Blood oxygen saturation detection method and blood oxygen saturation detection device

Abstract

The invention relates to the technical field of biomedical engineering, and discloses a blood oxygen saturation degree detection method which comprises the following steps: S1, a multi-wavelength light source is utilized to irradiate a target area, and the light source comprises light with different wavelengths in a visible spectrum range; s2, collecting intensity information of light reflected or transmitted by the tissue; s3, constructing an oxyhemoglobin saturation model according to the received light intensity ratios under different wavelengths, and further provides an oxyhemoglobin saturation detection device which comprises a multi-wavelength light source module used for emitting light with different wavelengths in a visible spectrum range; and the optical sensor array is used for collecting the intensity information of the light reflected or transmitted by the skin tissue. By means of the multi-wavelength light source, machine learning optimization, personalized calibration and the environment sensing module, the accuracy and stability of blood oxygen measurement are remarkably improved, different individual differences and complex environment conditions can be adapted, and accurate blood oxygen saturation monitoring is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of biomedical engineering, and in particular to a blood oxygen saturation detection method and a blood oxygen saturation detection device. Background Art

[0002] In daily life, blood oxygen saturation is an important indicator of human health. This is especially true for patients with heart disease, lung disease, or other chronic conditions. Blood oxygen monitoring can effectively help doctors assess their condition. Furthermore, athletes, the elderly, and those living at high altitudes often experience changes in blood oxygen levels that can affect their health. Therefore, convenient and accurate blood oxygen measurement has become a fundamental requirement for daily health monitoring.

[0003] Existing methods for measuring blood oxygen saturation typically use photoplethysmography (PPG). This technique calculates blood oxygen saturation by illuminating the skin's surface with a light source and measuring differences in light absorption. Traditional finger-clip oximeters are based on this principle and are widely used in hospitals, homes, and outdoor settings. Due to its non-invasive nature, simplicity, and low cost, this method has become a widely used health monitoring tool, particularly in hospitals and emergency rooms, where real-time blood oxygen data is now standard. This technology is suitable for measuring blood oxygen saturation in common situations and offers considerable convenience and affordability.

[0004] However, the application of existing technologies in complex environments has some shortcomings. First, a single-wavelength light source will produce large errors when faced with low blood flow or strenuous exercise, resulting in unstable measurement results. Second, traditional equipment lacks personalized calibration functions and cannot take into account differences in users' skin color, body shape and health status, resulting in large differences in measurement results of the same device in different populations. In addition, there is no effective compensation for the interference of environmental factors (such as temperature, humidity and light changes) on the measurement results, which often causes the blood oxygen measurement value to deviate from the actual value. Summary of the Invention

[0005] In response to the shortcomings of the existing technology, the present invention provides a blood oxygen saturation detection method and a blood oxygen saturation detection device, which solve the problems of inaccurate measurement in complex environments, failure to calibrate individual differences, and failure to compensate for environmental interference in existing blood oxygen measurement technology.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: A method for detecting blood oxygen saturation, comprising the following steps: S1. irradiating the target area with a multi-wavelength light source, the light source comprising light of different wavelengths within the visible spectrum; S2, collecting light intensity information after being reflected or transmitted through tissue; S3. Constructing a blood oxygen saturation model based on the ratio of the received light intensities at different wavelengths; S4. Apply machine learning algorithms to analyze the collected data, correct individual differences and other interference factors, and obtain blood oxygen saturation values; S5. Personalize the calibration of blood oxygen saturation based on health records to eliminate the influence of individual differences; S6. Use the environmental perception module to detect environmental changes and automatically adjust the light source intensity to compensate for the impact of environmental factors on the measurement.

[0007] Preferably, the multi-wavelength light source further comprises: Use multiple red light wavelengths and multiple infrared light wavelengths and illuminate the target area separately to achieve accurate measurement of blood oxygen saturation.

[0008] Preferably, the collected light intensity information includes: Absorption of red and infrared light, which is used to estimate the ratio of oxygenated to deoxygenated hemoglobin in the blood; The intensity of the light signal passing through the skin tissue is used to reflect changes in blood oxygen concentration.

[0009] Preferably, the blood oxygen saturation model constructed according to the received light intensity ratio further includes: A mathematical model is constructed based on the light intensity ratio under different wavelength light sources, wherein the model is used to describe the relationship between blood oxygen concentration and light absorption; Use machine learning methods to optimize the accuracy of blood oxygen saturation models.

[0010] Preferably, the machine learning algorithm includes: A combination of convolutional neural networks (CNN) and recurrent neural networks (RNN) to analyze and predict changes in blood oxygen saturation; Use the training dataset to calibrate and optimize the algorithm to improve the accuracy and stability of blood oxygen detection.

[0011] Preferably, the personalized calibration is based on: When the user uses it for the first time, a personalized baseline is established through blood oxygen data; Based on the user's health records and historical blood oxygen data, the calibration parameters are dynamically adjusted to improve the accuracy of blood oxygen measurement.

