Data calibration method, device, equipment and medium for medical-grade high-precision temperature sensor

Abstract

The present invention relates to a data calibration method, device, equipment and medium for a medical-grade high-precision temperature sensor, comprising the following steps: curve drawing based on the preprocessed temperature data set and the second collected temperature set to obtain a first temperature curve and a second temperature curve; time aligning the first temperature curve and the second temperature curve, identifying the offset point between the first temperature curve and the second temperature curve through a preset temperature offset detection algorithm, and marking the offset point on the first temperature curve to obtain a marked temperature curve; response analysis of the temperature sensor to be calibrated based on the marked temperature curve to obtain sensor response characteristic parameters, and calibrating the temperature sensor to be calibrated based on the sensor response characteristic parameters, thereby solving the technical problem that the temperature sensor may have errors of varying degrees due to differences in manufacturing processes, changes in the use environment, and aging caused by long-term use.

Description

Technical Field

[0001] The present invention relates to the technical field of temperature sensors, and in particular to a data calibration method, device, equipment and medium for a medical-grade high-precision temperature sensor. Background Art

[0002] In medical environments, accurate temperature measurement is essential to ensure patient safety and treatment effectiveness. With the advancement of modern medical technology, temperature-sensitive treatments such as hyperthermia, cryotherapy, and various in vitro diagnostic procedures rely on high-precision temperature sensors to provide reliable data support. However, due to differences in manufacturing processes, changes in the use environment, and aging caused by long-term use, temperature sensors may have varying degrees of errors, which poses a potential threat to the effectiveness and safety of medical operations. Therefore, developing a reliable calibration method to ensure that temperature sensors maintain their accuracy throughout their service life has become an important task in the field of medical engineering.

[0003] Existing temperature sensor calibration methods usually rely on standard reference comparison under static or semi-static conditions. Although this method can ensure the accuracy of calibration to a certain extent, it is difficult to adapt to the complex dynamic environmental changes in actual applications. For example, in medical scenarios such as operating rooms or intensive care units (ICUs), temperature sensors need to cope with rapidly changing temperature conditions and accurately reflect small temperature fluctuations. Traditional calibration techniques often ignore the impact of these dynamic factors, resulting in large deviations in actual use of calibrated sensors. In addition, traditional methods are mostly manual operations, which are not only time-consuming and labor-intensive, but also prone to human errors, reducing the consistency and reliability of the calibration results.

[0004] To address the above issues, this study proposed a medical-grade high-precision temperature sensor data calibration method based on a dynamic time warping algorithm and a temperature offset detection algorithm. This method synchronously collects data from the sensor to be calibrated and the target sensor of known accuracy, and uses advanced algorithms for data analysis and processing. It can effectively identify and correct problems such as nonlinearity and time delay in sensor response, thereby improving the measurement accuracy of the sensor in a dynamic environment. At the same time, the automated calibration model is used to accurately analyze the sensor response characteristics, further optimize the calibration process, reduce manual intervention, and enhance the stability and repeatability of the calibration results, thereby providing a more reliable technical guarantee for temperature monitoring in the medical field. Summary of the invention

[0005] The main purpose of the present invention is to provide a data calibration method, device, equipment and medium for a medical-grade high-precision temperature sensor, which solves the technical problem that the temperature sensor may have errors of varying degrees due to differences in manufacturing processes, changes in the use environment and aging caused by long-term use.

[0006] To achieve the above object, the present invention provides a data calibration method for a medical-grade high-precision temperature sensor, comprising the following steps:

[0007] The temperature of the designated medical area is collected by using a preset temperature sensor to be calibrated and a target sensor to obtain a corresponding first collected temperature set and a second collected temperature set;

[0008] Preprocessing the first collected temperature set to obtain a preprocessed temperature data set;

[0009] Plotting a curve based on the preprocessed temperature data set and the second collected temperature set to obtain a first temperature curve and a second temperature curve;

[0010] Using a preset dynamic time warping algorithm, the first temperature curve and the second temperature curve are time-aligned, and the offset point between the first temperature curve and the second temperature curve is identified by a preset temperature offset detection algorithm, and the offset point is marked on the first temperature curve to obtain a marked temperature curve;

[0011] By using a preset temperature sensor calibration model, a response analysis is performed on the temperature sensor to be calibrated based on the annotated temperature curve to obtain a sensor response characteristic parameter;

[0012] The temperature sensor to be calibrated is calibrated based on the sensor response characteristic parameter.

[0013] Furthermore, the temperature of the designated medical area is collected by using the preset temperature sensor to be calibrated and the target sensor to obtain the corresponding first collected temperature set and the second collected temperature set, including:

[0014] The designated medical area is spatially divided by a preset Voronoi diagram algorithm to obtain a plurality of divided sub-areas;

[0015] The temperature sensor to be calibrated and the target sensor are respectively deployed at the central coordinates of each of the divided sub-areas to obtain deployed sensors; wherein the deployed sensors include the deployed temperature sensor to be calibrated and the deployed target sensor;

[0016] Using a preset network time protocol to synchronize the time of the deployed sensors to obtain a uniformly set time reference;

[0017] Based on the time reference, synchronously start all the deployed sensors to collect temperature data to obtain a time-synchronized temperature data stream;

[0018] Through a multi-channel data acquisition system, the temperature data streams collected by each of the deployed sensors are constructed into a temperature matrix in a timestamp order to obtain a first temperature matrix and a second temperature matrix, wherein the first temperature matrix corresponds to the temperature sensor to be calibrated, the second temperature matrix corresponds to the target sensor, the first temperature matrix includes first temperature matrix rows, first temperature matrix columns and first temperature matrix elements, the first temperature matrix rows represent time series, the first temperature matrix columns represent the center coordinates of each of the divided sub-regions, the first temperature matrix elements represent the temperature values ​​of the temperature sensor to be calibrated, and the second temperature matrix includes second temperature matrix rows, second temperature matrix columns and second temperature matrix elements, the second temperature matrix rows represent time series, the second temperature matrix columns represent the center coordinates of each of the divided sub-regions, and the second temperature matrix elements represent the temperature values ​​of the target sensor;

[0019] The temperature data in the designated time period and the divided sub-areas are extracted from the first temperature matrix and the second temperature matrix respectively to obtain the corresponding first collected temperature set and the second collected temperature set.

[0020] Further, the preprocessing of the first collected temperature set to obtain a preprocessed temperature data set includes:

[0021] Performing signal noise reduction processing on the first collected temperature set through a preset Butterworth bandpass filter to obtain a filtered temperature data set;

[0022] Performing multi-scale decomposition on the filtered temperature data set based on a wavelet packet decomposition algorithm to obtain multi-scale decomposed temperature data;

[0023] By using a preset local outlier factor algorithm, identifying and detecting whether there are outliers in the multi-scale decomposition temperature data, and if so, removing the outliers in the multi-scale decomposition temperature data to obtain a temperature data set after removal; wherein the outliers are temperature values ​​that deviate from a normal range;

[0024] The cubic spline interpolation algorithm is used to interpolate and supplement the missing temperature data in the temperature data set after elimination, and the interpolation supplement temperature data set is obtained;

[0025] The interpolated supplementary temperature data set is subjected to data smoothing processing to obtain a preprocessed temperature data set.

[0026] Further, the first temperature curve and the second temperature curve are time-aligned by using a preset dynamic time warping algorithm, and the offset point between the first temperature curve and the second temperature curve is identified by using a preset temperature offset detection algorithm, and the offset point is marked on the first temperature curve to obtain the marked temperature curve, including:

[0027] Performing curve fitting on the first temperature curve and the second temperature curve respectively by using a preset Gaussian process regression technology to obtain a corresponding first temperature curve function and a corresponding second temperature curve function;

[0028] Based on a dynamic time warping algorithm, time series matching is performed on the first temperature curve function and the second temperature curve function respectively to obtain a time warping path matrix; wherein the time warping path matrix includes time warping path rows, time warping path columns and time warping path elements, the rows of the time warping paths represent the time points of the first temperature curve function, the columns of the time warping paths represent the time points of the second temperature curve function, and the elements of the time warping paths represent the measured temperatures of the temperature sensor to be calibrated and the target sensor at the same time;

[0029] Extracting matrix temperature values ​​on the diagonal from the time-warped path matrix to obtain a first temperature sequence and a second temperature sequence;

[0030] Calculating the difference between the first temperature sequence and the second temperature sequence to obtain a temperature difference sequence;

[0031] Perform sliding window processing on the temperature difference sequence according to a preset temperature offset detection algorithm to obtain a temperature difference mean sequence within the sliding window;

[0032] Performing offset point detection on the temperature difference mean sequence to obtain an offset point index set;

[0033] According to the offset point index set, offset points are marked on the first temperature curve to obtain a marked temperature curve.

[0034] Furthermore, extracting the matrix temperature values ​​on the diagonal from the time-warping path matrix to obtain the first temperature sequence and the second temperature sequence includes:

[0035] Based on the dynamic time warping algorithm, the main diagonal elements of the time warping path matrix are identified to obtain a main diagonal element set;

[0036] Extracting corresponding temperature values ​​of the first temperature curve from the main diagonal element set to form a first temperature sequence; wherein the first temperature sequence reflects the temperature change trend of the first temperature curve after time alignment;

[0037] The corresponding temperature values ​​of the second temperature curve are extracted from the main diagonal element set to form a second temperature sequence; wherein the second temperature sequence reflects the temperature change trend of the second temperature curve after time alignment.

