Intelligent blood oxygen saturation degree measuring method and device and electronic equipment

Abstract

The invention discloses an intelligent blood oxygen saturation measuring method and device and electronic equipment, and relates to the field of data processing. According to the method, an infrared light emitter is used for emitting a plurality of emitted light rays with different wavelengths; receiving reflected light rays corresponding to the emitted light rays by utilizing a photoelectric receiver, and converting the reflected light rays into photoelectric signals; extracting a time domain characteristic parameter and a frequency domain characteristic parameter of the photoelectric signal; inputting the time-domain characteristic parameters and the frequency-domain characteristic parameters into a preset blood oxygen saturation degree estimation model to obtain a blood oxygen saturation degree estimation value of the to-be-measured individual; and acquiring historical detection data of the to-be-detected individual, and correcting the blood oxygen saturation degree estimated value by adopting a Kalman filtering algorithm based on the historical detection data to obtain a final blood oxygen saturation degree estimated value. By implementing the technical scheme provided by the invention, the accuracy of blood oxygen saturation measurement is improved.

Description

Technical Field

[0001] This application relates to the field of data processing, and particularly to an intelligent blood oxygen saturation measurement method, device and electronic device. Background Art

[0002] With the development of technology and the improvement of people's health awareness, real-time monitoring of personal health status has become an important part of modern life. Blood oxygen saturation is one of the key physiological parameters for evaluating the functions of an individual's respiratory and circulatory systems. Traditional blood oxygen saturation measurement devices such as pulse oximeters have been widely used in medical institutions and households. These devices usually measure the ratio of oxyhemoglobin and reduced hemoglobin in the blood through non-invasive optical sensors, and then calculate the blood oxygen saturation. However, with the rise of wearable devices and the increasing demand for portable health monitoring devices by users, traditional blood oxygen measurement technologies are facing new challenges and higher requirements.

[0003] Currently, traditional blood oxygen saturation measurement technologies mainly have the following several defects: First, most devices can only measure using a single or limited number of light wavelengths, which limits their applicability and accuracy for different individuals (such as people with different skin colors or body types). Second, these devices are often highly sensitive to the external environment, such as motion interference, and are prone to inaccurate measurement results.

[0004] Therefore, there is an urgent need for an intelligent blood oxygen saturation measurement method, device and electronic device. Summary of the Invention

[0005] This application provides an intelligent blood oxygen saturation measurement method, device and electronic device, which improves the accuracy of blood oxygen saturation measurement.

[0006] In the first aspect of this application, an intelligent blood oxygen saturation measurement method is provided. The method includes: using an infrared light emitter to emit multiple emission light rays with different wavelengths; wherein, the emission light rays irradiate on the skin surface of the part to be measured and are reflected; using a photoelectric receiver to receive the reflected light rays corresponding to each of the emission light rays, and converting the reflected light rays into photoelectric signals; extracting the time-domain characteristic parameters and frequency-domain characteristic parameters of the photoelectric signals, the time-domain characteristic parameters include the peak value of the pulse wave, the trough value of the pulse wave and the pulse wave width, and the frequency-domain characteristic parameters include the ratio of the AC component and the DC component of the photoelectric signal; inputting the time-domain characteristic parameters and the frequency-domain characteristic parameters into a preset blood oxygen saturation estimation model to obtain an estimated value of the blood oxygen saturation of the individual to be measured; obtaining the historical detection data of the individual to be measured, and based on the historical detection data, using the Kalman filtering algorithm to correct the estimated value of the blood oxygen saturation to obtain a final estimated value of the blood oxygen saturation.

[0007] By adopting the above technical solution, by using an infrared light emitter to emit multiple emission light rays of different wavelengths, and using a photoelectric receiver to receive the reflected light rays corresponding to each emission light ray and convert them into photoelectric signals, the original data of the blood oxygen saturation of the part to be measured can be obtained. Extracting the time-domain characteristic parameters (pulse wave peak value, pulse wave valley value, pulse wave width) and frequency-domain characteristic parameters (ratio of alternating current component to direct current component) of the photoelectric signal can characterize the characteristic pattern of blood oxygen saturation from two dimensions of time and frequency, providing rich information for subsequent estimation. Inputting the extracted characteristic parameters into a preset blood oxygen saturation estimation model can quickly and accurately estimate the blood oxygen saturation value of the individual to be measured, realizing the quantitative evaluation of blood oxygen saturation. At the same time, by obtaining the historical detection data of the individual to be measured and using the Kalman filtering algorithm to dynamically correct the initial estimated value, the historical information of the individual can be fully utilized, improving the continuity and stability of the estimation and reducing the influence of random errors. The finally output blood oxygen saturation estimated value synthesizes the photoelectric signals of multiple wavelengths, time-domain and frequency-domain characteristics, and historical detection data, with high reliability and accuracy, providing an important basis for subsequent health assessment and medical decision-making.

[0008] Optionally, based on the historical detection data, using the Kalman filtering algorithm to correct the blood oxygen saturation estimated value to obtain the final blood oxygen saturation estimated value, specifically including: based on the historical detection data, using the Kalman filtering algorithm to establish a state transition equation and an observation equation for blood oxygen saturation; taking the blood oxygen saturation estimated value as the initial observation value and inputting it into the state transition equation and the observation equation to obtain the prior estimated value at the current moment; through the Kalman filtering algorithm, recursively correcting the prior estimated values at each moment to obtain a sequence of corrected prior estimated values, and determining the final blood oxygen saturation estimated value from the sequence of prior estimated values.

[0009] By adopting the above technical solution, when using the Kalman filtering algorithm to correct the estimated value of blood oxygen saturation, by establishing a state transition equation and an observation equation based on historical detection data, the dynamic change law of blood oxygen saturation and the observation relationship during the measurement process can be characterized, providing a mathematical model basis for the Kalman filter. Taking the initial estimated value as the observed value and inputting it into the state transition equation and the observation equation, the prior estimate at the current moment can be obtained, which reflects the prediction of the current blood oxygen saturation considering historical information. By recursively correcting the prior estimates at each moment through the Kalman filtering algorithm, the estimated value can be continuously updated to gradually approach the true value. The sequence of corrected prior estimated values contains the estimated values of blood oxygen saturation at different moments, reflecting the dynamic change trend of blood oxygen saturation. Determining the final estimated value of blood oxygen saturation from this sequence can obtain a stable estimation result, reducing the contingency of single estimation. The Kalman filtering algorithm makes full use of historical observation information, and through dynamic recursion, realizes the continuous optimization and correction of the estimated value, improving the accuracy of blood oxygen saturation estimation.

[0010] Optionally, the determining the final estimated value of blood oxygen saturation from the sequence of prior estimated values specifically includes: performing smoothing processing on the sequence of prior estimated values to obtain a target sequence of prior estimated values; extracting a preset number of target prior estimated values from the end of the target sequence of prior estimated values; calculating the average value of the preset number of target prior estimated values to obtain the final estimated value of blood oxygen saturation.

[0011] By adopting the above technical solution, when determining the final estimated value of blood oxygen saturation from the sequence of prior estimated values, first performing smoothing processing on the sequence of prior estimated values can remove high-frequency noise and mutation points in the sequence of prior estimated values, obtaining a stable and continuous target sequence of prior estimated values. Extracting a preset number of target prior estimated values from the end of this sequence can obtain the change situation of blood oxygen saturation in the recent period of time, reflecting the current blood oxygen saturation state. Calculating the average of the extracted multiple target prior estimated values can further reduce the influence of random errors, obtaining a stable and reliable final estimated value. This estimated value comprehensively considers the overall trend of the prior estimation sequence and the local characteristics at the end, balancing the long-term and short-term changes in blood oxygen saturation, and has high representativeness. Through smoothing processing, end extraction, and mean calculation, the finally obtained estimated value of blood oxygen saturation is more accurate and robust, reducing the influence of abnormal fluctuations and accidental errors, providing reliable data support for subsequent health assessments.