[0012] Preferably, the personalized calibration technique is optimized based on the user's health profile, which includes information on the user's age, gender, health status, and history of chronic diseases.

[0013] Preferably, the environment perception module includes: Temperature and humidity sensor, used to monitor ambient temperature and humidity; Air pressure sensor, used to monitor air pressure changes; Ambient light sensor, used to monitor the surrounding light intensity; The environmental parameters are used to adjust the intensity or wavelength of the light source to compensate for the impact of environmental changes on blood oxygen measurement.

[0014] Preferably, the automatically adjusting the light source intensity includes: Detect air pressure changes and adjust the wavelength and intensity of the light source according to the air pressure changes; Detects changes in ambient temperature and humidity and adjusts the device's operating mode to maintain measurement accuracy.

[0015] The present invention also provides a blood oxygen saturation detection device, comprising: A multi-wavelength light source module for emitting light of different wavelengths within the visible spectrum; An optical sensor array for collecting light intensity information after being reflected or transmitted through skin tissue; a data processing unit, configured to construct a blood oxygen saturation model based on the received light intensity information and apply a machine learning algorithm to analyze the data; a signal processing unit for filtering noise and enhancing the signal-to-noise ratio; Environmental sensing module, used to detect environmental parameters, including temperature, humidity, and air pressure, to adjust light sources and device operating modes; The display and feedback unit is used to display the blood oxygen saturation value in real time and provide personalized health feedback and suggestions based on the user's health profile.

[0016] The present invention provides a blood oxygen saturation detection method and blood oxygen saturation detection device, which have the following beneficial effects: 1. This invention introduces personalized calibration, combines each user's health profile, and automatically adjusts the blood oxygen measurement model, enabling accurate calculations for each individual's blood oxygen level. Unlike traditional methods, this approach eliminates the influence of factors such as skin color and health status, improving measurement consistency across individuals.

[0017] 2. The environmental perception module of the present invention can monitor environmental changes in real time and automatically adjust the light source intensity to compensate for environmental influences. This allows the system to maintain accurate measurements even in complex environments, avoiding measurement deviations caused by environmental factors in traditional methods.

[0018] 3. Through machine learning, this method can adjust the measurement model based on real-time data and user characteristics. Compared with traditional fixed algorithms, this method is more flexible and precise, can effectively adapt to various complex environments and the needs of different individuals, and improve the accuracy and stability of blood oxygen measurement. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a flow chart of the method of the present invention; Figure 2 This is a module architecture diagram of the present invention. DETAILED DESCRIPTION

[0020] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the specification of the present invention. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0021] Please see the attached Figure 1 , an embodiment of the present invention provides a blood oxygen saturation detection method, comprising the following steps: S1. irradiating the target area with a multi-wavelength light source, the light source comprising light of different wavelengths within the visible spectrum; S1 illuminates the target area with a multi-wavelength light source, laying the foundation for subsequent optical signal collection, data processing, and blood oxygen model construction. This step involves illuminating the human body surface (such as the skin or blood vessels) with light sources of different wavelengths and using the differences in blood absorption of light to provide data for subsequent analysis.

[0022] In this embodiment, the multi-wavelength light source module includes at least two light sources with different wavelengths, typically red and infrared. Specifically, the red wavelength is approximately 660nm, and the infrared wavelength is approximately 940nm. These two wavelengths absorb light to varying degrees with oxyhemoglobin (HbO2) and deoxyhemoglobin (Hb), respectively. Therefore, by measuring the difference in their absorbance, the concentration ratio of oxyhemoglobin to deoxyhemoglobin in the blood can be calculated. This ratio serves as an important reference for blood oxygen saturation (SpO2).

[0023] As an option, in some embodiments, additional wavelengths of light may be added based on application requirements. For example, these may include green light (520nm) or blue light (480nm). These wavelengths can further enhance blood oxygen saturation measurement accuracy, particularly in challenging environments (such as dark skin or poor blood flow). The use of these additional wavelengths helps provide more optical information, thereby increasing data diversity and improving the reliability of test results.

[0024] Specifically, multi-wavelength light sources utilize multiple light sources to illuminate the target area alternately or simultaneously. In some embodiments, to improve system adaptability and accuracy, the system can utilize light sources of different wavelengths in separate time periods, or simultaneously utilize light of different wavelengths at the same time. This design ensures sufficient data collection while minimizing errors caused by time differences or environmental changes.

[0025] The light source's illumination pattern is designed to ensure even distribution of light across the target area (e.g., skin, vascular area, or specific measurement location). By properly controlling the light source's illumination angle, emission intensity, and excitation time, sufficient illumination of the target area can be ensured, avoiding data errors caused by inappropriate light source angles or low light intensity.

[0026] In this embodiment, light sources of multiple wavelengths are directed onto the skin surface or measurement area through an optical system (e.g., lenses, reflectors, etc.), ensuring uniform illumination of the target area. By measuring the reflection or transmission of these light signals, the system can obtain absorption information of light at different wavelengths, providing basic data for subsequent blood oxygen concentration calculations.