[0038] Furthermore, the sensor response characteristic parameters include a temperature measurement sensitivity coefficient and a temperature drift compensation factor. The sensor response characteristic parameters obtained by performing response analysis on the temperature sensor to be calibrated based on the marked temperature curve through a preset temperature sensor calibration model include:

[0039] Obtaining a component topology diagram of the temperature sensor to be calibrated; wherein the component topology diagram describes the connection relationship and physical position of each component inside the temperature sensor to be calibrated;

[0040] Perform multi-scale feature extraction on the annotated temperature curve to obtain multi-scale temperature features; wherein the multi-scale temperature features include maximum temperature, minimum temperature, mean temperature, temperature variance, temperature peak time, temperature rise time and temperature drop time;

[0041] Reconstructing the topological structure of the multi-scale temperature characteristics by a complex network analysis algorithm to obtain a temperature sensor characteristic network;

[0042] Performing principal component analysis on the temperature sensor characteristic network to obtain a principal component characteristic network;

[0043] Mapping the principal component characteristic network onto the component topology map to obtain a mapping characteristic network; wherein the mapping process is based on the functions of the components in the temperature sensor to be calibrated and their influence on the temperature characteristics;

[0044] Based on the mapped feature network, the contribution of each component to the main component feature is analyzed to obtain the component health status index; wherein the component health status index includes stability, sensitivity and critical point;

[0045] Calculating the entropy of the multi-scale temperature feature to obtain entropy feature parameters, including information entropy and conditional entropy;

[0046] Performing correlation analysis on the entropy value characteristic parameters by using a preset least square method to obtain a correlation characteristic vector; wherein the correlation characteristic vector includes a correlation coefficient and a partial correlation coefficient;

[0047] Through a preset temperature sensor calibration model, based on the correlation feature vector and the component health status index, a response analysis is performed on the temperature sensor to be calibrated to obtain sensor response characteristic parameters; wherein the sensor response characteristic parameters include a temperature measurement sensitivity coefficient and a temperature drift compensation factor.

[0048] Furthermore, the multi-scale temperature feature is topologically reconstructed by a complex network analysis algorithm to obtain a temperature sensor feature network, including:

[0049] Performing Spearman rank correlation analysis on the multi-scale temperature features to obtain a feature correlation matrix;

[0050] The characteristic association matrix is ​​network constructed by a minimum spanning tree algorithm to obtain an initial temperature sensor characteristic network;

[0051] By using a preset PageRank algorithm, node importance evaluation is performed on the initial temperature sensor feature network to obtain a node importance ranking;

[0052] According to the node importance sorting, the edge weights of the initial temperature sensor feature network are adjusted to obtain an adjusted temperature sensor feature network;

[0053] Using a preset Louvain community discovery algorithm, community discovery is performed on the adjusted temperature sensor feature network to obtain divided temperature feature communities;

[0054] Based on the divided temperature feature communities, the adjusted temperature sensor feature network is topologically optimized to obtain a temperature sensor feature network; wherein the temperature sensor feature network includes network nodes and edges, the network nodes represent temperature feature points, the edges represent the association strength between the temperature feature points, and the temperature feature points represent the temperature measurement results at different physical locations.

[0055] The present invention also provides a data calibration device for a medical-grade high-precision temperature sensor, comprising:

[0056] The acquisition module is used to respectively acquire the temperature of the designated medical area through the preset temperature sensor to be calibrated and the target sensor to obtain the corresponding first acquisition temperature set and the second acquisition temperature set;

[0057] A preprocessing module, used for preprocessing the first collected temperature set to obtain a preprocessed temperature data set;

[0058] A drawing module, used for drawing a curve based on the preprocessed temperature data set and the second collected temperature set to obtain a first temperature curve and a second temperature curve;

[0059] A marking module, used to time-align the first temperature curve and the second temperature curve using a preset dynamic time warping algorithm, identify the offset point between the first temperature curve and the second temperature curve using a preset temperature offset detection algorithm, and mark the offset point on the first temperature curve to obtain a marked temperature curve;

[0060] An analysis module, used to analyze the annotated temperature curve through a preset temperature sensor calibration model to obtain calibration parameters; wherein the calibration parameters include an offset, a proportional factor, and a nonlinear correction coefficient;

[0061] A calibration module is used to calibrate the temperature sensor to be calibrated based on the calibration parameters.

[0062] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.

[0063] The present invention also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of any of the above-mentioned methods are implemented.

[0064] The data calibration method of the medical-grade high-precision temperature sensor provided by the present invention comprises the following steps: respectively collecting temperature of a designated medical area by using a preset temperature sensor to be calibrated and a target sensor to obtain a corresponding first collected temperature set and a second collected temperature set; preprocessing the first collected temperature set to obtain a preprocessed temperature data set; drawing curves based on the preprocessed temperature data set and the second collected temperature set to obtain a first temperature curve and a second temperature curve; using a preset dynamic time warping algorithm, time-aligning the first temperature curve and the second temperature curve, identifying the offset point between the first temperature curve and the second temperature curve by using a preset temperature offset detection algorithm, and marking the offset point on the first temperature curve to obtain a marked temperature curve; using a preset temperature sensor calibration model, based on the marked temperature curve, performing response analysis on the temperature sensor to be calibrated to obtain a sensor response characteristic parameter, and calibrating the temperature sensor to be calibrated based on the sensor response characteristic parameter. Through the above technical means, the technical problem that the temperature sensor will have different degrees of error due to differences in manufacturing processes, changes in the use environment, and aging caused by long-term use is solved, and it is not only applicable to calibration under static conditions, but also optimized for the rapidly changing temperature conditions that may be encountered in practical applications. By preprocessing the first collected temperature set and drawing a temperature curve based on the preprocessed data, the calibration process can better reflect the temperature changes in the actual usage scenario, thereby enhancing the adaptability and flexibility of the calibration method for different application scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 1 is a schematic diagram of the steps of a data calibration method for a medical-grade high-precision temperature sensor in one embodiment of the present invention;

[0066] Figure 2 It is a structural block diagram of a data calibration device for a medical-grade high-precision temperature sensor in one embodiment of the present invention;

[0067] Figure 3 It is a schematic block diagram of the structure of a computer device according to an embodiment of the present invention.

[0068] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0069] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0070] like Figure 1 As shown, Figure 1 This is a schematic diagram of the steps of a data calibration method for a medical-grade high-precision temperature sensor in one embodiment of the present invention;

[0071] In one embodiment of the present invention, a data calibration method for a medical-grade high-precision temperature sensor is provided, comprising the following steps:

[0072] Step S1 : temperature collection is performed on a designated medical area by using a preset temperature sensor to be calibrated and a target sensor respectively, so as to obtain a corresponding first collection temperature set and a second collection temperature set.

[0073] Specifically, in the data calibration method of medical-grade high-precision temperature sensors, it is first necessary to collect the temperature of the designated medical area through the preset temperature sensor to be calibrated and the target sensor respectively, and obtain the corresponding first collected temperature set and second collected temperature set. This step is the basis of the entire calibration process, which requires the use of two sensors to work simultaneously: one is the temperature sensor to be calibrated that needs to be calibrated, and the other is the target sensor that has been strictly verified and its accuracy is fully trusted. The two sensors are deployed in the same specific medical environment, such as in an operating room or a treatment area in an intensive care unit (ICU), to ensure that they can synchronously record temperature changes in the same environment. To achieve this goal, the data acquisition plan must be carefully designed and implemented. This means that not only the installation position of the sensors must be considered to ensure that they measure exactly the same temperature field, but also the synchronization between the two must be ensured, that is, to ensure that they start and end the collection of temperature data at the same time point. In addition, considering that there may be various interference factors in the medical environment, such as equipment heating, personnel activities, etc., it is also crucial to choose a suitable measurement time period, and these time periods that may affect the accuracy of temperature readings should be avoided as much as possible. For example, in a scenario where hyperthermia treatment is being performed, we can choose to collect temperature during a period when no other high-power equipment is started to reduce the impact of external factors on the measurement results. In this process, the temperature sensor to be calibrated and the target sensor will independently collect a series of temperature readings to form a first collection temperature set and a second collection temperature set. The first collection temperature set refers to the set of all temperature values ​​recorded by the temperature sensor to be calibrated, while the second collection temperature set represents the corresponding temperature data obtained by the target sensor. These two data sets will serve as the basis for subsequent analysis and processing to evaluate the performance of the temperature sensor to be calibrated and provide the necessary calibration basis for it. For example, in actual applications, if the calibration is performed in an environment such as an operating room, the data obtained from the two sensors will reflect the changes in the air temperature in the operating room over time. By comparing and analyzing these data, it is possible to identify whether the temperature sensor to be calibrated has deviations and the specific form of the deviation, thereby providing important reference information for the subsequent calibration work. In order to ensure that the collected data is sufficiently representative, multiple repeated measurements are usually required. This not only improves the quality of the data, but also helps detect and eliminate any accidental errors or outliers. In short, through this method, we can effectively establish a reliable data foundation, which lays a solid foundation for the subsequent temperature curve drawing, the application of dynamic time warping algorithm and the final analysis of sensor response characteristic parameters. The successful execution of this step plays a vital role in ensuring the effectiveness and accuracy of the entire calibration process.

[0074] Step S2: preprocessing the first collected temperature set to obtain a preprocessed temperature data set.

[0075] Specifically, in the data calibration method of the medical-grade high-precision temperature sensor, preprocessing the first collected temperature set to obtain a preprocessed temperature data set is a key step to ensure the accuracy and reliability of subsequent analysis. This process involves a series of technical processing of the first collected temperature set collected by the temperature sensor to be calibrated to eliminate possible noise, outliers and other abnormal factors, so as to obtain a purer and smoother temperature data set, namely, a preprocessed temperature data set. The purpose of preprocessing is to improve the data quality so that the subsequent curve drawing and dynamic time warping algorithm application based on these data can more accurately reflect the actual situation. The first step of preprocessing is usually to remove noise. In actual application scenarios, such as operating rooms or intensive care units (ICUs), due to the complex and changeable environment, temperature sensors may be affected by factors such as electromagnetic interference from surrounding equipment and airflow changes caused by people walking, resulting in the collected temperature data containing a certain amount of random noise. In order to reduce this effect, after obtaining the first collected temperature set, filtering technology can be used to smooth the data. Common filtering methods include moving average filtering, low-pass filtering, etc., which can effectively reduce the noise level while retaining the original signal characteristics. For example, in a scenario where hyperthermia treatment is being performed, by applying appropriate filters, we can filter out those instantaneous temperature fluctuations caused by the startup or shutdown of the device, so that the temperature data can more realistically reflect the actual temperature changes on the patient's body surface or in the environment. Next, outliers need to be identified and processed. Outliers refer to temperature readings that deviate significantly from the normal range, which may be caused by sensor failure, severe external interference, or other unknown reasons. If not handled, these outliers will seriously affect the final calibration results. Therefore, in the preprocessing stage, outliers can be detected by setting a reasonable threshold range or using statistical methods, and decide whether to remove or correct them. For example, in the above medical environment, if the temperature reading at a certain time point is far beyond the expected range of change, then we should further check it. If it is confirmed to be an outlier, it can be interpolated and filled based on the data of the adjacent time points before and after to ensure the continuity and consistency of the data. In addition, preprocessing may also involve data standardization or normalization operations. This is because there may be inconsistent measurement units between different sensors, or some specific algorithms require the input data to have specific distribution characteristics. By uniformly converting all temperature data, subsequent comparison and analysis can be made easier and more intuitive. For example, when comparing the first temperature set collected with the second temperature set collected, it is very important to ensure that both sets of data are within the same scale range. This not only helps improve the efficiency of the dynamic time warping algorithm, but also enhances the accuracy of the temperature excursion detection algorithm.In summary, by performing careful and comprehensive preprocessing on the first collected temperature set, we can obtain a more accurate and stable preprocessed temperature data set, which provides a solid foundation for drawing temperature curves, performing time alignment, and identifying offset points. At the same time, a high-quality preprocessed temperature data set is also conducive to improving the effect of the entire calibration process, ensuring that the calibrated temperature sensor can provide reliable temperature monitoring services in practical applications, and ensuring medical safety and treatment effects.