[0012] Optionally, the step of using the infrared light emitter to emit emission light rays of multiple different wavelengths specifically includes: determining the emission duration and emission interval corresponding to the emission light rays of each wavelength; generating a drive signal for the infrared light emitter according to the emission duration and the emission interval; transmitting the drive signal to the infrared light emitter, and controlling the infrared light emitter to alternately emit emission light rays of multiple wavelengths according to the drive signal.

[0013] By adopting the above technical solution, when using the infrared light emitter to emit emission light rays of multiple different wavelengths, by determining the emission duration and emission interval corresponding to the emission light rays of each wavelength, the emission process of light rays of different wavelengths can be finely controlled to achieve reasonable configuration in terms of time. Generating a drive signal for the infrared light emitter according to the emission duration and interval can accurately modulate the working state of the infrared light emitter to ensure that the emission timing and duration of the light rays meet the preset requirements. Transmitting the drive signal to the infrared light emitter and controlling it to alternately emit light rays of multiple wavelengths according to the drive signal can realize the automatic switching and cyclic emission of light rays of different wavelengths, reducing the trouble of manual operation. By setting the emission duration, interval, and drive signal, the emission efficiency and energy utilization rate of the light rays can be optimized, avoiding unnecessary energy waste. At the same time, alternately emitting light rays of different wavelengths can obtain blood oxygen saturation information at multiple wavelengths in a single measurement process, improving the comprehensiveness and reliability of information collection, and providing a rich data basis for subsequent feature extraction and estimation.

[0014] Optionally, before inputting the time-domain feature parameters and the frequency-domain feature parameters into a preset blood oxygen saturation estimation model to obtain the blood oxygen saturation estimation value of the individual to be measured, the method further includes: collecting the optoelectronic signals of multiple human samples and the reference blood oxygen saturation values of each of the human samples, and constructing a training data set for the preset blood oxygen saturation estimation model; extracting input feature vectors from the training data set; using the reference blood oxygen saturation value as the training target and the input feature vectors as training samples to train a support vector machine regression model to obtain a first-level model for blood oxygen saturation estimation; using the residual between the output result of the first-level model and the reference blood oxygen saturation value as the training target for a second-level model, and using the input feature vectors as training samples to train a BP neural network model to obtain a second-level model for blood oxygen saturation estimation; cascading the first-level model and the second-level model to form the preset blood oxygen saturation estimation model.

[0015] By adopting the above technical solution, when constructing the preset blood oxygen saturation estimation model, by collecting the optoelectronic signals and reference blood oxygen saturation values of multiple human samples, a training data set covering different individuals and different blood oxygen saturation levels can be obtained, providing rich and diverse samples for model training. Extracting input feature vectors from the training set samples can transform the original optoelectronic signals into a standardized feature representation that can be processed by the model, facilitating subsequent model training and estimation. Using the reference blood oxygen saturation value as the training target and the input feature vector as the training sample, training a support vector machine regression model to obtain the first-level model can make full use of the excellent performance of the support vector machine to establish a non-linear regression relationship between blood oxygen saturation and input features, realizing preliminary estimation. Using the residual of the first-level model as the training target and training a BP neural network with the same input features to obtain the second-level model can compensate and correct the estimation error of the first-level model, further improving the estimation accuracy. Cascade the first-level model and the second-level model to form the final estimation model, which can give play to the advantages of both models, realize multi-level optimization of blood oxygen saturation estimation, and significantly improve the estimation performance.

[0016] Optionally, after obtaining the historical detection data of the individual to be measured and based on the historical detection data, using the Kalman filter algorithm to correct the blood oxygen saturation estimation value to obtain the final blood oxygen saturation estimation value, the method further includes: cyclically obtaining multiple of the final blood oxygen saturation estimation values at a preset frequency to obtain a dynamic blood oxygen saturation estimation value sequence; judging the number of target blood oxygen saturation estimation values, where the target blood oxygen saturation estimation value is the target blood oxygen saturation estimation value in the dynamic blood oxygen saturation estimation value sequence that is not within the preset range; if it is determined that the number is greater than or equal to the preset number, trigger a blood oxygen saturation alarm to prompt the user that the blood oxygen saturation is abnormal.

[0017] By adopting the above technical solution, after obtaining the final blood oxygen saturation estimation value, cyclically obtaining multiple estimation values at a preset frequency to form a dynamic blood oxygen saturation estimation value sequence can track the change trend of blood oxygen saturation in real time and dynamically reflect the blood oxygen state of the individual. By judging the number of abnormal estimation values exceeding the preset normal range in the dynamic sequence, abnormal fluctuations and deviations of blood oxygen saturation can be found, and potential low oxygen or hypoxia risks can be detected in time. When the number of consecutive abnormal estimation values exceeds the preset threshold, triggering a blood oxygen saturation alarm can remind the user that the blood oxygen saturation is continuously abnormal, suggesting taking corresponding protective measures to intervene and maintain the user's physical health in time.

[0018] Optionally, before determining the number of estimated target blood oxygen saturation values, the method further includes: obtaining personal information of the individual to be measured, where the personal information includes age and gender; according to the personal information, extracting a corresponding reference blood oxygen saturation range from a preset threshold database, and using the reference blood oxygen saturation range as the preset range.

[0019] By adopting the above technical solution, before determining the number of abnormal estimated blood oxygen saturation values, by obtaining personal information such as the age and gender of the individual to be measured, the measurement process can be personalized and customized. Extracting the corresponding reference blood oxygen saturation range from the preset threshold database according to the personal information can obtain a personalized judgment criterion that matches the physiological characteristics of the individual to be measured. Using the personalized reference range as the preset range for abnormal judgment can improve the pertinence and accuracy of abnormal detection.

[0020] In a second aspect of the present application, there is provided an intelligent blood oxygen saturation measurement device, which includes a light emission module, a light reception module, a characteristic parameter extraction module, a blood oxygen saturation estimation module, and a blood oxygen saturation correction module, where: the light emission module is used to emit a plurality of emission light rays with different wavelengths by using an infrared light emitter; wherein, the emission light rays irradiate on the skin surface of the part to be measured and are reflected; the light reception module is used to receive the reflected light rays corresponding to each of the emission light rays by using a photoelectric receiver, and convert the reflected light rays into photoelectric signals; the characteristic parameter extraction module is used to extract the time-domain characteristic parameters and frequency-domain characteristic parameters of the photoelectric signals, the time-domain characteristic parameters include the peak value of the pulse wave, the valley value of the pulse wave, and the pulse wave width, and the frequency-domain characteristic parameters include the ratio of the alternating current component to the direct current component of the photoelectric signal; the blood oxygen saturation estimation module is used to input the time-domain characteristic parameters and the frequency-domain characteristic parameters into a preset blood oxygen saturation estimation model to obtain an estimated blood oxygen saturation value of the individual to be measured; the blood oxygen saturation correction module is used to obtain the historical detection data of the individual to be measured, and based on the historical detection data, use the Kalman filtering algorithm to correct the estimated blood oxygen saturation value to obtain the final estimated blood oxygen saturation value.