[0027] In one possible implementation, the light source module will utilize an integrated LED array. This array design allows for the integration of multiple wavelength LEDs within a single module, enabling illumination of different wavelengths through simple switching or combination. LEDs offer advantages such as low power consumption and long lifespan, while also providing high illumination intensity within a given size and power consumption.

[0028] Regarding the wavelength selection and emission intensity control of the light source, during implementation, the system takes into account the absorption and reflection mechanism of light. It will dynamically adjust the emission intensity of the light source according to different skin types, blood flow conditions, and the specific target area to be measured. The optical absorption formula is as follows: ; in, Indicates absorbance (degree of light absorption); is the molar absorptivity of the substance; is the concentration of the substance; is the path length of light through the sample.

[0029] This formula reflects the relationship between a substance's ability to absorb light, its concentration, and the length of the light path. When measuring blood oxygen saturation, oxyhemoglobin and deoxyhemoglobin in the blood have different absorption characteristics for light of different wavelengths. Therefore, by measuring changes in absorbance at different wavelengths, the concentrations of oxyhemoglobin and deoxyhemoglobin in the blood can be inferred. By selecting the appropriate wavelength and using the above formula, blood oxygen saturation can be effectively estimated.

[0030] In some embodiments, the application of the formula can be further optimized. For example, for the light absorption characteristics of skin, the skin's absorption coefficient can be obtained through experimental data and combined with the absorption coefficient of hemoglobin to further improve the accuracy of the relationship model between light intensity and absorbance.

[0031] In this embodiment, to improve measurement accuracy, additional wavelengths of light can be incorporated, such as yellow light (580nm) or ultraviolet light (350nm). These light sources can effectively interact with different blood components, providing more measurement data and more comprehensive information about blood oxygen status. By using these complementary wavelengths, the system can cover a wider range of blood composition variations and enhance its ability to accurately measure even in complex situations, such as low blood flow and motion artifacts.

[0032] Overall, the multi-wavelength light source illumination in the S1 is a key step in blood oxygen saturation detection. By providing rich optical data, it effectively identifies and analyzes changes in oxygenated and deoxygenated hemoglobin in the blood, thereby accurately estimating blood oxygen saturation. This design step not only improves measurement accuracy and reliability, but also provides sufficient data support for subsequent analysis and processing through the diversification and intelligent adjustment of the light source.

[0033] By applying this multi-wavelength light source, combined with appropriate optical models and machine learning algorithms, the blood oxygen saturation detection method of the present invention can achieve high-precision measurements in a variety of different environments and under different individual conditions.

[0034] S2, collecting light intensity information after being reflected or transmitted through tissue; The goal of collecting light intensity information is to obtain key information about the ratio of oxygenated hemoglobin to deoxygenated hemoglobin in the blood by accurately collecting light signals reflected or transmitted back from a target area (such as the skin surface or vascular tissue).

[0035] In step S2, the light intensity information collected primarily includes the intensity of light reflected from the target area after illuminating it with red (660nm) and infrared (940nm) light sources. This light intensity data is directly affected by the concentrations of oxyhemoglobin and deoxyhemoglobin in the blood. Because oxyhemoglobin and deoxyhemoglobin have different absorbances for different wavelengths of light, the optical sensor array collects this light intensity difference, providing the necessary data for calculating blood oxygen saturation.

[0036] Generally speaking, oxyhemoglobin absorbs red light more strongly, while deoxyhemoglobin absorbs infrared light more strongly. By collecting the reflected signals of red and infrared light, the relative concentrations of oxyhemoglobin and deoxyhemoglobin can be determined. Specifically, the intensity of the reflected light signal varies with the changes in these components in the blood. The system calculates the blood oxygen saturation (SpO2) by calculating the ratio of the light intensities.

[0037] Alternatively, in some embodiments, the optical sensor array may utilize higher-resolution sensors or additional wavelength light sources (e.g., green light, blue light, etc.). These additional wavelengths can enhance sensitivity to different blood components, especially in conditions with slow blood circulation or darker skin. Adding multiple wavelengths of light helps obtain more comprehensive light absorption information, thereby improving blood oxygen measurement accuracy.

[0038] In practice, an optical sensor array typically consists of multiple photodiodes or photosensors, which simultaneously collect reflection signals from light sources of different wavelengths. These sensor arrays may be distributed across the target area to ensure uniform light signal collection and reduce errors caused by localized light inhomogeneity. By collecting signals from multiple locations, data reliability and measurement accuracy can be further improved.

[0039] In one possible implementation, the signal acquisition process is real-time. Whenever a light source illuminates a target area, the sensor rapidly collects the reflected light signal and transmits it to a signal processing unit. This process occurs continuously, ensuring sufficient data for analysis within a short period of time. Furthermore, the sensor array design can be customized to accommodate the shape and size of different measurement areas, ensuring stability and data integrity under various environmental conditions.

