A body temperature detection method, device, system, equipment and readable storage medium

Abstract

The present invention discloses a body temperature detection method, device, system, equipment and readable storage medium, which are applied to the technical field of temperature measurement, and include: based on the ambient temperature, ambient humidity and body surface temperature, using a body temperature detection model based on temperature and humidity compensation for detection to obtain the first actual detected body temperature; determining the first-order differential actual detected body temperature difference corresponding to the first actual detected body temperature, and the temperature differences corresponding to the ambient temperature and the body surface temperature; based on the first-order differential actual detected body temperature difference, the temperature difference and the first actual detected body temperature, using a body temperature prediction model to obtain the second actual detected body temperature. The body temperature detection model based on temperature and humidity compensation in the present invention compensates the body surface temperature, and can obtain the actual detected body temperature closer to the actual human body temperature more quickly. Since the present application will perform the compensation most suitable for the current environment and shorten the equilibrium time based on the body temperature prediction model, the accuracy of the body temperature prediction value in the non-equilibrium state is improved while the equilibrium time is shortened.

Description

Technical Field

[0001] The present invention relates to the technical field of temperature measurement, and particularly relates to a body temperature detection method, device, system, equipment and readable storage medium. Background Art

[0002] In recent years, with the rapid development of sensor technology, wireless communication technology and mobile Internet technology, continuous electronic body temperature measurement devices have been gradually applied to multiple fields such as health management, disease prevention and assisted reproduction due to their advantages of small size, non-toxicity, convenience, intelligence and continuous temperature measurement. However, the current electronic thermometers or continuous temperature measurement devices have poor anti-interference ability, resulting in low accuracy of body temperature measurement. This may not only cause people to misjudge their own health conditions, but also delay the disease condition or make wrong decisions due to inaccurate body temperature data in key scenarios such as disease prevention and diagnosis.

[0003] The equilibration time of the existing body temperature monitoring devices in the market is relatively long, about 10 - 30 minutes. In medical diagnosis, especially for patients with suspected fever, the too long equilibration time leads to untimely diagnosis and seriously affects the user experience.

[0004] Therefore, how to improve the accuracy of body temperature measurement during the non-equilibrium state, shorten the equilibration time, avoid misdiagnosis due to inaccurate or delayed data, and provide a reliable basis for the formulation of subsequent treatment plans is a technical problem that needs to be urgently solved by those skilled in the art. Summary of the Invention

[0005] In view of this, the purpose of the present invention is to provide a body temperature detection method, device, equipment and computer-readable storage medium, which solves the technical problems of low accuracy and long equilibration time in existing body temperature detection.

[0006] To solve the above technical problems, the present invention provides a body temperature detection method, including:

[0007] Determine the closest reference temperature and reference humidity combination corresponding to the ambient temperature and ambient humidity in the current environment;

[0008] Determine a body temperature detection model based on temperature and humidity compensation based on the closest reference temperature and reference humidity combination;

[0009] Perform detection using the body temperature detection model based on temperature and humidity compensation based on the ambient temperature, the ambient humidity and the body surface temperature to obtain a first actual detected body temperature;

[0010] Determine the first-order difference actual detected body temperature difference corresponding to the first actual detected body temperature, and the temperature difference between the ambient temperature and the body surface temperature;

[0011] Using the body temperature prediction model based on the actually detected body temperature difference of the first-order difference, the temperature difference, and the first actually detected body temperature, obtain the second actually detected body temperature; the body temperature prediction model is a model determined based on the linear relationship between the first actually detected body temperature, the environmental temperature, the body surface temperature, and the second actually detected body temperature.

[0012] Optionally, the process of determining the body temperature prediction model includes:

[0013] Based on multiple second experimental environmental temperatures, multiple second experimental body surface temperatures, multiple first actually detected body temperatures of the second experiments, and multiple actual body temperatures of the second experiments, fit the body temperature prediction basic model to obtain various weight coefficients in the body temperature prediction basic model;

[0014] Construct the body temperature prediction model based on the various weight coefficients and the body temperature prediction basic model.

[0015] Optionally, based on multiple second experimental environmental temperatures, multiple second experimental body surface temperatures, multiple first actually detected body temperatures of the second experiments, and multiple actual body temperatures of the second experiments, fitting the body temperature prediction basic model to obtain various weight coefficients in the body temperature prediction basic model includes:

[0016] Determine the second experimental environmental temperature, the second experimental body surface temperature, the first actually detected body temperature of the second experiment, and the actual body temperature of the second experiment corresponding to each temperature measurement position;

[0017] Based on the second experimental environmental temperature, the second experimental body surface temperature, the first actually detected body temperature of the second experiment, and the actual body temperature of the second experiment corresponding to each temperature measurement position, fit the body temperature prediction basic model to obtain various weight coefficients corresponding to each temperature measurement position;

[0018] Correspondingly, constructing the body temperature prediction model based on the various weight coefficients and the body temperature prediction basic model includes:

[0019] Construct the body temperature prediction model corresponding to each temperature measurement position based on the various weight coefficients corresponding to each temperature measurement position and the body temperature prediction basic model.

[0020] Optionally, after obtaining the second actually detected body temperature by using the body temperature prediction model based on the actually detected body temperature difference of the first-order difference, the temperature difference, and the first actually detected body temperature, it further includes:

[0021] Based on the user data, the environmental temperature, and the environmental humidity, use the approximate search model to search and obtain a preset number of thermally balanced body temperatures; the user data is data related to the user's health;

[0022] Performing weighted processing on the preset number of thermostatic body temperatures to obtain an estimated thermostatic body temperature;

[0023] Based on the estimated thermostatic body temperature and the second actual detected body temperature, using a temperature prediction model adjusted based on the weights of similar users to determine the third actual detected body temperature; wherein, the temperature prediction model adjusted based on the weights of similar users is a model for adjusting the second actual detected body temperature based on the estimated thermostatic body temperature.

[0024] Optionally, the process of determining the approximate search model includes:

[0025] Processing the user data based on the body temperature value of the user's thermostasis to obtain processed user data;

[0026] Grouping the processed user data based on the measurement location to obtain grouped data;

[0027] Extracting features from the data in each grouped data to obtain user features and environmental features of each grouped data, and using the user features and the environmental features as target features;

[0028] Performing feature encoding on the target features to obtain feature vectors; wherein, numerical features are processed by maximum-minimum normalization, and binary categorical features are processed by one-hot encoding.

[0029] Constructing a K-D tree based on the feature vectors to obtain the approximate search model.

[0030] Optionally, performing weighted processing on the preset number of thermostatic body temperatures to obtain an estimated thermostatic body temperature includes:

[0031] Performing distance weighting on the preset number of thermostatic body temperatures to obtain the estimated thermostatic body temperature; wherein, the distance weighting is a method for determining the estimated thermostatic body temperature by weighting the body temperature values of neighbor users based on the feature distance between the current user and each neighbor user.

[0032] This application also provides a body temperature detection device, including:

[0033] The closest temperature and humidity combination determination module is used to determine the closest reference temperature and reference humidity combination corresponding to the environmental temperature and environmental humidity in the current environment;

[0034] The body temperature detection model determination module is used to determine a body temperature detection model based on temperature and humidity compensation based on the closest reference temperature and reference humidity combination;

[0035] The first actual detected body temperature determination module is configured to perform detection by using the body temperature detection model based on temperature and humidity compensation based on the environmental temperature, the environmental humidity, and the body surface temperature, so as to obtain the first actual detected body temperature;

[0036] The difference determination module is configured to determine the first-order difference actual detected body temperature difference corresponding to the first actual detected body temperature, and the temperature differences between the environmental temperature and the body surface temperature;

[0037] The second actual detected body temperature determination module is configured to obtain the second actual detected body temperature by using a body temperature prediction model based on the first-order difference actual detected body temperature difference, the temperature difference, and the first actual detected body temperature; the body temperature prediction model is a model determined based on the linear relationship between the first actual detected body temperature, the environmental temperature, the body surface temperature, and the second actual detected body temperature.

[0038] This application further provides a body temperature detection system, including:

[0039] A first temperature measurement element for detecting the environmental temperature;

[0040] A second temperature measurement element for detecting the body surface temperature;

[0041] A humidity measurement element for detecting the environmental humidity;

[0042] A body temperature detection device for obtaining a body temperature detection result by using the above body temperature detection method based on the environmental temperature, the body surface temperature, and the environmental humidity.

[0043] This application further provides a body temperature detection device, including:

[0044] A memory for storing a computer program;

[0045] A processor for executing the computer program to implement the steps of the above body temperature detection method.

[0046] This application further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above body temperature detection method are implemented.

[0047] This application further provides a computer program product, including computer programs / instructions, and when the computer programs / instructions are executed by a processor, the steps of the above body temperature detection method are implemented.