[0021] In a third aspect of the present application, there is provided an electronic device, including a processor, a memory, a user interface, and a network interface, where the memory is used to store instructions, the user interface and the network interface are both used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device executes the method as described in any one of the above.

[0022] In a fourth aspect of the present application, there is provided a computer-readable storage medium, where the computer-readable storage medium stores instructions, and when the instructions are executed, the method as described in any one of the above is executed.

[0023] In summary, one or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. By using an infrared light emitter to emit multiple emission light rays with different wavelengths, and using a photoelectric receiver to receive the reflected light rays corresponding to each emission light ray and convert them into photoelectric signals, the original data of the blood oxygen saturation of the part to be measured can be obtained. By extracting the time-domain characteristic parameters (pulse wave peak value, pulse wave valley value, pulse wave width) and frequency-domain characteristic parameters (ratio of alternating current component to direct current component) of the photoelectric signals, the characteristic pattern of blood oxygen saturation can be characterized from two dimensions of time and frequency, providing rich information for subsequent estimation. By inputting the extracted characteristic parameters into a preset blood oxygen saturation estimation model, the blood oxygen saturation value of the individual to be measured can be quickly and accurately estimated, realizing the quantitative evaluation of blood oxygen saturation. At the same time, by obtaining the historical detection data of the individual to be measured and using the Kalman filtering algorithm to dynamically correct the initial estimated value, the historical information of the individual can be fully utilized, improving the continuity and stability of the estimation and reducing the influence of random errors. The finally output blood oxygen saturation estimated value synthesizes the photoelectric signals of multiple wavelengths, time-domain and frequency-domain characteristics, and historical detection data, and has high reliability and accuracy, providing an important basis for subsequent health assessment and medical decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 is a schematic flowchart of an intelligent blood oxygen saturation measurement method disclosed in an embodiment of the present application; Figure 2 is a schematic block diagram of an intelligent blood oxygen saturation measurement device disclosed in an embodiment of the present application; Figure 3 is a schematic structural diagram of an electronic device disclosed in an embodiment of the present application.

[0025] Description of the reference numerals: 201, light emission module; 202, light reception module; 203, characteristic parameter extraction module; 204, blood oxygen saturation estimation module; 205, blood oxygen saturation correction module; 300, electronic device; 301, processor; 302, communication bus; 303, user interface; 304, network interface; 305, memory. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0026] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of this specification. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments.

[0027] In the description of the embodiments of the present application, words such as "for example" or "for illustration" are used to give examples, illustrations or explanations. Any embodiment or design solution described as "for example" or "for illustration" in the embodiments of the present application should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Rather, the use of words such as "for example" or "for illustration" is intended to present relevant concepts in a specific manner.

[0028] In the description of the embodiments of the present application, the term "plurality" means two or more. For example, a plurality of systems means two or more systems, and a plurality of screen terminals means two or more screen terminals. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly specifying the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. The terms "comprise", "include", "have" and their variants all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0029] The present application provides an intelligent blood oxygen saturation measurement method, with reference to Figure 1 , Figure 1 is a schematic flowchart of an intelligent blood oxygen saturation measurement method provided by an embodiment of the present application. This method is applied to a server, and the server is a server that executes an intelligent blood oxygen saturation measurement program. This method includes steps S101 to S105, and the above steps are as follows: Step S101: Use an infrared light emitter to emit a plurality of emission light rays with different wavelengths; wherein, the emission light rays irradiate on the skin surface of the part to be measured and are reflected.

[0030] In step S101, using an infrared light emitter to emit a plurality of emission light rays with different wavelengths specifically includes: determining the emission duration and emission interval corresponding to the emission light rays of each wavelength; generating a drive signal for the infrared light emitter according to the emission duration and emission interval; transmitting the drive signal to the infrared light emitter, and controlling the infrared light emitter to alternately emit a plurality of wavelengths of emission light rays according to the drive signal.

[0031] Specifically, the server determines the emission duration and emission interval corresponding to the emitted light of each wavelength according to the preset blood oxygen saturation measurement parameters. Usually, in order to obtain accurate blood oxygen saturation measurement results, at least two infrared lights of different wavelengths need to be selected, such as 660 nm and 940 nm. These two wavelengths of light have different absorption characteristics for hemoglobin and oxyhemoglobin. By measuring the absorption differences of light of different wavelengths, the blood oxygen saturation can be calculated. The server calculates the duration of continuous emission of the emitted light of each wavelength and the time interval between the emissions of the two wavelengths according to the sampling frequency and sampling time of blood oxygen saturation measurement. For example, if the sampling frequency is 100 Hz and the sampling time is 10 seconds, the emitted lights of 660 nm and 940 nm each need to be emitted 500 times, the duration of each emission is 10 milliseconds, and the time interval between the emissions of the two wavelengths is 10 milliseconds.

[0032] After determining the emission duration and emission interval, the server generates a drive signal for the infrared light emitter according to these parameters. The drive signal is usually a series of high and low level pulses. A high level indicates that the infrared light emitter is turned on and emits light, and a low level indicates that the infrared light emitter is turned off and does not emit light. The server generates the corresponding high and low level pulse sequences according to the emission duration and emission interval of 660 nm and 940 nm, and combines these pulse sequences into a complete drive signal. For example, if the emission duration of 660 nm is 10 milliseconds, the emission duration of 940 nm is also 10 milliseconds, and the emission interval is 10 milliseconds, the drive signal can be expressed as: "high level for 10 milliseconds, low level for 10 milliseconds, high level for 10 milliseconds, low level for 10 milliseconds...", where the first high level pulse corresponds to 660 nm, the second high level pulse corresponds to 940 nm, and so on.

[0033] Finally, the server transmits the generated drive signal to the infrared light emitter and controls the infrared light emitter to alternately emit infrared light of 660 nm and 940 nm according to the drive signal. Specifically, when the drive signal is at a high level, the infrared light emitter is turned on and emits infrared light of the corresponding wavelength; when the drive signal is at a low level, the infrared light emitter is turned off and stops emitting. By continuously cycling the output of high and low levels, the infrared light emitter can alternately emit infrared light of two wavelengths according to the preset emission duration and emission interval.

[0034] Step S102: Use a photoelectric receiver to receive the reflected light corresponding to each emitted light and convert the reflected light into a photoelectric signal.

[0035] In step S102, the server synchronously controls the receiving time window and receiving interval of the photoelectric receiver according to the emission time and emission interval of the infrared light emitter in step S101. Specifically, when the server emits each infrared light, it opens the receiving time window of the photoelectric receiver to allow the photoelectric receiver to receive the reflected light of the corresponding wavelength; during the interval between every two infrared light emissions, it closes the receiving time window of the photoelectric receiver to stop the receiving process of the photoelectric receiver. Through this synchronous control, it can be ensured that the photoelectric receiver only receives the reflected light of the target wavelength, avoiding the interference of light of other wavelengths.

[0036] When the infrared light irradiates the skin surface, part of the light is absorbed by the skin tissue and part of the light is reflected back. The reflected light carries the absorption information of the skin tissue, especially the absorption difference information of hemoglobin and oxyhemoglobin for light of different wavelengths. After the photoelectric receiver receives these reflected lights, it converts them into corresponding photoelectric signals. The photoelectric receiver is usually composed of a photodiode or a phototransistor. When the reflected light irradiates the photodiode or the phototransistor, a photocurrent or photovoltage proportional to the light intensity is generated at both ends of it.