[0040] In other embodiments, the optical signals collected by the sensor may be affected by external factors such as ambient lighting and user motion. Therefore, preliminary noise filtering and data smoothing are performed during signal processing. The signal processing unit improves signal quality by removing ambient light interference and motion artifacts, providing more accurate input data for subsequent blood oxygen calculations.

[0041] After collecting sufficient light intensity information, the system further processes the collected light intensity data and uses the known absorbance formula to perform calculations. In step S1, we mentioned the relationship between absorbance, substance concentration, and path length. For blood oxygen measurement, this formula is based on the Beer-Lambert law, which states that light absorbance is closely related to the concentrations of oxyhemoglobin and deoxyhemoglobin.

[0042] By measuring changes in light intensity at different wavelengths, the system can calculate the concentrations of oxyhemoglobin and deoxyhemoglobin in the blood based on this formula, providing data support for calculating blood oxygen saturation. Changes in light intensity at different wavelengths can reveal the relative concentrations of these hemoglobin components in the blood, further calculating the blood oxygen concentration.

[0043] Specifically, the system uses a ratio algorithm to calculate blood oxygen saturation based on the absorption of red and infrared light at different wavelengths. Because oxyhemoglobin and deoxyhemoglobin absorb light of different wavelengths differently, the system calculates blood oxygen saturation based on the difference in absorbance between red and infrared light. For example, the ratio of absorbance at red and infrared wavelengths is an important indicator and can be expressed by the following formula: ; in, and represent the absorbance of red light and infrared light respectively; and They are oxyhemoglobin and deoxyhemoglobin.

[0044] In another embodiment, comprehensive analysis of the absorption of more wavelengths of light can further improve measurement accuracy. The system incorporates the ratio of light intensities at different wavelengths into a model and uses a machine learning algorithm to optimize the calculation, ultimately yielding a more accurate blood oxygen saturation value.

[0045] Through precise light intensity information collection and analysis, the present invention can provide accurate blood oxygen saturation measurement results, providing a solid foundation for subsequent data processing and calculation.

[0046] S3. Constructing a blood oxygen saturation model based on the ratio of the received light intensities at different wavelengths; S3 is used to derive blood oxygen saturation from the collected light intensity data. In this step, the system constructs a mathematical model based on the absorbance ratio of light sources of different wavelengths to calculate blood oxygen saturation. This process relies not only on simple mathematical operations but also combines the light intensity data from the previous step (S2) with a machine learning algorithm to optimize the model's accuracy. Model establishment is a key step in data processing, directly determining the accuracy and stability of blood oxygen measurement.

[0047] In this embodiment, in step S3, the system first performs preliminary processing on the light intensity ratio data collected in step S2. Specifically, the system analyzes the reflected intensities of red and infrared wavelengths to infer the concentration ratio of oxyhemoglobin (HbO2) to deoxyhemoglobin (Hb) in the blood. This ratio is crucial for calculating blood oxygen saturation (SpO2). By modeling the light intensity ratio at different wavelengths, the system can derive the mathematical relationship between blood oxygen saturation and light absorption, thus supporting subsequent blood oxygen calculations.

[0048] Typically, collected light intensity data is converted to absorbance using a ratio formula. The difference in absorption of light of different wavelengths by oxyhemoglobin and deoxyhemoglobin in the blood determines the change in measured absorbance. Red light and infrared light have different absorption properties for these two hemoglobin components. Therefore, changes in the red to infrared absorbance ratio directly reflect the relative proportions of oxyhemoglobin and deoxyhemoglobin in the blood.

[0049] As an alternative, in addition to using the basic calculation of light intensity ratios, models can be optimized using methods such as polynomial fitting and regression analysis. These methods can further improve the accuracy of blood oxygen saturation calculations, especially in situations with complex blood composition or significant environmental interference, reducing errors and providing more accurate results.

[0050] Specifically, when building the blood oxygen saturation model, the system will first calculate the absorbance using the Beer-Lambert law. This formula has been introduced in step S2. Its function is to establish a mathematical relationship between the absorbance of light and physical quantities such as the concentration of the substance and the path length of light. Specifically, in the blood oxygen saturation measurement, the system calculates the corresponding absorbance based on the ratio of light intensities at different wavelengths, and then infers the concentration ratio of oxyhemoglobin and deoxyhemoglobin in the blood. This ratio can ultimately be used to calculate blood oxygen saturation.

[0051] In one embodiment, this method can be used to calculate the ratio of oxygenated to deoxygenated hemoglobin in the blood using the absorbance ratio of light sources at different wavelengths, combined with known absorbance formulas. Typically, oxygenated hemoglobin absorbs red light more strongly, while deoxygenated hemoglobin absorbs infrared light more strongly. Therefore, by calculating the ratio of light intensities at different wavelengths, the system can determine changes in blood oxygen concentration and, in turn, calculate blood oxygen saturation.