[0076] Step S3, plotting curves based on the preprocessed temperature data set and the second collected temperature set to obtain a first temperature curve and a second temperature curve.

[0077] Specifically, in the data calibration method of the medical-grade high-precision temperature sensor, the process of drawing a curve based on the preprocessed temperature data set and the second collected temperature set to obtain the first temperature curve and the second temperature curve is a key link in the entire calibration process. This process not only intuitively shows the relationship between the two temperature data sets, but also provides a visual basis for subsequent time alignment and offset point identification. In order to achieve this goal, it is necessary to convert the preprocessed first collected temperature set and the second collected temperature set directly obtained from the target sensor into a graphical representation, namely, a temperature curve. First of all, it should be understood that the "curve drawing" mentioned here is not simply connecting data points to form a line, but a process involving data analysis, selecting appropriate chart types, and ensuring accurate data display. In application scenarios such as operating rooms or intensive care units (ICUs), such as during hyperthermia treatment, the ambient temperature changes over time. Therefore, by drawing the preprocessed first collected temperature set into a curve, the trend of the temperature change over time recorded by the temperature sensor to be calibrated can be clearly reflected, which is the so-called first temperature curve. At the same time, the second temperature set collected by the target sensor is also plotted as another curve, namely the second temperature curve, which represents the change of the actual temperature in the area during the same time period. The two curves should be plotted in the same coordinate system to facilitate direct comparison of the similarities and differences between the two. When plotting the curves, the timestamp of the data must be taken into account to ensure the correspondence of the two curves on the time axis. Because even if the data is collected in the same environment, there may be a slight time difference due to the different sensor startup time and response speed. To this end, before plotting, the two data sets should be preliminarily aligned in time to ensure that they have corresponding temperature readings at the same time point. This step can be done by matching the time points with obvious characteristics in the two data sets, such as the moment when the temperature rises or falls significantly. For those cases where there are no clear feature points, interpolation or other mathematical methods can be used to estimate the missing time point data, so that the two curves can more accurately reflect the relationship between each other. In addition, in order to improve the quality of curve drawing, it is also necessary to select appropriate chart types and visual elements. For example, a line chart can be selected as the main chart form because it can well show the trend of continuous variables over time. At the same time, auxiliary elements such as grid lines and annotations can also be added to help observers interpret the curve information more easily. In addition, the choice of color is also very important. Different colors can help distinguish the first temperature curve from the second temperature curve, and can also be used to highlight certain specific temperature ranges or events, such as the critical temperature threshold reached during hyperthermia. These visual design details help enhance the expressiveness of the chart, so that even non-professionals can quickly understand the information conveyed by the two curves.In summary, by drawing carefully designed curves for the preprocessed temperature data set and the second collected temperature set, we can not only intuitively see the comparison of the measurement results between the temperature sensor to be calibrated and the target sensor, but also lay the foundation for a more in-depth time alignment analysis using the dynamic time warping algorithm. More importantly, this visual expression helps to identify possible system errors or abnormalities, thereby guiding subsequent calibration work and ensuring that the final temperature sensor can provide highly accurate and reliable temperature monitoring services in practical applications, ensuring medical safety and treatment effects.

[0078] Step S4, using a preset dynamic time warping algorithm, time-aligns the first temperature curve and the second temperature curve, and identifies the offset point between the first temperature curve and the second temperature curve through a preset temperature offset detection algorithm, and marks the offset point on the first temperature curve to obtain a marked temperature curve.

[0079] Specifically, in the data calibration method of the medical-grade high-precision temperature sensor, the first temperature curve and the second temperature curve are time-aligned using a preset dynamic time warping algorithm, and the offset point between the first temperature curve and the second temperature curve is identified by a preset temperature offset detection algorithm, and the offset point is marked on the first temperature curve to obtain the marked temperature curve. This process is an important step to ensure the accuracy of the calibration result. This process requires not only precise time alignment technology, but also an effective offset point identification mechanism to ensure that the measurement differences that may exist between the two sensors can be accurately captured. First, in application scenarios such as operating rooms or intensive care units (ICUs), such as during hyperthermia treatment, the ambient temperature changes over time. Because the temperature sensor to be calibrated and the target sensor may have slight differences in response speed, or because of installation location, equipment startup delay, etc., the two temperature curves may be inconsistent on the time axis. In order to eliminate this time deviation, the dynamic time warping (Dynamic Time Warping, DTW) algorithm is used to time-align the first temperature curve and the second temperature curve. The DTW algorithm is a method that can handle nonlinear time series. It finds the best matching path between two sequences so that the most similar parts between the two can be found even if there is temporal stretching or compression. Specifically, the algorithm constructs a distance matrix between the two temperature curves, and then finds a path that minimizes the accumulated distance along this path to achieve the best match between the two. This step ensures that the two temperature curves can remain consistent in the time dimension even in a complex dynamic environment, providing a reliable basis for subsequent analysis. After completing the time alignment, the next task is to identify the offset points between the first temperature curve and the second temperature curve. This is because even after time alignment, the two sensors may still record different temperature values ​​due to their own characteristics or external factors. For this reason, the preset temperature offset detection algorithm comes into play. The algorithm compares the difference between the two temperature curves at each time point based on a certain mathematical model or statistical rule. For example, during the hyperthermia treatment process, if the temperature displayed by the first temperature curve at a certain moment is significantly higher or lower than that of the second temperature curve, and this difference exceeds the set threshold range, then the time point is determined to be an offset point. These offset points may be caused by sensor aging, decreased sensitivity, improper installation position or other unknown reasons. Identifying these offset points is crucial to understanding the error pattern of the sensor, which helps to make targeted calibration adjustments later. Finally, after all the offset points are identified, the offset points will be marked on the first temperature curve to form a marked temperature curve. This operation is not just a simple mark, but it intuitively reflects the offset information on the graph, which is convenient for technicians to further analyze.For example, the offset points can be highlighted by changing the color, adding symbols or annotations, so that the observer can see at a glance which parts have measurement differences. In addition, other metadata, such as the environmental conditions at the time, the operating status of surrounding equipment, etc., can be combined to help more deeply understand the cause of the offset. The annotated temperature curve provides a visual reference for the final calibration, making the subsequent analysis of sensor response characteristic parameters more targeted and also pointing out the direction for improving sensor performance. In short, by using the dynamic time warping algorithm and the temperature offset detection algorithm, we not only achieved the precise alignment of the first temperature curve and the second temperature curve in time, but also successfully identified and annotated the offset points between the two, laying a solid foundation for subsequent calibration work. This method ensures that even in complex and changing medical environments, temperature sensors can provide accurate and reliable temperature monitoring services, ensuring the safety and effectiveness of medical services.

[0080] Step S5, using a preset temperature sensor calibration model, based on the annotated temperature curve, a response analysis is performed on the temperature sensor to be calibrated to obtain sensor response characteristic parameters.

[0081] Specifically, in the data calibration method of medical-grade high-precision temperature sensors, the response analysis of the temperature sensor to be calibrated is performed based on the annotated temperature curve through a preset temperature sensor calibration model, and obtaining the sensor response characteristic parameters is the core link in the entire calibration process. This step aims to use the first temperature curve that has been previously processed by time alignment and offset point annotation to deeply understand the behavior pattern of the temperature sensor to be calibrated and quantify its response characteristics to ensure the accuracy and effectiveness of subsequent calibration work. First, in application scenarios such as operating rooms or intensive care units (ICUs), such as during hyperthermia treatment, the ambient temperature changes over time, and the annotated temperature curve provides technicians with an intuitive and detailed reference framework. All identified offset points are marked on this curve, which reflect the measurement differences between the temperature sensor to be calibrated and the target sensor. In order to further analyze the reasons behind these differences, it is necessary to use a preset temperature sensor calibration model to perform response analysis. This model is a tool built on mathematical algorithms and technical theories. It can simulate and calculate the behavior of the sensor to be calibrated under different conditions based on known standard sensor data (i.e., the second temperature curve) and the first temperature curve that has been preprocessed, time-aligned and annotated. Next, the response analysis process involves inputting the annotated temperature curve into the calibration model to examine the temperature readings at each moment in detail. The model considers a variety of factors, such as the sensor's time constant, linearity, sensitivity, etc., which are key properties that affect the accuracy of temperature measurement. By evaluating these characteristics, the degree of deviation of the temperature sensor to be calibrated relative to the standard sensor can be determined and quantified as specific response characteristic parameters. For example, during hyperthermia treatment, if the sensor to be calibrated is found to exhibit hysteresis during the rapid heating phase, it may mean that it has a slow response speed; or if the sensor shows a nonlinear response within a certain temperature range, it indicates that its sensitivity is unstable within this range. This information is very important for understanding the working principle of the sensor and its potential limitations. In addition, in order to improve the accuracy of response analysis, machine learning or statistical methods can also be introduced as a supplementary means. By training the model using a large amount of historical data, including temperature curves collected under different environmental conditions, the calibration model can be made to more intelligently predict and explain the current sensor's performance. For example, using supervised learning algorithms, model parameters can be automatically adjusted based on the input-output relationship in existing cases to more accurately fit the actual measurement results. This method not only enhances the understanding and description of sensor response characteristics, but also provides a scientific basis for developing personalized calibration strategies. Finally, based on the sensor response characteristic parameters obtained from the above response analysis, the settings or structures of the temperature sensor to be calibrated can be adjusted in a targeted manner to reduce its error and optimize its performance.For example, in a hyperthermia treatment scenario, if a certain type of sensor is found to have a large measurement deviation in the low temperature area, this problem can be corrected by software compensation or hardware improvement. At the same time, these parameters can also be used to guide the design and manufacture of future sensors, helping engineers to better control product quality and ensure that every medical-grade temperature sensor shipped meets strict accuracy requirements. In summary, response analysis through the preset temperature sensor calibration model enables us to fully and deeply understand the response characteristics of the temperature sensor to be calibrated, and then obtain a set of reliable response characteristic parameters. This step not only provides solid data support for the final calibration, but also promotes the continuous advancement and improvement of temperature sensor technology, ensuring that they can provide accurate and reliable temperature monitoring services in complex medical environments to protect the health and safety of patients.

