Abnormity early warning and intelligent supervision platform for non-contact vital sign monitoring

Abstract

The invention, which relates to the technical field of data intelligent monitoring, discloses a non-contact vital sign monitoring abnormity early warning and intelligent supervision platform comprising a sign monitoring module, a feature analysis module, a real-time monitoring module, and an abnormity early warning module. According to the method, sensor equipment is installed, monitoring data are extracted and classified, a defecation image comparison library is established, comfort level correlation interaction analysis is carried out in combination with an environment monitoring data set, a target environment parameter influence value is obtained, infant defecation health degree analysis is carried out, and infant defecation health coefficients are obtained; an associated data model is constructed, monitoring target parameters are output, real-time defecation health state analysis is performed, a real-time defecation physiological health coefficient is generated, defecation stage change analysis is performed, an infant physiological cycle state report is generated, and then a defecation early warning signal is generated, so that the abnormal defecation condition of an infant can be quickly found, and the defecation early warning efficiency is improved. And generating a physiological health early warning signal based on the infant physiological cycle state report and the real-time defecation physiological health coefficient.

Description

Technical Field

[0001] This invention relates to the field of data intelligent monitoring technology, and in particular to a non-contact vital sign monitoring platform for abnormal early warning and intelligent supervision. Background Technology

[0002] Non-contact vital sign monitoring technology is a technology that captures human physiological signals or behavioral characteristics through non-physical contact methods (such as radar, cameras, infrared sensors, etc.) to achieve real-time monitoring of vital signs such as heart rate, respiration, body temperature, and blood oxygenation. It indirectly infers physiological signals by sensing electromagnetic waves, infrared radiation, or changes in environmental pressure reflected by the human body. This includes radar wave monitoring, camera monitoring, pressure and vibration sensor monitoring, and acoustic monitoring. Its core advantage lies in avoiding interference from contact sensors (such as stimulation during defecation or dislodgement). It is especially suitable for sensitive groups such as infants, the elderly, and postoperative patients, and has become an important research direction in the fields of smart healthcare and health management. It has been gradually implemented in the following scenarios: hospital scenarios: ICU, postoperative wards, monitoring patients' respiration, heart rate, and fall risk; home scenarios: infant monitoring, and safety of elderly people living alone.

[0003] Currently, when monitoring infants, contactless data collection is achieved through cameras, radar, or environmental sensors. This method can only independently monitor environmental parameters or physiological signals. When monitoring and caring for infants, the physiological state of their defecation is encountered. During the defecation process, the health of the infant's defecation is not analyzed, and it is impossible to analyze the defecation state and physiological state after the infant defecation. This results in poor health management and monitoring effects when monitoring infants.

[0004] To address the aforementioned technical shortcomings, a solution is proposed. Summary of the Invention

[0005] The purpose of this invention is to: use various analytical methods, including defecation characteristic analysis, environmental correlation analysis, and health quantification assessment, to conduct real-time health analysis to promptly grasp the current health status of infants, combine environmental parameters to make the analysis more comprehensive, and generate periodic status reports to understand the trend of physiological changes, thereby improving the accuracy and comprehensiveness of early warning and ensuring the health of infants.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a non-contact vital sign monitoring abnormality early warning and intelligent supervision platform, including a vital sign monitoring module, a feature analysis module, a real-time monitoring module, and an abnormality early warning module; The vital signs monitoring module monitors the infant's vital signs and external environment through installed sensor devices, acquires parameter data during the infant monitoring process, and generates environmental monitoring datasets, image monitoring datasets, and vital signs datasets after extracting the monitoring data. The feature analysis module acquires the monitored image dataset, extracts and analyzes defecation feature data, establishes a defecation image comparison library, and performs comfort correlation and interaction analysis in conjunction with the environmental monitoring dataset to obtain the impact value of the target environmental parameters. Based on the vital signs dataset, it extracts key physiological features, performs infant defecation health analysis, obtains the infant defecation health coefficient, and constructs a correlation data model. Based on the correlation data model, it outputs the target parameters for infant vital signs monitoring. The real-time monitoring module acquires parameter data during real-time infant monitoring, performs real-time bowel health status analysis to generate a real-time bowel physiological health coefficient, combines real-time external environmental parameters to obtain an external environmental influence coefficient, and analyzes bowel phase changes based on the infant's bowel health coefficient over a period of time to generate an infant physiological cycle status report. The abnormal warning module acquires parameter data from real-time monitoring of the infant, performs a comparison with a defecation image library to generate a feature contrast of the defecation image, generates a defecation warning signal based on the feature contrast, and generates a physiological health warning signal based on the infant's physiological cycle status report and real-time defecation physiological health coefficient, which is then sent to the intelligent supervision platform.