[0052] Furthermore, to improve model accuracy, especially in the presence of interfering factors such as motion artifacts and low blood flow, the system can further optimize the model using machine learning algorithms. Machine learning algorithms, such as support vector machines (SVM), random forests (RF), or neural networks (NN), can analyze large amounts of training data to identify the complex relationship between blood oxygen saturation and light intensity ratio. Specifically, the system first uses the training data set to train the optimal model parameters through a machine learning algorithm, and then predicts blood oxygen saturation based on the new light intensity ratio data.

[0053] In some embodiments, during data processing, the system can also incorporate additional physiological characteristic data, such as the user's heart rate, respiratory rate, and body temperature, to further optimize blood oxygen saturation prediction. This additional data can be input into the model along with the light intensity ratio data through multivariate regression analysis for comprehensive analysis, thereby improving the accuracy of blood oxygen saturation calculation.

[0054] In one possible implementation, to further enhance the adaptability of the blood oxygen saturation model, the system will perform personalized calibration based on the user's health profile. By dynamically updating the user's blood oxygen saturation model, the system can adjust the parameters used in the blood oxygen calculation process based on individual differences such as age, gender, health status, and long-term health data. This process can effectively improve the accuracy of measurement results, especially for patients with chronic diseases or other special populations.

[0055] By analyzing the absorbance ratio of red and infrared light wavelengths, combined with the Beer-Lambert law and machine learning algorithms, the system can accurately calculate the ratio of oxygenated hemoglobin to deoxygenated hemoglobin in the blood.

[0056] S4. Apply machine learning algorithms to analyze the collected data, correct individual differences and other interference factors, and obtain blood oxygen saturation values; By analyzing the light intensity ratio data collected in the previous steps using a machine learning algorithm, the system can effectively correct for individual differences and interference from the environment or equipment, thereby improving the accuracy of blood oxygen saturation calculations. The machine learning algorithm not only processes data based on the light intensity ratio in step S3 but also further optimizes the model based on individual health data, ensuring accurate blood oxygen saturation predictions across a wide range of environments and individual differences.

[0057] In this embodiment, the application of machine learning algorithms is primarily reflected in two aspects: first, by establishing a mathematical model between blood oxygen saturation and light intensity ratio, and second, by integrating individual health records for personalized calibration. In step S3, a preliminary blood oxygen saturation model has been constructed using light intensity ratio. This model needs to be further optimized using machine learning algorithms to adapt to individual physiological differences, environmental factors, and other interfering factors.

[0058] Generally, machine learning algorithms can use regression analysis methods, such as linear regression or nonlinear regression models, to fit the relationship between blood oxygen saturation and input data (light intensity ratio, health records, etc.). The model first learns from the training data set and automatically adjusts various model parameters to ensure that the predicted blood oxygen saturation results are as close to the actual value as possible. In certain embodiments, algorithms such as support vector machine regression (SVR) or decision tree regression can also be used to build the prediction model. These algorithms can handle data with highly nonlinear characteristics and are suitable for complex blood oxygen saturation prediction tasks.

[0059] Alternatively, more complex machine learning methods, such as deep learning algorithms like convolutional neural networks (CNNs) and recurrent neural networks (RNNs), can also be used to train and optimize blood oxygen saturation models. CNNs are particularly adept at extracting spatial features from large-scale data, while RNNs are capable of capturing temporal relationships in time series data. Through these deep learning methods, the system can automatically learn the complex relationship between blood oxygen saturation and light intensity data. This is particularly true in situations such as unstable environmental conditions and user motion artifacts. Deep learning methods can better adapt to complex data patterns, thereby improving prediction accuracy.

[0060] Specifically, the machine learning algorithm uses the light intensity ratio data collected from steps S2 and S3 as input and is trained in combination with the existing data in the blood oxygen saturation model. The goal of the training process is to make the model's output blood oxygen saturation value (SpO2) match the actual blood oxygen value as accurately as possible. In order to improve the algorithm's prediction accuracy, the system calculates the error between the predicted value and the actual value through a loss function and adjusts the model parameters based on these errors. Common loss functions include mean square error (MSE) or root mean square error (RMSE), which are calculated as follows: ; in, is the value of the loss function; is the sample size; It is The actual blood oxygen value of each sample; is the blood oxygen value predicted by the model.

[0061] In some embodiments, the system can also incorporate individual health records (such as age, gender, heart rate, and other physiological data) to optimize the blood oxygen saturation prediction model. Due to individual physiological differences, a simple calculation of light intensity ratios may not accurately reflect an individual's blood oxygen saturation. Therefore, by incorporating individual health data into the machine learning model, the system can be personalized for different users, thereby improving measurement accuracy.

[0062] As an option, during the personalized calibration process, the system performs a baseline calibration for each user upon first use, building a personalized model based on the user's health record and blood oxygen measurement data. By tracking each user's health record over time, the system automatically adjusts the model parameters with each measurement, achieving more accurate blood oxygen saturation predictions.