[0082] Step S6: calibrating the temperature sensor to be calibrated based on the sensor response characteristic parameters.

[0083] Specifically, calibrating the temperature sensor to be calibrated based on the sensor response characteristic parameters is the final step of the entire medical-grade high-precision temperature sensor calibration process, and is also a key step to ensure that the sensor provides accurate measurement results in practical applications. In this process, the sensor response characteristic parameters obtained by the previous response analysis, such as time constant, linearity and sensitivity, are used to adjust the behavior of the temperature sensor to be calibrated so that its output is closer to the ideal value. Specifically, in application scenarios such as operating rooms or intensive care units (ICUs), such as during hyperthermia treatment, the ambient temperature changes are complex and changeable, and these response characteristic parameters provide detailed calibration guidance for technicians. Through a deep understanding of these parameters, the specific deviation of the sensor to be calibrated relative to the standard sensor can be determined, and corresponding correction measures can be formulated accordingly. For example, if a sensor is found to exhibit a significant hysteresis effect in the rapid heating stage, indicating that its time constant is large, then its response speed can be accelerated by adjusting the algorithm or hardware settings inside the sensor. For example, if the sensor shows a nonlinear response in a certain temperature range, this means that the sensitivity of the range is unstable. At this time, a software compensation mechanism can be introduced to correct it according to the known nonlinear characteristics so that the sensor can maintain a consistent sensitivity throughout the entire working range. In addition, in order to achieve accurate calibration, mathematical models can also be used to convert response characteristic parameters into specific correction formulas or coefficients. These formulas or coefficients can be directly applied to the data processing logic of the sensor to automatically correct the results of each measurement. For example, in a hyperthermia treatment scenario, when the sensor detects the temperature, the system automatically applies a pre-calculated correction factor to adjust the raw data to eliminate errors caused by sensor characteristics. This method not only improves the calibration efficiency, but also ensures the consistency and reliability of the calibration results. In summary, the process of calibration based on the sensor response characteristic parameters combines theoretical analysis with practical application, ensuring that the temperature sensor to be calibrated can provide highly accurate and stable temperature monitoring services in complex medical environments. This process is crucial to ensuring the safety and effectiveness of medical services, and it also helps to promote the continuous improvement and development of temperature sensing technology.

[0084] In a specific embodiment, the temperature of the designated medical area is collected by the preset temperature sensor to be calibrated and the target sensor respectively to obtain the corresponding first collected temperature set and the second collected temperature set, including:

[0085] The designated medical area is spatially divided by a preset Voronoi diagram algorithm to obtain a plurality of divided sub-areas;

[0086] The temperature sensor to be calibrated and the target sensor are respectively deployed at the central coordinates of each of the divided sub-areas to obtain deployed sensors; wherein the deployed sensors include the deployed temperature sensor to be calibrated and the deployed target sensor;

[0087] Using a preset network time protocol to synchronize the time of the deployed sensors to obtain a uniformly set time reference;

[0088] Based on the time reference, synchronously start all the deployed sensors to collect temperature data to obtain a time-synchronized temperature data stream;

[0089] Through a multi-channel data acquisition system, the temperature data streams collected by each of the deployed sensors are constructed into a temperature matrix in a timestamp order to obtain a first temperature matrix and a second temperature matrix, wherein the first temperature matrix corresponds to the temperature sensor to be calibrated, and the second temperature matrix corresponds to the target sensor; wherein the first temperature matrix includes first temperature matrix rows, first temperature matrix columns and first temperature matrix elements, the first temperature matrix rows represent time series, the first temperature matrix columns represent the center coordinates of each of the divided sub-areas, and the first temperature matrix elements represent the temperature values ​​of the temperature sensors to be calibrated, and the second temperature matrix includes second temperature matrix rows, second temperature matrix columns and second temperature matrix elements, the second temperature matrix rows represent time series, the second temperature matrix columns represent the center coordinates of each of the divided sub-areas, and the second temperature matrix elements represent the temperature values ​​of the target sensors;

[0090] The temperature data in the designated time period and the divided sub-areas are extracted from the first temperature matrix and the second temperature matrix respectively to obtain the corresponding first collected temperature set and the second collected temperature set.

[0091] Specifically, first of all, in application scenarios such as operating rooms or intensive care units (ICUs), such as during hyperthermia treatment, changes in ambient temperature are crucial to the treatment effect of patients. In order to ensure the accuracy of the measurement results, a scientific method must be used to optimize the spatial layout of sensors. The Voronoi diagram algorithm is introduced here, which is an effective tool for spatial partitioning. By applying the Voronoi diagram algorithm to the specified medical area, the entire monitoring area can be divided into multiple non-overlapping sub-areas, each of which contains a unique center point. This method can ensure that each sensor can cover the temperature changes within a certain range around it, while avoiding the problem of redundant data caused by the close distance between sensors. For example, in a standard-sized operating room, the optimal number of sub-areas calculated by the Voronoi diagram algorithm may be six, each of which has a clear boundary, and the distance between each other is carefully designed to ensure the comprehensiveness and representativeness of temperature data collection. Next, the temperature sensor to be calibrated and the target sensor need to be deployed at the center coordinates of each sub-area to form a deployed sensor. This means that each sub-area will be equipped with a pair of sensors-one as the object to be calibrated and the other as a reference standard. This arrangement ensures that each pair of sensors works in the same environment, thereby reducing the impact of external factors on the calibration results. For example, in the above example of the operating room, each of the six divided sub-areas will be installed with a temperature sensor to be calibrated and a target sensor, which are precisely placed at the geometric center of their respective sub-areas. This not only enables the two sensors to record temperature changes at the same location at the same time, but also facilitates subsequent data analysis and comparison. In order to further improve the consistency of data collection, it is necessary to use the preset network time protocol (NTP) to synchronize the time of all deployed sensors and establish a unified time base. Time synchronization is the basis for ensuring that all sensors start and end data collection at the same time. Only in this way can the data recorded by different sensors be comparable. With the help of the network time protocol, the internal clocks of all sensors can be calibrated to a common time source, such as the standard time provided by the Internet time server. This step eliminates errors caused by clock drift or setting differences and ensures that the timestamps of the temperature data stream are completely consistent. For example, in the application scenario of the operating room, once all sensors have completed time synchronization, no matter which sensor receives the temperature change signal first, they will record the time of occurrence of this event according to the unified time base, providing a reliable time reference for subsequent data processing. Based on time synchronization, all deployed sensors will start collecting temperature data synchronously, generating a time-synchronized temperature data stream. This means that when the predetermined start time is reached, all sensors will immediately enter the working state and continuously collect the surrounding temperature information.These data streams contain the temperature value detected by each sensor at a specific time point, as well as the corresponding timestamp, forming a complete record of temperature changes. For example, during the hyperthermia treatment, as the equipment starts and gradually heats up, all sensors will respond to temperature changes in real time, sending the latest temperature readings to the central control system every second. These time-synchronized temperature data streams not only reflect the evolution of ambient temperature over time, but also lay the foundation for the subsequent construction of the temperature matrix. Subsequently, through the multi-channel data acquisition system, the temperature data streams collected by each deployed sensor are constructed into a temperature matrix in the order of timestamps, and finally the first temperature matrix and the second temperature matrix are obtained. This is a process of organizing discrete data points into structured tables, making a large amount of raw data more orderly and easier to analyze. For the first temperature matrix, each row represents a time series, that is, the data recorded by all sensors within a fixed time interval; each column represents the center coordinates of each divided sub-area, which means the position of the sensor in each sub-area; and each element in the matrix stores the temperature value of the temperature sensor to be calibrated. Similarly, the second temperature matrix follows the same rules, but it reflects the measurement results of the target sensor. For example, in the example of the operating room, if every minute is selected as a time series unit, the 10th row of the first temperature matrix will contain the temperature values ​​recorded by all sensors at the 10th minute, and the 2nd column corresponds to the temperature data of the temperature sensor to be calibrated in the second divided sub-area at each time point. Finally, in order to obtain the first acquisition temperature set and the second acquisition temperature set, it is necessary to extract the temperature data of the specified time period and the unified divided sub-area from the first temperature matrix and the second temperature matrix. This means that a time period of interest should be determined, and the measurement values ​​of the sensors in each sub-area should be selected on this basis. For example, during the process of hyperthermia treatment, technicians may pay attention to the temperature changes from 30 minutes before treatment to 60 minutes after treatment. Therefore, they can extract the data of the corresponding rows from the temperature matrix to form new sets, which are respectively used as the first acquisition temperature set and the second acquisition temperature set. These two sets will be used for subsequent operations such as curve drawing, time alignment and offset point identification, providing solid data support for the final accurate calibration of the temperature sensor. In summary, through the above detailed steps, we not only achieve the accurate acquisition of the temperature in the specified medical area, but also lay a solid foundation for subsequent data processing and calibration. This approach ensures that temperature sensors can provide highly accurate and reliable temperature monitoring services even in complex dynamic environments, ensuring the safety and effectiveness of medical services.