[0007] Furthermore, the intelligent monitoring platform receives defecation warning signals and physiological health warning signals, and sends them to the user terminal and the medical monitoring terminal. The user terminal intelligently pushes events to users based on defecation warning signals. The push notifications are based on the confidence level of the defecation warning signals. Push notifications include generating pending confirmation message pushes, pop-up sound reminder pushes, and intelligent alarm pushes. The message push pending confirmation time threshold is set, and a reminder push is sent again based on the pending confirmation time threshold. Based on physiological health warning signals, the medical staff will analyze the infant's long-term defecation trend. When the number of defecation periods continues to decline and the fluctuation exceeds the set standard threshold, the medical staff will automatically trigger the intervention strategy recommended by the medical staff. By establishing a user practice feedback and evaluation mechanism between the medical staff and the user end, the system obtains the number of times push messages sent by users on the intelligent monitoring platform are received and confirmed. Based on the number of confirmed receipts, the system makes intelligent adjustments to the user end, and promptly adjusts and optimizes service content and processes.

[0008] Furthermore, environmental monitoring datasets, image monitoring datasets, and vital sign datasets are generated, specifically including the following: S100. Using non-contact sensor devices, including temperature sensors, camera devices, and millimeter-wave radar sensors, the sampling frequency and sampling time are set according to the monitoring requirements and the performance characteristics of the sensors. S101. Acquire images through a camera, locate and identify the baby in the image, extract image feature data related to the baby's defecation, and construct an image dataset; S102. Acquire external environmental data through temperature sensors, extract key data, organize and store the extracted external environmental parameter data in chronological order, and construct an environmental monitoring dataset; integrate the number of times and time of infant defecation to construct a vital signs dataset.

[0009] Furthermore, the impact values ​​of the target environmental parameters are obtained, specifically including the following: S200. Obtain the monitored image dataset, extract and analyze defecation feature data to obtain defecation images, classify the extracted defecation images into diaper image features before defecation and diaper image features after defecation, store the labeled defecation images in the database, and establish a defecation image comparison library. S201. Integrate the environmental monitoring dataset and the defecation image dataset, align the timestamps of the two datasets, and perform a correlation analysis between external temperature and temperature during defecation. S202. Based on the external temperature and the temperature at the time of defecation, the deviation between the two is obtained. Then, based on the infant's body temperature and a personal comfort temperature setting, an external temperature control analysis is performed to obtain the target environmental parameter influence value R. The calculation process is as follows. ; In the above formula, For the baby's personal comfort temperature, Environmental monitoring data is concentrated on the external temperature at timestamp t. The defecation image data focuses on the temperature at the time of defecation at timestamp d.

[0010] Furthermore, the above analysis of infant bowel health yields an infant bowel health coefficient, which includes the following: S300: Acquire parameter data during infant monitoring, obtain the infant's physiological characteristic dataset, perform bowel health status analysis, classify bowel movements into normal, diarrhea, and constipation types, organize the classified bowel images, and add detailed annotation information to each image, including bowel movement type, collection time, and basic infant information. S301. Based on the frequency, timing, and type of the infant's bowel movements, conduct an analysis of the infant's bowel health to obtain an estimated range for healthy bowel movement frequency, as well as a range for the frequency of diarrhea and constipation in the infant. ]; S302. Based on the obtained defecation health status, and combined with the external influence coefficient, a defecation physiological health coefficient is generated. The calculation process is as follows: The average defecation interval T is initially calculated. ; In the above formula, These are the start and end times of the periodic time interval. The time of the first bowel movement preceding the first bowel movement. Let N be the time of the i-th defecation, where N = {1, 2, 3, ..., N}; ; In the above formula, H represents bowel health. These are the number of bowel movements, the number of times constipated, and the number of times diarrhea occur within a given period. This represents the maximum number of bowel movements within a given cycle. This represents the minimum number of bowel movements within a cycle. ; In the above formula, C is the physiological health coefficient of defecation, and W is the custom external influence coefficient, which takes the value [0, 1].