[0063] Furthermore, to further improve the accuracy of blood oxygen saturation prediction, the system can employ ensemble learning methods. For example, the Random Forest or XGBoost algorithms can train different base learners multiple times and combine the prediction results of these learners to obtain the final prediction value. These ensemble methods are generally effective in improving model stability, especially in cases of high data noise, reducing overfitting and improving model generalization.

[0064] In certain embodiments, to adapt to changing environmental conditions, the system can adjust the model's predictions by monitoring external environmental factors (such as temperature, humidity, and air pressure) in real time. For example, at high altitudes, where oxygen concentrations are lower, the calculation of blood oxygen saturation may require appropriate adjustments. The system can adjust the algorithm parameters in real time based on environmental changes to ensure the accuracy of predictions.

[0065] Step S4 applies a machine learning algorithm to optimize the blood oxygen saturation calculation model, effectively correcting for individual differences and environmental interference. By combining multiple dimensions of information, such as light intensity ratio, user health records, and environmental data, the system can provide a more personalized and accurate blood oxygen saturation prediction for each user.

[0066] S5. Personalize the calibration of blood oxygen saturation based on health records to eliminate the influence of individual differences; Step S5 performs personalized calibration of blood oxygen saturation based on each user's health profile. This personalized calibration effectively eliminates errors caused by individual differences or external environmental factors by combining the user's physiological characteristics with the machine learning model, ensuring accurate blood oxygen saturation calculations. This step is closely related to step S4, relying on a blood oxygen model optimized through machine learning and comprehensively adjusting it with the user's health profile.

[0067] In this embodiment, in step S5, the first step of personalized calibration is to collect the user's health record data. This health record includes, but is not limited to, physiological and health data such as the user's age, gender, weight, height, presence of chronic diseases (such as heart disease, diabetes, lung disease, etc.), and exercise habits. The system uses this data to make preliminary adjustments to the blood oxygen saturation model. The personalized model will make corresponding adjustments to data processing and algorithm model parameters based on the user's specific health status, lifestyle, and other factors, thereby reducing or eliminating errors caused by individual differences.

[0068] Typically, when a user first uses a blood oxygen monitoring device, the system asks them to enter their personal health information. This information enables the system to perform a personalized calibration of the blood oxygen saturation measurement model. During this process, the system compares collected health record data with historical measurement data (if available) based on a preset calibration algorithm. Through this comparative analysis, the system automatically calculates a personalized calibration factor and applies it to the blood oxygen saturation measurement model, thereby improving prediction accuracy.

[0069] Alternatively, in some embodiments, a user's health profile may also include their long-term blood oxygen measurement data. By tracking this data over time, the system can more accurately understand the user's normal blood oxygen range and dynamically update the model based on this data. For example, if a user's long-term blood oxygen level is below the normal range, the system can make appropriate adjustments to this special situation in subsequent measurements. By incorporating long-term data, the system can optimize the personalized model based on the user's health trends.

[0070] Specifically, personalized calibration involves adjusting parameters in multiple areas. For example, the user's age and gender may have a certain impact on blood oxygen levels. Older users typically have lower blood oxygen saturation, and there may also be certain differences in blood oxygen concentration between men and women. Based on this, the system uses statistical analysis to factor these influencing factors into the model parameters. In this way, when the user enters their age, gender, and other data, the system can dynamically adjust the model to suit their specific physiological characteristics.

[0071] In one possible implementation, the blood oxygen model is further calibrated based on data such as the user's weight and height. Heavier users may face a higher burden on their respiratory system, resulting in different blood oxygen saturation. Therefore, the system can infer a specific blood oxygen standard value based on weight and height and adjust the model parameters accordingly. In addition, the user's health record may also include lifestyle factors such as smoking and drinking, which can have a significant impact on blood oxygen levels. The system should also effectively model this data.

[0072] Furthermore, the implementation of personalized calibration requires dynamic optimization based on a machine learning algorithm. In step S4, the machine learning algorithm has already performed preliminary data processing and output a predicted blood oxygen saturation value. Based on this, personalized calibration fine-tunes the algorithm according to the user's health record. For example, using algorithms such as support vector machines (SVM), decision tree regression, or random forests, the system can reprocess the prediction results based on individual differences. The core purpose of the machine learning model is to dynamically adjust the prediction parameters by learning from the health records and blood oxygen measurement data of different users, so that the blood oxygen saturation prediction results of each user are as close to the actual value as possible.

[0073] In some embodiments, the system can also incorporate incremental learning techniques during the personalized calibration process to further improve model accuracy. Incremental learning allows the system to gradually adjust the existing model as new data arrives, without having to retrain the entire dataset. Incremental learning is particularly useful for long-term health monitoring, as it can adapt to changes in the user's health status in real time. For example, if a user's health status changes significantly (such as developing a respiratory disease or a significant change in weight), the incremental learning mechanism can quickly adjust the model to ensure accurate blood oxygen measurement.