[0037] The server collects the photoelectric signals output by the photoelectric receiver in real time through the intelligent blood oxygen saturation measurement program, and processes the photoelectric signals such as amplification and filtering. Amplification is to amplify the weak photoelectric signals to a level range suitable for collection and processing; filtering is to remove high-frequency noise and power interference in the photoelectric signals, etc. The photoelectric signals after amplification and filtering can accurately reflect the intensity changes of the reflected lights of different wavelengths, providing reliable raw data for the subsequent calculation of blood oxygen saturation.

[0038] In the entire step S102, the server precisely controls the receiving time and receiving process of the photoelectric receiver through the intelligent blood oxygen saturation measurement program, automatically collects and processes the photoelectric signals, ensuring the integrity and accuracy of the blood oxygen saturation measurement data. By precisely synchronizing the receiving process of the photoelectric receiver with the emission process of the infrared light emitter, it can effectively eliminate the interference of ambient light and light of other wavelengths, improving the signal-to-noise ratio and stability of the measurement. At the same time, the server can also automatically adjust the amplification factor and filtering parameters according to the quality of the photoelectric signals to adapt to different measurement environments and conditions, further improving the adaptability and robustness of the blood oxygen saturation measurement.

[0039] For example, when the infrared light emitter alternately emits infrared light of 660 nm and 940 nm, the server will synchronously control the photoelectric receiver to only receive the reflected light of 660 nm during the emission of 660 nm, and only receive the reflected light of 940 nm during the emission of 940 nm. The photoelectric receiver converts the received reflected light into corresponding photoelectric signals and transmits them to the server. The server amplifies the photoelectric signals by 100 times and uses a band-pass filter to filter out the noise interference below 0.5 Hz and above 10 Hz to obtain pure photoelectric signals. These photoelectric signals accurately reflect the intensity changes of the reflected light of 660 nm and 940 nm, laying a data foundation for subsequent blood oxygen saturation calculation.

[0040] Step S103: Extract the time-domain characteristic parameters and frequency-domain characteristic parameters of the photoelectric signal. The time-domain characteristic parameters include the peak value of the pulse wave, the trough value of the pulse wave, and the pulse wave width. The frequency-domain characteristic parameters include the ratio of the AC component to the DC component of the photoelectric signal.

[0041] In step S103, for the extraction of the time-domain characteristic parameters, the server first preprocesses the photoelectric signal obtained in step S102, including removing baseline drift and high-frequency noise. After removing the baseline drift and high-frequency noise, the server extracts features from the preprocessed photoelectric signal. For the peak value of the pulse wave and the trough value of the pulse wave, the server uses a peak detection algorithm to automatically search for the local maximum point and local minimum point of the photoelectric signal and marks them as the peak value of the pulse wave and the trough value of the pulse wave. For the pulse wave width, the server uses a threshold method or a derivative method to detect the starting point and ending point of the pulse wave and calculates the time interval between the starting point and the ending point to obtain the pulse wave width. The server takes the extracted peak value of the pulse wave, the trough value of the pulse wave, and the pulse wave width as the time-domain characteristic parameters of the photoelectric signal for subsequent blood oxygen saturation calculation.

[0042] For the extraction of the frequency-domain characteristic parameters, the server first performs a frequency-domain transformation on the preprocessed photoelectric signal to convert the time-domain signal into a frequency-domain signal. The frequency-domain transformation methods include Fourier transform, wavelet transform, etc. The server uses the fast Fourier transform (FFT) algorithm to decompose the photoelectric signal into sine wave components of different frequencies and calculates the amplitude and phase of each frequency component. In the frequency domain, the photoelectric signal can be divided into two parts: the DC component and the AC component. The DC component reflects the average value of the photoelectric signal and corresponds to the low-frequency components in the photoelectric signal; the AC component reflects the dynamic changes of the photoelectric signal and corresponds to the high-frequency components in the photoelectric signal. The changes in the photoelectric signal caused by heartbeat and breathing are mainly reflected in the AC component.

[0043] The server analyzes the frequency-domain signal after FFT transformation and calculates the ratio of the DC component to the AC component. The DC component is obtained by calculating the amplitude of the frequency-domain signal at 0 Hz; the AC component is obtained by calculating the sum of the squares of the amplitudes of the frequency-domain signal within a specific frequency range (such as 0.5 - 4 Hz). Dividing the sum of the squares of the amplitudes of the AC component by the amplitude of the DC component gives the ratio of the AC component to the DC component of the optoelectronic signal. This ratio reflects the relative intensity of the dynamically changing components in the optoelectronic signal, is closely related to the concentration ratio of oxyhemoglobin and reduced hemoglobin in the blood, and is an important parameter for calculating blood oxygen saturation.

[0044] Step S104: Input the time-domain characteristic parameters and frequency-domain characteristic parameters into a preset blood oxygen saturation estimation model to obtain an estimated value of the blood oxygen saturation of the individual to be measured.

[0045] In a possible implementation manner, before step S104, the method further includes: collecting optoelectronic signals of multiple human samples and the reference blood oxygen saturation values of each human sample, and constructing a training data set for the preset blood oxygen saturation estimation model; extracting input feature vectors from the training data set; using the reference blood oxygen saturation value as the training target and the input feature vectors as the training samples to train a support vector machine regression model to obtain a first-level model for blood oxygen saturation estimation; using the residual between the output result of the first-level model and the reference blood oxygen saturation value as the training target of the second-level model, and using the input feature vectors as the training samples to train a BP neural network model to obtain a second-level model for blood oxygen saturation estimation; cascading the first-level model and the second-level model to form a preset blood oxygen saturation estimation model.

[0046] Specifically, the server collects optoelectronic signals of multiple human samples and the reference blood oxygen saturation values of each human sample, and constructs a training data set for the preset blood oxygen saturation estimation model. Specifically, the server selects a group of volunteers with good physical health and a wide age distribution as human samples, and simultaneously collects optoelectronic signals and reference blood oxygen saturation values for each volunteer. The process of collecting optoelectronic signals is similar to steps S101 and S102, using an infrared light emitter to alternately emit infrared light at 660 nm and 940 nm, and using a photoelectric receiver to receive the reflected light and convert it into an optoelectronic signal. Collecting the reference blood oxygen saturation value requires using a medical blood oxygen meter, and through spectrophotometric analysis of the arterial blood sample of the volunteer, an accurate measured value of blood oxygen saturation is obtained. The server pairs the collected optoelectronic signals and reference blood oxygen saturation values according to the volunteers to form samples of the training data set.

[0047] For each sample in the training dataset, the server extracts the time-domain characteristic parameters (pulse wave peak value, pulse wave valley value, pulse wave width) and frequency-domain characteristic parameters (ratio of AC component to DC component) of the optoelectronic signal according to the method in step S103, and combines these characteristic parameters into a feature vector. The feature vector contains both the characteristic parameters of the 660nm optoelectronic signal and the characteristic parameters of the 940nm optoelectronic signal. These feature vectors are used as the input of the preset blood oxygen saturation estimation model, and the corresponding reference blood oxygen saturation values are used as the training targets of the preset blood oxygen saturation estimation model.

