Real-time heart rate monitoring method and system with adaptive spectrum correction and peak location

Abstract

The present application relates to a method, device, computer equipment, storage medium and computer program product for real-time monitoring of heart rate with adaptive spectrum correction and peak positioning. The method includes: classifying the quality of the photoplethysmogram signal based on the timing characteristics of the photoplethysmogram signal and the energy of the three-axis acceleration signal, and the quality classification includes absolute correct value, relative correct value, normal value and abnormal value; constructing an adaptive spectrum correction model based on the parameters of the three-axis acceleration signal energy, and performing correction operations on the photoplethysmogram signal based on the adaptive spectrum correction model; and performing correction operations on the heart rate corresponding to the photoplethysmogram signal based on the historical heart rate. The use of this method can reduce the complexity of the algorithm and improve the real-time performance through accurate classification of the PPG signal, and better meet the real-time requirements while improving the monitoring accuracy.

Description

Technical Field

[0001] The present application relates to the technical field of biomedical signal processing, and in particular to a real-time heart rate monitoring method and system with adaptive spectrum correction and peak location. Background Art

[0002] As one of the four major physiological indicators of the human body, the dynamic changes of heart rate reflect the health status of the heart in daily life and exercise, which is crucial for individual health management. Real-time monitoring of heart rate not only helps athletes better understand their own body functions, develop scientific and reasonable exercise plans, and reduce the risk of disease, but also becomes more urgent and important for patients who already have cardiovascular diseases.

[0003] At present, common sports heart rate monitoring equipment is based on photoplethysmography (PPG), which calculates the heart rate by detecting the changes in photoplethysmography signals after absorption by human tissue and blood. Due to its high continuity, non-invasiveness and comfort, the PPG method has become the most commonly used heart rate detection method. Although the PPG signal is strongly correlated with the heart cycle and pulse changes, its signal quality is easily affected by motion artifacts (MA), resulting in a large error between the calculated heart rate and the actual heart rate. For this reason, a sensor that receives a three-axis acceleration (ACC) signal is often added near the sensor that receives the PPG signal. The ACC signal is used as a reference to remove the MA interference in the PPG signal.

[0004] In the prior art, the key assumption of the Independent Component Analysis (ICA) method, namely the statistical independence or uncorrelation of signals, is usually not valid in PPG signals contaminated by MA, and ICA requires multiple PPG sensors, which may be difficult to implement in some wearable devices; the Empirical Mode Decomposition (EMD) method can adaptively decompose several intrinsic mode components (IMF) of different frequencies according to the characteristics of the input signal itself, but the EMD method needs to rely on artificial experience to set parameters such as the number of iterations and the amplitude of white noise, and is not universal. The selection basis of IMF is insufficient, and it cannot solve problems such as modal aliasing and reconstruction error. The variational mode decomposition (VMD) model determines the center frequency and bandwidth of the modal component by iteratively searching for the optimal solution of the variational model, identifies and removes the components containing noise, and reconstructs the remaining modes. However, it is not effective for PPG signals with strong motion artifacts and is sensitive to noise. When there is noise in the signal, modal aliasing may occur in the decomposition. ANC is very sensitive to the reference signal, and it is very difficult to reconstruct a qualified reference signal. In addition, the deep learning-based method has not been used in actual scenarios due to its computational complexity and unexplainability. Summary of the invention

[0005] Based on this, it is necessary to provide a real-time heart rate monitoring method, device, computer equipment, computer-readable storage medium and computer program product that improves the denoising effect of the PPG signal and further enhances the accuracy of real-time heart rate monitoring through adaptive spectrum correction and peak positioning in order to address the above-mentioned technical problems.

[0006] In a first aspect, the present application provides a real-time heart rate monitoring method with adaptive spectrum correction and peak location. The method comprises:

[0007] Classifying the quality of the photoplethysmography signal based on the timing characteristics of the photoplethysmography signal and the energy of the triaxial acceleration signal, wherein the quality classification includes an absolutely correct value, a relatively correct value, a normal value, and an abnormal value;

[0008] An adaptive spectrum correction model is constructed based on the parameters of the triaxial acceleration signal energy, and a correction operation is performed on the spectrum of the photoplethysmography signal based on the adaptive spectrum correction model;

[0009] The current heart rate corresponding to the photoplethysmogram signal is corrected based on the historical heart rate.