[0074] In some embodiments, the system can also use ensemble learning to combine the prediction results of multiple personalized models. By weighted averaging or voting the results of multiple models, the system can reduce the overfitting problem that may arise from a single model and improve the stability of predictions. For example, both random forest and XGBoost algorithms can effectively integrate multiple prediction models to improve the reliability of blood oxygen saturation measurements.

[0075] The core of personalized calibration is to dynamically adjust the blood oxygen saturation model by combining the user's health records, long-term data, and machine learning algorithms. In this way, the system can eliminate the influence of individual differences and improve the accuracy of blood oxygen measurement.

[0076] S6. Use the environmental perception module to detect environmental changes and automatically adjust the light source intensity to compensate for the impact of environmental factors on the measurement; The primary task of the environmental perception module is to adjust light source intensity by real-time monitoring of environmental changes (such as temperature, humidity, air pressure, and light intensity) to compensate for their impact on blood oxygen saturation measurements. This adjustment process ensures that the system can provide highly accurate blood oxygen saturation measurements under varying environmental conditions. Environmental changes can alter the way light propagates through the air, causing variations in signal strength and path, which in turn can affect blood oxygen saturation calculations. Therefore, automatically adjusting light source intensity is an effective compensation method to reduce interference from these factors and ensure measurement stability and accuracy.

[0077] In this embodiment, the environmental perception module uses multiple sensors to monitor environmental parameters in real time, primarily including temperature, humidity, air pressure, and light intensity. Various environmental factors can affect the performance of light sources. For example, rising ambient temperature can cause changes in the brightness of a light source, increased humidity can affect the refractive index of light, and changes in air pressure can have a certain impact on air density and the speed of light propagation. By monitoring these environmental factors in real time, the system can calculate the appropriate light source intensity adjustment to ensure measurement accuracy.

[0078] Generally speaking, temperature, humidity, and air pressure are important factors that influence the intensity of light sources and the path of light propagation. Temperature changes can cause variations in the light source's luminous efficiency, which in turn affects light intensity. Humidity changes can affect the air's refractive index, altering the speed and path of light propagation. Pressure changes can affect air density, which in turn affects the degree of light refraction. Therefore, the system uses these sensor data to monitor environmental changes in real time and calculate the necessary adjustments to the light source's intensity.

[0079] Alternatively, an ambient light intensity sensor could be used to monitor ambient lighting conditions, particularly in outdoor environments where strong sunlight can affect blood oxygen measurement accuracy. By monitoring ambient light intensity in real time, the system can adjust the light source output appropriately to ensure that the optical signal is not disturbed by external light. The light sensor input, along with other environmental parameters, serves as a reference for adjusting the light source.

[0080] Specifically, the system uses the temperature, humidity, air pressure, and light intensity data collected by the environmental sensing module as input and uses a preset algorithm to calculate the intensity of the light source adjustment. The goal of adjusting the light source intensity is to compensate for the impact of environmental factors on the measurement results, making the final blood oxygen saturation value more accurate. The impact of environmental changes on blood oxygen measurement can be quantitatively described by the following formula, which includes the relationship between environmental parameters and the amount of light source intensity adjustment: ; in, Indicates the adjusted light intensity; is the intensity of the light source under standard ambient conditions; is the environmental impact coefficient; is the temperature change (unit: °C); is the humidity change (unit: %); is the change in air pressure (unit: Pa); is the change in ambient light intensity (unit: lux).

[0081] Through this formula, the system can dynamically adjust the intensity of the light source according to changes in ambient temperature, humidity, air pressure and light intensity to ensure accurate blood oxygen saturation measurement values in various environments. This formula reflects how environmental factors affect the output of the light source, and by adjusting the coefficient , the system can control the adjustment range of the light source intensity to adapt to different environmental changes.

[0082] In some embodiments, in order to improve the accuracy of the adjustment, the system can further refine the environmental impact coefficient , so that it is more compatible with specific environmental changes and light source characteristics. The influence of factors such as temperature and humidity on light source intensity may not be linear, and the system can be optimized through experimental data or historical data sets. By learning from actual data, the system can gradually adjust the light source intensity adjustment strategy to improve measurement stability.

[0083] In some embodiments, if the system detects a dramatic change in the external environment (e.g., a sudden temperature increase or strong sunlight), it can use more complex compensation models (such as predictive models based on machine learning) to calculate and adjust the light source intensity. These models analyze environmental data to predict future environmental changes and adjust light source settings in advance to reduce the impact of sudden environmental changes on measurement results.

[0084] Furthermore, the system can also use adaptive algorithms to dynamically adjust light intensity. For example, in dim or weak lighting conditions, the system can increase light intensity to compensate. In the case of excessive light, the system will reduce light output to prevent excessive light from interfering with the sensor. This adaptive adjustment allows for more flexible response to lighting adjustment needs in different environments.

[0085] The environmental sensing module monitors environmental changes in real time, and combined with a light intensity adjustment algorithm, effectively compensates for the effects of temperature, humidity, air pressure, and lighting on blood oxygen measurement. By calculating and adjusting light intensity using a formula, the system ensures high-precision blood oxygen measurement in a variety of environmental conditions.