[0048] It can be seen that the present invention determines the closest reference temperature and reference humidity combination corresponding to the ambient temperature and ambient humidity in the current environment; determines a body temperature detection model based on temperature and humidity compensation based on the closest reference temperature and reference humidity combination; performs detection using the body temperature detection model based on temperature and humidity compensation based on the ambient temperature, ambient humidity, and body surface temperature to obtain the first actual detected body temperature; determines the first-order differential actual detected body temperature difference corresponding to the first actual detected body temperature, and the temperature differences between the ambient temperature and the body surface temperature; uses the body temperature prediction model based on the first-order differential actual detected body temperature difference, the temperature difference, and the first actual detected body temperature to obtain the second actual detected body temperature; the body temperature prediction model is a model determined based on the linear relationship between the first actual detected body temperature, the ambient temperature, the body surface temperature, and the second actual detected body temperature.

[0049] The beneficial effects of the present invention are as follows: Compared with directly using the measured temperature as the actual body temperature of the user currently, the present application determines the closest reference temperature and reference humidity combination corresponding to the ambient temperature and ambient humidity in the current environment, determines a body temperature detection model based on temperature and humidity compensation corresponding to the closest reference temperature and reference humidity combination, and compensates the measured value of the body surface temperature detected based on the body temperature detection model based on temperature and humidity compensation, so as to obtain the actual body temperature closer to the actual human body temperature faster. Since the present application performs the most suitable compensation for the current environment based on the current ambient temperature and ambient humidity, and uses the body temperature prediction model to shorten the equilibrium time, thereby improving the accuracy of the body temperature prediction value in the non-equilibrium state while shortening the equilibrium time.

[0050] In addition, the present invention also provides a body temperature detection device, system, equipment, and readable storage medium, which also have the above beneficial effects. Description of the Drawings

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0052] Figure 1 It is the flowchart of the first body temperature detection method provided by the embodiment of the present invention;

[0053] Figure 2 It is a method for realizing a body temperature detection model based on temperature and humidity compensation provided by the embodiment of the present invention;

[0054] Figure 3 It is the flowchart of the second body temperature detection method provided by the embodiment of the present invention;

[0055] Figure 4It is a schematic diagram of the thermal equilibrium time;

[0056] Figure 5 It is a flow chart example of a body temperature detection method provided by an embodiment of the present invention;

[0057] Figure 6 It is a schematic structural diagram of the first body temperature detection system provided by an embodiment of the present invention;

[0058] Figure 7 It is a schematic structural framework diagram of the second body temperature detection system provided by an embodiment of the present invention;

[0059] Figure 8 It is a schematic diagram of a continuous body temperature detection device provided by an embodiment of the present invention;

[0060] Figure 9 It is a schematic structural diagram of a body temperature detection device provided by an embodiment of the present invention;

[0061] Figure 10 It is a schematic structural diagram of a body temperature detection device provided by an embodiment of the present invention. Specific embodiments

[0062] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0063] Please refer to Figure 1 , Figure 1 It is a flow chart of the first body temperature detection method provided by an embodiment of the present invention. The method may include:

[0064] S101, determine the closest reference temperature and reference humidity combination corresponding to the ambient temperature and ambient humidity in the current environment.

[0065] The execution subject of this embodiment is an electronic device. The electronic device in this embodiment can be a computer, a mobile phone, etc. The environmental temperature and environmental humidity in the current environment in this embodiment refer to the real-time environmental temperature and real-time environmental humidity in the current environment collected by the environmental temperature sensor and the environmental humidity sensor. There are multiple combinations of reference temperature and reference humidity in this embodiment. The environmental temperature, environmental humidity, reference temperature, and reference humidity can be compared to determine the closest combination of reference temperature and reference humidity. The reference temperature and reference humidity in this embodiment are combinations of reference humidity and reference temperature obtained based on the actual temperature and humidity in different regions and different seasons. For example, the possible range of indoor temperature throughout the year is generally between 18°C and 22°C in spring, the humidity in the south is generally between 60% and 80%RH, and the humidity in the north is generally between 30% and 50%; it is generally between 24°C and 28°C in summer, the humidity in the south is generally between 70% and 90%, and the humidity in the north is generally between 40% and 60%; it is generally between 18°C and 22°C in autumn, and the humidity is between 50% - 70%RH; in winter, heating is provided in the north, the indoor temperature is generally between 18 - 22°C, there is no centralized heating in the south, the indoor temperature is between 10°C and 18°C, the humidity in the north is between 20% and 40%RH, and the humidity in the south is between 40% and 70%RH. Therefore, the environmental temperature can be set to three reference temperatures: low, medium, and high, and the humidity can be set to three reference humidities: low, medium, and high. Or the environmental temperature can be set to five levels: lowest, low, medium, high, and highest, and the humidity can also be set to five levels: lowest, low, medium, high, and highest. Or the environmental temperature can be set to three reference temperatures: low, medium, and high, and the environmental humidity can be set to five levels: lowest, low, medium, high, and highest. For example, when the environmental temperature is set to three reference temperatures: low, medium, and high, and the humidity is set to three reference humidities: low, medium, and high, the low, medium, and high values of the environmental temperature can be 15°C, 23°C, and 30°C respectively, and the low, medium, and high values of the environmental humidity can be 30%RH, 50%RH, and 80%RH respectively.

[0066] S102. Determine a body temperature detection model based on temperature and humidity compensation based on the closest combination of reference temperature and reference humidity.

[0067] The body temperature detection model based on temperature and humidity compensation in this embodiment is a body temperature detection model obtained by fitting based on a reference temperature, a reference humidity, and multiple first experimental body surface temperatures. The reason why the body temperature detection model is not limited in this embodiment is that S103 writes that based on the ambient temperature, the ambient humidity, and the body surface temperature, the body temperature is detected using the body temperature detection model based on temperature and humidity compensation to obtain the first actual detected body temperature. This step can be understood as obtaining the output of the first actual detected body temperature based on the input ambient temperature, ambient humidity, and body surface temperature. That is, this body temperature detection model can be a body temperature detection formula. The reason why this embodiment can determine the body temperature detection model based on temperature and humidity compensation based on the combination of the closest reference temperature and reference humidity is that different fittings can be performed under different reference temperatures and reference humidities, thereby obtaining fitting parameters with different parameters, that is, different body temperature detection models. The fitting based on the reference temperature, the reference humidity, and multiple first experimental body surface temperatures in this embodiment can be understood as follows: in order to implement different temperature and humidity environments, a simulated constant temperature source can be placed in a temperature and humidity chamber with a settable temperature, and the ambient temperature, ambient humidity, and body surface temperature collected under different reference temperatures, reference humidities, and simulated constant temperature sources (actual body temperature) can be obtained. Then, based on the multiple ambient temperatures, ambient humidities, and body surface temperatures collected, the body temperature detection model based on temperature and humidity compensation is obtained. This embodiment does not limit the specific fitting method. For example, this embodiment can perform fitting based on the least squares method, or this embodiment can also perform fitting based on optimization methods such as the gradient descent method.

[0068] It should be further noted that based on any of the above embodiments, in order to improve the accuracy of the design of the body temperature detection model based on temperature and humidity compensation during the non-equilibrium state, before determining the body temperature detection model based on temperature and humidity compensation based on the combination of the closest reference temperature and reference humidity, an implementation method of the body temperature detection model based on temperature and humidity compensation can also be included. For specific details, please refer to Figure 2 , Figure 2 An implementation method of a body temperature detection model based on temperature and humidity compensation provided by an embodiment of the present invention may include:

[0069] S201, determining multiple combinations of reference temperatures and reference humidities based on the actual environmental conditions, and constructing a corresponding fitting model based on each combination of reference temperature and reference humidity.

[0070] The actual environmental conditions of this embodiment refer to the actual environmental humidity and environmental temperature in different regions and different seasons, so that multiple reference temperatures and multiple reference humidities can be obtained. Then, the reference temperatures and reference humidities are combined in pairs to obtain combinations of reference temperature and reference humidity. The reason for constructing a corresponding fitting model based on each combination of reference temperature and reference humidity in this embodiment is that environmental temperature and environmental humidity will have different effects on the actually detected body temperature under different reference temperatures and reference humidities. The fitting model in this embodiment is a model based on reference temperature and reference humidity and composed of multiple numerical parameters to be determined. Environmental temperature, environmental humidity, and body surface temperature are independent variables, and the actually detected body temperature is the dependent variable.

[0071] S202. Based on multiple first experimental body surface temperatures, multiple first experimental environmental temperatures, and multiple first experimental environmental humidities corresponding to each combination of reference temperature and reference humidity, fit the corresponding fitting model to obtain a body temperature detection model based on temperature and humidity compensation corresponding to each combination of reference temperature and reference humidity.

