

[0040] Furthermore, by using facial images to capture changes in employees' facial features in real time, fatigue, stress or other adverse conditions can be detected in a timely manner, which helps to quickly intervene and prevent safety hazards. By extracting and comparing eigenvalues ​​and eigenvectors, a refined analysis of facial expressions is achieved, which can capture subtle changes in expressions and thus more accurately reflect facial features. Qualitative facial expression analysis is converted into a quantitative facial feature index, which makes it easier for managers to intuitively understand and use the evaluation results.

[0041] Furthermore, by combining vital capacity, body temperature and blood alcohol content to obtain a health index, a multi-angle assessment of the health status of employees is achieved, providing objective and quantifiable health status assessment results, which helps managers quickly grasp the overall health status of employees, identify potential health risks, and take timely preventive measures.

[0042] Furthermore, by comparing the actual number of safety items carried with the rated number of items carried, the precise match between the assessment criteria and actual work needs is ensured.

[0043] Furthermore, by integrating four key indexes, a multi-dimensional and comprehensive evaluation of employee status is achieved. The weighted sum method allows the weight of each index to be flexibly adjusted according to different working environments, job requirements or management priorities, making the evaluation more adaptive and targeted. It integrates multiple qualitative and quantitative indicators into a single status value, which is convenient for managers to grasp the overall status of employees intuitively and quickly.

[0044] Furthermore, by continuously receiving and processing new raw data, it can automatically adapt to the ever-changing working environment and production conditions, maintain the timeliness and accuracy of the model, and by establishing a functional relationship between state values ​​and total output, the system can predict future production trends based on current state values, and provide forward-looking guidance for production decisions. By using reinforcement learning algorithms, it can dynamically adjust the learning rate and discount factor according to actual production conditions, continuously optimize the functional relationship between state values ​​and total output, and improve the accuracy of predictions.

[0045] Furthermore, according to the specific performance of the vigilance index, facial feature index, health index and safety index, targeted optimization measures are implemented to ensure the accuracy and effectiveness of the intervention. According to the changes in each index and status value after the optimization measures, the weight parameters are dynamically adjusted so that the evaluation system can continuously optimize itself and improve its accuracy and adaptability.

[0046] Furthermore, by increasing the weight of the index of insignificant changes, the evaluation focus can be automatically adjusted according to the actual improvement effect, thereby improving the flexibility and adaptability of the evaluation system, facilitating a more reasonable allocation of resources, investing more resources in areas where the improvement effect is not obvious, improving resource utilization efficiency, and helping to discover and resolve potential problems early. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a flow chart of a data processing method for a team accounting safety management system according to an embodiment of the present invention;

[0048] Figure 2 A flowchart of calculating an index according to an embodiment of the present invention;

[0049] Figure 3 A flow chart of obtaining a functional relationship for an embodiment of the present invention;

[0050] Figure 4 The flowchart of the adjustment calculation is shown in the embodiment of the present invention. DETAILED DESCRIPTION

[0051] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0052] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0053] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0054] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" 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 it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0055] See also Figure 1 ,like Figure 1As shown, it is a flow chart of a data processing method for a team accounting safety management system according to an embodiment of the present invention;

[0056] Specifically, an embodiment of the present invention provides a data processing method for a team accounting safety management system, including:

[0057] Step S1, obtaining the original data of employees from the team accounting safety management system, which includes questionnaire scores, facial images, lung capacity, body temperature, blood alcohol content, and the number of safety items carried by employees;

[0058] Step S2, calculating the employee's alertness index, facial feature index, health index and safety index based on the employee's original data;

[0059] Step S3, adding weight parameters to the alertness index, the facial feature index, the health index and the safety index, and performing weighted summation to calculate the employee's status value;

[0060] Step S4, inputting the state value into a correlation model between the state value and the total output for learning, and obtaining a functional relationship between the state value and the total output;

[0061] Step S5, predicting the overall status of the employees and taking optimization measures according to the functional relationship between the status value and the total output, and adjusting the method for calculating the status value.