[0010] In one embodiment, classifying the quality of the photoplethysmography signal based on the timing characteristics of the photoplethysmography signal and the energy of the three-axis acceleration signal includes:

[0011] Acquire a photoplethysmogram signal after filtering operation;

[0012] Calculating the peak point of the photoplethysmography signal and segmenting it according to the peak period;

[0013] Calculate the Pearson correlation coefficient between corresponding periods;

[0014] The mean of the Pearson correlation coefficient is calculated and compared with a preset standard threshold. If the mean is within the preset standard threshold, the quality classification is determined to be an absolutely correct value.

[0015] In one embodiment, classifying the quality of the photoplethysmography signal based on the timing characteristics of the photoplethysmography signal and the energy of the three-axis acceleration signal includes:

[0016] Obtain the three-axis acceleration signal after filtering;

[0017] Calculating the energy of the three-axis acceleration signal based on the three-axis acceleration signal;

[0018] Processing the photoplethysmogram signal by using an adaptive spectrum correction algorithm to obtain a spectrum diagram of the photoplethysmogram signal;

[0019] Corresponding relatively correct values, normal values ​​and abnormal values ​​are determined based on the frequency spectrum and the energy of the three-axis acceleration signal.

[0020] In one embodiment, an adaptive spectrum correction model is constructed based on the parameters of the triaxial acceleration signal energy, and the correction operation of the photoplethysmography signal based on the adaptive spectrum correction model includes:

[0021] Performing filtering processing on the photoplethysmography signal by using a preset signal preprocessing algorithm;

[0022] Adaptive parameters are introduced to construct a spectrum correction model based on the energy of triaxial acceleration signals;

[0023] The spectrum of the photoplethysmography signal and the spectrum of the triaxial acceleration signal are normalized.

[0024] In one embodiment, the heart rate corresponding to the photoplethysmography signal is corrected based on the historical heart rate, including:

[0025] Calculate the current heart rate based on the historical heart rate, the frequency corresponding to the spectrum of the photoplethysmography signal, and the correction parameters corresponding to the energy of the triaxial acceleration signal;

[0026] The relationship between the correction parameter and the heart rate change threshold is set based on the difference between the current heart rate average and the historical heart rate value within the preset time and the preset heart rate threshold range, and the heart rate change threshold is calculated based on the maximum value of the heart rate change in the intense exercise data set and the mean absolute error of the data set.

[0027] In one of the embodiments, a priority level is established based on a quality classification of the photoplethysmography signal;

[0028] Correcting the historical heart rate based on the priority of the quality classification of the photoplethysmography signal;

[0029] When the quality of the photoplethysmography signal corresponding to the historical heart rate is classified as an abnormal value, the photoplethysmography signal is reset.

[0030] In a second aspect, the present application also provides a real-time heart rate monitoring device with adaptive spectrum correction and peak location. The device comprises:

[0031] A signal quality classification module, used to classify the quality of the photoplethysmography signal based on the timing characteristics of the photoplethysmography signal and the energy of the three-axis acceleration signal, wherein the quality classification includes an absolutely correct value, a relatively correct value, a normal value, and an abnormal value;

[0032] A spectrum module correction module, used to construct an adaptive spectrum correction model based on the parameters of the triaxial acceleration signal energy, and to perform correction operations on the spectrum of the photoplethysmography signal based on the adaptive spectrum correction model;

[0033] The historical heart rate correction module is used to correct the current heart rate corresponding to the photoplethysmogram signal based on the historical heart rate.

[0034] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the following steps when executing the computer program:

[0035] Classifying the quality of the photoplethysmography signal based on the timing characteristics of the photoplethysmography signal and the energy of the triaxial acceleration signal, wherein the quality classification includes an absolutely correct value, a relatively correct value, a normal value, and an abnormal value;

[0036] An adaptive spectrum correction model is constructed based on the parameters of the triaxial acceleration signal energy, and a correction operation is performed on the spectrum of the photoplethysmography signal based on the adaptive spectrum correction model;

[0037] The current heart rate corresponding to the photoplethysmogram signal is corrected based on the historical heart rate.