[0011] Furthermore, based on the correlation data model, the target parameters for infant vital sign monitoring are output, specifically including the following: S400: Acquire monitored image datasets, environmental monitoring datasets, and physiological characteristic datasets as input data for the correlation data model; S401. Set up the processing layer of the associated data model, and use the established defecation image comparison library, the influence value of target environmental parameters, and the infant defecation health coefficient as the key physical constraints of the processing layer. S402. Construct a neural network architecture based on physical constraints, and construct the loss function of the initial network model. Embed the residual gradient of the equation into the loss function and introduce the gradient penalty term of the equation relating environment and physiology. The monitored image dataset, environmental monitoring dataset, and physiological characteristic dataset are divided into training set and dataset. The range of hidden layer nodes is set according to the sample size of the training set. After training, an association data model is obtained. Based on the association data model, the target parameters for infant vital sign monitoring are output. The target parameters are defecation health coefficient and environmental regulation target.

[0012] Furthermore, a report on the infant's menstrual cycle status is generated, specifically including the following: S500: Acquire parameter data during real-time infant monitoring, perform real-time bowel health status analysis, and generate real-time bowel physiological health coefficient. S501. Based on real-time external environmental parameters, obtain the external environmental influence coefficient, and collect the infant defecation health coefficient according to the set periodic time to obtain the infant defecation health coefficient set. S502. Based on the infant's bowel health coefficient, analyze the changes in bowel movement stages, construct a trend staging curve, analyze the upward and downward trends of the infant's bowel health coefficient, and generate an infant physiological cycle status report. Furthermore, generate physiological health early warning signals, specifically including the following: S600: Acquire parameter data from real-time infant monitoring, perform stool image comparison against a database to generate feature contrast of the stool images, and generate a stool warning signal based on the feature contrast, which is divided into the following two cases: When the matching degree is greater than the preset standard matching degree, a defecation warning signal is generated; When the matching degree is less than the preset standard matching degree, a similar defecation warning signal is generated; S601. Based on the infant's physiological cycle status report and combined with the real-time defecation physiological health coefficient, a physiological health early warning signal is generated and sent to the intelligent monitoring platform, including the following two situations: When the real-time defecation physiological health coefficient is greater than the preset standard defecation physiological health coefficient, a physiological health warning signal is generated. When the fluctuation value of the physiological health coefficient of periodic defecation exceeds the preset fluctuation standard threshold, a physiological health warning signal is generated.

[0013] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: This non-contact vital sign monitoring platform for abnormal early warning and intelligent supervision effectively extracts and classifies monitoring data by installing sensor devices, establishes a defecation image comparison library, and performs comfort correlation and interaction analysis with environmental monitoring datasets to obtain the impact values ​​of target environmental parameters. It also analyzes infant defecation health to obtain an infant defecation health coefficient, constructs a correlation data model, outputs monitoring target parameters, performs real-time defecation health status analysis, generates a real-time defecation physiological health coefficient, and combines real-time external environmental parameters to obtain an external environmental impact coefficient. This comprehensive approach considers the dynamic impact of environmental factors on infant health, making health analysis more comprehensive. Based on the health coefficient over a period of time, it analyzes changes in defecation stages and generates an infant physiological cycle status report, helping medical staff or guardians understand the changing trends of the infant's physiological state and prepare accordingly for nursing and health management. By matching features from the defecation image comparison library, it generates defecation early warning signals, enabling rapid detection of abnormal infant defecation. Based on the infant physiological cycle status report and the real-time defecation physiological health coefficient, it generates physiological health early warning signals, comprehensively considering multiple factors to improve the accuracy and comprehensiveness of early warnings. Attached Figure Description