[0062] Specifically, the questionnaire is a series of multiple-choice questions regularly distributed by the safety management system to all employees. The multiple-choice questions are divided into psychological problems and safety knowledge problems. The psychological problems are based on the Hamilton Depression Rating Scale and work content, and the safety knowledge comes from the work-related safety knowledge question bank. They are used to examine the mental health of employees and their emphasis on safety, respectively, to avoid accidents caused by mental illness and disregard for safety. The safety management system obtains the facial image of the employee through face recognition, realizes the basic function of identifying identity and is used to identify the facial expression of the employee to judge the current facial features of the employee. The disease that affects the working status is characterized by large temperature fluctuations and affecting respiratory function. By testing the body temperature and lung capacity of employees, employees with common diseases are prevented from working, avoiding accidents among employees. The questionnaire scores, facial images, lung capacity, body temperature, blood alcohol content, and the number and category of items carried by employees cover the areas of employees' mental health, mental health, physical health, and safety preparation related to safety, respectively, to ensure the safety of employees to the greatest extent.

[0063] Specifically, by converting various indicators into indexes to calculate employee status values, a quantitative assessment of employee status is achieved, the adaptability and accuracy of the assessment is improved, and it is convenient for managers to intuitively grasp the employee status. By establishing a correlation model between status values ​​and total output, not only can the employee status be evaluated, but also its potential impact on production efficiency can be predicted, providing a scientific basis for management decisions. By predicting the overall status of employees and taking timely optimization measures, a management model transformation from passive response to active prevention is achieved, which effectively reduces safety risks and improves production efficiency.

[0064] Please continue reading Figure 2 ,like Figure 2 As shown, it is a flow chart of calculating the index according to an embodiment of the present invention;

[0065] Specifically, the step S2 includes:

[0066] Step S21, calculating the employee's vigilance index according to the questionnaire score;

[0067] Step S22, extracting features from the facial image to calculate the employee's facial feature index;

[0068] Step S23, determining the employee's health index according to the vital capacity, body temperature and blood alcohol content;

[0069] Step S24, determining the employee's safety index according to the number of safety items carried.

[0070] Specifically, the measurement of body temperature, vital capacity and blood alcohol content can screen out employees who are in abnormal physical conditions. The body temperature is quickly measured by an electronic thermometer, and the vital capacity and alcohol content are measured simultaneously by a special two-in-one gas detector. When employees engage in dangerous work, they need to carry safety items that meet the requirements to avoid accidents. Before employees enter the workplace, the types and quantities of safety items must be checked and registered. For employees with too few items, they will be provided with additional safety items or restricted from entering the workplace.

[0071] Specifically, by subdividing employee status into four indexes, a comprehensive assessment of the overall status of employees is achieved, avoiding the one-sidedness that may be caused by a single indicator, and greatly improving the objectivity and accuracy of the assessment.

[0072] Specifically, in step S21, the vigilance index is calculated based on the scores of different categories of questions in the questionnaire. The questions in the questionnaire include psychological problems and safety knowledge problems. The scores of employees on these two types of problems are respectively called psychological problem scores and safety knowledge problem scores. The vigilance index is calculated based on the psychological problem scores and safety knowledge problem scores.

[0073] Specifically, the employee's vigilance index is calculated from the scores of psychological problems and safety knowledge. The proportion of the two problems in the test paper is adjusted according to actual needs. At the same time, the total score of the questionnaire remains unchanged. The scores of employees on these two problems are called work safety knowledge scores and psychological problem scores respectively. The formula for calculating the vigilance index is:

[0074] η1=α α / (α+β) +β β / (α+β)

[0075] Among them, α is the work safety knowledge score, β is the psychological problem score, and η1 is the vigilance index.

[0076] During the specific implementation process, if an employee’s work safety knowledge score is 50 and his psychological problem score is 50, then according to the formula for calculating the vigilance index, the vigilance index is 14.

[0077] Specifically, by calculating the psychological problem score and the safety knowledge problem score separately and combining these two scores to derive the vigilance index, the assessment results are more accurate and reliable, and can better reflect the actual status of employees, thereby tracking the changing trends of employee vigilance, promptly discovering and resolving potential problems, and formulating more effective management strategies.

[0078] Specifically, in step S22, the facial feature index is calculated as follows:

[0079] Extracting feature values ​​and feature vectors from the facial image, comparing the feature values ​​and feature vectors with the facial image of the corresponding employee stored in the security management system to obtain facial expression parameters of the employee, and then obtaining facial features corresponding to the employee according to the facial expression parameters of the employee;

[0080] Searching for a facial expression reference that is highly similar to the eigenvalue, the eigenvector, and the corresponding facial image in the security management system, and obtaining facial feature parameters corresponding to the facial expression reference;

[0081] The facial feature index is obtained according to the facial feature parameters, wherein the facial feature index is positively correlated with the corresponding facial feature parameters.