[0038] In a fourth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0039] Classifying the quality of the photoplethysmography signal based on the timing characteristics of the photoplethysmography signal and the energy of the triaxial acceleration signal, wherein the quality classification includes an absolutely correct value, a relatively correct value, a normal value, and an abnormal value;

[0040] An adaptive spectrum correction model is constructed based on the parameters of the triaxial acceleration signal energy, and a correction operation is performed on the spectrum of the photoplethysmography signal based on the adaptive spectrum correction model;

[0041] The current heart rate corresponding to the photoplethysmogram signal is corrected based on the historical heart rate.

[0042] In a fifth aspect, the present application further provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0043] Classifying the quality of the photoplethysmography signal based on the timing characteristics of the photoplethysmography signal and the energy of the triaxial acceleration signal, wherein the quality classification includes an absolutely correct value, a relatively correct value, a normal value, and an abnormal value;

[0044] An adaptive spectrum correction model is constructed based on the parameters of the triaxial acceleration signal energy, and a correction operation is performed on the spectrum of the photoplethysmography signal based on the adaptive spectrum correction model;

[0045] The current heart rate corresponding to the photoplethysmogram signal is corrected based on the historical heart rate.

[0046] The above-mentioned real-time heart rate monitoring method, device, computer equipment, storage medium and computer program product with adaptive spectrum correction and peak positioning classify the quality of the photoelectric volume pulse wave signal based on the timing characteristics of the photoelectric volume pulse wave signal and the energy of the three-axis acceleration signal, and the quality classification includes absolute correct value, relative correct value, normal value and abnormal value; construct an adaptive spectrum correction model based on the parameters of the three-axis acceleration signal energy, and perform correction operations on the spectrum of the photoelectric volume pulse wave signal based on the adaptive spectrum correction model; perform correction operations on the current heart rate corresponding to the photoelectric volume pulse wave signal based on the historical heart rate. The present application adopts the above-mentioned method, and the complexity of the algorithm can be reduced and the real-time performance can be improved through the accurate classification of the PPG signal. While improving the monitoring accuracy, it can better meet the real-time requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 A diagram showing an application environment of a real-time heart rate monitoring method with adaptive spectrum correction and peak location in one embodiment;

[0048] Figure 2 It is a flow chart of a real-time heart rate monitoring method with adaptive spectrum correction and peak location in one embodiment;

[0049] Figure 3 is a schematic diagram of a PPG signal waveform in one embodiment;

[0050] Figure 4 A schematic diagram of the influence of different MAs on a PPG signal in one embodiment;

[0051] Figure 5 A schematic diagram of constructing a first-order linear model in one embodiment;

[0052] Figure 6 A statistical diagram of a public data set in an embodiment;

[0053] Figure 7 A statistical diagram of the data set in this article in an embodiment;

[0054] Figure 8 It is a structural block diagram of a real-time heart rate monitoring device with adaptive spectrum correction and peak location in one embodiment;

[0055] Fig. 9 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0057] The real-time heart rate monitoring method with adaptive spectrum correction and peak location provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the terminal communicates with the server through the network. The data storage system can store the data that the server needs to process. The data storage system can be integrated on the server, or it can be placed on the cloud or other network servers. Among them, the terminal can be but not limited to various personal computers, laptops, smart phones, tablets, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car-mounted devices, etc. Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server can be implemented as an independent server or a server cluster consisting of multiple servers.

[0058] In one embodiment, Figure 2 As shown, in this embodiment, the method includes the following steps:

[0059] Step 202: classify the quality of the photoplethysmography signal based on the timing characteristics of the photoplethysmography signal and the energy of the three-axis acceleration signal.

[0060] Among them, signal feature extraction aims to deal with special situations in a targeted manner to improve prediction accuracy and reduce the amount of calculation; four results are set through signal features: absolute correct value, relative correct value, normal value and abnormal value. The priorities of the above results decrease in turn, and the interference of incorrect historical heart rate is eliminated by setting priorities.

[0061] Step 204: construct an adaptive spectrum correction model based on the parameters of the triaxial acceleration signal energy, and perform correction operations on the photoplethysmography signal based on the adaptive spectrum correction model.

[0062] Among them, there are usually two types of PPG signal waveforms without MA interference. Figure 3 As shown, signal 1 is a typical PPG signal without MA interference after 0.5-4Hz bandpass filtering, recorded as waveform 1; signal 2 is the result of 0.5-4Hz bandpass filtering of the PPG signal without MA interference in the public data set, recorded as waveform 2. Obviously, both signals can be regarded as periodic signals without MA interference. In waveform 1, there is a maximum value in each cycle, but in waveform 2, the number of extreme values ​​in each cycle cannot be determined. For waveform 1, the heart rate can be obtained by a method similar to the ECG signal to calculate the heart rate, as follows:

[0063]

[0064] Among them, F s is the sampling rate, L en is the signal length, and R is the number of peaks in the signal.