[0072] The multiple first experimental environmental temperatures and multiple first experimental environmental humidities in this embodiment can be determined according to the approximate range of indoor temperature and humidity, and are used to quantify the interference of external conditions on the temperature measurement of the constant temperature source. In this embodiment, a body temperature detection model based on temperature and humidity compensation corresponding to each combination of reference temperature and reference humidity will be obtained. Thus, when in use, a body temperature detection model based on temperature and humidity compensation can be directly determined based on the combination of reference temperature and reference humidity determined by the current environmental temperature and environmental humidity. For ease of understanding, the fitting model (basic temperature and humidity compensation model) is: ; where represents the sensor measurement value (the first actually detected body temperature) after temperature and humidity compensation at time t, represents the body surface temperature measurement value, represents the measured value of environmental temperature, represents the reference environmental temperature value (reference temperature), represents the measured value of environmental humidity, represents the environmental humidity reference value (reference humidity); 、 、 、 are the weight coefficients of the model, and these coefficients are obtained by fitting (such as fitting by the least squares method). For each body part, the 、 values are all different. According to historical data, maximum likelihood estimation is used, and the mean value of the data is used as and . Among them, when fitting, The value is the measured value of the sensor at the reference temperature and humidity with a simulated constant temperature source (the set temperature value can simulate the human body temperature). In different temperature and humidity environments, the simulated constant temperature source can be placed in a constant temperature and humidity chamber with adjustable temperature and humidity. A batch of body temperature stickers can be pasted on the simulated constant temperature source to simulate the process of body temperature heat conduction. The measurement data can be recorded according to the following table. Continuously collect data for 15 minutes at each temperature gradient and humidity gradient, and take the average value of the stable section of each batch of sensors as the recorded value for this temperature and humidity gradient. As shown in Table 1, Table 1 is a schematic table of a temperature and humidity gradient provided by an embodiment of the present invention.

[0073] Table 1 Schematic table of a temperature and humidity gradient

[0074]

[0075] Furthermore, the temperature of the simulated constant temperature source can be 36°C, 37°C, 38°C, 39°C, 40°C, 41°C. Possible combinations of reference temperature and humidity can be selected. For example, the possible range of indoor temperature throughout the year is usually between 18°C - 22°C in spring, the humidity in the south is usually between 60% - 80%RH, the humidity in the north is usually between 30% - 50%, between 24°C - 28°C in summer, the humidity in the south is usually between 70% - 90%, the humidity in the north is usually between 40% - 60%, between 18°C - 22°C in autumn, the humidity is between 50% - 70%RH, in winter, the north has heating, the indoor temperature is usually between 18 - 22°C, in the south without central heating, the indoor temperature is between 10°C - 18°C, the humidity in the north is between 20% - 40%RH, and the humidity in the south is between 40% - 70%RH. Therefore, the environmental temperature can be set to three reference temperatures: low, medium, and high, and the humidity can be set to three reference humidities: low, medium, and high. After combining the reference temperature and humidity as much as possible, the coefficient fitting of the model is performed. Then, in the body temperature detection model based on temperature and humidity compensation, and are set to possible combinations of low, medium, and high. For example, The low, medium, and high values are 15°C, 23°C, and 30°C respectively, The low, medium, and high values of

[0076] Table 2 Schematic table of a combination of reference temperature and reference humidity

[0077]

[0078] Furthermore, each fitting model is as follows:

[0079] ;

[0080] ;

[0081] ;

[0082] ;

[0083] ;

[0084] ;

[0085] ;

[0086] ;

[0087] 。

[0088] This embodiment can use the ambient temperature, ambient humidity, and body surface temperature measured by the sensor to select the above and combinations and perform parameter fitting using the least squares method. For example, is set to 23 °C, is set to 50% RH, and least squares fitting is used to fit . The fitting target is to set the constant temperature source at a reference temperature of 23 °C and a reference humidity of 50% RH to 37 °C ( ) for the body surface temperature . Then, the body surface temperature values measured by the lower shell are shown in Table 3. Table 3 is an example table of body surface temperatures under different ambient temperatures and ambient humidities provided by the embodiment of the present invention.

[0089] Table 3 Example Table of Body Surface Temperatures under Different Ambient Temperatures and Ambient Humidities

[0090]

[0091] According to the model, after fitting, the values of , , , are 0.712, 0.029, -0.005, and 10.373 respectively, and their corresponding temperature and humidity compensation formula is: . The remaining fitting formulas can be implemented according to the above method. This embodiment gives the specific fitting process of the body temperature detection model based on temperature and humidity compensation corresponding to the fitted reference temperature and reference humidity combination, improving the fitting accuracy.

[0092] S103, based on the ambient temperature, ambient humidity, and body surface temperature, use a body temperature detection model based on temperature and humidity compensation to perform detection, and obtain the first actual detected body temperature.

[0093] The body surface temperature in this embodiment is the body surface temperature measured based on a temperature measuring element (an element for measuring the body surface temperature). The input of the body temperature detection model based on temperature and humidity compensation in this embodiment is the ambient temperature, ambient humidity, and body surface temperature, and the output is the first actual detected body temperature. It should be noted that the ambient temperature is also affected by the human body temperature. Due to the different paste positions of the temperature sticker, when the measured ambient temperature is higher than a certain threshold, environmental temperature compensation may not be performed. For example, when it is higher than the set threshold (35°C), temperature and humidity compensation is not performed. At this time, the first actual detected body temperature is equal to the body surface temperature. Therefore, before using the body temperature detection model based on temperature and humidity compensation for detection in this embodiment, it can be determined whether the ambient temperature is higher than the set ambient temperature threshold. When it is not higher, the body temperature detection model based on temperature and humidity compensation is used for detection; otherwise, the directly detected body surface temperature is used as the first actual detected body temperature.

[0094] It should be further noted that, in order to improve the user experience, after using the body temperature detection model based on temperature and humidity compensation to detect based on the ambient temperature, ambient humidity, and body surface temperature and obtain the first actual detected body temperature, it may further include: using a health analysis large model to perform health analysis based on the data of the current user, and obtaining health analysis and suggestions; where the data of the current user includes health data, the first actual detected body temperature data, living habits, and environmental data, and the health analysis large model is a model obtained by fine-tuning the large model based on the target input and target output. The target input is the data of the current user, and the target output is the health results and health suggestions given based on each item of the user's data. In this embodiment, the user can open the help assistant of the APP to have an interactive conversation, and call the AIGC (health analysis large model) according to the conversation information to give an analysis of the body temperature fluctuation trend, health intervention suggestions, and generate a personalized customized health report. The health analysis large model can be used for analyzing the body temperature fluctuation trend, giving health intervention suggestions, and generating a personalized customized health report to help users understand the body temperature changes and their impact on health. The steps of its model fine-tuning can include: Step 1: Dataset production. Write the user's basic information, health data, body temperature monitoring data, living habits, environmental data, and labels of key health events such as disease or emergency event labels into the dataset and convert it into the JSONL format required for large model fine-tuning.

[0095] This embodiment can train the health analysis large model in the "prompt-response" format, that is, each pair of input (prompt) and target output (response) data will be used to fine-tune the model. The dataset example is as follows: {"prompt": "The user's body temperature has fluctuated greatly in the past 24 hours, with a maximum of 38.2°C and a minimum of 36.5°C. The user is overweight, with a BMI of 29, and has hypertension. The user has been under a lot of stress and has insufficient sleep recently. Please provide health intervention suggestions based on this information.", "response": "It is recommended that the user try to reduce strenuous exercise and maintain sufficient water intake when the body temperature fluctuates greatly. Due to hypertension and overweight, it is recommended to control the intake of salt and fat and increase low-intensity exercise. When the sleep quality is poor, it is recommended to carry out relaxation training and improve the work and rest time."} If the dialogue format is used for fine-tuning, including the dialogue history, the dataset example is as follows: {"messages": [{"role": "user", "content": "My body temperature has often fluctuated between 37.5°C and 38.0°C recently. What's going on?"}, {"role": "assistant", "content": "Body temperature fluctuations may be related to multiple factors, such as infection, overfatigue, or changes in environmental temperature. If this fluctuation persists or is accompanied by other symptoms, it is recommended to seek medical advice as soon as possible. Maintaining regular work and rest and a healthy diet helps to stabilize body temperature."}, {"role": "user", "content": "I've been under a lot of stress recently. Could that affect my body temperature?"}, {"role": "assistant", "content": "Yes, excessive stress can affect body temperature, especially by regulating body temperature through hormone secretion. It is recommended to reduce stress through meditation, deep breathing, and light exercise, which helps to stabilize body temperature."}]. Step 2: Fine-tuning of the health analysis large model. Use the large model fine-tuning interface to upload the dataset in the required format and the data pair id used as the training set, and select the pre-trained model for fine-tuning. Step 3: Verification of the fine-tuned model. After fine-tuning, verify whether the fine-tuned model can give reasonable health analysis and suggestions based on the new dialogue data. The health analysis large model in this embodiment can help analyze and predict the trend of body temperature fluctuations, so as to identify potential health problems. It can extract the change trend of the user's body temperature during the day, such as in the morning, at noon, and in the evening, or analyze whether there are periodic fluctuations in body temperature on a weekly or monthly basis, analyze the historical fluctuations of body temperature, and predict the body temperature change trend in the next few days; it can also identify abnormal fluctuations in body temperature based on the fluctuation characteristics of the given body temperature data, and remind the user to pay attention to potential health problems; it can also process on-site device data in real time, and if abnormal body temperature fluctuations are found, it can provide push or warning.Alternatively, the health analysis large model can provide health intervention suggestions. Based on the body temperature change trend, the user's health status, and individual characteristics, such as integrating multi-dimensional data including the user's exercise data, sleep quality, diet, medical history, etc., it comprehensively evaluates the health status, recommends diet, exercise, or drug use, and helps the user adjust the body temperature through behavioral intervention. Further, the health analysis large model can generate detailed and personalized health reports, helping the user fully understand the impact of body temperature changes on health and providing customized suggestions according to individual differences. For example, during the pregnancy preparation period, the report will specifically focus on the basal body temperature changes, generate a body temperature monitoring report, and give an estimated ovulation date; if the user is pregnant, it will identify abnormal body temperature fluctuations based on the user's own health data and pay attention to the impact of body temperature fluctuations on the fetus; for patients with chronic diseases, such as heart disease or diabetes patients, the model will analyze the potential impact of body temperature fluctuations on these diseases by combining these factors and give specialized intervention suggestions. This embodiment can perform intelligent data analysis and interaction, and can use the health analysis large model for body temperature data analysis, health intervention, and personalized health report generation, further enhancing user interaction and experience.