[0082] Specifically, the facial feature index of the corresponding employee is obtained from the processing of the facial image. The facial features contained in the face image contain the corresponding facial identity information and expression. By analyzing the image, the identity information and expression of the specific person are obtained. The facial image obtained by the processed safety management system is compared with the standard facial image to analyze the mood of the employee in the recent period and conduct targeted counseling for the employee according to the actual situation of the employee. The feature extraction method is used to extract the eigenvalue and eigenvector from the facial image. After comparing the face database and obtaining the corresponding identity information, the eigenvalue and eigenvector are compared with the standard face recognition library stored in the team accounting safety management system to obtain the employee's facial expression corresponding to the employee's facial image. The facial features and mood corresponding to the employee's facial image are inferred based on the employee's facial expression. The feature extraction formula is:

[0083]

[0084] c=(u*u) / (n-1)

[0085] c*v=θ*v

[0086] Where u is the data matrix, n is the number of pixels in the image, a, b, c are weight parameters, R is the proportion of red in the color of a pixel, G is the proportion of green in the color of a pixel, B is the proportion of blue in the color of a pixel, θ is the eigenvalue, and v is the eigenvector;

[0087] Find the facial expression reference with the corresponding eigenvalue and eigenvector that are highly similar to the facial image, and obtain the facial features and facial feature parameters corresponding to the facial expression reference; the facial feature index is positively correlated with the corresponding facial feature parameters, and the formula is:

[0088] η2=m-δ

[0089] Where η2 is the facial feature index, δ is the facial feature parameter, and m is the correlation coefficient.

[0090] In the specific implementation process, the facial feature parameter is 40, the correlation coefficient is 0.5, and the facial feature index is calculated to be 20.

[0091] The recognition and retention of employee information ensures that the various indexes of specific employees can be queried when improving the status value later. The recognition and quantification of facial expressions provide a facial feature index with strong correlation for the calculation of the status value without taking up almost any additional resources.

[0092] Specifically, by using facial images to capture changes in employees' facial features in real time, fatigue, stress or other adverse conditions can be detected in a timely manner, which helps to quickly intervene and prevent safety hazards. By extracting and comparing eigenvalues ​​and eigenvectors, a refined analysis of facial expressions is achieved, which can capture subtle changes in expressions and thus more accurately reflect facial features. Qualitative facial expression analysis is converted into a quantitative facial feature index, which makes it easier for managers to intuitively understand and use the evaluation results.

[0093] Specifically, in step S23, the value of the health index is calculated by vital capacity, body temperature, and blood alcohol content. The health index is positively correlated with vital capacity, and negatively correlated with the square difference between body temperature and standard human temperature, and blood alcohol content.

[0094] Specifically, vital capacity, body temperature, and blood alcohol content are used as indexes to reflect the physical condition of employees. The health index is calculated to obtain the health index, which reflects the overall physical condition of employees. The health index is proportional to the degree of physical health. The lung capacity of the human body is positively correlated with the health level of the human body. The surface temperature of a healthy human body is maintained at a standard temperature of about 36.5 degrees Celsius at room temperature. The blood alcohol content is at a low level in a human body that does not consume alcohol. The health index is proportional to the lung capacity, inversely proportional to the square difference between the body temperature and the standard temperature, and inversely proportional to the blood alcohol content.

[0095] The formula for calculating the health index is:

[0096]

[0097] Among them, l is vital capacity, t is body temperature, h is blood alcohol content, λ is standard vital capacity, μ is standard human body surface temperature, ο is the maximum blood alcohol content of humans without drinking, and η3 is the health index.

[0098] In the specific implementation process, the standard vital capacity is 4000, the standard human body surface temperature is 37, the highest blood alcohol content of humans without drinking is 0.01, and an employee's vital capacity is 3000, body temperature is 38, and blood alcohol content is 0.05. The health index calculated by the health index formula is -1.29.

[0099] Specifically, by combining lung capacity, body temperature and blood alcohol content to obtain a health index, a multi-angle assessment of employee health status is achieved, providing objective and quantifiable health status assessment results, which helps managers quickly grasp the overall health status of employees, identify potential health risks, and take timely preventive measures.