[0065] Step 206: Correct the heart rate corresponding to the photoplethysmography signal based on the historical heart rate.

[0066] Among them, spectrum peak location is the most important part of the heart rate monitoring algorithm. The PPG spectrum seriously interfered by MA will still have multiple peaks after the denoising algorithm. The frequency corresponding to the correct heart rate peak is the frequency of the heart rate (heartbeats per second). Multiplying it by 60 seconds per minute can get the heart rate (heartbeats per minute), that is:

[0067] BPM = 60 × loc;

[0068] Among them, BPM is the heart rate, and loc is the frequency corresponding to the peak heart rate.

[0069] Therefore, spectrum peak positioning can significantly improve the accuracy of overall monitoring by selecting the peak closest to the actual heart rate.

[0070] In the above-mentioned real-time heart rate monitoring method with adaptive spectrum correction and peak location, the connection between historical heart rate and current heart rate: determining the range of current heart rate through historical heart rate is the most important part of spectrum peak location. Based on human physiological characteristics: the speed of change of human heart rate has a certain range, that is, the difference between the current measured heart rate and the historical measured heart rate is within a certain range. The selection of current heart rate basically depends on the historical heart rate. Usually, the selection of heart rate is based on the following:

[0071]

[0072] Among them, B prev is the historical heart rate value, P is the set of loc, loc is the frequency corresponding to the peak heart rate, and argmin function is the variable value when the objective function reaches the minimum value.

[0073] In one embodiment, Figure 3 As shown, the determination of the absolutely correct value can be performed as follows:

[0074] Assume there are 2n segments represented as and In two signal sets S a and S b Medium, signal 1 medium Signal 2 in Each segment has the same data length L en . Pearson correlation coefficient r i , i=1,2,...,n is calculated as:

[0075]

[0076] in and yes and The corresponding means are calculated by Pearson correlation coefficient With the predetermined threshold r min Compare and distinguish waveform S a and S b If Then determine S a is the absolute correct value. min ∈[0.9,0.99].

[0077] In this example, the threshold r is used min To calculate the PPG signal without the dicrotic notch, the signal is classified as an absolutely correct value. a and S b , through segmentation and obtain the corresponding segment and By locating two consecutive peaks of a waveform, the waveform data can be stretched or scaled to obtain signals of equal length. and

[0078] In this embodiment, the heart rate is calculated using a time domain method, which can significantly reduce the amount of calculation and obtain a more accurate heart rate measurement value.

[0079] In one embodiment, Figure 4 As shown, the determination of the relative correct value, the normal value and the abnormal value can be specifically performed as follows:

[0080] The triaxial acceleration signal largely reflects the intensity of the person's exercise and is related to heart rate changes. Therefore, its characteristics, such as energy, can be used to guide the location of the spectrum peak. The relative correct value, normal value, and abnormal value are confirmed by the triaxial acceleration signal energy. Triaxial acceleration signal energy E Acc The calculation is:

[0081]

[0082] Among them, Acc x is the result of bandpass filtering of the x-axis acceleration signal in the unit window, Acc y is the result of bandpass filtering of the y-axis acceleration signal in the unit window, Acc z is the result of bandpass filtering of the z-axis acceleration signal in the unit window, E A is the energy of the three-axis acceleration signal.

[0083] E A The threshold setting should be comprehensively classified from the following two aspects:

[0084] Hand Activities: Indirect Instruction E A At different intensity levels, E A There are significant differences in the values, and E A It is positively correlated with the intensity of hand activity. The intensity of hand activity is also positively correlated with the interference degree of MA. Therefore, hand activity can be divided into three categories: nearly static state, weak activity state (no continuous hand movement) and strong activity state (continuous hand movement).

[0085] PPG signal spectrum: Direct guidance E A The spectrum can intuitively reflect the interference degree of MA, and by setting the threshold E A Combining the hand activity and the three states of the signal spectrum, this paper uses the ACC signal energy E corresponding to each segmented signal spectrum AAs a classification standard, PPG signals are divided into three levels: high-quality PPG signals (H H ), medium quality PPG signal (H m ) and low quality PPG signals (H l ), corresponding to the relative correct value, normal value, and abnormal value respectively. ACC signal energy threshold H H , H m and H l It is based on the individual activity states (known labels) in the training set and the corresponding spectrogram results (such as Figure 4 The mean values ​​of each group were obtained.