[0092] In a specific embodiment, the preprocessing of the first collected temperature set to obtain a preprocessed temperature data set includes:

[0093] Performing signal noise reduction processing on the first collected temperature set through a preset Butterworth bandpass filter to obtain a filtered temperature data set;

[0094] Performing multi-scale decomposition on the filtered temperature data set based on a wavelet packet decomposition algorithm to obtain multi-scale decomposed temperature data;

[0095] By using a preset local outlier factor algorithm, identifying and detecting whether there are outliers in the multi-scale decomposition temperature data, and if so, removing the outliers in the multi-scale decomposition temperature data to obtain a temperature data set after removal; wherein the outliers are temperature values ​​that deviate from a normal range;

[0096] The cubic spline interpolation algorithm is used to interpolate and supplement the missing temperature data in the temperature data set after elimination, and the interpolation supplement temperature data set is obtained;

[0097] The interpolated supplementary temperature data set is subjected to data smoothing processing to obtain a preprocessed temperature data set.

[0098] Specifically, first, in application scenarios such as operating rooms or intensive care units (ICUs), such as during hyperthermia treatment, changes in ambient temperature are crucial to the treatment effect of patients. In order to ensure the accuracy of the measurement results, a series of preprocessing operations must be performed on the first collected temperature set obtained from the temperature sensor to be calibrated. These operations are intended to eliminate possible noise, outliers, and other abnormal factors, so as to obtain a purer and smoother temperature data set, namely, a preprocessed temperature data set. The first step of preprocessing is to perform signal noise reduction processing on the first collected temperature set through a preset Butterworth bandpass filter to obtain a filtered temperature data set. The Butterworth filter is a linear filter that can effectively reduce the noise level while retaining the original signal characteristics. In practical applications, such as during hyperthermia treatment, due to the instantaneous temperature fluctuations generated at the moment of device startup or the influence of external factors such as electromagnetic interference, the first collected temperature set may be mixed with various random noises. By applying the Butterworth bandpass filter, we can filter out high-frequency noise that does not belong to the target temperature change frequency range, so that the temperature data more truly reflects the actual temperature changes of the patient's body surface or environment. Next, the filtered temperature data set is multi-scale decomposed based on the wavelet packet decomposition algorithm to obtain multi-scale decomposed temperature data. Wavelet packet decomposition is a method that can analyze signals in both time and frequency domains. It can decompose complex temperature time series into multiple sub-signals at different scales, each of which reflects the characteristics of the original signal in a specific frequency band. This is particularly important for processing temperature data in medical environments with complex dynamic characteristics. For example, in an operating room, temperature fluctuations caused by factors such as personnel movement and equipment opening and closing often have different time scales. Through wavelet packet decomposition, these temperature change components of different scales can be separated for further analysis. This multi-scale decomposition not only helps to reveal the potential patterns in temperature data, but also provides a more detailed basis for subsequent outlier detection. Then, through a preset local outlier factor algorithm, identify and detect whether there are outliers in the multi-scale decomposed temperature data. If so, remove the outliers in the multi-scale decomposed temperature data to obtain a temperature data set after removal. The outliers here refer to temperature values ​​that are significantly deviated from the normal range, which may be caused by sensor failure, severe external interference or other unknown reasons. The local outlier factor algorithm is a statistical method that can evaluate whether each data point is an outlier by calculating the relative density of each data point relative to its neighboring points. For example, in a hyperthermia treatment scenario, if the temperature reading at a certain moment is far beyond the expected range of variation, then we should further examine it. If it is confirmed to be an outlier, it can be removed based on the data of the adjacent time points before and after to ensure the continuity and consistency of the data.This method can effectively remove inaccurate measurements caused by sensor problems or sudden changes in the environment, and improve the quality of the data. In addition, in order to ensure the integrity of the data, the cubic spline interpolation algorithm is used to interpolate and supplement the missing temperature data in the temperature data set after elimination, and obtain the interpolated supplemented temperature data set. In the actual data acquisition process, it is inevitable that data at certain time points are lost or marked as outliers and eliminated. Cubic spline interpolation is a commonly used numerical analysis method that can construct a smooth curve between known data points to estimate the value of the missing position. For example, in the application of the operating room, if the temperature value is not recorded due to sensor failure for a period of time, we can use the cubic spline interpolation algorithm to fill this gap based on the temperature data of the surrounding time points to keep the entire temperature time series intact. This not only improves the availability of the data, but also provides necessary support for subsequent curve drawing and time alignment. Finally, the interpolated supplemented temperature data set is smoothed to obtain the preprocessed temperature data set. Data smoothing is a technique to reduce data fluctuations and highlight the main trends. It can improve the visual presentation of data and simplify subsequent analysis. Common smoothing methods include moving average, exponential smoothing, etc. In this step, choosing a suitable method to smooth the interpolated temperature data can make the temperature change curve smoother and easier to observe long-term trends. For example, during hyperthermia treatment, the smoothed temperature curve can more clearly show the overall temperature changes during treatment, helping medical staff to better monitor the progress of treatment. At the same time, smoothing also helps to eliminate small fluctuations introduced by interpolation, ensuring that the final temperature data set used for calibration is both accurate and stable. In summary, by preprocessing the first collected temperature set, we not only achieve effective purification of the original data, but also lay a solid foundation for subsequent operations such as temperature curve drawing, time alignment, and offset point identification. This series of carefully designed preprocessing steps ensures that even in complex medical environments, temperature sensors can provide highly accurate and reliable temperature monitoring services, ensuring the safety and effectiveness of medical services. Through the above methods, we can maximize the quality of temperature data and provide reliable data support for the ultimate realization of accurate calibration of sensors.

[0099] In a specific embodiment, the first temperature curve and the second temperature curve are time-aligned by using a preset dynamic time warping algorithm, and the offset point between the first temperature curve and the second temperature curve is identified by using a preset temperature offset detection algorithm, and the offset point is marked on the first temperature curve to obtain the marked temperature curve, including:

[0100] Performing curve fitting on the first temperature curve and the second temperature curve respectively by using a preset Gaussian process regression technology to obtain a corresponding first temperature curve function and a corresponding second temperature curve function;

[0101] Based on a dynamic time warping algorithm, time series matching is performed on the first temperature curve function and the second temperature curve function respectively to obtain a time warping path matrix; wherein the time warping path matrix includes time warping path rows, time warping path columns and time warping path elements, the rows of the time warping paths represent the time points of the first temperature curve function, the columns of the time warping paths represent the time points of the second temperature curve function, and the elements of the time warping paths represent the measured temperatures of the temperature sensor to be calibrated and the target sensor at the same time;

[0102] Extracting matrix temperature values ​​on the diagonal from the time-warped path matrix to obtain a first temperature sequence and a second temperature sequence;

[0103] Calculating the difference between the first temperature sequence and the second temperature sequence to obtain a temperature difference sequence;

[0104] Perform sliding window processing on the temperature difference sequence according to a preset temperature offset detection algorithm to obtain a temperature difference mean sequence within the sliding window;

[0105] Performing offset point detection on the temperature difference mean sequence to obtain an offset point index set;

[0106] According to the offset point index set, offset points are marked on the first temperature curve to obtain a marked temperature curve.

[0107] Specifically, in the data calibration method of the medical-grade high-precision temperature sensor, the first temperature curve and the second temperature curve are time-aligned using a preset dynamic time warping algorithm, and the offset point between the first temperature curve and the second temperature curve is identified by a preset temperature offset detection algorithm, and the offset point is marked on the first temperature curve. The process of obtaining the marked temperature curve is a key step to ensure the accuracy and reliability of the calibration result. This process requires not only precise time alignment technology, but also an effective offset point identification mechanism to ensure that the measurement differences that may exist between the two sensors can be accurately captured. First, in application scenarios such as operating rooms or intensive care units (ICUs), such as during hyperthermia treatment, the ambient temperature changes with time. In order to achieve this goal, it is necessary to curve fit the first temperature curve and the second temperature curve respectively through a preset Gaussian process regression technology to obtain the corresponding first temperature curve function and second temperature curve function. Gaussian process regression is a powerful non-parametric regression method that can construct a smooth probability distribution model based on observed data without a clear function form. In this process, for the temperature data obtained from the temperature sensor to be calibrated and the target sensor, i.e., the first temperature curve and the second temperature curve, we apply Gaussian process regression to fit each curve, thereby obtaining the first temperature curve function and the second temperature curve function that mathematically represent their behavior. This method can not only capture the trend of temperature change over time, but also allow us to predict temperature values ​​at unobserved time points, providing a solid foundation for subsequent time series matching. Next, based on the dynamic time warping algorithm, the first temperature curve function and the second temperature curve function are matched for time series respectively to obtain the time warping path matrix. The dynamic time warping (DTW) algorithm is a method for processing nonlinear time series. It finds the best matching path between two sequences so that the most similar parts between the two can be found even if there is temporal stretching or compression. Specifically, the algorithm constructs a distance matrix between the first temperature curve function and the second temperature curve function, and then finds a path that minimizes the accumulated distance along this path to achieve the optimal match between the two. This step ensures that the two temperature curves can be consistent in the time dimension even in a complex dynamic environment. For example, during the hyperthermia treatment, the temperature recording is not synchronized due to equipment startup delay or other factors. The DTW algorithm can effectively eliminate this time deviation, providing a reliable basis for subsequent analysis. After completing the time alignment, the matrix temperature values ​​on the diagonal are extracted from the time-warped path matrix to obtain the first temperature sequence and the second temperature sequence. The diagonal elements here represent the measured temperatures of the temperature sensor to be calibrated and the target sensor at the same time point after time warping.In this way, we can ensure that the temperature data at the same time point are compared, thereby reducing the error caused by time inconsistency. For example, in the hyperthermia treatment scenario, the extracted diagonal temperature values ​​reflect the temperature values ​​recorded by the two sensors at the same time on the adjusted time axis, and these values ​​constitute the first temperature sequence and the second temperature sequence that can be directly compared. Subsequently, the difference between the first temperature sequence and the second temperature sequence is calculated to obtain a temperature difference sequence. This step aims to quantify the measurement difference between the two sensors for further analysis. Each value in the temperature difference sequence represents the degree of deviation of the temperature sensor to be calibrated relative to the target sensor at the corresponding time point. For example, during the hyperthermia treatment process, if the temperature displayed by the first temperature sequence at a certain moment is significantly higher than that of the second temperature sequence, it means that the sensor to be calibrated may have a positive deviation at this time; vice versa. These temperature differences provide a direct basis for us to identify potential offset points. Then, the temperature difference sequence is subjected to sliding window processing according to the preset temperature offset detection algorithm to obtain the temperature difference mean sequence within the sliding window. Sliding window is a commonly used data processing technique that smoothes the original sequence and highlights long-term trends by averaging a series of continuous data points. In this example, by setting a suitable window size, the window can be moved on the temperature difference sequence, and the average value of the temperature difference in each window can be calculated to form a new temperature difference mean sequence. This method not only helps to reduce the impact of short-term fluctuations, but also better reveals the temperature deviation pattern over a long period of time. For example, in a hyperthermia treatment scenario, using a sliding window with a length of 5 minutes can filter out instantaneous temperature changes and more clearly show the stability of sensor performance during the entire treatment process. Finally, the temperature difference mean sequence is subjected to offset point detection to obtain an offset point index set. Offset points refer to those temperature differences that significantly deviate from the normal range, which may be caused by sensor aging, decreased sensitivity, improper installation position or other unknown reasons. By applying specific statistical rules or machine learning models, these abnormal points can be identified and their positions in the temperature difference mean sequence, i.e., the offset point index set, can be recorded. For example, during hyperthermia treatment, if it is found that the temperature difference mean in a certain time period continues to exceed the set threshold, the time period is determined as an offset point. Once the positions of all offset points are determined, the offset points can be marked on the first temperature curve according to the offset point index set to obtain the marked temperature curve. This step is not just a simple mark, but it intuitively reflects the offset information in the graph, helping technicians to more easily understand and analyze the behavioral characteristics of the temperature sensor. In summary, through the above detailed steps, we not only achieved the precise alignment of the first temperature curve and the second temperature curve in time, but also successfully identified and marked the offset point between the two, laying a solid foundation for the final calibration work.This approach ensures that even in complex and changing medical environments, temperature sensors can provide accurate and reliable temperature monitoring services, ensuring the safety and effectiveness of medical services. Through these carefully designed operations, we can maximize the quality of temperature data and provide reliable data support for the ultimate realization of accurate calibration of sensors.