[0048] After constructing the training dataset, the server trains the first-level model for blood oxygen saturation estimation based on Support Vector Machine Regression (SVR). SVR is a non-linear regression algorithm that realizes the mapping from the input feature vector to the blood oxygen saturation value by finding the optimal hyperplane in the high-dimensional feature space. The server uses the feature vectors in the training dataset as the training samples of SVR, and the corresponding reference blood oxygen saturation values as the training targets of SVR. By adjusting the hyperparameters such as the kernel function type of SVR, the optimal SVR model is trained. This SVR model can output the corresponding estimated blood oxygen saturation value according to the input feature vector. However, since SVR is a global regression algorithm, its estimation accuracy may be affected by abnormal samples.

[0049] To further improve the accuracy of blood oxygen saturation estimation, the server trains the second-level BP neural network model based on the first-level SVR model. The BP neural network is a multi-layer perceptron network with forward propagation and backward correction, which realizes the non-linear mapping from input to output by adjusting the connection weights of neurons layer by layer. The server uses the residual between the output result of the first-level SVR model and the reference blood oxygen saturation value as the training target of the second-level BP neural network, and the feature vector as the input of the BP neural network. By setting the number of neurons in the hidden layer, the type of activation function, and the learning rate parameter, the optimal BP neural network model is trained. This BP neural network model can compensate and correct the estimation result of the first-level SVR model, and improve the fitting degree between the estimated blood oxygen saturation value and the true value.

[0050] Finally, the server cascades the trained first-level SVR model (the first-level model) and the second-level BP neural network model (the second-level model) to form a complete preset blood oxygen saturation estimation model.

[0051] In step S104, when a new optoelectronic signal is input, the server first extracts its feature vector, and then inputs the feature vector into the first-level SVR model of the preset blood oxygen saturation estimation model to obtain a preliminary blood oxygen saturation estimation value. Then, the preliminary estimation value and the feature vector are input into the second-level BP neural network model of the preset blood oxygen saturation estimation model to obtain the final blood oxygen saturation estimation value.

[0052] Step S105: Obtain the historical detection data of the individual to be measured, and based on the historical detection data, use the Kalman filtering algorithm to correct the blood oxygen saturation estimation value to obtain the final blood oxygen saturation estimation value.

[0053] In step S105, based on the historical detection data, using the Kalman filtering algorithm to correct the blood oxygen saturation estimation value to obtain the final blood oxygen saturation estimation value specifically includes: based on the historical detection data, using the Kalman filtering algorithm to establish the state transition equation and the observation equation of blood oxygen saturation; taking the blood oxygen saturation estimation value as the initial observation value and inputting it into the state transition equation and the observation equation to obtain the prior estimation value at the current moment; through the Kalman filtering algorithm, recursively correcting the prior estimation values at each moment to obtain a sequence of corrected prior estimation values, and determining the final blood oxygen saturation estimation value from the sequence of prior estimation values.

[0054] Specifically, the server obtains the historical detection data of the individual to be measured, including the historical blood oxygen saturation estimation values and the corresponding measurement times. These historical data can be from multiple blood oxygen saturation measurements of the individual to be measured in the past period of time, reflecting the dynamic change trend of the blood oxygen saturation of this individual. The server arranges the historical detection data in chronological order to form a time series.

[0055] Next, the server uses the Kalman filtering algorithm to establish the state transition equation and the observation equation of blood oxygen saturation. The state transition equation describes the evolution relationship of blood oxygen saturation between two adjacent moments, representing the functional relationship between the blood oxygen saturation state at the current moment and the state at the previous moment. The observation equation describes the relationship between the blood oxygen saturation state and the actual measurement value, representing the functional relationship between the blood oxygen saturation measurement value at the current moment and the true state. The server estimates the parameters in the state transition equation and the observation equation, including the state transition matrix, the observation matrix, the covariance matrices of the process noise and the measurement noise, etc., according to the historical detection data by using methods such as the least squares method or the maximum likelihood estimation. The general forms of the state transition equation and the observation equation are as follows: State transition equation: x(k)=A*x(k - 1)+B*u(k)+w(k) Observation equation: z(k)=H*x(k)+v(k) Among them, \(x(k)\) represents the blood oxygen saturation state variable at the \(k\)th moment, which is a quantity to be estimated; \(x(k - 1)\) represents the blood oxygen saturation state variable at the \((k - 1)\)th moment; \(u(k)\) represents the control input at the \(k\)th moment, usually set to 0; \(w(k)\) represents the process noise at the \(k\)th moment, which follows a normal distribution with a mean of 0; \(z(k)\) represents the measured value of blood oxygen saturation at the \(k\)th moment; \(v(k)\) represents the measurement noise at the \(k\)th moment, which also follows a normal distribution with a mean of 0; \(A\), \(B\), and \(H\) are the state transition matrix, control input matrix, and observation matrix respectively, and they determine the linear relationship between the state variable and the measured value.

[0056] After establishing the state transition equation and the observation equation, the server takes the blood oxygen saturation estimated value obtained in step S104 as the initial observation value and inputs it into the Kalman filtering algorithm. The Kalman filtering algorithm is a recursive Bayesian estimation algorithm that continuously corrects the prior estimated value of blood oxygen saturation through two steps: prediction and update. In the prediction step, the server uses the posterior estimated value at the previous moment and the state transition equation to calculate the prior estimated value and the prior estimated error covariance matrix at the current moment. In the update step, the server uses the observed value at the current moment and the observation equation to calculate the Kalman gain, and uses the Kalman gain to correct the prior estimated value to obtain the posterior estimated value and the posterior estimated error covariance matrix at the current moment. The server continuously repeats the prediction and update steps, recursively processes the entire historical detection data sequence, and obtains a series of corrected prior estimated values. Finally, the server determines the final blood oxygen saturation estimated value from the sequence of prior estimated values.

[0057] For example, the individual to be tested, Xiao Wang, received 7 blood oxygen saturation measurements in the past week, and the historical detection data is {95.1%, 94.8%, 95.3%, 95.0%, 94.9%, 95.2%, 95.4%}. The server established a first-order autoregressive model as the state transition equation and a linear Gaussian model as the observation equation based on these historical data. In the current measurement, the blood oxygen saturation estimated value obtained in step S104 is 94.5%. The server takes this estimated value as the initial observation value and inputs it into the Kalman filtering algorithm. After recursive calculation, a series of corrected prior estimated values {95.2%, 95.1%, 95.0%, 94.9%, 94.8%, 94.7%, 94.6%} are obtained. Finally, the server outputs 94.6% as the optimal estimated value of Xiao Wang's current blood oxygen saturation. Compared with directly using the estimated value of 94.5%, the Kalman filtering algorithm makes full use of the blood oxygen saturation change trend of Xiao Wang in the past week, smooths and corrects the estimated value, making the final estimated value more stable and reliable.

[0058] In a possible implementation, the final blood oxygen saturation estimate value is determined from the sequence of prior estimate values, which specifically includes: smoothing the sequence of prior estimate values to obtain a target sequence of prior estimate values; extracting a preset number of target prior estimate values from the end of the target sequence of prior estimate values; calculating the average value of the preset number of target prior estimate values to obtain the final blood oxygen saturation estimate value.

[0059] Specifically, the server smooths the sequence of prior estimate values obtained by the Kalman filtering algorithm to obtain a target sequence of prior estimate values. The server can use methods such as moving average filtering, exponential smoothing filtering, and wavelet denoising to smooth the sequence of prior estimate values. Taking moving average filtering as an example, the server selects an appropriate sliding window length (such as 5 data points), slides this window on the sequence of prior estimate values, and calculates the arithmetic mean of the data points within the window as the smoothed target estimate value. By continuously moving the window until it reaches the end of the sequence, the complete target sequence of prior estimate values is obtained. Compared with the original sequence of prior estimate values, the volatility of the target sequence of prior estimate values is reduced, and the trend characteristics are more obvious.