[0096] S104. Determine the first-order difference actual body temperature difference corresponding to the first actual detected body temperature, the temperature difference between the environmental temperature and the body surface temperature.

[0097] In this embodiment, the first-order difference actual body temperature difference corresponding to the first actual detected body temperature is , where is the first actual detected body temperature at time t, is the first actual detected body temperature at time

[0098] S105. Based on the first-order difference actual body temperature difference, the temperature difference, and the first actual detected body temperature, use the body temperature prediction model to obtain the second actual detected body temperature; the body temperature prediction model is a model determined based on the linear relationship between the first actual detected body temperature, the environmental temperature, the body surface temperature, and the second actual detected body temperature.

[0099] In this embodiment, the independent variables of the body temperature prediction model are the first-order difference actual body temperature difference, the temperature difference, and the first actual detected body temperature, and the dependent variable is the second actual detected body temperature. The body temperature prediction model in this embodiment is a model determined based on the first-order linear relationship between the first actual detected body temperature, the environmental temperature, the body surface temperature, and the second actual detected body temperature.

[0100] It should be further noted that, based on any of the above embodiments, the process of determining the above body temperature prediction model may include: fitting a body temperature prediction basic model based on multiple second experimental environmental temperatures, multiple second experimental body surface temperatures, multiple first actually detected body temperatures of multiple second experiments, and multiple actual body temperatures of multiple second experiments (the actual body temperature values of the human body are measured by a mercury thermometer, etc.) to obtain various weight coefficients in the body temperature prediction basic model; constructing a body temperature prediction model based on the various weight coefficients and the body temperature prediction basic model. In this embodiment, the numbers of the multiple second experimental environmental temperatures, the multiple second experimental body surface temperatures, the multiple first actually detected body temperatures of multiple second experiments, and the multiple actual body temperatures of multiple second experiments are the same. The body temperature prediction basic model in this embodiment is a model with three parameters, namely the first actually detected body temperature, the first-order difference actual detected body temperature difference corresponding to the first actually detected body temperature, and the temperature difference corresponding to the environmental temperature and the body surface temperature, as independent variables, and certain weight coefficients are assigned, and the second actually detected body temperature is used as the dependent variable. Since the body temperature prediction model is obtained by fitting based on the actual body temperature values of the human body, it is possible to quickly determine a more accurate actual body temperature of the human body based on the relationship between the actual body temperature of the human body and the first actually detected body temperature, thereby shortening the equilibration time. This embodiment provides a specific method for constructing a body temperature prediction model, improving the accuracy of constructing the body temperature prediction model.

[0101] It should be further noted that, in order to improve the accuracy of determining the body temperature prediction model, the above process of fitting the body temperature prediction basic model based on multiple second experimental environmental temperatures, multiple second experimental body surface temperatures, multiple first actually detected body temperatures of multiple second experiments, and multiple actual body temperatures of multiple second experiments to obtain various weight coefficients in the body temperature prediction basic model may include: determining the second experimental environmental temperature, the second experimental body surface temperature, the first actually detected body temperature of the second experiment, and the actual body temperature of the second experiment corresponding to each temperature measurement position; fitting the body temperature prediction basic model based on the second experimental environmental temperature, the second experimental body surface temperature, the first actually detected body temperature of the second experiment, and the actual body temperature of the second experiment corresponding to each temperature measurement position to obtain various weight coefficients corresponding to each temperature measurement position; correspondingly, constructing a body temperature prediction model based on the various weight coefficients and the body temperature prediction basic model includes: constructing a body temperature prediction model corresponding to each temperature measurement position based on the various weight coefficients corresponding to each temperature measurement position and the body temperature prediction basic model. Different temperature measurement positions in this embodiment may be, for example, the armpit, the forehead, etc. This embodiment takes into account that the equilibration times corresponding to different temperature measurement positions are inconsistent, so when constructing the body temperature prediction model, body temperature prediction models corresponding to different temperature measurement positions will be constructed. This embodiment assigns different prediction model weights based on different strategic positions, and flexibly combines additional models to comprehensively improve the prediction accuracy.

[0102] For ease of understanding, for example, the body temperature prediction basic model in this embodiment may be ; wherein, represents the predicted temperature value output at time (the second actual detected body temperature), represents the sensor measurement value after temperature and humidity compensation at time (the first actual detected body temperature at time t), represents the sensor measurement value after temperature and humidity compensation at time t-1, represents the body surface temperature, represents the ambient temperature, and and are the weight coefficients of the model, and these coefficients are obtained by least squares fitting. When performing least squares fitting, is the actual body temperature of the human body. Please refer to Table 4, which is a schematic table of a second experimental data provided by an embodiment of the present invention.

[0103] Table 4 Schematic Table of a Second Experimental Data

[0104]

[0105] According to the data in Table 4, according to the model, after fitting and and take the values of -0.773, 0.630, -0.846 respectively, and the corresponding temperature prediction model output is: .

[0106] A body temperature detection method provided by an embodiment of the present invention may include: S101, determining the closest reference temperature and reference humidity combination corresponding to the environmental temperature and environmental humidity in the current environment; S102, determining a body temperature detection model based on temperature and humidity compensation based on the closest reference temperature and reference humidity combination; S103, detecting based on the environmental temperature, environmental humidity, and body surface temperature using the body temperature detection model based on temperature and humidity compensation to obtain a first actual detected body temperature. S104, determining the first-order difference actual detected body temperature difference corresponding to the first actual detected body temperature, and the temperature difference between the environmental temperature and the body surface temperature. S105, using a body temperature prediction model based on the first-order difference actual detected body temperature difference, the temperature difference, and the first actual detected body temperature to obtain a second actual detected body temperature; the body temperature prediction model is a model determined based on the linear relationship between the first actual detected body temperature, the environmental temperature, the body surface temperature, and the second actual detected body temperature. Compared with directly using the measured temperature as the user's actual body temperature currently, this application will determine the body temperature detection model based on temperature and humidity compensation corresponding to the closest reference temperature and reference humidity corresponding to the environmental temperature and environmental humidity in the current environment, and compensate the measured value of the body surface temperature detected by the body temperature detection model based on temperature and humidity compensation to obtain an actual body temperature closer to the actual human body temperature. Since this application will perform compensation most suitable for the current environment based on the current environmental temperature and environmental humidity, it reduces the interference of the external environment on temperature measurement, thereby improving the accuracy of body temperature detection, and will obtain a second actual detected body temperature closer to the actual human body temperature based on the body temperature prediction model, thereby improving the accuracy of the body temperature prediction value in a non-equilibrium state while shortening the equilibrium time.

[0107] For the convenience of understanding the present invention, please specifically refer to Figure 3 , Figure 3 which is a flowchart of the second body temperature detection method provided by an embodiment of the present invention, and may specifically include:

[0108] S301, determining the closest reference temperature and reference humidity combination corresponding to the environmental temperature and environmental humidity in the current environment.

[0109] For the explanations of S301 - 3405 in this embodiment, please refer to the explanations of S101 - S105 above.