[0108] In a specific embodiment, extracting the matrix temperature values ​​on the diagonal from the time-warped path matrix to obtain the first temperature sequence and the second temperature sequence includes:

[0109] Based on the dynamic time warping algorithm, the main diagonal elements of the time warping path matrix are identified to obtain a main diagonal element set;

[0110] Extracting corresponding temperature values ​​of the first temperature curve from the main diagonal element set to form a first temperature sequence; wherein the first temperature sequence reflects the temperature change trend of the first temperature curve after time alignment;

[0111] The corresponding temperature values ​​of the second temperature curve are extracted from the main diagonal element set to form a second temperature sequence; wherein the second temperature sequence reflects the temperature change trend of the second temperature curve after time alignment.

[0112] Specifically, first, in application scenarios such as operating rooms or intensive care units (ICUs), such as during hyperthermia treatment, the ambient temperature changes over time. In order to achieve accurate time alignment, we use the Dynamic Time Warping (DTW) algorithm to process the time difference between the first temperature curve function and the second temperature curve function. Through the DTW algorithm, a time warping path matrix is ​​constructed, each element in which represents the measured temperature of the temperature sensor to be calibrated and the target sensor at the same time. In this matrix, the main diagonal elements are particularly important because they represent the most similar parts of the two temperature curves after time warping, that is, the temperature values ​​recorded by the two sensors at the same time point. Therefore, identifying the main diagonal elements of the time warping path matrix to obtain the main diagonal element set is the key to extracting synchronized temperature data. Next, the corresponding temperature values ​​of the first temperature curve are extracted from the main diagonal element set to form a first temperature sequence. The first temperature sequence here not only contains the temperature data after time alignment, but also reflects the temperature change trend of the first temperature curve on the adjusted time axis. In other words, through the temperature values ​​in the diagonal element set, we can reconstruct a new temperature curve, which eliminates the influence of time delay or other asynchronous factors in the original data, and more realistically shows the temperature response of the temperature sensor to be calibrated during the entire monitoring period. For example, during the hyperthermia treatment process, the first temperature sequence can clearly show the actual temperature readings of the sensor to be calibrated at each moment during the treatment, helping technicians to evaluate its performance and find potential problems. Similarly, the corresponding temperature values ​​of the second temperature curve are extracted from the main diagonal element set to form a second temperature sequence. The second temperature sequence here also reflects the temperature change trend of the second temperature curve after time alignment. Since the target sensor has been calibrated and its accuracy has been verified, the second temperature sequence can be used as a standard reference for comparative analysis of the performance of the sensor to be calibrated. For example, in the above-mentioned hyperthermia treatment scenario, the second temperature sequence provides the real change of the ambient temperature during the entire treatment process, allowing technicians to intuitively compare the differences between the two temperature curves and identify any possible deviations or anomalies. In this way, we not only achieve accurate alignment of the first temperature curve and the second temperature curve in the time dimension, but also construct two sets of temperature sequences that can be directly compared. These temperature series not only retain the time characteristics of the original data, but also remove the time differences caused by factors such as sensor response speed and installation position, providing a solid foundation for subsequent temperature difference calculation, sliding window processing, and offset point detection. In addition, by extracting diagonal elements, we can ensure that each pair of temperature values ​​comes from the same time point, thereby enhancing the reliability and effectiveness of the comparison results.In summary, by identifying and extracting the main diagonal elements in the time-warped path matrix, we successfully constructed the first temperature series and the second temperature series that reflect the respective temperature change trends. This step not only improves the quality of the data, but also lays a good foundation for subsequent analysis. In a complex medical environment, this precise time alignment and data extraction method is crucial to ensure the calibration accuracy of the temperature sensor, which helps to provide more accurate and reliable temperature monitoring services and ensure the safety and effectiveness of medical services. It can be understood that the main diagonal elements include multiple (x, y), where x represents the corresponding temperature value of the first temperature curve at a certain point in time, and y represents the corresponding temperature value of the second temperature curve at a certain point in time.

[0113] In a specific embodiment, the sensor response characteristic parameters include a temperature measurement sensitivity coefficient and a temperature drift compensation factor. The response analysis of the temperature sensor to be calibrated is performed based on the marked temperature curve through a preset temperature sensor calibration model to obtain the sensor response characteristic parameters, including:

[0114] Obtaining a component topology diagram of the temperature sensor to be calibrated; wherein the component topology diagram describes the connection relationship and physical position of each component inside the temperature sensor to be calibrated;

[0115] Perform multi-scale feature extraction on the annotated temperature curve to obtain multi-scale temperature features; wherein the multi-scale temperature features include maximum temperature, minimum temperature, mean temperature, temperature variance, temperature peak time, temperature rise time and temperature drop time;

[0116] Reconstructing the topological structure of the multi-scale temperature characteristics by a complex network analysis algorithm to obtain a temperature sensor characteristic network;

[0117] Performing principal component analysis on the temperature sensor characteristic network to obtain a principal component characteristic network;

[0118] Mapping the principal component characteristic network onto the component topology map to obtain a mapping characteristic network; wherein the mapping process is based on the functions of the components in the temperature sensor to be calibrated and their influence on the temperature characteristics;

[0119] Based on the mapped feature network, the contribution of each component to the main component feature is analyzed to obtain the component health status index; wherein the component health status index includes stability, sensitivity and critical point;

[0120] Calculating the entropy of the multi-scale temperature feature to obtain entropy feature parameters, including information entropy and conditional entropy;

[0121] Performing correlation analysis on the entropy value characteristic parameters by using a preset least square method to obtain a correlation characteristic vector; wherein the correlation characteristic vector includes a correlation coefficient and a partial correlation coefficient;

[0122] Through a preset temperature sensor calibration model, based on the correlation feature vector and the component health status index, a response analysis is performed on the temperature sensor to be calibrated to obtain sensor response characteristic parameters; wherein the sensor response characteristic parameters include a temperature measurement sensitivity coefficient and a temperature drift compensation factor.

[0123] Specifically, first of all, in application scenarios such as operating rooms or intensive care units (ICUs), such as during hyperthermia treatment, changes in ambient temperature are crucial to the treatment effect of patients. In order to ensure the measurement accuracy of the temperature sensor, its internal structure must be deeply understood. Therefore, the first step is to obtain the component topology diagram of the temperature sensor to be calibrated. This diagram describes in detail the connection relationship and physical location of the various components inside the sensor, providing the necessary hardware foundation for subsequent analysis. For example, in a typical temperature sensor, the topology diagram may show how resistors, capacitors, and other sensitive components are connected, and indicate their specific positions relative to the sensor housing. This information is very important for understanding the function of each component and their response mechanism to temperature changes. Next, multi-scale feature extraction is performed on the annotated temperature curve to obtain multi-scale temperature features. The multi-scale temperature features here include temperature maximum value, temperature minimum value, temperature mean value, temperature variance, temperature peak time, temperature rise time, and temperature drop time. By comprehensively analyzing the annotated temperature curve, the temperature change pattern on different time scales can be captured, thereby more accurately reflecting the behavior of the sensor. For example, during the process of hyperthermia treatment, if the temperature rises rapidly at a certain moment and then slowly decreases, we can quantify the key points in this process through multi-scale feature extraction, such as the time when the temperature reaches the peak, the rate of rise and the rate of decrease. These features not only help to identify the dynamic response characteristics of the sensor, but also provide rich data support for subsequent analysis. Subsequently, the multi-scale temperature features are topologically reconstructed by complex network analysis algorithm to obtain the temperature sensor feature network. Complex network analysis is a method to study the interaction between elements in a system, which can help us reveal the complex relationships hidden behind the temperature data. In this step, we construct the multi-scale temperature features extracted from the temperature curve into a network model, in which each node represents a feature point and the edge represents the strength of the association between the feature points. For example, in the constructed temperature sensor feature network, the maximum temperature and the temperature peak time may be closely connected nodes because they usually reflect different aspects of the same event. This method can provide a new perspective to observe the behavior pattern of the temperature sensor and lay the foundation for further analysis. Then, the temperature sensor feature network is subjected to principal component analysis to obtain the principal component feature network. Principal component analysis (PCA) is a commonly used dimensionality reduction technique that can transform multiple related variables into a few unrelated comprehensive variables, namely principal components. In this process, we use PCA to extract the main components that best represent the original temperature characteristics to form a new feature network. This step not only simplifies the data structure, but also highlights the most important temperature change characteristics.For example, in the principal component feature network, it may be found that some temperature features can better explain the trend of overall temperature change than other features, thus helping us focus on the most critical factors. Then, the principal component feature network is mapped to the component topology map to obtain a mapped feature network. This mapping process is based on the functions of the components in the temperature sensor to be calibrated and their impact on the temperature characteristics. In this way, we can intuitively see which components are closely related to specific temperature characteristics and then evaluate their status. For example, in the mapped feature network, if a certain resistor is found to have a strong correlation with the temperature rise time, then we can infer that the resistor may play an important role in the sensor heating process. This mapping not only deepens our understanding of the internal working principle of the sensor, but also provides clues for subsequent fault diagnosis and performance optimization. On this basis, based on the mapped feature network, the contribution of each component to the principal component feature is analyzed to obtain the component health status indicators. These indicators include stability, sensitivity, and critical point, which are used to measure the working status of each component and its impact on temperature measurement. For example, if a capacitor shows low stability or high sensitivity, then this may mean that it is prone to problems when the temperature changes and needs special attention. In this way, we can identify potential weak links and take corresponding measures to improve the reliability of the entire sensor system. In addition, the entropy value of the multi-scale temperature feature is calculated to obtain entropy feature parameters, including information entropy and conditional entropy. Entropy calculation is a method to measure data uncertainty, which is used here to evaluate the randomness and regularity of temperature features. For example, information entropy can tell us how complex the temperature changes in a certain period of time are, while conditional entropy can reveal the dependencies between different temperature features. These entropy feature parameters provide a quantitative basis for subsequent correlation analysis. Finally, the entropy feature parameters are subjected to correlation analysis by the preset least squares method to obtain a correlation feature vector; wherein the correlation feature vector includes a correlation coefficient and a partial correlation coefficient. The least squares method is a technique widely used in regression analysis, which can help us find the linear relationship between entropy feature parameters. For example, the correlation coefficient can measure the degree of linear correlation between two temperature features, while the partial correlation coefficient can evaluate the direct relationship between them while controlling other factors. Through this analysis, we can further refine our understanding of the behavior of the temperature sensor and determine which features have significant correlations. In summary, through the above detailed steps, we not only achieved an in-depth analysis of the internal structure and external behavior of the temperature sensor, but also obtained a series of important response characteristic parameters, such as temperature measurement sensitivity coefficient and temperature drift compensation factor. These parameters are of great significance for improving the design of temperature sensors, improving their measurement accuracy, and ensuring long-term stable operation.In complex medical environments, this approach ensures that temperature sensors can provide reliable data support even in the face of changing temperature conditions, ensuring the safety and effectiveness of medical services.