[0014] Figure 1 A schematic diagram of the overall system flow structure of the present invention is shown. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] Example 1: like Figure 1 As shown, a non-contact vital sign monitoring abnormality early warning and intelligent supervision platform includes a vital sign monitoring module, a feature analysis module, a real-time monitoring module, and an abnormality early warning module; The vital signs monitoring module monitors the infant's vital signs and external environment through installed sensor devices, acquires parameter data during the infant monitoring process, and generates environmental monitoring datasets, image monitoring datasets, and vital signs datasets after extracting the monitoring data. The feature analysis module acquires the monitored image dataset, extracts and analyzes defecation feature data, establishes a defecation image comparison library, and performs comfort correlation and interaction analysis in conjunction with the environmental monitoring dataset to obtain the impact value of the target environmental parameters. Based on the vital signs dataset, it extracts key physiological features, performs infant defecation health analysis, obtains the infant defecation health coefficient, and constructs a correlation data model. Based on the correlation data model, it outputs the target parameters for infant vital signs monitoring. The real-time monitoring module acquires parameter data during real-time infant monitoring, performs real-time bowel health status analysis to generate a real-time bowel physiological health coefficient, combines real-time external environmental parameters to obtain an external environmental influence coefficient, and analyzes bowel phase changes based on the infant's bowel health coefficient over a period of time to generate an infant physiological cycle status report. The abnormal warning module acquires parameter data from real-time monitoring of the infant, performs a comparison with a defecation image library to generate a feature contrast of the defecation image, generates a defecation warning signal based on the feature contrast, and generates a physiological health warning signal based on the infant's physiological cycle status report and real-time defecation physiological health coefficient, which is then sent to the intelligent supervision platform.

[0017] The intelligent monitoring platform receives defecation warning signals and physiological health warning signals, and sends them to the user terminal and the medical monitoring terminal. The user terminal intelligently pushes events to users based on defecation warning signals. The push notifications are based on the confidence level of the defecation warning signals. Push notifications include generating pending confirmation message pushes, pop-up sound reminder pushes, and intelligent alarm pushes. The message push pending confirmation time threshold is set, and a reminder push is sent again based on the pending confirmation time threshold. Based on physiological health warning signals, the medical staff will analyze the infant's long-term defecation trend. When the number of defecation periods continues to decline and the fluctuation exceeds the set standard threshold, the medical staff will automatically trigger the intervention strategy recommended by the medical staff. By establishing a user practice feedback and evaluation mechanism between the medical staff and the user end, the system obtains the number of times push messages sent by users on the intelligent monitoring platform are received and confirmed. Based on the number of confirmed receipts, the system makes intelligent adjustments to the user end, and promptly adjusts and optimizes service content and processes.

[0018] Generate environmental monitoring datasets, image monitoring datasets, and vital sign datasets, specifically including the following: S100. Using non-contact sensor devices, including temperature sensors, camera devices, and millimeter-wave radar sensors, the sampling frequency and sampling time are set according to the monitoring requirements and the performance characteristics of the sensors. S101. Acquire images through a camera, locate and identify the baby in the image, extract image feature data related to the baby's defecation, and construct an image dataset; S102. Acquire external environmental data through temperature sensors, extract key data, organize and store the extracted external environmental parameter data in chronological order, and construct an environmental monitoring dataset; integrate the number of times and time of infant defecation to construct a vital signs dataset.

[0019] The impact values ​​of the target environmental parameters are obtained, specifically including the following: S200. Obtain the monitored image dataset, extract and analyze defecation feature data to obtain defecation images, classify the extracted defecation images into diaper image features before defecation and diaper image features after defecation, store the labeled defecation images in the database, and establish a defecation image comparison library. S201. Integrate the environmental monitoring dataset and the defecation image dataset, align the timestamps of the two datasets, and perform a correlation analysis between external temperature and temperature during defecation. S202. Based on the external temperature and the temperature at the time of defecation, the deviation between the two is obtained. Then, based on the infant's body temperature and a personal comfort temperature setting, an external temperature control analysis is performed to obtain the target environmental parameter influence value R. The calculation process is as follows. ; In the above formula, For the baby's personal comfort temperature, Environmental monitoring data is concentrated on the external temperature at timestamp t. The defecation image data focuses on the temperature at the time of defecation at timestamp d.