[0100] Specifically, in step S24, the method for determining the safety index is:

[0101] The number of safety items carried by employees is determined based on data provided by the safety management system, which is set with a rated carrying quantity. The safety index is positively correlated with the number of safety items. When the number of safety items exceeds the rated carrying quantity, the safety index does not change even if the number of safety items continues to increase.

[0102] Specifically, the safety index is positively correlated with the number of safety items. When the number of safety items exceeds the rated carrying amount, a marginal effect will occur, and the safety index will not change even if the number continues to increase.

[0103] The calculation formula of the safety index is:

[0104] η4=q-(1-p / g)

[0105] Among them, q is the safety factor, p is the rated number of safety items carried, g is the actual number of safety items carried, and η4 is the safety index.

[0106] In the specific implementation process, the safety factor is 100, the rated number of safety items carried is 5, and the actual number of safety items carried by an employee is 3, then the safety index of the employee is 40.

[0107] Specifically, by comparing the actual number of safety items carried with the rated number of items carried, the assessment criteria are accurately matched to actual work needs.

[0108] Specifically, in step S3, the state value is obtained by weighted summation of the alertness index, the facial feature index, the health index and the safety index.

[0109] Specifically, the four indexes can reflect the status of employees in four aspects. The status value is intuitive and linearly correlated with the four data. It reflects the overall status of employees. The status value is obtained by weighted summation of the vigilance index, facial feature index, health index and safety index. The calculation formula of the status value is:

[0110] s=ε1-η1+ε2*η2+ε3-η3+ε4*η4

[0111] Among them, s is the state value, η1 represents the vigilance index, η2 represents the facial feature index, η3 represents the health index, and η4 represents the safety index; ε1 is the vigilance index weight parameter, ε2 is the facial feature index weight parameter, ε3 is the health index weight parameter, and ε4 is the safety index weight parameter.

[0112] In the specific implementation process, the vigilance index weight parameter, facial feature index weight parameter, health index weight parameter and safety index weight parameter are all 25%. An employee's vigilance index is 14, facial feature index is 20, health index is -1.29, and safety index is 40, then the status value is 18.2.

[0113] Specifically, by integrating four key indexes, a multi-dimensional and comprehensive assessment of employee status is achieved. The weighted summation method allows the weight of each index to be flexibly adjusted according to different working environments, job requirements or management priorities, making the assessment more adaptable and targeted. It integrates multiple qualitative and quantitative indicators into a single status value, making it easier for managers to grasp the overall status of employees intuitively and quickly.

[0114] Please continue reading Figure 3 ,like Figure 3 As shown, it is a flow chart of obtaining a functional relationship according to an embodiment of the present invention;

[0115] Specifically, in step S4, it includes:

[0116] S41, inputting the calculated state value into the association model;

[0117] S42, using reinforcement learning to obtain a functional relationship between the state value and the total output, wherein:

[0118] Reinforcement learning is that the safety management system continuously receives raw data and processes the raw data. As the data processing proceeds, the total output and the state value are continuously updated according to the actual situation. The total output after reaching the new state is used as a standard, and the learning rate in the function is adjusted according to the standard. The optimal learning rate is obtained through continuous iteration, and then the state value of the new state is introduced into the function of the state value and the total output for calculation, and the discount factor is continuously adjusted to obtain the functional relationship between the state value and the total output that predicts the trend of total output changes;

[0119] S43, predicting the change trend of the total output according to the state value.

[0120] Specifically, the operating principle of the association model is as follows: first, the inverse of the difference between the actual total output at the same time and the total output predicted by the function of the state value and the total output is taken as the reward for the function of the state value and the total output, and the weighted sum of the rewards in the future period is calculated to obtain the cumulative reward. The cumulative reward represents the accuracy of the output result of the function of the current state value and the total output. The closer the output result is to the actual situation, the higher the cumulative reward. The function of the state value and the total output with the most accurate output result is obtained by maximizing the cumulative reward.

[0121] The formula for calculating the cumulative reward is:

[0122] G_t=R_{t+1}+γR_{t+2}+γ^2R_{t+3}+...=∑_{k

[0123] =0}^∞γ^k R_{t+k+1}

[0124] Where G_t is the cumulative reward, R_{t} is the reward obtained at time t, and γ is the discount factor (usually a value between 0 and 1, indicating the degree of decay of future rewards).