[0124] In a specific embodiment, the topological structure of the multi-scale temperature feature is reconstructed by a complex network analysis algorithm to obtain a temperature sensor feature network, including:

[0125] Performing Spearman rank correlation analysis on the multi-scale temperature features to obtain a feature correlation matrix;

[0126] The characteristic association matrix is ​​network constructed by a minimum spanning tree algorithm to obtain an initial temperature sensor characteristic network;

[0127] By using a preset PageRank algorithm, node importance evaluation is performed on the initial temperature sensor feature network to obtain a node importance ranking;

[0128] According to the order of node importance, the edge weights of the initial temperature sensor feature network are adjusted to obtain an adjusted temperature sensor feature network;

[0129] The preset Louvain community discovery algorithm is used to discover the community of the adjusted temperature sensor feature network, and the temperature feature community is divided;

[0130] Based on the divided temperature feature communities, the adjusted temperature sensor feature network is topologically optimized to obtain a temperature sensor feature network; wherein the temperature sensor feature network includes network nodes and edges, the network nodes represent temperature feature points, the edges represent the association strength between the temperature feature points, and the temperature feature points represent the temperature measurement results at different physical locations.

[0131] Specifically, first, in application scenarios such as operating rooms or intensive care units (ICUs), such as during hyperthermia treatment, changes in ambient temperature are crucial to the treatment effect of patients. To achieve this goal, we first need to perform Spearman rank correlation analysis on the multi-scale temperature features to obtain a feature correlation matrix. Spearman rank correlation analysis is a non-parametric statistical method that does not rely on the specific distribution of the data, but measures the monotonic relationship between two variables by calculating the correlation between their rankings. For example, during hyperthermia treatment, we may observe a strong positive correlation between the two feature points of maximum temperature and peak temperature time. This correlation means that when one feature increases, the other feature also tends to increase. Through Spearman rank correlation analysis, we can quantify these relationships and organize the results into a feature correlation matrix, in which each element represents the strength of association between two features. This step helps to reveal the intrinsic connection between different temperature features and provides a basis for subsequent network construction. Next, based on the obtained feature correlation matrix, we use the minimum spanning tree algorithm to construct the network to obtain the initial temperature sensor feature network. The minimum spanning tree algorithm is a graph theory algorithm, which aims to find a tree with all nodes and the minimum total weight in an undirected graph. In this application context, we regard the feature association matrix as a complete graph, in which each node represents a temperature feature point, and the weight of the edge is determined by the feature association. The minimum spanning tree algorithm can ensure that all temperature feature points are connected while avoiding the formation of loops, so that an initial temperature sensor feature network with simple structure but rich information is obtained. For example, for the maximum temperature and the peak temperature time mentioned above, if the correlation between them is very high, then in the minimum spanning tree, these two nodes are likely to be directly connected, indicating a close connection between them. This step not only simplifies the network structure, but also retains the most important feature association information. In order to further evaluate the importance of each node in this network, the preset PageRank algorithm is used to evaluate the node importance of the initial temperature sensor feature network, and then the node importance ranking is obtained. PageRank was originally designed as an algorithm to solve the problem of web page ranking, but it is also applicable to any type of network structure. In this context, PageRank considers the links between each node and other nodes and the quality of the links, so as to give an importance score for each temperature feature point. For example, if a temperature feature point is not only strongly associated with multiple other feature points, but is also the main associated object of other high-importance feature points, then it will receive a higher importance score according to the PageRank algorithm. This step helps to identify the most critical feature points in the entire temperature sensor feature network and provides guidance for subsequent network optimization.According to the order of node importance, we adjust the edge weights of the initial temperature sensor feature network to obtain the adjusted temperature sensor feature network. This is to reflect the true influence relationship between nodes, because some nodes may be more important than others, so the connection between them should be given a greater weight. For example, if the maximum temperature is found to be a very important node, then all edges connected to it should increase the weight accordingly to emphasize its core position in the network. This adjustment makes the final temperature sensor feature network more accurately reflect the actual temperature feature interaction pattern. In addition, by adjusting the edge weights, the importance of the key path in the network can also be enhanced, thereby improving the robustness and stability of the entire system. Subsequently, through the preset Louvain community discovery algorithm, we perform community discovery on the adjusted temperature sensor feature network to obtain the divided temperature feature community. The Louvain algorithm is an effective method for detecting community structure in complex networks. It can divide the network into several relatively independent but highly interconnected subsets, namely communities. In the temperature sensor feature network, this means that we cluster a group of similar or closely related temperature feature points together to form different temperature feature communities. For example, a group of feature points including temperature rise time and temperature drop time may be classified into the same community because they both reflect the temporal characteristics of temperature change. Such a division can help us better understand the behavior pattern of the temperature sensor and provide a basis for further topology optimization. Finally, based on the divided temperature feature community, we optimize the topology of the adjusted temperature sensor feature network and finally obtain the temperature sensor feature network. In this optimization process, we will comprehensively consider the connectivity within the community and the association between communities to ensure that the entire network is both compact and flexible. For example, we should pay special attention to those important connections across different communities to ensure that they are not lost due to oversimplification. The optimized temperature sensor feature network includes network nodes and edges, where network nodes represent temperature feature points and edges represent the strength of association between temperature feature points. These temperature feature points represent temperature measurement results at different physical locations, such as temperature changes in different divided sub-areas in the operating room. Such a carefully designed and optimized temperature sensor feature network can provide us with deep insights into the response characteristics of temperature sensors and help us achieve more accurate data calibration and performance improvement in the field of medical-grade high-precision temperature measurement. In summary, through a series of complex network analysis and optimization processing of multi-scale temperature features, we not only achieve a deep understanding of the internal structure and external behavior of the temperature sensor, but also obtain a network model that accurately reflects the interaction of temperature features. This approach ensures that even in the face of changing temperature conditions, temperature sensors can provide reliable data support to ensure the safety and effectiveness of medical services.Through the above steps, we can maximize the quality of temperature data and provide a solid foundation for the ultimate accurate calibration of the sensor.

[0132] The above describes the data calibration method of the medical-grade high-precision temperature sensor in the embodiment of the present invention. The following describes the data calibration system of the medical-grade high-precision temperature sensor in the embodiment of the present invention. Figure 2 , an embodiment of a data calibration system for a medical-grade high-precision temperature sensor in an embodiment of the present invention includes:

[0133] The acquisition module 21 is used to respectively acquire the temperature of the designated medical area through the preset temperature sensor to be calibrated and the target sensor to obtain the corresponding first acquisition temperature set and the second acquisition temperature set;

[0134] A preprocessing module 22, configured to preprocess the first collected temperature set to obtain a preprocessed temperature data set;

[0135] A drawing module 23, used for drawing a curve based on the preprocessed temperature data set and the second collected temperature set to obtain a first temperature curve and a second temperature curve;

[0136] A marking module 24, configured to time-align the first temperature curve and the second temperature curve using a preset dynamic time warping algorithm, identify an offset point between the first temperature curve and the second temperature curve using a preset temperature offset detection algorithm, and mark the offset point on the first temperature curve to obtain a marked temperature curve;

[0137] An analysis module 25 is used to analyze the annotated temperature curve through a preset temperature sensor calibration model to obtain calibration parameters; wherein the calibration parameters include an offset, a proportional factor, and a nonlinear correction coefficient;

[0138] The calibration module 26 is configured to calibrate the temperature sensor to be calibrated based on the calibration parameters.

[0139] In this embodiment, for the specific implementation of each unit in the above system embodiment, please refer to the above method embodiment, which will not be repeated here.

[0140] Reference Figure 3 The present invention also provides a computer device in an embodiment, wherein the internal structure of the computer device can be as follows: Figure 3As shown. The computer device includes a processor, a memory, a display screen, an input device, a network interface and a database connected through a system bus. Among them, the processor designed by the computer is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the corresponding data in this embodiment. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the above method is implemented.

[0141] Those skilled in the art will understand that Figure 3 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied.