[0060] Next, the server extracts a preset number of target prior estimate values from the end of the target sequence of prior estimate values. The preset number can be set according to the specific application scenario and real-time requirements, and usually the last 3 - 10 target estimate values are selected. The reason for extracting the end estimate values is that the dynamic change of blood oxygen saturation usually has a certain time correlation, and the closer the estimate value is to the current moment, the higher its credibility and representativeness. By focusing on a small segment of data at the end of the sequence, the current blood oxygen saturation state of the individual to be measured can be more accurately reflected.

[0061] Finally, the server calculates the arithmetic mean of the extracted preset number of target prior estimate values and outputs the average result as the final blood oxygen saturation estimate value. The purpose of calculating the average value is to further reduce the influence of random errors and improve the robustness of the estimate value. Since the target sequence of prior estimate values has been smoothed, its volatility is very small, so the preset number of estimate values are usually relatively close, and their arithmetic mean can represent the overall level of this small segment of data. At the same time, the calculation of the arithmetic mean is simple and efficient, easy to implement on embedded devices or mobile terminals, which is conducive to real-time estimation and monitoring.

[0062] In a possible implementation, after step S105, the method further includes: cyclically obtaining a plurality of final blood oxygen saturation estimate values at a preset frequency to obtain a dynamic sequence of blood oxygen saturation estimate values; judging the number of target blood oxygen saturation estimate values, where the target blood oxygen saturation estimate value is the target blood oxygen saturation estimate value in the dynamic sequence of blood oxygen saturation estimate values that is not within the preset range; if it is determined that the number is greater than or equal to the preset number, then trigger a blood oxygen saturation alarm to prompt the user that the blood oxygen saturation is abnormal.

[0063] Specifically, a preset frequency is set inside the server, and multiple final blood oxygen saturation estimated values are cyclically obtained at this frequency to form a dynamic blood oxygen saturation estimated value sequence. The selection of the preset frequency needs to comprehensively consider factors such as the change speed of blood oxygen saturation, the performance of the measurement device, and the bandwidth of data transmission. Usually, the preset frequency can be set to 15 times per minute, which can not only timely reflect the dynamic changes of blood oxygen saturation but also will not bring too much burden to the measurement system and communication network.

[0064] Next, the server determines the number of outliers in the dynamic blood oxygen saturation estimated value sequence. The server first sets a preset range, which corresponds to the physiological value range of blood oxygen saturation of the normal population. Then, the server traverses each estimated value in the dynamic blood oxygen saturation estimated value sequence to determine whether it falls within the preset range. If an estimated value is lower than the lower limit of the preset range (such as 90%) or higher than the upper limit of the preset range (such as 100%), it is marked as an outlier, which is called the target blood oxygen saturation estimated value. The server counts the number of all target blood oxygen saturation estimated values in the sequence to obtain an outlier count result. This count result reflects the abnormal degree and duration of the blood oxygen saturation of the individual to be measured.

[0065] Finally, the server determines whether the number of target blood oxygen saturation estimated values is greater than or equal to the preset number to decide whether to trigger a blood oxygen saturation alarm. The selection of the preset number needs to balance sensitivity and specificity, which should not only be able to timely detect abnormal fluctuations in blood oxygen saturation but also try to avoid false alarms due to accidental factors. Usually, the preset number can be set to 3 - 5 consecutive outliers. If the server finds that in the dynamic blood oxygen saturation estimated value sequence, there are continuously more than or equal to the preset number of target blood oxygen saturation estimated values, it is considered a continuous blood oxygen saturation abnormal event, and the blood oxygen saturation alarm is triggered. The server can send alarm information to the user through methods such as mobile phone APP, SMS, voice call, etc., to prompt the user that the current blood oxygen saturation level is abnormal and recommend that the user take corresponding measures, such as deep breathing, oxygen inhalation, or seeking medical treatment.

[0066] In a possible implementation manner, before determining the number of target blood oxygen saturation estimated values, the method further includes: obtaining the personal information of the individual to be measured, where the personal information includes age and gender; according to the personal information, extracting the corresponding reference blood oxygen saturation range from the preset threshold database and using the reference blood oxygen saturation range as the preset range.

[0067] Specifically, before measuring the blood oxygen saturation, the server obtains the personal information of the individual to be measured through the human-computer interaction interface, mainly including two characteristics: age and gender. The age can be obtained by inputting the date of birth or directly filling in the number of full years, and the gender can be obtained by selecting the male / female option. These personal information are crucial for determining the appropriate normal range of blood oxygen saturation because there are certain differences in the physiological values of blood oxygen saturation among different age groups and genders. Generally speaking, the blood oxygen saturation of newborns and the elderly is relatively low, while that of young and middle-aged people is relatively high; the blood oxygen saturation of women is slightly higher than that of men of the same age group.

[0068] After obtaining the personal information, the server accesses the preset threshold database, which stores the normal ranges of blood oxygen saturation for different populations. The preset threshold database is obtained based on large-scale population surveys and medical research. According to characteristics such as age and gender, the population is divided into several groups, and the corresponding normal ranges of blood oxygen saturation are given for each group. For example, for adult men aged 18 to 40, the normal range of their blood oxygen saturation may be 95% to 99%; for elderly women over 60 years old, the normal range of their blood oxygen saturation may be 92% to 97%. These ranges reflect the physiological characteristics of each population and can be used as a reference basis for judging whether the individual's blood oxygen saturation is abnormal.

[0069] The server searches for the corresponding group in the preset threshold database according to the age and gender of the individual to be measured, extracts the normal range of blood oxygen saturation of this group, and uses it as the preset range for subsequent outlier judgment. The extraction process can be achieved through exact matching or fuzzy matching. Exact matching means finding the group that is exactly the same as the age and gender of the individual to be measured and directly using the normal range of this group. Fuzzy matching means finding several groups that are closest to the age and gender of the individual to be measured, and performing weighted averaging or taking the intersection of their normal ranges to obtain a normal range applicable to the individual to be measured. No matter which matching method is adopted, the final preset range can better reflect the physiological characteristics of the individual to be measured and help improve the accuracy of outlier judgment.

[0070] Refer to Figure 2, this application also provides an intelligent blood oxygen saturation measurement device, which is a server. The server includes a light emission module 201, a light reception module 202, a characteristic parameter extraction module 203, a blood oxygen saturation estimation module 204, and a blood oxygen saturation correction module 205, where: The light emission module 201 is used to emit multiple emission lights with different wavelengths by using an infrared light emitter; among them, the emission lights are irradiated on the skin surface of the part to be measured and reflected; The light reception module 202 is used to receive the reflected lights corresponding to the respective emission lights by using a photoelectric receiver and convert the reflected lights into photoelectric signals; The characteristic parameter extraction module 203 is used to extract the time-domain characteristic parameters and frequency-domain characteristic parameters of the photoelectric signals. The time-domain characteristic parameters include the peak value of the pulse wave, the valley value of the pulse wave, and the pulse wave width. The frequency-domain characteristic parameters include the ratio of the AC component to the DC component of the photoelectric signal; The blood oxygen saturation estimation module 204 is used to input the time-domain characteristic parameters and frequency-domain characteristic parameters into a preset blood oxygen saturation estimation model to obtain an estimated value of the blood oxygen saturation of the individual to be measured; The blood oxygen saturation correction module 205 is used to obtain the historical detection data of the individual to be measured and, based on the historical detection data, use the Kalman filter algorithm to correct the estimated value of the blood oxygen saturation to obtain the final estimated value of the blood oxygen saturation.