[0086] The threshold of the Pearson correlation coefficient is used to classify the absolute correct value, and E A Used to identify the remaining three categories.

[0087] In one embodiment, Figure 5 As shown, an adaptive spectrum correction model is constructed based on the parameters of the triaxial acceleration signal energy, and the correction operation of the photoplethysmogram signal based on the adaptive spectrum correction model includes:

[0088] After the input signal is filtered for the first time (bandpass filtering in the frequency range of 0.5Hz-4Hz) and then filtered for the second time (sparse signal reconstruction filtering), the motion artifacts in the PPG signal are removed using the three-axis acceleration signal through the adaptive spectrum correction model. There are two types of peaks in the PPG signal spectrum after sparse signal reconstruction: 1. The true heart rate peak or the true heart rate peak offset by the influence of MA, and 2. The noise peak generated by the influence of MA on the PPG.

[0089] In order to further remove MA and obtain a clean PPG signal spectrum, one method is to use the three-axis acceleration signal spectrum to perform spectrum correction to remove MA in the PPG spectrum. Through spectrum correction, the PPG spectrum peak can be corrected to the real heart rate value. Since the size difference between the PPG spectrum and the three-axis acceleration spectrum is too large, and the parameter sizes collected by different hardware are different, the direct spectrum correction effect is not ideal, so the PPG spectrum and the three-axis acceleration spectrum are normalized:

[0090]

[0091] in and Normalization is performed at the same time. Normalize separately. max(S pec ) is S pec The maximum value S in pec After normalization, it is expressed as Π={S PPG , S x , Sy , S z}, S∈Π normalization not only eliminates the difference in data amplitude between PPG and ACC spectra, but also reduces the variation in data amplitude between ACC spectra under different intensities. This paper proposes an adaptive spectrum correction model to complete the task of adaptively removing MA in the spectrum under different interference intensities. As shown in the following formula:

[0092]

[0093] Where S k represents the kth element of the set S, and Similarly, M is the spectral resolution. λ∈[0,1] is based on E A It is worth mentioning that 20 sets of data under motion scenes are selected as experimental objects. The least square method LSM is used according to λ and E A To model the intrinsic relationship between signals, all the following steps are performed in each segment between signals.

[0094] Parameter sampling, select N parameters in the i-th window and get:

[0095]

[0096] Where W is the number of signal segments, and in the i-th segment Get N spectra

[0097] For each spectrum Perform HR estimation.

[0098]

[0099] in, is the estimated HR, B rel is the real heart rate obtained from the ECG signal. P is obtained from the spectrum The set of candidate heart rates obtained from n = 1, 2, ..., N. N HR estimates Draw from the i-th window segment.

[0100] Select the correct t in the i-th window i make Closest to B rel .

[0101]

[0102] The parameter t i represents n=1,2,…,N. i With the corresponding fragment Combination

[0103] Compute the parameter mean. To obtain a uniform distribution of D i , and [0, E max ] is evenly divided into T segments, at this time the qth segment q=1,2,...,T. According to E q , D i Divide the y-axis into T groups evenly, and calculate the mean of the qth group as L is for D i Satisfied The number of , j = 1, 2,…, W.

[0104] like Figure 5 As shown, LSM is constructed to perform first-order linear fitting. q , t q ), q=1,2,...,T, establish a first-order linear model E A =at q +b.

[0105]

[0106] Where a and b are unknown coefficients, and the point set (E q , t q ) is fitted by the LSM model, such as Figure 5 As shown, the MA interference in the PPG signal spectrum is dynamically removed.

[0107] Note: By spectrum The number of subsets of P obtained is different, and the candidate heart rate P is not only determined by the real heart rate.

[0108] In one embodiment, in practice, the noise is often closer to the historical heart rate than the real heart rate. Therefore, the formula is improved in this embodiment, including:

[0109]

[0110] Among them B prev is the heart rate in the last second, B est is the current heart rate to be estimated. μ is a correction parameter based on the current activity state and HR variation. P is the set of candidate heart rates derived from the ASC algorithm.