[0110] S302, determining a body temperature detection model based on temperature and humidity compensation based on the closest reference temperature and reference humidity combination; wherein, the body temperature detection model based on temperature and humidity compensation is a body temperature detection model obtained by fitting based on the reference temperature, reference humidity, and multiple first experimental body surface temperatures.

[0111] S303, detecting based on the environmental temperature, environmental humidity, and body surface temperature using the body temperature detection model based on temperature and humidity compensation to obtain a first actual detected body temperature.

[0112] S304. Determine the first-order difference of the actually detected body temperature corresponding to the first actually detected body temperature, the temperature difference corresponding to the ambient temperature and the body surface temperature.

[0113] S305. Based on the first-order difference of the actually detected body temperature, the temperature difference, and the first actually detected body temperature, use the body temperature prediction model to obtain the second actually detected body temperature; wherein, the body temperature prediction model is a model determined based on the linear relationship between the first actually detected body temperature, the ambient temperature, the body surface temperature, and the second actually detected body temperature.

[0114] S306. Based on the user data, the ambient temperature, and the ambient humidity, use the approximate search model to perform a search to obtain a preset number of thermoneutral body temperatures; the user data is data related to the user's health.

[0115] This embodiment does not limit the specific user data. For example, the user data in this embodiment is the body temperature value when each user reaches thermoneutrality, the user's gender, age, height, weight, etc. The approximate search model in this embodiment is to find the thermoneutral body temperatures of the K nearest neighbors of the current user. This embodiment does not limit the specific preset number. For example, the preset number in this embodiment can be 3; or the preset number in this embodiment can be 4, etc. When performing the search in this embodiment, it can start from the root node and perform a recursive search, using the Euclidean distance of the mixed features to measure the similarity degree, and using the body temperature values of the thermoneutral body temperatures of the K nearest neighbors found as the prediction basis.

[0116] Further, based on any of the above embodiments, the determination process of the above approximate search model may include:

[0117] S1: Process the user data based on the body temperature value of the user's thermoneutrality to obtain the processed user data.

[0118] S2: Group the processed user data based on the measurement location to obtain grouped data.

[0119] This embodiment can clean, group, and extract features of historical data (including the body temperature value of the user's thermoneutrality, and excluding the body temperature value if the user does not have a thermoneutrality) in a data preprocessing container, and group the data with the same measurement location, such as supporting locations like the armpit, chest, groin, wrist, etc.

[0120] S3: Extract features from the data in each grouped data to obtain user features and the environmental features of each grouped data, and use the user features and the environmental features as target features.

[0121] The user characteristic data in this embodiment includes at least one of user gender, user age, and BMI index (Body Mass Index), and the environmental characteristics include the average environmental temperature, the average environmental humidity, the maximum body surface temperature, the minimum body surface temperature, and the body surface temperature change rate. For example, the characteristic data extracted in this embodiment may include user gender, age, BMI index (obtained by converting height and weight), the average environmental temperature and environmental humidity within a preset time (20 s), and the maximum, minimum, and change rate of the body surface temperature within 1 min (which can be set according to requirements).

[0122] S4: Perform feature encoding on the target features to obtain feature vectors; among them, numerical features are processed by maximum-minimum normalization, and binary categorical features are processed by one-hot encoding.

[0123] This embodiment can perform feature transformation on the target features, perform maximum-minimum normalization on numerical features, and mark binary categorical features such as gender as 0, 1 features and perform one-hot encoding (unique hot encoding).

[0124] S5: Construct a K-D tree based on the feature vectors to obtain an approximate search model.

[0125] When constructing, these feature vectors can be used to construct a K-D tree to organize user data for fast nearest neighbor search. For each user grouped according to the measurement location, the above 8 features are extracted. When constructing each layer of the tree, a specific dimension is selected. For one-hot encoded features such as gender, direct splitting is performed. For example, gender 0 is in the left subtree and gender 1 is in the right subtree. For numerical features, after sorting, the median value of the numerical values is used as the current node, and the left and right subtrees are constructed in this way until all data is processed to obtain an approximate search model. This embodiment gives a specific method for constructing the approximate search model, improving the accuracy of constructing the approximate search model, and since multiple user features and environmental features can be used, the integrity of constructing the approximate search model can be improved.

[0126] S307: Perform weighted processing based on a preset number of heat balance body temperatures to obtain an estimated heat balance body temperature.

[0127] This embodiment is not limited to the specific method of weighting. This embodiment can perform distance weighting, or this embodiment can also perform weight weighting.

[0128] It should be further noted that, in order to improve the accuracy of weighted processing, the above-mentioned weighted processing based on a preset number of thermostatic body temperatures to obtain the estimated thermostatic body temperature may include: performing distance weighting based on a preset number of thermostatic body temperatures to obtain the estimated thermostatic body temperature; where distance weighting is a method of determining the estimated thermostatic body temperature by weighting the body temperature values of neighbor users based on the feature distance between the current user and each neighbor user. This embodiment can use distance weighting to obtain the final body temperature estimate of the target user according to the thermostatic body temperature values of K neighbors. The formula for obtaining the thermostatic body temperature estimate using distance weighting can be: ; where is the estimated value of the target user's thermostatic body temperature, is the th neighbor's body temperature value, is the distance between the target user's feature vector value and the th neighbor. The Euclidean distance of the mixed features is as follows: ; where is the Euclidean distance of the mixed features, is the Euclidean distance used for numerical features, is the Hamming distance used for binary categorical features, is the weight factor of the binary categorical features, takes 0.3. Assume that the body temperatures of the three nearest neighbors found by the target user (current user) and the distances from the target user's feature vector are shown in Table 5, and Table 5 is a schematic table of the Euclidean distance provided by the embodiment of the present invention.

[0129] Table 5 Schematic table of Euclidean distance

[0130]

[0131] At this time, the estimated thermostatic body temperature obtained based on distance weighting is:

[0132] .

[0133] S308. Based on the estimated thermostatic body temperature and the second actual detected body temperature, use the temperature prediction model adjusted based on the weights of similar users to determine the third actual detected body temperature; where the temperature prediction model adjusted based on the weights of similar users is a model that adjusts the second actual detected body temperature based on the estimated thermostatic body temperature.

[0134] In this embodiment, the independent variables of the temperature prediction model based on the adjustment of the weights of similar users are the estimated thermostatic body temperature and the second actually detected body temperature, and the dependent variable is the third actually detected body temperature. In order to shorten the thermostatic time in this embodiment, the body temperature data at the time of reaching thermostasis of the historical user most similar to the current user can be found in the historical database according to features such as the saved user information and measured temperature data, and a weight factor is added to calibrate the above model (body temperature prediction model). The temperature prediction model with the adjustment of the weights of similar users added is (the temperature prediction model based on the adjustment of the weights of similar users): ; where is the body temperature data at the time of reaching thermostasis of the historical user most similar to the current user found in the historical database using the KNN method (approximate search model), is the output body temperature value according to the temperature prediction model. is the weight coefficient factor. Further, can be adjusted according to the actual situation. If it is selected not to attach the adjustment of the weights of similar users, then .

[0135] It should be noted that through experiments, in a specific embodiment, for example, the estimated thermostatic body temperature = 36.8 °C, Take 0.8, and the actual body temperature value is 36.9 °C. Then, the body surface temperature curve, the body temperature prediction model curve, and the curve of the temperature prediction model based on the adjustment of the weights of similar users are as Figure 4 shown. Figure 4 This is a schematic diagram of a thermostatic time: the horizontal axis of the curve represents the time point, the time interval is 5 s, the thermostatic time of the body surface temperature curve is about 12 min at the longest, the thermostatic time of the body temperature prediction model is 6.8 min (corresponding label 81), and the thermostatic time of the temperature prediction model based on the adjustment of the weights of similar users is 4.1 min (corresponding label 49). It can be seen that when determining the actually detected body temperature according to the temperature prediction model based on the adjustment of the weights of similar users, the thermostatic time is the shortest.

[0136] The current electronic thermometers and continuous temperature measurement devices take a long time to measure, generally taking 15 - 20 min to reach thermostasis, and the vast majority of prediction models cannot adapt to multi-site temperature measurement and have a large anti-interference ability to the environment. In addition, the user interaction experience is poor, and users do not have a more comprehensive understanding of their own temperature data. The main technical problem to be solved by this application is to improve the detection accuracy of the continuous body temperature detection device and shorten the equilibration time.

[0137] For the convenience of understanding the present invention, please specifically refer to Figure 5 . Figure 5 This is a flow chart example of a body temperature detection method provided by an embodiment of the present invention, which may specifically include:

[0138] S501. Obtain the ambient temperature and ambient humidity, and determine a body temperature detection model based on temperature and humidity compensation based on the reference temperature and reference humidity corresponding to the ambient temperature and ambient humidity.

[0139] S502. Based on the ambient temperature, ambient humidity, and body surface temperature, use the body temperature detection model based on temperature and humidity compensation to obtain the first actual detected body temperature.