[0142] An embodiment of the present invention 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 above method is implemented. It can be understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0143] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided by the present invention and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double-speed data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM.

[0144] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.

[0145] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A data calibration method for a medical-grade high-precision temperature sensor, characterized in that: The following steps are involved: The temperature of the designated medical area is collected by using a preset temperature sensor to be calibrated and a target sensor to obtain a corresponding first collected temperature set and a second collected temperature set; Preprocessing the first collected temperature set to obtain a preprocessed temperature data set; Plotting a curve based on the preprocessed temperature data set and the second collected temperature set to obtain a first temperature curve and a second temperature curve; Using a preset dynamic time warping algorithm, the first temperature curve and the second temperature curve are time-aligned, and the offset point between the first temperature curve and the second temperature curve is identified by a preset temperature offset detection algorithm, and the offset point is marked on the first temperature curve to obtain a marked temperature curve; By using a preset temperature sensor calibration model, a response analysis is performed on the temperature sensor to be calibrated based on the annotated temperature curve to obtain a sensor response characteristic parameter; calibrating the temperature sensor to be calibrated based on the sensor response characteristic parameter; The sensor response characteristic parameters include a temperature measurement sensitivity coefficient and a temperature drift compensation factor. The sensor response characteristic parameters obtained by performing response analysis on the temperature sensor to be calibrated based on the marked temperature curve through a preset temperature sensor calibration model include: Obtaining a component topology diagram of the temperature sensor to be calibrated; wherein the component topology diagram describes the connection relationship and physical position of each component inside the temperature sensor to be calibrated; Perform multi-scale feature extraction on the annotated temperature curve to obtain multi-scale temperature features; wherein the multi-scale temperature features include maximum temperature, minimum temperature, mean temperature, temperature variance, temperature peak time, temperature rise time and temperature drop time; Reconstructing the topological structure of the multi-scale temperature characteristics by a complex network analysis algorithm to obtain a temperature sensor characteristic network; Performing principal component analysis on the temperature sensor characteristic network to obtain a principal component characteristic network; Mapping the principal component characteristic network onto the component topology map to obtain a mapping characteristic network; wherein the mapping process is based on the functions of the components in the temperature sensor to be calibrated and their influence on the temperature characteristics; Based on the mapped feature network, the contribution of each component to the main component feature is analyzed to obtain the component health status index; wherein the component health status index includes stability, sensitivity and critical point; Calculating the entropy of the multi-scale temperature feature to obtain entropy feature parameters, including information entropy and conditional entropy; Performing correlation analysis on the entropy value characteristic parameters by using a preset least square method to obtain a correlation characteristic vector; wherein the correlation characteristic vector includes a correlation coefficient and a partial correlation coefficient; Through a preset temperature sensor calibration model, based on the correlation feature vector and the component health status index, a response analysis is performed on the temperature sensor to be calibrated to obtain sensor response characteristic parameters; wherein the sensor response characteristic parameters include a temperature measurement sensitivity coefficient and a temperature drift compensation factor.

2. The data calibration method for a medical-grade high-precision temperature sensor according to claim 1, characterized in that: The temperature of the designated medical area is collected by using the preset temperature sensor to be calibrated and the target sensor to obtain the corresponding first collected temperature set and the second collected temperature set, including: The designated medical area is spatially divided by a preset Voronoi diagram algorithm to obtain a plurality of divided sub-areas; The temperature sensor to be calibrated and the target sensor are respectively deployed at the central coordinates of each of the divided sub-areas to obtain deployed sensors; wherein the deployed sensors include the deployed temperature sensor to be calibrated and the deployed target sensor; Using a preset network time protocol to synchronize the time of the deployed sensors to obtain a uniformly set time reference; Based on the time reference, synchronously start all the deployed sensors to collect temperature data to obtain a time-synchronized temperature data stream; Through a multi-channel data acquisition system, the temperature data streams collected by each of the deployed sensors are constructed into a temperature matrix in a timestamp order to obtain a first temperature matrix and a second temperature matrix, wherein the first temperature matrix corresponds to the temperature sensor to be calibrated, the second temperature matrix corresponds to the target sensor, the first temperature matrix includes first temperature matrix rows, first temperature matrix columns and first temperature matrix elements, the first temperature matrix rows represent time series, the first temperature matrix columns represent the center coordinates of each of the divided sub-regions, the first temperature matrix elements represent the temperature values ​​of the temperature sensor to be calibrated, and the second temperature matrix includes second temperature matrix rows, second temperature matrix columns and second temperature matrix elements, the second temperature matrix rows represent time series, the second temperature matrix columns represent the center coordinates of each of the divided sub-regions, and the second temperature matrix elements represent the temperature values ​​of the target sensor; The temperature data in the designated time period and the divided sub-areas are extracted from the first temperature matrix and the second temperature matrix respectively to obtain the corresponding first collected temperature set and the second collected temperature set.

3. The data calibration method for a medical-grade high-precision temperature sensor according to claim 1, characterized in that: The preprocessing of the first collected temperature set to obtain a preprocessed temperature data set includes: Performing signal noise reduction processing on the first collected temperature set through a preset Butterworth bandpass filter to obtain a filtered temperature data set; Performing multi-scale decomposition on the filtered temperature data set based on a wavelet packet decomposition algorithm to obtain multi-scale decomposed temperature data; By using a preset local outlier factor algorithm, identifying and detecting whether there are outliers in the multi-scale decomposition temperature data, and if so, removing the outliers in the multi-scale decomposition temperature data to obtain a temperature data set after removal; wherein the outliers are temperature values ​​that deviate from a normal range; The cubic spline interpolation algorithm is used to interpolate and supplement the missing temperature data in the temperature data set after elimination, and the interpolation supplement temperature data set is obtained; The interpolated supplementary temperature data set is subjected to data smoothing processing to obtain a preprocessed temperature data set.

4. The data calibration method for a medical-grade high-precision temperature sensor according to claim 1, characterized in that: The method uses a preset dynamic time warping algorithm to time-align the first temperature curve and the second temperature curve, identifies the offset point between the first temperature curve and the second temperature curve through a preset temperature offset detection algorithm, and marks the offset point on the first temperature curve to obtain the marked temperature curve, including: Performing curve fitting on the first temperature curve and the second temperature curve respectively by using a preset Gaussian process regression technology to obtain a corresponding first temperature curve function and a corresponding second temperature curve function; Based on a dynamic time warping algorithm, time series matching is performed on the first temperature curve function and the second temperature curve function respectively to obtain a time warping path matrix; wherein the time warping path matrix includes time warping path rows, time warping path columns and time warping path elements, the rows of the time warping paths represent the time points of the first temperature curve function, the columns of the time warping paths represent the time points of the second temperature curve function, and the elements of the time warping paths represent the measured temperatures of the temperature sensor to be calibrated and the target sensor at the same time; Extracting matrix temperature values ​​on the diagonal from the time-warped path matrix to obtain a first temperature sequence and a second temperature sequence; Calculating the difference between the first temperature sequence and the second temperature sequence to obtain a temperature difference sequence; Perform sliding window processing on the temperature difference sequence according to a preset temperature offset detection algorithm to obtain a temperature difference mean sequence within the sliding window; Performing offset point detection on the temperature difference mean sequence to obtain an offset point index set; According to the offset point index set, offset points are marked on the first temperature curve to obtain a marked temperature curve.

5. The data calibration method for a medical-grade high-precision temperature sensor according to claim 4, characterized in that: The step of extracting the matrix temperature values ​​on the diagonal from the time-warped path matrix to obtain the first temperature sequence and the second temperature sequence includes: Based on the dynamic time warping algorithm, the main diagonal elements of the time warping path matrix are identified to obtain a main diagonal element set; Extracting corresponding temperature values ​​of the first temperature curve from the main diagonal element set to form a first temperature sequence; wherein the first temperature sequence reflects the temperature change trend of the first temperature curve after time alignment; The corresponding temperature values ​​of the second temperature curve are extracted from the main diagonal element set to form a second temperature sequence; wherein the second temperature sequence reflects the temperature change trend of the second temperature curve after time alignment.

6. The data calibration method for a medical-grade high-precision temperature sensor according to claim 1, characterized in that: The topological structure of the multi-scale temperature feature is reconstructed by a complex network analysis algorithm to obtain a temperature sensor feature network, including: Performing Spearman rank correlation analysis on the multi-scale temperature features to obtain a feature correlation matrix; The characteristic association matrix is ​​network constructed by a minimum spanning tree algorithm to obtain an initial temperature sensor characteristic network; By using a preset PageRank algorithm, node importance evaluation is performed on the initial temperature sensor feature network to obtain a node importance ranking; According to the node importance sorting, the edge weights of the initial temperature sensor feature network are adjusted to obtain an adjusted temperature sensor feature network; Using a preset Louvain community discovery algorithm, community discovery is performed on the adjusted temperature sensor feature network to obtain divided temperature feature communities; Based on the divided temperature feature communities, the adjusted temperature sensor feature network is topologically optimized to obtain a temperature sensor feature network; wherein the temperature sensor feature network includes network nodes and edges, the network nodes represent temperature feature points, the edges represent the association strength between the temperature feature points, and the temperature feature points represent the temperature measurement results at different physical locations.

7. A device based on the data calibration method of the medical-grade high-precision temperature sensor according to claim 1, characterized in that: include: The acquisition module is used to respectively acquire the temperature of the designated medical area through the preset temperature sensor to be calibrated and the target sensor to obtain the corresponding first acquisition temperature set and the second acquisition temperature set; A preprocessing module, used for preprocessing the first collected temperature set to obtain a preprocessed temperature data set; A drawing module, used for drawing a curve based on the preprocessed temperature data set and the second collected temperature set to obtain a first temperature curve and a second temperature curve; A marking module, used to time-align the first temperature curve and the second temperature curve using a preset dynamic time warping algorithm, identify the offset point between the first temperature curve and the second temperature curve using a preset temperature offset detection algorithm, and mark the offset point on the first temperature curve to obtain a marked temperature curve; An analysis module, used to analyze the annotated temperature curve through a preset temperature sensor calibration model to obtain calibration parameters; wherein the calibration parameters include an offset, a proportional factor, and a nonlinear correction coefficient; A calibration module is used to calibrate the temperature sensor to be calibrated based on the calibration parameters.

8. A computer device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images