A heart rate estimation method

Abstract

The present invention discloses a heart rate estimation method, comprising: obtaining a cardiac impact signal; dividing the cardiac impact signal into a non-motion interference segment and a motion interference segment; performing J peak detection on the non-motion interference segment to obtain a first J peak sequence; constructing a J peak detection model according to the first J peak sequence and a point process model; performing J peak detection on the motion interference segment using the J peak detection model to obtain a second J peak sequence; and obtaining a heart rate estimation value according to the first J peak sequence and the second J peak sequence. The present invention can effectively identify the non-motion interference segment and the motion interference segment in the cardiac impact signal, and can construct a J peak detection model based on the non-motion interference segment and the point process model to perform J peak detection on the motion interference segment, thereby obtaining an accurate heart rate estimation value.

Description

Technical Field

[0001] The present invention relates to the technical field of heart rate detection, and in particular to a heart rate estimation method based on heart shock signals. Background Art

[0002] Ballistocardiography (BCG) signals are signals formed by measuring tiny changes in human body surface pressure through sensors, and contain information such as human heartbeat, breathing, and movement. Compared with traditional commonly used ECG signals, ballistocardiography signals can be obtained by installing sensors in places such as mattresses and seat cushions, and have extremely strong detection convenience, which makes ballistocardiography signals the main method for detecting cardiac cycles and calculating heart rates. Figure 1 As shown, the maximum value of the cardiac impulse signal in a cardiac cycle is recorded as J peak, and the nearest maximum values ​​on the left and right are recorded as H peak and L peak. The HJL wave group is the embodiment of the response of blood vessels to cardiac ejection; if the position of each J peak in a signal can be obtained, each cardiac cycle and heart rate can be calculated.

[0003] However, since the cardiac signal is often measured by an external sensor, and the signal generated by the heartbeat is very small compared to the signal generated by breathing and possible body movement, the waveform of the cardiac signal is susceptible to greater interference, especially when the body movement amplitude of the measured object is large. At present, for the cardiac signal, the interference with a clear frequency range such as breathing can be removed by a bandpass filter, but the possible human movement interference cannot be removed well because its amplitude is large relative to the amplitude of the cardiac information and there is no prior frequency domain feature, resulting in that the heart rate measurement based on the cardiac signal cannot be extended to many application scenarios. Summary of the invention

[0004] The purpose of the present invention is to provide a heart rate estimation method, which can effectively identify the motion-free interference segment and the motion-interference segment in the cardiac impulse signal, and construct a J-peak detection model based on the motion-free interference segment and the point process model to perform J-peak detection on the motion-interference segment, thereby obtaining an accurate heart rate estimation value.

[0005] In order to achieve the above object, the present invention is implemented by the following technologies:

[0006] A heart rate estimation method, comprising:

[0007] Obtaining cardiac shock signal;

[0008] Dividing the heartbeat signal into a segment without motion interference and a segment with motion interference;

[0009] Performing J peak detection on the non-action interference segment to obtain a first J peak sequence;

[0010] Constructing a J peak detection model according to the first J peak sequence and the point process model;

[0011] Using the J-peak detection model to perform J-peak detection on the segment with motion interference to obtain a second J-peak sequence; and

[0012] A heart rate estimation value is obtained according to the first J peak sequence and the second J peak sequence.

[0013] Preferably, the step of dividing the cardiac impulse signal into a segment without motion interference and a segment with motion interference comprises:

[0014] Dividing the cardiac impulse signal into a plurality of evaluation signals according to a preset evaluation window length;

[0015] Using an interference evaluation function to calculate each of the evaluation signals to obtain an evaluation function value corresponding to each of the evaluation signals;

[0016] Dividing all the evaluation signals into a plurality of evaluation groups according to a preset cardiac window length;

[0017] Selecting a maximum value from the evaluation function values ​​corresponding to all the evaluation signals in each evaluation group as the maximum value of each evaluation group;

[0018] Obtaining an action interference threshold according to the maximum values ​​of all the evaluation groups;

[0019] According to the motion interference threshold, segmentally process the evaluation function values ​​corresponding to all the evaluation signals in the heartbeat signal to obtain a first threshold segment and a second threshold segment; and

[0020] The first threshold segment and the second threshold segment are respectively mapped to the cardiac impulse signal to obtain the motion-free interference segment and the motion-interference segment.