[0020] The above analysis of infant bowel health yields an infant bowel health coefficient, which includes the following: S300: Acquire parameter data during infant monitoring, obtain the infant's physiological characteristic dataset, perform bowel health status analysis, classify bowel movements into normal, diarrhea, and constipation types, organize the classified bowel images, and add detailed annotation information to each image, including bowel movement type, collection time, and basic infant information. S301. Based on the frequency, timing, and type of the infant's bowel movements, conduct an analysis of the infant's bowel health to obtain an estimated range for healthy bowel movement frequency, as well as a range for the frequency of diarrhea and constipation in the infant. ]; S302. Based on the obtained defecation health status, and combined with the external influence coefficient, a defecation physiological health coefficient is generated. The calculation process is as follows: The average defecation interval T is initially calculated. ; In the above formula, These are the start and end times of the periodic time interval. The time of the first bowel movement preceding the first bowel movement. Let N be the time of the i-th defecation, where N = {1, 2, 3, ..., N}; ; In the above formula, H represents bowel health. These are the number of bowel movements, the number of times constipated, and the number of times diarrhea occur within a given period. This represents the maximum number of bowel movements within a given cycle. This represents the minimum number of bowel movements within a cycle. ; In the above formula, C is the physiological health coefficient of defecation, and W is the custom external influence coefficient, which takes the value [0, 1].

[0021] The target parameters for infant vital sign monitoring are output based on a correlation data model, specifically including the following: S400: Acquire monitored image datasets, environmental monitoring datasets, and physiological characteristic datasets as input data for the correlation data model; S401. Set up the processing layer of the associated data model, and use the established defecation image comparison library, the influence value of target environmental parameters, and the infant defecation health coefficient as the key physical constraints of the processing layer. S402. Construct a neural network architecture based on physical constraints, and construct the loss function of the initial network model. Embed the residual gradient of the equation into the loss function and introduce the gradient penalty term of the equation relating environment and physiology. The monitored image dataset, environmental monitoring dataset, and physiological characteristic dataset are divided into training set and dataset. The range of hidden layer nodes is set according to the sample size of the training set. After training, an association data model is obtained. Based on the association data model, the target parameters for infant vital sign monitoring are output. The target parameters are defecation health coefficient and environmental regulation target.

[0022] Generate an infant's menstrual cycle status report, which includes the following: S500: Acquire parameter data during real-time infant monitoring, perform real-time bowel health status analysis, and generate real-time bowel physiological health coefficient. S501. Based on real-time external environmental parameters, obtain the external environmental influence coefficient, and collect the infant defecation health coefficient according to the set periodic time to obtain the infant defecation health coefficient set. S502. Based on the infant's bowel health coefficient, analyze the changes in bowel movement stages, construct a trend staging curve, analyze the upward and downward trends of the infant's bowel health coefficient, and generate an infant physiological cycle status report. Generate physiological health early warning signals, specifically including the following: S600: Acquire parameter data from real-time infant monitoring, perform stool image comparison against a database to generate feature contrast of the stool images, and generate a stool warning signal based on the feature contrast, which is divided into the following two cases: When the matching degree is greater than the preset standard matching degree, a defecation warning signal is generated; When the matching degree is less than the preset standard matching degree, a similar defecation warning signal is generated; S601. Based on the infant's physiological cycle status report and combined with the real-time defecation physiological health coefficient, a physiological health early warning signal is generated and sent to the intelligent monitoring platform, including the following two situations: When the real-time defecation physiological health coefficient is greater than the preset standard defecation physiological health coefficient, a physiological health warning signal is generated. When the fluctuation value of the physiological health coefficient of periodic defecation exceeds the preset fluctuation standard threshold, a physiological health warning signal is generated.

[0023] The size of the interval and threshold is set to facilitate comparison. The size of the threshold depends on the amount of sample data and the number of bases set by those skilled in the art for each set of sample data; as long as it does not affect the ratio between the parameter and the quantized value.