[0111] It is worth mentioning that HR and exercise trends are closely related, and there are two relationship patterns: (i) in activities, the rate of change of HR increases slowly as HR increases; (ii) in non-activity, the rate of change of HR decreases slowly as HR decreases. In order to reduce the error of current HR estimation and the different effects of various activity patterns on human resources, this paper adds a flow adjustment parameter μ.

[0112]

[0113] Among them I low and I high For counting E A The number of times the low quality or high quality PPG signal threshold is met. C low and C high Yes and I low and I high The associated count threshold. and are the thresholds corresponding to low HR, medium HR, and high HR. δ is the calibration value for HR change, defined as:

[0114] δ=B MAX +B MAE ;

[0115] Among them B MAX is the maximum change in HR per second in the high-intensity exercise dataset. MAE is the mean absolute error of the data set. low and I high The iterative adjustment rule is as follows:

[0116]

[0117] Among them I MAX and I MIN are the upper and lower limits of the activity status change indicator counter. low >I MIN When it means the activity state is starting, adjust μ accordingly. The same goes for the opposite.

[0118] In one embodiment, a priority is constructed based on the quality classification of the photoplethysmography signal; the historical heart rate is corrected based on the priority of the quality classification of the photoplethysmography signal; when the quality classification of the photoplethysmography signal corresponding to the historical heart rate is an abnormal value, the photoplethysmography signal is reset.

[0119] Among them, after the historical heart rate deviates, the current heart rate needs to be corrected in time: As mentioned in the previous section, the accuracy of HR largely determines the overall accuracy of HR. In the above algorithm, it is assumed that the historical HR value is accurate. In practice, if the noise peak is very close to the true HR peak or the noise peak is close to the historical HR, measurement errors may occur, and therefore the current HR estimate obtained will also be inaccurate.

[0120] The purpose of the above four signal quality classifications is to handle the current deviation state, that is, to execute the reset mechanism. In order to be more standardized, two additional constraints are added to further distinguish common values ​​from abnormal values. If the following condition 1 is met, it is classified as a common value, and if 2 is met, it is classified as an abnormal value.

[0121] |B est -B prev |≤ΔS min ;

[0122] |B est -B prev |>ΔS max ;

[0123] Where ΔS min and ΔS max are the lower and upper bounds of the error, respectively. By considering the potential noise interference in the estimated and actual HR, the rate of change of the actual HR may exceed the limit B MAX Therefore, ΔS min and ΔS max It is set to δ and 2δ to solve the situation where the maximum error is reached in both historical HR and current HR.

[0124] It is worth mentioning that the performance comparison of different algorithms on public data sets is shown in Table 1:

[0125]

[0126]

[0127] Table 1;

[0128] The results of this algorithm on the dataset are shown in Table 1:

[0129] Scene Type MAE Accuracy at 3 bpm Accuracy at 5 bpm Indoor walking 3.8 74.8 81.8 Outdoor walking 2.6 80.5 86.3 Indoor Cycling 1.0 95.9 97.3 Outdoor Cycling 1.6 92.5 94.3 Indoor running 1.3 91.9 94.6 Outdoor running 1.8 90.3 93.6 Rest 1.7 88.1 95.7 Climb stairs 1.5 92.8 95.0 Free time 4.8 67.3 76.1 Mean 2.2 86.0 90.5

[0130] Table 2;

[0131] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0132] Based on the same inventive concept, the embodiment of the present application also provides a real-time heart rate monitoring device with adaptive spectrum correction and peak location for implementing the real-time heart rate monitoring method with adaptive spectrum correction and peak location mentioned above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in the embodiments of one or more real-time heart rate monitoring devices with adaptive spectrum correction and peak location provided below can refer to the limitations of the real-time heart rate monitoring method with adaptive spectrum correction and peak location above, and will not be repeated here.

[0133] In one embodiment, Figure 8 As shown, a real-time heart rate monitoring device with adaptive spectrum correction and peak location is provided, including: a signal quality classification module, a spectrum module correction module and a historical heart rate correction module, wherein:

[0134] A signal quality classification module, used to classify the quality of the photoplethysmography signal based on the timing characteristics of the photoplethysmography signal and the energy of the three-axis acceleration signal, wherein the quality classification includes an absolutely correct value, a relatively correct value, a normal value, and an abnormal value;

[0135] A spectrum module correction module, used for constructing an adaptive spectrum correction model based on the parameters of the triaxial acceleration signal energy, and performing correction operations on the photoplethysmogram signal based on the adaptive spectrum correction model;

[0136] The historical heart rate correction module is used to correct the heart rate corresponding to the photoplethysmogram signal based on the historical heart rate.