[0021] Preferably, the interference evaluation function is expressed as follows:

[0022]

[0023] Wherein, Q(i) represents the evaluation function value corresponding to the i-th evaluation signal; n represents the number of sampling points in the i-th evaluation signal; D(k) represents the signal value corresponding to the k-th sampling point in the i-th evaluation signal;

[0024] The motion interference threshold is calculated using the following formula:

[0025] threshold=3×min{Y′(1), Y′(2),…,Y′(h),…}

[0026] Among them, threshold represents the action interference threshold; {Y'(1), Y'(2), ···, Y'(h), ···} represents the set of maximum values ​​of all the evaluation groups; Y'(h) represents the maximum value of the hth evaluation group.

[0027] Preferably, before executing the step of performing J peak detection on the non-action interference segment to obtain a first J peak sequence, the step further includes:

[0028] The heartbeat signal is processed to remove respiratory interference by using empirical wavelet transform to obtain a signal without respiratory interference.

[0029] Preferably, the step of performing J peak detection on the non-action interference segment to obtain a first J peak sequence comprises:

[0030] Dividing the de-respiratory interference signal into a plurality of transformed signals according to a window length of a preset length transformation window;

[0031] Performing length transformation on each of the transformation signals to obtain a length transformation result of each of the transformation signals;

[0032] Obtaining a length transformation result of the de-respiratory interference signal according to the length transformation results of all the transformation signals;

[0033] Mapping all peaks in the length transformation result of the de-respiratory interference signal to the heartbeat signal to obtain an initial J peak sequence of the heartbeat signal; and

[0034] An initial J peak sequence corresponding to the motion-free interference segment in the cardiac impulse signal is selected as the first J peak sequence.

[0035] Preferably, the length of each of the transformed signals is transformed using the following formula:

[0036]

[0037] Among them, L(t) represents the length transformation result of the t-th transformed signal; w represents the window length of the preset length transformation window; m represents the number of sampling points in the t-th transformed signal; Z(b) represents the signal value corresponding to the b-th sampling point in the t-th transformed signal.

[0038] Preferably, the J peak detection model is a normal probability model, and the probability density function of the J peak detection model is:

[0039]

[0040] Wherein, μ is the mean of the J-peak detection model, and σ is the variance of the J-peak detection model.

[0041] Preferably, the step of using the J peak detection model to perform J peak detection on the segment with motion interference to obtain a second J peak sequence includes:

[0042] Selecting an initial J peak sequence corresponding to the motion interference segment in the heartbeat signal as a candidate J peak sequence;

[0043] Obtaining a plurality of candidate J peak distances according to the candidate J peak sequence, and all the candidate J peak distances constitute a candidate J peak distance sequence;

[0044] Obtaining the probability of the candidate J peak spacing sequence according to the probability density function of the J peak detection model; and

[0045] The second J peak sequence is selected from the candidate J peak sequences according to the probability of the candidate J peak interval sequences.

[0046] Preferably, the probability of the candidate J peak spacing sequence is calculated using the following formula:

[0047]

[0048] Wherein, {JJ} represents the candidate J peak spacing sequence; P({JJ}) represents the probability of the candidate J peak spacing sequence; JJ r represents the rth candidate J peak distance; f(JJ r ) represents the probability density function value of the rth candidate J peak distance; s represents the number of the candidate J peak distances; δ=1.

[0049] Preferably, the step of obtaining a heart rate estimation value according to the first J peak sequence and the second J peak sequence comprises:

[0050] Obtaining a final J peak sequence of the ballistocardi signal according to the first J peak sequence and the second J peak sequence; and

[0051] The heart rate estimation value is obtained according to the number of final J peaks in the final J peak sequence of the cardiac impulse signal within a preset time.

[0052] Compared with the prior art, the present invention has at least one of the following advantages:

[0053] The present invention provides a heart rate estimation method, which can divide a cardiac impulse signal into a segment without motion interference and a segment with motion interference. Based on the segment without motion interference, a first J peak sequence can be obtained, and a J peak detection model can be constructed according to the first J peak sequence and a point process model; a second J peak sequence can be obtained by performing J peak detection on the segment with motion interference through the J peak detection model, and a heart rate estimation value can be obtained according to the first J peak sequence and the second J peak sequence.