[0086] The blood oxygen saturation detection device described below and the blood oxygen saturation detection method described above can refer to each other.

[0087] Please see the attached Figure 2 The present invention also provides a blood oxygen saturation detection device, comprising: A multi-wavelength light source module for emitting light of different wavelengths within the visible spectrum; An optical sensor array for collecting light intensity information after being reflected or transmitted through skin tissue; a data processing unit, configured to construct a blood oxygen saturation model based on the received light intensity information and apply a machine learning algorithm to analyze the data; a signal processing unit for filtering noise and enhancing the signal-to-noise ratio; Environmental sensing module, used to detect environmental parameters, including temperature, humidity, and air pressure, to adjust light sources and device operating modes; The display and feedback unit is used to display the blood oxygen saturation value in real time and provide personalized health feedback and suggestions based on the user's health profile.

[0088] The device of this embodiment can be used to execute the above method embodiment, and its principles and technical effects are similar, so they will not be repeated here.

[0089] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A method for detecting blood oxygen saturation, characterized in that: The following steps are involved: S1. irradiating the target area with a multi-wavelength light source, the light source comprising light of different wavelengths within the visible spectrum; S2, collecting light intensity information after being reflected or transmitted through tissue; S3. Constructing a blood oxygen saturation model based on the ratio of the received light intensities at different wavelengths; S4. Apply machine learning algorithms to analyze the collected data, correct individual differences and other interference factors, and obtain blood oxygen saturation values; S5. Personalize the calibration of blood oxygen saturation based on health records to eliminate the influence of individual differences; S6. Use the environmental perception module to detect environmental changes and automatically adjust the light source intensity to compensate for the impact of environmental factors on the measurement.

2. A blood oxygen saturation detection method according to claim 1, characterized in that: The multi-wavelength light source further comprises: Use multiple red light wavelengths and multiple infrared light wavelengths and illuminate the target area separately to achieve accurate measurement of blood oxygen saturation.

3. The method for detecting blood oxygen saturation according to claim 1, wherein: The collected light intensity information includes: Absorption of red and infrared light, which is used to estimate the ratio of oxygenated to deoxygenated hemoglobin in the blood; The intensity of the light signal passing through the skin tissue is used to reflect changes in blood oxygen concentration.

4. The method for detecting blood oxygen saturation according to claim 1, wherein: The blood oxygen saturation model constructed according to the received light intensity ratio further includes: A mathematical model is constructed based on the light intensity ratio under different wavelength light sources, wherein the model is used to describe the relationship between blood oxygen concentration and light absorption; Use machine learning methods to optimize the accuracy of blood oxygen saturation models.

5. The method for detecting blood oxygen saturation according to claim 1, wherein: The machine learning algorithm includes: A combination of convolutional neural networks and recurrent neural networks to analyze and predict changes in blood oxygen saturation; Use the training dataset to calibrate and optimize the algorithm to improve the accuracy and stability of blood oxygen detection.

6. The method for detecting blood oxygen saturation according to claim 1, wherein: The personalized calibration is based on: When the user uses it for the first time, a personalized baseline is established through blood oxygen data; Based on the user's health records and historical blood oxygen data, the calibration parameters are dynamically adjusted to improve the accuracy of blood oxygen measurement.

7. The method for detecting blood oxygen saturation according to claim 6, wherein: The personalized calibration technology is optimized based on the user's health profile, which includes information on the user's age, gender, health status, and history of chronic diseases.

8. The method for detecting blood oxygen saturation according to claim 1, wherein: The environment perception module includes: Temperature and humidity sensor, used to monitor ambient temperature and humidity; Air pressure sensor, used to monitor air pressure changes; Ambient light sensor, used to monitor the surrounding light intensity; The environmental parameters are used to adjust the intensity or wavelength of the light source to compensate for the impact of environmental changes on blood oxygen measurement.

9. The method for detecting blood oxygen saturation according to claim 1, wherein: The automatic adjustment of light source intensity comprises: Detect air pressure changes and adjust the wavelength and intensity of the light source according to the air pressure changes; Detects changes in ambient temperature and humidity and adjusts the device's operating mode to maintain measurement accuracy.

10. A blood oxygen saturation detection device, characterized in that: A method for detecting blood oxygen saturation according to any one of claims 1 to 9, comprising: A multi-wavelength light source module for emitting light of different wavelengths within the visible spectrum; An optical sensor array for collecting light intensity information after being reflected or transmitted through skin tissue; a data processing unit, configured to construct a blood oxygen saturation model based on the received light intensity information and apply a machine learning algorithm to analyze the data; a signal processing unit for filtering noise and enhancing the signal-to-noise ratio; Environmental sensing module, used to detect environmental parameters, including temperature, humidity, and air pressure, to adjust light sources and device operating modes; The display and feedback unit is used to display the blood oxygen saturation value in real time and provide personalized health feedback and suggestions based on the user's health profile.

Images