[0137] In one embodiment, the signal quality classification module is also used to: obtain a photoplethysmogram signal that has undergone filtering; calculate the peak value of the photoplethysmogram signal and the corresponding Pearson correlation coefficient; calculate the mean of the Pearson correlation coefficient and compare it with a preset standard threshold, and if the mean is within the preset standard threshold, determine that the quality classification is an absolutely correct value.

[0138] In one embodiment, the signal quality classification module is also used to: obtain a three-axis acceleration signal after filtering; calculate the energy of the three-axis acceleration signal based on the three-axis acceleration signal; process the photoelectric volumetric pulse wave signal through the ASC algorithm to obtain a spectrum diagram of the photoelectric volumetric pulse wave signal; determine the corresponding relatively correct value, normal value and abnormal value based on the spectrum diagram and the hand activity intensity.

[0139] In one embodiment, the spectrum module correction module is also used to: filter the photoplethysmogram signal by preset spectrum and coefficient signal respectively; introduce adaptive parameters to construct a spectrum correction model based on the energy of the three-axis acceleration signal; and normalize the spectrum of the photoplethysmogram signal and the spectrum of the three-axis acceleration signal.

[0140] In one embodiment, the historical heart rate correction module is also used to: calculate the current heart rate based on the historical heart rate, the frequency corresponding to the photoplethysmography signal spectrum and the correction parameters; set the relationship between the correction parameters and the heart rate change threshold based on the difference between the current heart rate average and the historical heart rate value within a preset time and a preset heart rate threshold range, and calculate the heart rate change threshold based on the maximum value of the heart rate change in the strenuous exercise data set and the mean absolute error of the data set.

[0141] In one embodiment, the historical heart rate correction module is also used to: establish a priority based on the quality classification of the photoelectric volumetric pulse wave signal; correct the historical heart rate based on the priority of the quality classification of the photoelectric volumetric pulse wave signal; when the quality classification of the photoelectric volumetric pulse wave signal corresponding to the historical heart rate is an abnormal value, reset the photoelectric volumetric pulse wave signal.

[0142] Each module in the above-mentioned real-time heart rate monitoring device with adaptive spectrum correction and peak location can be implemented in whole or in part by software, hardware and their combination. Each of the above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the corresponding operations of each of the above modules.

[0143] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Fig. 9As shown. The computer device includes a processor, a memory and a network interface connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a real-time heart rate monitoring method with adaptive spectrum correction and peak positioning is implemented.

[0144] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Fig. 9 As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a real-time heart rate monitoring method with adaptive spectrum correction and peak positioning is implemented.

[0145] Those skilled in the art will understand that Fig. 9 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0146] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0147] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0148] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0149] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0150] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0151] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0152] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A real-time heart rate monitoring method with adaptive spectrum correction and peak location, characterized in that: The method comprises: Acquire a triaxial acceleration signal and a pre-processed photoplethysmogram signal; Normalizing the frequency spectrum corresponding to the three-axis acceleration signal and the frequency spectrum corresponding to the preprocessed photoplethysmography signal; According to the three-axis acceleration signal, the energy E of the three-axis acceleration signal is obtained. A ; The energy E of the three-axis acceleration signal A ∈(0,H h ), the quality classification result of the photoelectric volumetric pulse wave signal is a relatively correct value; E A ∈(H m ,H l ), the quality classification result of the photoelectric volumetric pulse wave signal is a normal value; E A ≥H l In the case of H, the quality classification result of the photoelectric volume pulse wave signal is an abnormal value; h , H m and H l The classification criteria are high-quality signals, medium-quality signals, and low-quality signals, respectively; The adaptive spectrum correction model is used to correct the normalized spectrum of the photoplethysmography signal and the spectrum of the triaxial acceleration signal to obtain S k , the adaptive spectrum correction model can be expressed as: in, is the frequency corresponding to the kth frequency point in the spectrum of the photoplethysmographic signal, and is the frequency corresponding to the kth frequency point in the spectrum of the three-axis acceleration signal on the x, y and z axes, M is the spectrum resolution, λ∈[0,1], is the energy E of the three-axis acceleration signal A Changing adaptive parameters; For the S k Perform peak location processing, and obtain a candidate heart rate set P according to the frequencies corresponding to each located heart rate peak; Based on the candidate heart rate set P, the historical heart rate B prev Perform heart rate estimation processing to obtain the current heart rate B est , current heart rate B est It can be expressed as: The value of the flow adjustment parameter μ is related to the quality classification result of the photoplethysmogram signal during the estimation of each historical heart rate. The flow adjustment parameter μ can be expressed as: If the quality classification result of the photoelectric volume pulse wave signal in the estimation process of the historical heart rate is a relatively correct value, then I high The count is increased by one; if the quality classification result of the photoelectric volume pulse wave signal in the historical heart rate estimation process is an abnormal value, then I low Count up by one; C low and C high Yes and I low and I high Related counting threshold; B est’ Indicates the B corresponding to the previous heart rate estimation process of the current heart rate estimation process est , and are the preset thresholds corresponding to low heart rate, medium heart rate and high heart rate; δ is the preset calibration value of heart rate change.