[0054] After the present invention divides the cardiac impact signal into several evaluation signals, the evaluation function value corresponding to each evaluation signal can be obtained through the interference evaluation function, thereby obtaining the motion interference threshold, and then effectively identifying the motion interference-free segment and the motion interference segment in the cardiac impact signal according to the motion interference threshold.

[0055] The present invention can filter out respiratory interference in the cardiac impact signal based on empirical wavelet transform and obtain a respiratory interference-free signal; by performing length transformation on the respiratory interference-free signal, the initial J peak sequence of the cardiac impact signal can be obtained, and then the first J peak sequence corresponding to the motion-free interference segment can be obtained.

[0056] The present invention can construct a J peak detection model according to the first J peak sequence and the point process model, and can accurately detect the second J peak sequence from the motion interference segment through the probability density function of the J peak detection model and the candidate J peak interval.

[0057] The present invention can obtain a final J peak sequence of a cardiac shock signal by combining the first J peak sequence and the second J peak sequence, and can obtain a heart rate estimation value by counting the number of final J peaks in the final J peak sequence of the cardiac shock signal within a preset time. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a schematic diagram of the J peak, H peak and L peak in the cardiac impulse signal in a cardiac cycle;

[0059] Figure 2 is a flow chart of a heart rate estimation method provided by an embodiment of the present invention;

[0060] Figure 3 is a cardiac impulse signal diagram of a heart rate estimation method provided by an embodiment of the present invention;

[0061] Figure 4 is an evaluation function value diagram corresponding to all evaluation signals of a heart rate estimation method provided by an embodiment of the present invention;

[0062] Figure 5 is a segmented diagram of a cardiac impulse signal of a heart rate estimation method provided by an embodiment of the present invention;

[0063] Figure 6 is a signal diagram of a heart rate estimation method without respiratory interference provided by an embodiment of the present invention;

[0064] Figure 7 It is a length transformation result diagram of a heart rate estimation method without respiratory interference signal provided by an embodiment of the present invention;

[0065] Figure 8 It is a final J peak sequence diagram of a cardiac impulse signal of a heart rate estimation method provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0066] The following is a further detailed description of a heart rate estimation method proposed in the present invention in combination with the accompanying drawings and specific embodiments. According to the following description, the advantages and features of the present invention will be clearer. It should be noted that the accompanying drawings are in a very simplified form and use non-precise proportions, which are only used to conveniently and clearly assist in explaining the purpose of the embodiments of the present invention. In order to make the purposes, features and advantages of the present invention more obvious and easy to understand, please refer to the accompanying drawings. It should be noted that the structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification for people familiar with this technology to understand and read, and are not used to limit the limiting conditions for the implementation of the present invention, so they have no technical substantive significance. Any structural modification, change in proportional relationship or adjustment of size, without affecting the effects that the present invention can produce and the purposes that can be achieved, should still fall within the scope of the technical content disclosed by the present invention.

[0067] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0068] Combined with Figures 2 to 8 As shown, this embodiment provides a heart rate estimation method, including: step S110, obtaining a cardiac impact signal; step S120, dividing the cardiac impact signal into a segment without motion interference and a segment with motion interference; step S130, performing J peak detection on the segment without motion interference to obtain a first J peak sequence; step S140, constructing a J peak detection model based on the first J peak sequence and a point process model; step S150, using the J peak detection model to perform J peak detection on the segment with motion interference to obtain a second J peak sequence; and step S160, obtaining a heart rate estimation value based on the first J peak sequence and the second J peak sequence.

[0069] Please also refer to Figures 2 to 5, the step S120 includes: dividing the cardiac impact signal into several evaluation signals according to the window length of the preset evaluation window; using the interference evaluation function to calculate each of the evaluation signals to obtain the evaluation function value corresponding to each of the evaluation signals; dividing all the evaluation signals into several evaluation groups according to the window length of the preset cardiac window; selecting the maximum value from the evaluation function values ​​corresponding to all the evaluation signals in each evaluation group as the maximum value of each evaluation group; obtaining the motion interference threshold according to the maximum values ​​of all the evaluation groups; segmenting the evaluation function values ​​corresponding to all the evaluation signals in the cardiac impact signal according to the motion interference threshold to obtain the first threshold segment and the second threshold segment; and mapping the first threshold segment and the second threshold segment to the cardiac impact signal respectively to obtain the segment without motion interference and the segment with motion interference.