[0140] S503. Determine a body temperature prediction model based on the temperature measurement location.

[0141] S504. Determine the first-order difference actual detected body temperature difference corresponding to the first actual detected body temperature, and the temperature differences corresponding to the ambient temperature and body surface temperature.

[0142] S505. Based on the first-order difference actual detected body temperature difference, temperature difference, and the first actual detected body temperature, use the body temperature prediction model to obtain the second actual detected body temperature.

[0143] S506. Based on the user data, ambient temperature, and ambient humidity, use an approximate search model to perform a search to obtain a preset number of heat balance body temperatures; the user data is data related to the user's health.

[0144] The device corresponding to the body temperature detection method in this embodiment is a body temperature detection device. The continuous body temperature monitoring device can be pasted onto the selected part to collect the body surface temperature 1 minute before the start of the rising section (which can be set according to requirements) and send it to the cloud. The cloud calls the KNN approximate search model based on the user data, ambient temperature and humidity data, and body surface temperature to obtain a preset number of heat balance body temperatures.

[0145] S507. Perform a weighted process on the preset number of heat balance body temperatures to obtain an estimated heat balance body temperature.

[0146] S508. Based on the estimated heat balance body temperature and the second actual detected body temperature, use a temperature prediction model based on the adjustment of similar user weights to determine the third actual detected body temperature.

[0147] S509. Perform a health analysis on the data of the current user using a large health analysis model to obtain a health analysis and suggestions; the data of the current user includes the actual detected body temperature at each time point and personal basic data.

[0148] Compared with the prior art, the present invention has the following advantages and effects:

[0149] Based on the body temperature detection model with temperature and humidity compensation, the collected body surface temperature is compensated for the environmental temperature and humidity to reduce the interference of the external environment on temperature measurement; different body temperature prediction models are determined based on different measurement positions to comprehensively improve the prediction accuracy; the thermal equilibrium body temperature is estimated using similar users, and a body temperature weight term of similar users is added to further shorten the equilibrium time; intelligent data analysis and interaction can use a large health analysis model for body temperature data analysis, health intervention, and generation of personalized health reports to further enhance user interaction and experience.

[0150] The body temperature detection system provided by the embodiments of the present invention will be introduced below. The body temperature detection system described below can be correspondingly referred to the body temperature detection method described above.

[0151] Specifically, please refer to Figure 6 , Figure 6 which is a schematic structural diagram of the first body temperature detection system provided by the embodiments of the present invention, and may include:

[0152] The first temperature measurement element 100 is used to detect the environmental temperature;

[0153] The second temperature measurement element 200 is used to detect the body surface temperature;

[0154] The humidity measurement element 300 is used to detect the environmental humidity;

[0155] The body temperature detection device 400 is used to perform body temperature detection based on the environmental temperature, body surface temperature, and environmental humidity in the above body temperature detection method to obtain a body temperature detection result.

[0156] To make the present invention easier to understand, specifically, please refer to Figure 7 , Figure 7 which is a schematic structural framework diagram of the second body temperature detection system provided by the embodiments of the present invention (the system corresponding to a body temperature detection method), and specifically may include: 1 - Continuous body temperature detection device (body temperature sensor), 2 - Data transmission component, 3 - Container management platform (containerized application deployment), 4 - Front-end component, 5 - Container image repository, 6 - Device terminal.

[0157] The continuous body temperature monitoring device in this embodiment can collect user data, environmental information, and body temperature data through a mobile terminal application software (APP). The data transmission component is responsible for transmitting the collected data to the cloud service platform for data feature extraction, analysis, and the construction of models (including the above-mentioned body temperature detection model based on temperature and humidity compensation, body temperature prediction model, and temperature prediction model based on similar user weight adjustment). In addition to supporting basic data communication interfaces and protocols (serial port and Bluetooth), the data transmission component also supports high-speed data transmission protocols (such as LoRa, 5G, Wi-Fi). The container management platform supports Docker Swarm (a native container orchestration tool), OpenShift (a container platform), and Kubernetes (a container orchestration engine), and is suitable for container orchestration and management in different scales, demand levels, and scenarios for automated deployment, expansion, and management of container operations.

[0158] Furthermore, the image repository is used to store container images with different functions. The container images encapsulate applications with different functions and are used to store, calculate, analyze, and process the data transmitted from the continuous body temperature monitoring device to the cloud platform.

[0159] Furthermore, the types of container images include data transmission middleware containers, data storage containers, data preprocessing containers, modeling containers, inference containers, and log monitoring containers. Among them, the data transmission middleware container includes Apache Kafka (a distributed stream processing platform) for user body temperature data streams and message passing to ensure the rapid transmission of data from the device to the cloud platform. Among them, the data storage container includes various database containers such as the PostgreSQL database container for recording user information and device information, the MongoDB database container for storing logs, and the InfluxDB database specifically for storing body temperature time series data. Among them, the data preprocessing container includes the cleaning, grouping, feature extraction, and structured data generation of user data, environmental data, and body temperature data. Among them, the modeling container encapsulates the environments required for various model training and testing, such as components of scikit-learn (a machine learning library), pandas (a data analysis library), numpy (a numerical calculation library), and pytorch (a deep learning framework) for AI modeling calculations, as well as a model fine-tuning interface adaptable to the knowledge in the field of body temperature measurement.

[0160] Furthermore, when fine-tuning the large model, training data in the field of body temperature measurement needs to be prepared, such as health reports associated with body temperature data, text data of body temperature records and anomaly analysis. And the data needs to conform to the form of question and answer and dialogue. The training data is uploaded through the interface to start model fine-tuning. Among them, the inference container includes a KNN model for similarity search according to user and body temperature data characteristics, a body temperature detection model based on a temperature and humidity compensation model, a temperature prediction model for different measurement parts, and a temperature prediction model with adjusted weights of similar users (temperature prediction model based on adjusted weights of similar users), and an API interface of the above-mentioned fine-tuned body temperature data analysis AIGC model (health analysis large model). Among them, the log monitoring container integrates monitoring tools such as Prometheus (open source monitoring system) and Grafana (data visualization platform) to monitor the health status of the continuous body temperature monitoring system in real time and record logs. The front-end component integrates a user interaction interface, using React and WebAssembly technologies, supporting dynamic charts and real-time data display. The device terminal is a mobile device or a PC terminal. The terminal cooperates with the front-end component to receive and display data from the system and control the operation of the entire system.

[0161] Further, the user views and controls the container management platform through the device terminal, copies the containers required for continuous body temperature monitoring analysis from the image repository according to business needs, and initializes and runs them. The continuous body temperature monitoring device uploads user data, environmental data, and body temperature data to the container management platform through the data transmission component. The container management platform receives the data through the data transmission middleware container and stores the data in the data storage container. The data preprocessing container obtains data from the data storage for preprocessing according to business requirements to generate structured data required for model training. The modeling container uses the structured data to train various business requirement models. The trained models are verified through the model inference container. The updated weight parameters of the verified models are transmitted to the APP software (mobile terminal application software) of the continuous body temperature monitoring analysis device for displaying and receiving body temperature detection data. The APP software can call the model inference service link of the inference container to perform body temperature fluctuation trend analysis, health intervention suggestions, and generate personalized customized health reports.

[0162] Figure 8Schematic diagram of a device for continuous body temperature measurement and analysis provided for an invention embodiment. A device for continuous body temperature measurement and analysis in this embodiment may include a housing, a circuit board 1-2 placed inside the housing 1-1, a temperature measurement element 1 (1-3), a temperature measurement element 2 (1-4), a humidity measurement element 1-5 connected to the circuit board 1-2, a temperature-sensitive sheet 1-6 connected to the temperature measurement element 2 (1-4), a power supply module 1-7, and a mobile terminal application software (APP) 1-8 (for implementing the steps of the above body temperature detection method to achieve actual body temperature detection). The circuit board includes an MCU (Microcontroller Unit), a Bluetooth communication module, a flash memory, a Hall switch, etc. Further, the MCU is responsible for coordinating functions such as data acquisition, processing, storage, and transmission of various sensors (such as temperature measurement elements and humidity measurement elements). Further, the Bluetooth communication module is used for transmitting temperature and humidity data and sensor abnormality information. The flash memory is used for storing temperature and humidity data within a set storage period (such as within 7 days). The Hall switch is used to activate the temperature measurement device. The temperature measurement element 1 and the temperature measurement element 2, the temperature measurement element 1 is placed on the upper shell of the housing for measuring the ambient temperature, and the temperature measurement element 2 is placed on the skin side of the lower shell of the housing for measuring the body surface temperature. The humidity measurement element probe is embedded in the housing surface for measuring the ambient humidity. The power supply module is used to supply power to the body temperature measurement circuit, and the power supply module supplies power to a polymer lithium battery of 100 mAh. The mobile terminal application software is used to receive the temperature and humidity data transmitted by the continuous body temperature measurement device and perform data interaction with the second body temperature detection system.