[0125] Next, when optimizing the function of state value and total output, the updated state value function can be used to filter the optimization content, and the state value and total output functions modified according to different methods are used to calculate the maximum cumulative reward. The state value function is calculated by maximizing the cumulative reward, and the quality of the method is determined according to the output value of the state value function. The method with the highest output value is identified as the effective method, and the state value and total output function is modified according to this method;

[0126] The formula for calculating the state value function is:

[0127] vπ(s)=Eπ[G t |

[0128] Where vπ(s) is the state value function, which represents the expected cumulative reward obtained by taking actions according to strategy π starting from state s;

[0129] Finally, the Q-learning algorithm is used to feed back the change of total output to the state value, so as to maximize the expectation of total output increased by the growth of the state value, thereby establishing a functional relationship between the state value and the total output. The future total output value is predicted based on the functional relationship between the state value and the total output, and the real state value and the real total output value are used as the input and output of the function of the state value and the total output, and the parameters of the function of the state value and the total output are updated;

[0130] The method to obtain the functional relationship between the state value and the total output through the association model is:

[0131] Set the initial learning rate and change value for predicting the total output, obtain the optimal learning rate and change value through continuous iteration, and establish the relationship between the state value and the total output by fitting the functional relationship between the total output and the state value;

[0132] The formula for the functional relationship between the state value and the total output is:

[0133] Q(s)=(1-α)Q(s)+α(r+γQ(s'))

[0134] Q(s) represents the total output, s represents the state value, α is the learning rate, r is the change in total output after reaching the new state, γ is the discount factor, s' represents the new state value, and Q(s') represents the new output value;

[0135] The association model takes the function between the state value and the total output as input. After the function of the state value and the total output is optimized by the Q-learning algorithm, the optimization effect of the Q-learning algorithm is determined according to the structure by calculating the maximum cumulative reward and substituting it into the state value function.

[0136] If the optimization effect is good, the function of the current optimized state value and total output will be used as the new state value and total output function. If the effect is poor, the current state value and total output function will still be used, and the function between the state value and total output will be optimized according to the state value function. The model will then be repeatedly run until a function with good optimization effect is obtained.

[0137] Specifically, by continuously receiving and processing new raw data, it can automatically adapt to the ever-changing working environment and production conditions, maintain the timeliness and accuracy of the model, and by establishing a functional relationship between state values ​​and total output, the system can predict future production trends based on current state values, and provide forward-looking guidance for production decisions. By using reinforcement learning algorithms, it can dynamically adjust the learning rate and discount factor according to actual production conditions, continuously optimize the functional relationship between state values ​​and total output, and improve the accuracy of predictions.

[0138] Please continue reading Figure 4 ,like Figure 4 As shown, it is a flow chart of adjustment calculation according to an embodiment of the present invention;

[0139] Specifically, the step S51 includes:

[0140] S51, taking corresponding optimization measures according to the values ​​of the alertness index, the facial feature index, the health index and the safety index of the calculated state value to improve the state value;

[0141] S52, adjusting the weight parameter corresponding to each index when calculating the state value according to the degree of change of different indexes and state values ​​after the optimization measures are taken.

[0142] Specifically, after taking optimization measures, the changes in different indexes will affect the changes in state values, and then the changed predicted total output will be obtained through the association model.

[0143] Specifically, based on the specific performances of the vigilance index, facial feature index, health index and safety index, targeted optimization measures are implemented to ensure the accuracy and effectiveness of the intervention. According to the changes in each index and status value after the optimization measures, the weight parameters are dynamically adjusted so that the evaluation system can continuously optimize itself and improve its accuracy and adaptability.

[0144] Specifically, in step S52, when the change degree of a certain index after taking the optimization measures is less than the preset change degree threshold, the weight parameter corresponding to it when calculating the state value is increased, and the weight parameter of the index with the largest change degree is reduced.

[0145] Specifically, the preset change degree threshold is 20%. When the change degree of the alertness index in each index is 10% less than the change degree threshold, and the change degree of the safety index is the largest, the weight of the alertness index is increased by 10% and the weight of the safety index is reduced by 10%.

[0146] Specifically, by increasing the weight of the index of unclear changes, the evaluation focus can be automatically adjusted according to the actual improvement effect, thereby improving the flexibility and adaptability of the evaluation system, facilitating a more rational allocation of resources, investing more resources in areas where the improvement effect is not obvious, improving resource utilization efficiency, and helping to detect and resolve potential problems early.