[0070] It can be understood that, in some other embodiments, the interference evaluation function is expressed as follows:

[0071]

[0072] Wherein, Q(i) represents the evaluation function value corresponding to the i-th evaluation signal; n represents the number of sampling points in the i-th evaluation signal; D(k) represents the signal value corresponding to the k-th sampling point in the i-th evaluation signal;

[0073] The motion interference threshold is calculated using the following formula:

[0074] threshold=3×min{Y′(1), Y′(2),…,Y′(h),…} (2)

[0075] Among them, threshold represents the action interference threshold; {Y'(1), Y'(2), ···, Y'(h), ···} represents the set of maximum values ​​of all the evaluation groups; Y'(h) represents the maximum value of the hth evaluation group.

[0076] Specifically, the cardiac impact signal is measured by an in vitro sensor, and the measured object will inevitably breathe and move the body during the measurement process, and the breathing and body movement of the measured object will bring respiratory interference and motion interference to the cardiac impact signal. Since the respiratory information in the cardiac impact signal has a clear frequency range, it can be effectively filtered out. For motion interference, when the change amplitude of a certain signal segment in the cardiac impact signal is too large, it means that the signal segment may be affected by motion interference and may be the segment with motion interference; when calculating the heart rate based on the cardiac impact signal, in order to ensure the accuracy of the J peak detection model and the accuracy of the calculated heart rate estimate, it is necessary to accurately identify the segment without motion interference and the segment with motion interference in the cardiac impact signal, but the present invention is not limited to this.

[0077] In this embodiment, when the cardiac impact signal is divided into several evaluation signals, the window length of the preset evaluation window can be set to 16 sampling points, and the number of sampling points in each evaluation signal is also 16. The evaluation function value corresponding to each evaluation signal calculated by formula (1) is the differential energy of each evaluation signal; and when the change amplitude of the evaluation signal is large, its differential energy will be larger, that is, its corresponding evaluation function value will be larger. When all the evaluation signals are divided into several evaluation groups, considering the actual value range of heart rate, the time range of 2s must contain at least one cardiac cycle; based on the 50Hz sampling frequency of the external sensor, the window length of the preset cardiac window can be set to 100 sampling points; since the number of sampling points in each evaluation signal is 16, each evaluation group can contain 6 evaluation signals. The maximum value is selected from the evaluation function values ​​corresponding to the 6 evaluation signals in each evaluation group, and is used as the maximum value of each evaluation group, so the maximum value of each evaluation group is a local maximum value. By selecting the minimum value from the maximum values ​​of all the evaluation signal groups, the motion interference base value, i.e., min{Y'(1), Y'(2), ···, Y'(h), ···}, can be obtained; the motion interference threshold can be calculated based on the motion interference base value and formula (2), and the motion interference threshold is 145. According to the motion interference threshold, the evaluation function values ​​corresponding to all the evaluation signals in the heartbeat signal can be divided into the first threshold segment ( Figure 4 No shaded area) and the second threshold segment ( Figure 4The evaluation function value corresponding to each evaluation signal in the first threshold segment is not greater than the motion interference threshold, and the evaluation function value corresponding to each evaluation signal in the second threshold segment is greater than the motion interference threshold. The sampling points corresponding to the first threshold segment are mapped one by one to the sampling points of the cardiac impact signal, and the non-motion interference segment ( Figure 5 Similarly, the sampling points corresponding to the second threshold segment are mapped one by one to the sampling points of the cardiac impact signal, and the motion interference segment ( Figure 5 There is a shaded portion in the figure), but the present invention is not limited thereto.

[0078] Please also refer to Figure 2 , Figure 6 and Figure 7 Before executing the step S130, the method further includes: using empirical wavelet transform to remove respiratory interference from the heart beat signal to obtain a signal without respiratory interference.

[0079] It can be understood that, in some other embodiments, the step S130 includes: dividing the de-respiratory interference signal into a plurality of transformed signals according to the window length of a preset length transformation window; performing a length transformation on each of the transformed signals to obtain a length transformation result of each of the transformed signals; obtaining a length transformation result of the de-respiratory interference signal according to the length transformation results of all the transformed signals; mapping all peaks in the length transformation result of the de-respiratory interference signal to the cardiac impact signal to obtain an initial J peak sequence of the cardiac impact signal; and selecting the initial J peak sequence corresponding to the motion-free interference segment in the cardiac impact signal as the first J peak sequence.