[0024] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation. In the two embodiments provided in this application, it should be understood that the disclosed apparatus and system can be implemented in other ways; for example, the apparatus embodiments described above are merely illustrative, and the division of modules is merely a logical functional division. In actual implementation, there may be other division methods, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed; furthermore, the coupling or direct coupling or communication connection between the shown or discussed mutuals can be through some interfaces, and the indirect coupling or communication connection between the apparatus or modules can be electrical, mechanical or other forms. The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A non-contact vital sign monitoring platform for abnormal early warning and intelligent supervision, characterized in that, It includes a vital signs monitoring module, a feature analysis module, a real-time monitoring module, and an anomaly early warning module; The vital signs monitoring module monitors the infant's vital signs and external environment through installed sensor devices, acquires parameter data during the infant monitoring process, and generates environmental monitoring datasets, image monitoring datasets, and vital signs datasets after extracting the monitoring data. The feature analysis module acquires the monitored image dataset, extracts and analyzes defecation feature data, establishes a defecation image comparison library, and performs comfort correlation and interaction analysis in conjunction with the environmental monitoring dataset to obtain the impact value of the target environmental parameters. Based on the vital signs dataset, it extracts key physiological features, performs infant defecation health analysis, obtains the infant defecation health coefficient, and constructs a correlation data model. Based on the correlation data model, it outputs the target parameters for infant vital signs monitoring. The real-time monitoring module acquires parameter data during real-time infant monitoring, performs real-time bowel health status analysis to generate a real-time bowel physiological health coefficient, combines real-time external environmental parameters to obtain an external environmental influence coefficient, and analyzes bowel phase changes based on the infant's bowel health coefficient over a period of time to generate an infant physiological cycle status report. The abnormal warning module acquires parameter data from real-time monitoring of the infant, performs a comparison with a defecation image library to generate a feature contrast of the defecation image, generates a defecation warning signal based on the feature contrast, and generates a physiological health warning signal based on the infant's physiological cycle status report and real-time defecation physiological health coefficient, which is then sent to the intelligent supervision platform.

2. The non-contact vital sign monitoring abnormality early warning and intelligent supervision platform according to claim 1, characterized in that, The intelligent monitoring platform receives defecation warning signals and physiological health warning signals, and sends them to the user terminal and the medical monitoring terminal. The user terminal intelligently pushes events to users based on defecation warning signals. The push notifications are based on the confidence level of the defecation warning signals. Push notifications include generating pending confirmation message pushes, pop-up sound reminder pushes, and intelligent alarm pushes. The message push pending confirmation time threshold is set, and a reminder push is sent again based on the pending confirmation time threshold. Based on physiological health warning signals, the medical staff will analyze the infant's long-term defecation trend. When the number of defecation periods continues to decline and the fluctuation exceeds the set standard threshold, the medical staff will automatically trigger the intervention strategy recommended by the medical staff. By establishing a user practice feedback and evaluation mechanism between the medical staff and the user end, the system obtains the number of times push messages sent by users on the intelligent monitoring platform are received and confirmed. Based on the number of confirmed receipts, the system makes intelligent adjustments to the user end, and promptly adjusts and optimizes service content and processes.

3. The non-contact vital sign monitoring abnormality early warning and intelligent supervision platform according to claim 1, characterized in that, Generate environmental monitoring datasets, image monitoring datasets, and vital sign datasets, specifically including the following: S100. Using non-contact sensor devices, including temperature sensors, camera devices, and millimeter-wave radar sensors, the sampling frequency and sampling time are set according to the monitoring requirements and the performance characteristics of the sensors. S101. Acquire images through a camera, locate and identify the baby in the image, extract image feature data related to the baby's defecation, and construct an image dataset; S102. Acquire external environmental data through temperature sensors, extract key data, organize and store the extracted external environmental parameter data in chronological order, and construct an environmental monitoring dataset; integrate the number of times and time of infant defecation to construct a vital signs dataset.

4. The non-contact vital sign monitoring abnormality early warning and intelligent supervision platform according to claim 1, characterized in that, The impact values ​​of the target environmental parameters are obtained, specifically including the following: S200. Obtain the monitored image dataset, extract and analyze defecation feature data to obtain defecation images, classify the extracted defecation images into diaper image features before defecation and diaper image features after defecation, store the labeled defecation images in the database, and establish a defecation image comparison library. S201. Integrate the environmental monitoring dataset and the defecation image dataset, align the timestamps of the two datasets, and perform a correlation analysis between external temperature and temperature during defecation. S202. Based on the external temperature and the temperature at the time of defecation, the deviation between the two is obtained. Then, based on the infant's body temperature and a personal comfort temperature setting, an external temperature control analysis is performed to obtain the target environmental parameter influence value R. The calculation process is as follows. ； In the above formula, For the baby's personal comfort temperature, Environmental monitoring data is concentrated on the external temperature at timestamp t. The defecation image data focuses on the temperature at the time of defecation at timestamp d.