[0147] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0148] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A data processing method for a team accounting safety management system, characterized in that: include: Step S1, obtaining the original data of employees from the team accounting safety management system, which includes questionnaire scores, facial images, lung capacity, body temperature, blood alcohol content, and the number of safety items carried by employees; Step S2, calculating the employee's alertness index, facial feature index, health index and safety index based on the employee's original data; Step S3, adding weight parameters to the alertness index, the facial feature index, the health index and the safety index, and performing weighted summation to calculate the employee's status value; Step S4, inputting the state value into a correlation model between the state value and the total output for learning, and obtaining a functional relationship between the state value and the total output; Step S5, predicting the overall status of the employees and taking optimization measures according to the functional relationship between the status value and the total output, and adjusting the method for calculating the status value.

2. The data processing method for the team accounting safety management system according to claim 1 is characterized in that: In the step S2, it includes: Step S21, calculating the employee's vigilance index according to the questionnaire score; Step S22, extracting features from the facial image to calculate the employee's facial feature index; Step S23, determining the employee's health index according to the vital capacity, body temperature and blood alcohol content; Step S24, determining the employee's safety index according to the number of safety items carried.

3. The data processing method for the team accounting safety management system according to claim 2 is characterized in that: In step S21, the vigilance index is calculated by the scores of different categories of questions in the questionnaire score. The questions in the questionnaire include psychological problems and safety knowledge problems. The scores of employees on these two types of problems are respectively psychological problem scores and safety knowledge problem scores. The vigilance index is calculated based on the psychological problem scores and safety knowledge problem scores.

4. The data processing method for the team accounting safety management system according to claim 2 is characterized in that: In step S22, the facial feature index is calculated as follows: Extracting feature values ​​and feature vectors from the facial image, comparing the feature values ​​and feature vectors with the facial image of the corresponding employee stored in the security management system to obtain facial expression parameters of the employee, and then obtaining facial features corresponding to the employee according to the facial expression parameters of the employee; Searching for a facial expression reference that is highly similar to the eigenvalue, the eigenvector, and the corresponding facial image in the security management system, and obtaining facial feature parameters corresponding to the facial expression reference; The facial feature index is obtained according to the facial feature parameters, wherein the facial feature index is positively correlated with the corresponding facial feature parameters.

5. The data processing method for the team accounting safety management system according to claim 2 is characterized in that: In step S23, the value of the health index is calculated by vital capacity, body temperature, and blood alcohol content. The health index is positively correlated with vital capacity, and negatively correlated with the square difference between body temperature and standard human temperature, and blood alcohol content.

6. The data processing method for the team accounting safety management system according to claim 2 is characterized in that: In step S24, the method for determining the safety index is: The number of safety items carried by employees is determined based on data provided by the safety management system, which is set with a rated carrying quantity. The safety index is positively correlated with the number of safety items. When the number of safety items exceeds the rated carrying quantity, the safety index does not change even if the number of safety items continues to increase.

7. The data processing method for the team accounting safety management system according to claim 1 is characterized in that: In step S3, the state value is obtained by weighted summation of the alertness index, the facial feature index, the health index and the safety index.

8. The data processing method for the team accounting safety management system according to claim 5 is characterized in that: In the step S4, it includes: S41, inputting the calculated state value into the association model; S42, using reinforcement learning to obtain a functional relationship between the state value and the total output, wherein: Reinforcement learning is that the safety management system continuously receives raw data and processes the raw data. As the data processing proceeds, the total output and the state value are continuously updated according to the actual situation. The total output after reaching the new state is used as a standard, and the learning rate in the function is adjusted according to the standard. The optimal learning rate is obtained through continuous iteration, and then the state value of the new state is introduced into the function of the state value and the total output for calculation, and the discount factor is continuously adjusted to obtain the functional relationship between the state value and the total output that predicts the trend of total output changes; S43, predicting the change trend of the total output according to the state value.

9. The data processing method for the team accounting safety management system according to claim 1, characterized in that: In the step S51, it includes: S51, taking corresponding optimization measures according to the values ​​of the alertness index, the facial feature index, the health index and the safety index of the calculated state value to improve the state value; S52, adjusting the weight parameter corresponding to each index when calculating the state value according to the degree of change of different indexes and state values ​​after the optimization measures are taken.

10. The data processing method for the team accounting safety management system according to claim 9, characterized in that: In step S52, when the change degree of a certain index is less than a preset change degree threshold after taking the optimization measures, the weight parameter corresponding to it when calculating the state value is increased, and the weight parameter of the index with the largest change degree is reduced.

Images