[0071] In a possible implementation manner, the blood oxygen saturation correction module 205 corrects the estimated value of the blood oxygen saturation by using the Kalman filter algorithm based on the historical detection data to obtain the final estimated value of the blood oxygen saturation, which specifically includes: The blood oxygen saturation correction module 205 establishes a state transition equation and an observation equation of the blood oxygen saturation by using the Kalman filter algorithm based on the historical detection data; The blood oxygen saturation correction module 205 inputs the estimated value of the blood oxygen saturation as the initial observation value into the state transition equation and the observation equation to obtain the prior estimated value at the current moment; The blood oxygen saturation correction module 205 recursively corrects the prior estimated values at each moment through the Kalman filter algorithm to obtain a sequence of corrected prior estimated values, and determines the final estimated value of the blood oxygen saturation from the sequence of prior estimated values.

[0072] In a possible implementation manner, the blood oxygen saturation correction module 205 determines the final estimated value of the blood oxygen saturation from the sequence of prior estimated values, which specifically includes: The blood oxygen saturation correction module 205 performs a smoothing process on the sequence of prior estimated values to obtain a target sequence of prior estimated values; The blood oxygen saturation correction module 205 extracts a preset number of target prior estimated values from the end of the target sequence of prior estimated values; The blood oxygen saturation correction module 205 calculates the average value of the preset number of target prior estimated values to obtain the final estimated value of the blood oxygen saturation.

[0073] In a possible implementation, the light emission module 201 uses an infrared light emitter to emit emission lights of multiple different wavelengths. Specifically, the light emission module 201 determines the emission duration and emission interval corresponding to the emission lights of each wavelength; the light emission module 201 generates a drive signal for the infrared light emitter according to the emission duration and emission interval; the light emission module 201 transmits the drive signal to the infrared light emitter and controls the infrared light emitter to alternately emit emission lights of multiple wavelengths according to the drive signal.

[0074] In a possible implementation, before the blood oxygen saturation estimation module 204 inputs the time-domain characteristic parameters and frequency-domain characteristic parameters into a preset blood oxygen saturation estimation model to obtain the blood oxygen saturation estimation value of the individual to be measured, the method further includes: the blood oxygen saturation estimation module 204 collects the optoelectronic signals of multiple human samples and the reference blood oxygen saturation values of each human sample, and constructs a training data set for the preset blood oxygen saturation estimation model; the blood oxygen saturation estimation module 204 extracts input feature vectors from the training data set; the blood oxygen saturation estimation module 204 uses the reference blood oxygen saturation value as the training target and the input feature vector as the training sample to train a support vector machine regression model to obtain a first-level model for blood oxygen saturation estimation; the blood oxygen saturation estimation module 204 uses the residual between the output result of the first-level model and the reference blood oxygen saturation value as the training target of the second-level model, and uses the input feature vector as the training sample to train a BP neural network model to obtain a second-level model for blood oxygen saturation estimation; the blood oxygen saturation estimation module 204 cascades the first-level model and the second-level model to form a preset blood oxygen saturation estimation model.

[0075] In a possible implementation, after the blood oxygen saturation correction module 205 obtains the historical detection data of the individual to be measured and corrects the blood oxygen saturation estimation value by using the Kalman filtering algorithm based on the historical detection data to obtain the final blood oxygen saturation estimation value, the method further includes: the blood oxygen saturation correction module 205 cyclically obtains multiple final blood oxygen saturation estimation values at a preset frequency to obtain a dynamic blood oxygen saturation estimation value sequence; the blood oxygen saturation correction module 205 determines the number of target blood oxygen saturation estimation values, and the target blood oxygen saturation estimation value is the target blood oxygen saturation estimation value in the dynamic blood oxygen saturation estimation value sequence that is not within the preset range; if the blood oxygen saturation correction module 205 determines that the number is greater than or equal to the preset number, a blood oxygen saturation alarm is triggered to prompt the user that the blood oxygen saturation is abnormal.

[0076] In a possible implementation, before the blood oxygen saturation correction module 205 determines the number of target blood oxygen saturation estimated values, the method further includes: the blood oxygen saturation correction module 205 obtains personal information of the individual to be measured, where the personal information includes age and gender; the blood oxygen saturation correction module 205 extracts a corresponding reference blood oxygen saturation range from a preset threshold database according to the personal information, and uses the reference blood oxygen saturation range as a preset range.

[0077] It should be noted that: when the device provided in the above embodiment realizes its functions, only the division of the above function modules is used for illustration. In actual applications, the above functions can be allocated to different function modules according to needs, that is, the internal structure of the device is divided into different function modules to complete all or part of the functions described above. In addition, the device and method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0078] This application also provides an electronic device. Referring to Figure 3 , Figure 3 is a schematic structural diagram of an electronic device provided in an embodiment of this application. The electronic device 300 may include: at least one processor 301, at least one network interface 304, a user interface 303, a memory 305, and at least one communication bus 302.

[0079] Among them, the communication bus 302 is used to realize the connection and communication between these components.

[0080] Among them, the user interface 303 may include a display screen (Display) and a camera (Camera). Optionally, the user interface 303 may further include a standard wired interface and a wireless interface.

[0081] Among them, the network interface 304 may optionally include a standard wired interface and a wireless interface (such as a Wi-Fi interface).

[0082] Among them, the processor 301 may include one or more processing cores. The processor 301 connects various parts within the entire server through various interfaces and lines, and executes various functions of the server and processes data by running or executing instructions, programs, code sets, or instruction sets stored in the memory 305, and by invoking the data stored in the memory 305. Optionally, the processor 301 may be implemented in at least one hardware form of digital signal processing (DSP), field-programmable gate array (FPGA), or programmable logic array (PLA). The processor 301 may integrate a combination of one or more of a central processing unit (CPU), a graphics processing unit (GPU), and a modem, etc. Among them, the CPU mainly processes the operating system, user interface, application programs, etc.; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communication. It can be understood that the above-mentioned modem may not be integrated into the processor 301 and may be implemented separately by a single chip.

[0083] Among them, the memory 305 may include random access memory (RAM) and may also include read-only memory. Optionally, the memory 305 includes a non-transitory computer-readable storage medium. The memory 305 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 305 may include a program storage area and a data storage area. Among them, the program storage area may store instructions for implementing the operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store the data involved in the above-mentioned various method embodiments. Optionally, the memory 305 may also be at least one storage device located far from the aforementioned processor 301. Refer to Figure 3 , the memory 305, as a computer storage medium, may include an operating system, a network communication module, a user interface module, and an application program for an intelligent blood oxygen saturation measurement method.

[0084] In Figure 3In the electronic device 300 shown, the user interface 303 is mainly used to provide an interface for the user to input and obtain the data input by the user; and the processor 301 can be used to call the application program stored in the memory 305 for an intelligent blood oxygen saturation measurement method. When executed by one or more processors 301, the electronic device 300 is caused to execute one or more of the methods as described in the above embodiments. It should be noted that for the foregoing method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that this application is not limited by the described action sequence, because according to this application, certain steps can be in other sequences or performed simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0085] This application also provides a computer-readable storage medium storing instructions. When executed by one or more processors 301, the electronic device 300 is caused to execute one or more of the methods as described in the above embodiments.