5. The non-contact vital sign monitoring abnormality early warning and intelligent supervision platform according to claim 1, characterized in that, The above analysis of infant bowel health yields an infant bowel health coefficient, which includes the following: S300: Acquire parameter data during infant monitoring, obtain the infant's physiological characteristic dataset, perform bowel health status analysis, classify bowel movements into normal, diarrhea, and constipation types, organize the classified bowel images, and add detailed annotation information to each image, including bowel movement type, collection time, and basic infant information. S301. Based on the frequency, timing, and type of the infant's bowel movements, conduct an analysis of the infant's bowel health to obtain an estimated range for healthy bowel movement frequency, as well as a range for the frequency of diarrhea and constipation in the infant. ]; S302. Based on the obtained defecation health status, and combined with the external influence coefficient, a defecation physiological health coefficient is generated. The calculation process is as follows: The average defecation interval T is initially calculated. ； In the above formula, These are the start and end times of the periodic time interval. The time of the first bowel movement preceding the first bowel movement. Let N be the time of the i-th defecation, where N = {1, 2, 3, ..., N}; ； In the above formula, H represents bowel health. These are the number of bowel movements, the number of times constipated, and the number of times diarrhea occur within a given period. This represents the maximum number of bowel movements within a given cycle. This represents the minimum number of bowel movements within a cycle. ； In the above formula, C is the physiological health coefficient of defecation, and W is the custom external influence coefficient, which takes the value [0, 1].

6. The non-contact vital sign monitoring abnormality early warning and intelligent supervision platform according to claim 1, characterized in that, The target parameters for infant vital sign monitoring are output based on a correlation data model, specifically including the following: S400: Acquire monitored image datasets, environmental monitoring datasets, and physiological characteristic datasets as input data for the correlation data model; S401. Set up the processing layer of the associated data model, and use the established defecation image comparison library, the influence value of target environmental parameters, and the infant defecation health coefficient as the key physical constraints of the processing layer. S402. Construct a neural network architecture based on physical constraints, and construct the loss function of the initial network model. Embed the residual gradient of the equation into the loss function and introduce the gradient penalty term of the equation relating environment and physiology. The monitored image dataset, environmental monitoring dataset, and physiological characteristic dataset are divided into training set and dataset. The range of hidden layer nodes is set according to the sample size of the training set. After training, an association data model is obtained. Based on the association data model, the target parameters for infant vital sign monitoring are output. The target parameters are defecation health coefficient and environmental regulation target.

7. The non-contact vital sign monitoring abnormality early warning and intelligent supervision platform according to claim 1, characterized in that, Generate an infant's menstrual cycle status report, which includes the following: S500: Acquire parameter data during real-time infant monitoring, perform real-time bowel health status analysis, and generate real-time bowel physiological health coefficient. S501. Based on real-time external environmental parameters, obtain the external environmental influence coefficient, and collect the infant defecation health coefficient according to the set periodic time to obtain the infant defecation health coefficient set. S502. Based on the infant's defecation health coefficient, conduct a defecation stage change analysis, construct a trend stage curve, analyze the rising and falling trends of the infant's defecation health coefficient, and generate an infant physiological cycle status report.

8. The non-contact vital sign monitoring abnormality early warning and intelligent supervision platform according to claim 1, characterized in that, Generate physiological health early warning signals, specifically including the following: S600: Acquires real-time monitoring parameter data of the infant, performs stool image comparison against a database to generate feature contrast of the stool images, and generates a stool warning signal based on the feature contrast, which is divided into the following two cases: When the matching degree is greater than the preset standard matching degree, a defecation warning signal is generated; When the matching degree is less than the preset standard matching degree, a similar defecation warning signal is generated; S601. Based on the infant's physiological cycle status report and combined with the real-time defecation physiological health coefficient, a physiological health early warning signal is generated and sent to the intelligent monitoring platform, including the following two situations: When the real-time defecation physiological health coefficient is greater than the preset standard defecation physiological health coefficient, a physiological health warning signal is generated. When the fluctuation value of the physiological health coefficient of periodic defecation exceeds the preset fluctuation standard threshold, a physiological health warning signal is generated.

Images