[0163] The body temperature detection system in this embodiment, for implementing a method for continuous body temperature monitoring and analysis, can: establish a user profile. The user opens the APP (mobile terminal application software) to fill in personal basic information and health information; initialize environmental data. After the APP successfully connects to the Bluetooth of the temperature measurement device, it collects 20s of environmental data and uploads it to the cloud. The cloud returns the weights of the body temperature detection model based on temperature and humidity compensation adapted to the APP; the user selects the temperature measurement location in the APP, and the cloud returns the weights of the temperature prediction model corresponding to this location to the APP; the user pastes the continuous body temperature monitoring device to the selected part, and the temperature sensor collects the temperature data near the skin side 1 minute before the starting rising section and uploads it to the cloud. The cloud calls the KNN approximate search model based on the user data, environmental temperature and humidity data, and measured temperature data to obtain the estimated heat balance body temperature value; the user selects a model. The user can choose whether to add a temperature and humidity compensation model and a temperature prediction model with similar user weight adjustment as needed; output the predicted temperature value (actual detected body temperature) according to the given model and model weights; save the predicted temperature value locally and synchronously transmit it to the cloud for storage; the user can open the help assistant of the APP for interactive dialogue, and call the AIGC interface according to the dialogue information to give an analysis of body temperature fluctuation trends, health intervention suggestions, and generate a personalized customized health report.

[0164] The temperature detection device provided by the embodiments of the present invention will be introduced below. The temperature detection device described below can be correspondingly referred to the temperature detection method described above.

[0165] Specifically, please refer to Figure 9 , Figure 9 which is a schematic structural diagram of a temperature detection device provided by an embodiment of the present invention, and may include:

[0166] The closest temperature and humidity combination determination module 500 is configured to determine the closest reference temperature and reference humidity combination corresponding to the ambient temperature and ambient humidity in the current environment;

[0167] The body temperature detection model determination module 600 is configured to determine a body temperature detection model based on temperature and humidity compensation based on the closest reference temperature and reference humidity combination;

[0168] The first actual detected body temperature determination module 700 is configured to perform detection using the body temperature detection model based on temperature and humidity compensation based on the ambient temperature, the ambient humidity, and the body surface temperature to obtain a first actual detected body temperature;

[0169] The difference determination module 800 is configured to determine the first-order differential actual detected body temperature difference corresponding to the first actual detected body temperature, the temperature difference between the ambient temperature and the body surface temperature;

[0170] The second actual detected body temperature determination module 900 is configured to obtain a second actual detected body temperature using a body temperature prediction model based on the first-order differential actual detected body temperature difference, the temperature difference, and the first actual detected body temperature; the body temperature prediction model is a model determined based on the linear relationship between the first actual detected body temperature, the ambient temperature, the body surface temperature, and the second actual detected body temperature.

[0171] Further, based on any of the above embodiments, the above temperature detection device may further include:

[0172] The weight coefficient determination module is configured to fit a body temperature prediction basic model based on a plurality of second experimental ambient temperatures, a plurality of second experimental body surface temperatures, a plurality of first actual detected body temperatures in the second experiments, and a plurality of actual body temperatures in the second experiments to obtain various weight coefficients in the body temperature prediction basic model;

[0173] The body temperature prediction model determination module is configured to construct the body temperature prediction model based on the various weight coefficients and the body temperature prediction basic model.

[0174] Further, based on the above embodiment, the above weight coefficient determination module may include:

[0175] A second experimental data determination unit, configured to determine a second experimental ambient temperature, a second experimental body surface temperature, a first actual detected body temperature of the second experiment, and an actual body temperature of the second experiment corresponding to each temperature measurement position;

[0176] A body temperature prediction basic model fitting unit, configured to fit the body temperature prediction basic model based on the second experimental ambient temperature, the second experimental body surface temperature, the first actual detected body temperature of the second experiment, and the actual body temperature of the second experiment corresponding to each temperature measurement position, to obtain various weight coefficients corresponding to each temperature measurement position;

[0177] Correspondingly, a body temperature prediction model determination module includes:

[0178] A body temperature prediction model determination unit, configured to construct the body temperature prediction model corresponding to each temperature measurement position based on the various weight coefficients and the body temperature prediction basic model corresponding to each temperature measurement position.

[0179] Further, based on any of the above embodiments, the above body temperature detection device may further include:

[0180] A preset number of heat balance body temperature determination modules, configured to search for a preset number of heat balance body temperatures by using an approximate search model based on user data, the ambient temperature, and the ambient humidity; the user data is data related to user health;

[0181] An estimated heat balance body temperature determination module, configured to perform weighted processing on the preset number of heat balance body temperatures to obtain an estimated heat balance body temperature;

[0182] A third actual detected body temperature determination module, configured to determine a third actual detected body temperature by using a temperature prediction model adjusted based on similar user weights based on the estimated heat balance body temperature and the second actual detected body temperature; wherein, the temperature prediction model adjusted based on similar user weights is a model for adjusting the second actual detected body temperature based on the estimated heat balance body temperature.

[0183] Further, based on the above embodiments, the above body temperature detection device may further include:

[0184] A data processing module, configured to process the user data based on the body temperature value of the user's heat balance to obtain processed user data;

[0185] A grouping module, configured to group the processed user data based on the measurement position to obtain grouped data;

[0186] A feature extraction module for extracting features from the data in each grouped data to obtain user features and environmental features of each grouped data, and using the user features and the environmental features as target features; wherein, the user feature data includes at least one of user gender, user age, and BMI index, and the environmental features include the average environmental temperature, the average environmental humidity, the maximum body surface temperature, the minimum body surface temperature, and the body surface temperature change rate;

[0187] A normalization processing module for performing feature encoding on the target features to obtain feature vectors; wherein, numerical features are subjected to maximum-minimum normalization processing, and binary categorical features are subjected to one-hot encoding processing;

[0188] An approximate search model construction module for constructing a K-D tree based on the feature vectors to obtain the approximate search model.

[0189] Further, based on any of the above embodiments, the above estimated thermoneutral body temperature determination module may include:

[0190] An estimated thermoneutral body temperature determination unit for performing distance weighting based on the preset number of thermoneutral body temperatures to obtain the estimated thermoneutral body temperature; wherein, the distance weighting is a method of determining the estimated thermoneutral body temperature by weighting the body temperature values of neighbor users based on the feature distance between the current user and each neighbor user.

[0191] Further, based on any of the above embodiments, the above body temperature detection device may further include:

[0192] A health analysis module for performing health analysis on the data of the current user using a health analysis large model to obtain health analysis and suggestions; wherein, the data of the current user includes health data, the first actual detected body temperature data, living habits, and environmental data, and the health analysis large model is a model obtained by fine-tuning the large model based on the target input and the target output, the target input is the data of the current user, and the target output is the health results and health suggestions given based on each item of the user's data.

[0193] It should be noted that the order of the modules and units in the above body temperature detection device can be changed before and after without affecting the logic.

[0194] A body temperature detection device provided by an embodiment of the present invention may include: a closest temperature and humidity combination determination module 500, configured to determine the closest reference temperature and reference humidity combination corresponding to the ambient temperature and ambient humidity in the current environment; a body temperature detection model determination module 600, configured to determine a body temperature detection model based on temperature and humidity compensation based on the closest reference temperature and reference humidity combination; a first actual detected body temperature determination module 700, configured to perform detection using the body temperature detection model based on temperature and humidity compensation based on the ambient temperature, the ambient humidity, and the body surface temperature to obtain a first actual detected body temperature; a difference determination module 800, configured to determine the first-order differential actual detected body temperature difference corresponding to the first actual detected body temperature, the temperature difference between the ambient temperature and the body surface temperature; a second actual detected body temperature determination module 900, configured to obtain a second actual detected body temperature based on the first-order differential actual detected body temperature difference, the temperature difference, and the first actual detected body temperature using a body temperature prediction model; the body temperature prediction model is a model determined based on the linear relationship between the first actual detected body temperature, the ambient temperature, the body surface temperature, and the second actual detected body temperature. Compared with directly using the measured temperature as the actual body temperature of the user currently, the present application will determine the body temperature detection model based on temperature and humidity compensation corresponding to the closest reference temperature and reference humidity corresponding to the ambient temperature and ambient humidity in the current environment, and compensate the measured value of the body surface temperature using the body temperature detection model based on temperature and humidity compensation to obtain an actual body temperature closer to the actual human body temperature. Since the present application will perform compensation most suitable for the current environment based on the current ambient temperature and ambient humidity, the accuracy of body temperature detection is improved, and a second actual detected body temperature closer to the actual human body temperature will be obtained based on the body temperature prediction model, thereby improving the accuracy of the body temperature prediction value in a non-equilibrium state while shortening the equilibrium time.