[0086] In the above embodiments, the descriptions of the respective embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0087] In several implementation manners provided by this application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are only illustrative. For example, the division of units is only a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection to each other can be through some service interfaces. The indirect coupling or communication connection of the device or unit can be in an electrical or other form.

[0088] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place, or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0089] In addition, in each embodiment of this application, the functional units can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0090] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in various embodiments of this application. And the aforementioned memory includes: various media such as USB flash drives, mobile hard disks, magnetic disks, or optical discs that can store program codes.

[0091] The foregoing are only exemplary embodiments of the present disclosure, and the scope of the present disclosure cannot be limited thereby. That is, any equivalent changes and modifications made in accordance with the teachings of the present disclosure still fall within the scope covered by the present disclosure. Those skilled in the art will readily think of other implementation schemes of the present disclosure after considering the specification and the disclosure of the practical truth.

[0092] This application aims to cover any variations, uses, or adaptive changes of the present disclosure that follow the general principles of the present disclosure and include the common general knowledge or conventional technical means in the technical field not recorded in the present disclosure. The specification and the embodiments are only regarded as exemplary, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. An intelligent blood oxygen saturation measurement method, characterized in that: The method comprises: Using an infrared light emitter to emit a plurality of light beams of different wavelengths; wherein the light beams are irradiated on the skin surface of the part to be tested and are reflected; Using a photoelectric receiver to receive the reflected light corresponding to each of the emitted light rays, and converting the reflected light rays into photoelectric signals; Extracting time domain characteristic parameters and frequency domain characteristic parameters of the photoelectric signal, wherein the time domain characteristic parameters include a pulse wave peak value, a pulse wave trough value, and a pulse wave width, and the frequency domain characteristic parameters include a ratio of an AC component to a DC component of the photoelectric signal; Inputting the time domain characteristic parameter and the frequency domain characteristic parameter into a preset blood oxygen saturation estimation model to obtain a blood oxygen saturation estimation value of the individual to be measured; The historical test data of the individual to be tested is obtained, and based on the historical test data, the blood oxygen saturation estimate is corrected by using a Kalman filter algorithm to obtain a final blood oxygen saturation estimate.

2. The method according to claim 1, characterized in that The method of correcting the blood oxygen saturation estimation value based on the historical detection data by using a Kalman filter algorithm to obtain a final blood oxygen saturation estimation value specifically includes: Based on the historical detection data, a state transfer equation and an observation equation of blood oxygen saturation are established using a Kalman filter algorithm; Input the estimated value of blood oxygen saturation as an initial observation value into the state transfer equation and the observation equation to obtain a priori estimated value at the current moment; The prior estimation value at each moment is recursively corrected by using the Kalman filter algorithm to obtain a corrected prior estimation value sequence, and the final blood oxygen saturation estimation value is determined from the prior estimation value sequence.

3. The method according to claim 2, characterized in that The determining the final blood oxygen saturation estimation value from the prior estimation value sequence specifically includes: Smoothing the prior estimation value sequence to obtain a target prior estimation value sequence; Extracting a preset number of target a priori estimate values ​​from the end of the target a priori estimate value sequence; An average value is calculated for a preset number of the target a priori estimated values ​​to obtain the final blood oxygen saturation estimated value.

4. The method according to claim 1, characterized in that The method of using an infrared light emitter to emit a plurality of light beams of different wavelengths specifically includes: Determine the emission duration and emission interval corresponding to the emission light of each wavelength; generating a driving signal for the infrared light emitter according to the emission duration and the emission interval; The driving signal is transmitted to the infrared light emitter, and the infrared light emitter is controlled to alternately emit light beams of multiple wavelengths according to the driving signal.

5. The method according to claim 1, characterized in that Before inputting the time domain characteristic parameter and the frequency domain characteristic parameter into a preset blood oxygen saturation estimation model to obtain the blood oxygen saturation estimation value of the individual to be measured, the method further includes: Collecting photoelectric signals of a plurality of human body samples and reference blood oxygen saturation values ​​of each of the human body samples to construct a training data set for the preset blood oxygen saturation estimation model; Extracting an input feature vector from the training data set; Taking the reference blood oxygen saturation value as a training target and the input feature vector as a training sample, training a support vector machine regression model to obtain a first-level model for blood oxygen saturation estimation; The residual between the output result of the first-level model and the reference blood oxygen saturation value is used as the training target of the second-level model, and the input feature vector is used as a training sample to train the BP neural network model to obtain the second-level model for blood oxygen saturation estimation; The first-level model and the second-level model are cascaded to form the preset blood oxygen saturation estimation model.

6. The method according to claim 1, characterized in that After acquiring the historical test data of the individual to be tested and correcting the blood oxygen saturation estimate value based on the historical test data using a Kalman filter algorithm to obtain a final blood oxygen saturation estimate value, the method further includes: cyclically acquiring a plurality of the final blood oxygen saturation estimation values ​​at a preset frequency to obtain a dynamic blood oxygen saturation estimation value sequence; Determining the number of target blood oxygen saturation estimation values, the target blood oxygen saturation estimation values ​​being target blood oxygen saturation estimation values ​​in the dynamic blood oxygen saturation estimation value sequence that are not within a preset range; If it is determined that the number is greater than or equal to the preset number, a blood oxygen saturation alarm is triggered to prompt the user that the blood oxygen saturation is abnormal.

7. The method according to claim 6, characterized in that Before determining the number of target blood oxygen saturation estimation values, the method further includes: Obtaining personal information of the individual to be tested, wherein the personal information includes age and gender; According to the personal information, a corresponding reference blood oxygen saturation range is extracted from a preset threshold database, and the reference blood oxygen saturation range is used as the preset range.

8. An intelligent blood oxygen saturation measurement device, characterized in that: The device comprises a light emitting module (201), a light receiving module (202), a characteristic parameter extraction module (203), a blood oxygen saturation estimation module (204) and a blood oxygen saturation correction module (205), wherein: The light emitting module (201) is used to emit a plurality of light beams of different wavelengths using an infrared light emitter; wherein the light beams are irradiated on the skin surface of the part to be tested and reflected; The light receiving module (202) is used to receive the reflected light corresponding to each of the emitted light using a photoelectric receiver, and convert the reflected light into a photoelectric signal; The characteristic parameter extraction module (203) is used to extract the time domain characteristic parameters and frequency domain characteristic parameters of the photoelectric signal, wherein the time domain characteristic parameters include the pulse wave peak value, the pulse wave trough value and the pulse wave width, and the frequency domain characteristic parameters include the ratio of the AC component to the DC component of the photoelectric signal; The blood oxygen saturation estimation module (204) is used to input the time domain characteristic parameter and the frequency domain characteristic parameter into a preset blood oxygen saturation estimation model to obtain a blood oxygen saturation estimation value of the individual to be measured; The blood oxygen saturation correction module (205) is used to obtain the historical test data of the individual to be tested, and based on the historical test data, use a Kalman filter algorithm to correct the blood oxygen saturation estimate to obtain a final blood oxygen saturation estimate.

9. An electronic device, characterized in that: The electronic device (300) comprises a processor (301), a memory (305), a user interface (303) and a network interface (304), wherein the memory (305) is used to store instructions, the user interface (303) and the network interface (304) are used to communicate with other devices, and the processor (301) is used to execute the instructions stored in the memory (305) so that the electronic device (300) executes the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed, the method according to any one of claims 1 to 7 is executed.

Images