2. The method according to claim 1, characterized in that: The method further comprises: Processing the photoplethysmography signal in segments, and calculating the corresponding Pearson correlation coefficient for each segment obtained by the segment processing; If the mean value of each of the Pearson correlation coefficients is greater than or equal to a preset standard threshold, the quality classification result of the photoplethysmography signal is an absolutely correct value.

3. The method according to claim 2, characterized in that The method further comprises: Based on the quality classification result of the photoplethysmography signal, determining the priority order corresponding to the current heart rate and the historical heart rate, wherein the priority order corresponding to the quality classification result is: absolutely correct value>relatively correct value>normal value>abnormal value; The current heart rate is corrected according to the priority order to obtain a corrected current heart rate.

4. A real-time heart rate monitoring device with adaptive spectrum correction and peak location, characterized in that: The device comprises: The spectrum module correction module is used to obtain the three-axis acceleration signal and the photoelectric volume pulse wave signal after the preprocessing operation; normalize the spectrum corresponding to the three-axis acceleration signal and the photoelectric volume pulse wave signal after the preprocessing operation; obtain the energy E of the three-axis acceleration signal according to the three-axis acceleration signal A ; The energy E of the three-axis acceleration signal A ∈(0,H h ), the quality classification result of the photoelectric volumetric pulse wave signal is a relatively correct value; E A ∈(H m ,H l ), the quality classification result of the photoelectric volumetric pulse wave signal is a normal value; E A ≥H l In the case of H, the quality classification result of the photoelectric volume pulse wave signal is an abnormal value; h , H m and H l The classification criteria are high-quality signals, medium-quality signals, and low-quality signals, respectively; The adaptive spectrum correction model is used to correct the normalized spectrum of the photoplethysmography signal and the spectrum of the triaxial acceleration signal to obtain S k , the adaptive spectrum correction model can be expressed as: in, is the frequency corresponding to the kth frequency point in the spectrum of the photoplethysmographic signal, and is the frequency corresponding to the kth frequency point in the spectrum of the three-axis acceleration signal on the x, y and z axes, M is the spectrum resolution, λ∈[0,1], is the energy E of the three-axis acceleration signal A Changing adaptive parameters; For the S k Perform peak location processing, and obtain a candidate heart rate set P according to the frequencies corresponding to each located heart rate peak; Based on the candidate heart rate set P, the historical heart rate B prev Perform heart rate estimation processing to obtain the current heart rate B est , current heart rate B est It can be expressed as: The value of the flow adjustment parameter μ is related to the quality classification result of the photoplethysmogram signal during the estimation of each historical heart rate. The flow adjustment parameter μ can be expressed as: If the quality classification result of the photoelectric volume pulse wave signal in the estimation process of the historical heart rate is a relatively correct value, then I high The count is increased by one; if the quality classification result of the photoelectric volume pulse wave signal in the historical heart rate estimation process is an abnormal value, then I low Count up by one; C low and C high Yes and I low and I high Related counting threshold; B est’ Indicates the B corresponding to the previous heart rate estimation process of the current heart rate estimation process est , and are the preset thresholds corresponding to low heart rate, medium heart rate and high heart rate; δ is the preset calibration value of heart rate change.

5. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 3 are implemented.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 3 are implemented.

7. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 3 are implemented.

Images