[0195] The following introduces a body temperature detection device provided by an embodiment of the present invention. The body temperature detection device described below can be mutually corresponding and referred to with the body temperature detection method described above.

[0196] Please refer to Figure 10 , Figure 10 which is a schematic structural diagram of a body temperature detection device provided by an embodiment of the present invention and may include:

[0197] A memory 10, configured to store a computer program;

[0198] A processor 20, configured to execute the computer program to implement the above body temperature detection method.

[0199] The memory 10, the processor 20, and the communication interface 30 all complete communication with each other through a communication bus 40.

[0200] In an embodiment of the present invention, the memory 10 is used to store one or more programs. The program may include program code, and the program code includes computer operation instructions. In an embodiment of the present invention, a program for implementing the above-mentioned body temperature detection method may be stored in the memory 10.

[0201] In a possible implementation manner, the memory 10 may include a program storage area and a data storage area. Among them, the program storage area may store an operating system and application programs required for at least one function, etc.; the data storage area may store data created during use.

[0202] In addition, the memory 10 may include a read-only memory and a random access memory, and provide instructions and data to the processor. A part of the memory may also include NVRAM. The memory stores an operating system and operation instructions, executable modules or data structures, or subsets thereof, or extended sets thereof. Among them, the operation instructions may include various operation instructions for implementing various operations. The operating system may include various system programs for implementing various basic tasks and processing hardware-based tasks.

[0203] The processor 20 may be a central processing unit (CPU), an application-specific integrated circuit, a digital signal processor, a field programmable gate array or other programmable logic devices. The processor 20 may be a microprocessor or any conventional processor, etc. The processor 20 may call the program stored in the memory 10.

[0204] The communication interface 30 may be an interface of a communication module for connecting to other devices or systems.

[0205] Of course, it should be noted that Figure 10 the shown structure does not constitute a limitation on the body temperature detection device in the embodiment of the present invention. In actual applications, the body temperature detection device may include more or fewer components than Figure 10 those shown, or combine some components.

[0206] Next, the readable storage medium provided by the embodiment of the present invention will be introduced. The readable storage medium described below can be mutually corresponded and referred to with the above-mentioned body temperature detection method.

[0207] The present invention also provides a readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned body temperature detection method are implemented.

[0208] The readable storage medium may include: various media that can store program codes, such as USB flash drives, external hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0209] The embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. For the same or similar parts among the embodiments, reference can be made to each other. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple. For the relevant parts, reference can be made to the description in the method section.

[0210] Those skilled in the art can further realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in this article can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the components and steps of each example have been generally described according to functions in the above description. Whether these functions are executed in the form of hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0211] Finally, it should also be noted that in this article, relationships such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including", or any other variant is intended to cover non-exclusive inclusion, so that a process, method, article, or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or further includes elements inherent to such process, method, article, or device.

[0212] The above has introduced in detail a body temperature detection method, device, system, equipment, and readable storage medium provided by the present invention. Specific examples are used in this article to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A body temperature detection method, characterized in that: include: Determine the closest reference temperature and reference humidity combination corresponding to the ambient temperature and ambient humidity in the current environment; Determining a body temperature detection model based on temperature and humidity compensation based on the closest reference temperature and reference humidity combination; Based on the ambient temperature, the ambient humidity and the body surface temperature, the body temperature detection model based on temperature and humidity compensation is used to perform detection to obtain a first actually detected body temperature; Determine a first-order difference actual detected body temperature difference corresponding to the first actual detected body temperature, a temperature difference between the ambient temperature and the body surface temperature; Based on the first-order difference actual detected body temperature difference, the temperature difference and the first actual detected body temperature, a body temperature prediction model is used to obtain a second actual detected body temperature; the body temperature prediction model is a model determined based on a linear relationship between the first actual detected body temperature, the ambient temperature, the body surface temperature and the second actual detected body temperature; Based on the user data, the ambient temperature and the ambient humidity, an approximate search model is used to search and obtain a preset number of thermal equilibrium body temperatures; The user data is data related to the user's health; wherein the determination process of the approximate search model includes: processing the user data based on the body temperature value of the user's thermal balance to obtain the processed user data; The processed user data is grouped based on the measurement position to obtain grouped data; feature extraction is performed on the data in each grouped data to obtain user features and environmental features of each grouped data, and the user features and environmental features are used as target features; feature encoding is performed on the target features to obtain feature vectors; wherein, the numerical features are subjected to maximum and minimum normalization processing, and the binary category features are subjected to one-hot encoding processing; a KD tree is constructed based on the feature vector to obtain the approximate search model; Distance weighting is performed based on the preset number of thermal equilibrium body temperatures to obtain an estimated thermal equilibrium body temperature; wherein the distance weighting is a method for weighting the body temperature values ​​of neighboring users to determine the estimated thermal equilibrium body temperature based on the characteristic distance between the current user and each neighboring user; Based on the estimated thermal equilibrium body temperature and the second actual detected body temperature, a third actual detected body temperature is determined using a temperature prediction model adjusted based on similar user weights; wherein the temperature prediction model adjusted based on similar user weights is a model that adjusts the second actual detected body temperature based on the estimated thermal equilibrium body temperature.

2. The body temperature detection method according to claim 1, characterized in that: The process of determining the temperature prediction model includes: Based on multiple second experiment environment temperatures, multiple second experiment body surface temperatures, multiple first actually detected body temperatures of the second experiments, and multiple actual body temperatures of the second experiments, a body temperature prediction basic model is fitted to obtain various weight coefficients in the body temperature prediction basic model; The body temperature prediction model is constructed based on the weight coefficients and the body temperature prediction basic model.

3. The body temperature detection method according to claim 2, characterized in that: Based on the multiple second experiment environment temperatures, the multiple second experiment body surface temperatures, the multiple second experiment first actual detected body temperatures and the multiple second experiment actual body temperatures, the body temperature prediction basic model is fitted to obtain various weight coefficients in the body temperature prediction basic model, including: Determine the second experiment environment temperature, the second experiment body surface temperature, the first actual detected body temperature of the second experiment, and the actual body temperature of the second experiment corresponding to each temperature measurement position; Based on the second experimental environment temperature, the second experimental body surface temperature, the first actually detected body temperature of the second experiment, and the actual body temperature of the second experiment corresponding to each of the temperature measurement positions, the body temperature prediction basic model is fitted to obtain various weight coefficients corresponding to each of the temperature measurement positions; Accordingly, constructing the body temperature prediction model based on the weight coefficients and the body temperature prediction basic model includes: Based on the various weight coefficients corresponding to each temperature measurement position and the body temperature prediction basic model, the body temperature prediction model corresponding to each temperature measurement position is constructed.

4. A body temperature detection device, characterized in that: The body temperature detection method according to any one of claims 1 to 3 comprises: The closest temperature and humidity combination determination module is used to determine the closest reference temperature and reference humidity combination corresponding to the ambient temperature and ambient humidity in the current environment; A body temperature detection model determination module, used to determine a body temperature detection model based on temperature and humidity compensation based on the closest reference temperature and reference humidity combination; A first actually detected body temperature determination module, configured to perform detection based on the ambient temperature, the ambient humidity and the body surface temperature using the temperature detection model based on temperature and humidity compensation to obtain a first actually detected body temperature; a difference determination module, used to determine a first-order difference actual detected body temperature difference corresponding to the first actual detected body temperature, a temperature difference between the ambient temperature and the body surface temperature; a second actual detected body temperature determination module, configured to obtain a second actual detected body temperature using a body temperature prediction model based on the first-order difference actual detected body temperature difference, the temperature difference and the first actual detected body temperature; the body temperature prediction model is a model determined based on a linear relationship between the first actual detected body temperature, the ambient temperature, the body surface temperature and the second actual detected body temperature; A third actual detected body temperature determination module is used to determine the third actual detected body temperature based on the estimated thermal equilibrium body temperature and the second actual detected body temperature, using a temperature prediction model adjusted based on similar user weights; wherein the temperature prediction model adjusted based on similar user weights is a model that adjusts the second actual detected body temperature based on the estimated thermal equilibrium body temperature.

5. A body temperature detection system, characterized in that: include: A first temperature measuring element, used for detecting the ambient temperature; A second temperature measuring element is used to detect body surface temperature; Humidity measuring element, used to detect ambient humidity; A body temperature detection device, used to obtain a body temperature detection result based on the ambient temperature, the body surface temperature and the ambient humidity using the body temperature detection method according to any one of claims 1 to 3.

6. A body temperature detection device, characterized in that: include: Memory for storing computer programs; A processor, used to execute the computer program to implement the steps of the body temperature detection method as claimed in any one of claims 1 to 3.

7. A readable storage medium, characterized in that: The readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the body temperature detection method according to any one of claims 1 to 3 are implemented.

Images