[0080] In some embodiments, the length of each of the transformed signals is transformed using the following formula:

[0081]

[0082] Among them, L(t) represents the length transformation result of the t-th transformed signal; w represents the window length of the preset length transformation window; m represents the number of sampling points in the t-th transformed signal; Z(b) represents the signal value corresponding to the b-th sampling point in the t-th transformed signal.

[0083] Specifically, in this embodiment, since the respiratory information in the cardiac impact signal has a clear frequency range, the respiratory interference in the cardiac impact signal can be filtered out based on the prior respiratory information frequency and using the empirical wavelet transform to obtain the respiratory interference-free signal, but the present invention is not limited to this.

[0084] Specifically, in this embodiment, when the de-respiratory interference signal is divided into a plurality of the transformation signals, taking into account the duration of the HJL wave group in the cardiac impact signal, the window length of the preset length transformation window can be set to 16 sampling points, and the number of sampling points in each of the transformation signals is also 16. The length transformation result of each of the transformation signals can be calculated using formula (3), and the length transformation result of each of the transformation signals characterizes the degree of change of each of the transformation signals. The length transformation result of the de-respiratory interference signal can be obtained by combining the length transformation results of all the transformation signals, and there are several peaks in the length transformation result of the de-respiratory interference signal. According to the definition of J peak in the cardiac impact signal: the peak with the largest amplitude in a cardiac cycle, the sampling points corresponding to all peaks in the length transformation result of the de-respiratory interference signal can be mapped one by one to the sampling points of the cardiac impact signal, so as to obtain the initial J peak sequence of the cardiac impact signal, and the initial J peak sequence of the cardiac impact signal includes several initial J peaks; more specifically, if the sampling point corresponding to a peak in the length transformation result of the de-respiratory interference signal is mapped to a certain sampling point of the cardiac impact signal, and the cardiac impact signal does not have an initial J peak at the sampling point, then the maximum value of other signals near the sampling point can be taken as the initial J peak, so as to obtain the initial J peak sequence of the cardiac impact signal. In the initial J peak sequence of the cardiac impact signal, several initial J peaks corresponding to the non-action interference segment can be selected to form the first J peak sequence, but the present invention is not limited to this.

[0085] Please continue to refer to Figure 2 , the J peak detection model is a normal probability model, and the probability density function of the J peak detection model is:

[0086]

[0087] Wherein, μ is the mean of the J-peak detection model, and σ is the variance of the J-peak detection model.

[0088] Specifically, in the present embodiment, in the step S140, when the J peak detection model is constructed based on the point process model and the first J peak sequence is used as sample data, in order to ensure the accuracy of the J peak detection model, the maximum 5% and the minimum 5% of the samples in the first J peak sequence can be removed, and the middle 90% of the samples are taken to construct the sample set of the J peak detection model, and the sample mean and sample variance in the sample set are taken as the mean and variance of the J peak detection model, but the present invention is not limited to this.

[0089] Please continue to refer to Figure 2, the step S150 includes: selecting the initial J peak sequence corresponding to the motion interference segment in the cardiac impulse signal as a candidate J peak sequence; obtaining a plurality of candidate J peak intervals according to the candidate J peak sequence, and all the candidate J peak intervals constitute a candidate J peak interval sequence; obtaining the probability of the candidate J peak interval sequence according to the probability density function of the J peak detection model; and selecting the second J peak sequence from the candidate J peak sequence according to the probability of the candidate J peak interval sequence.

[0090] It is understandable that, in some other embodiments, the probability of the candidate J peak spacing sequence is calculated using the following formula:

[0091]

[0092] Wherein, {JJ} represents the candidate J peak spacing sequence; P({JJ}) represents the probability of the candidate J peak spacing sequence; JJ r represents the rth candidate J peak distance; f(JJ r ) represents the probability density function value of the rth candidate J peak distance; s represents the number of the candidate J peak distances; δ=1.

[0093] Specifically, in this embodiment, the candidate J peak sequence may include several candidate J peaks, and the distance between the sampling points of two adjacent candidate J peaks is a candidate J peak spacing, so several candidate J peak spacings can be obtained based on several candidate J peaks, and the candidate J peak spacing sequence can be obtained based on several candidate J peak spacings; the probability of the candidate J peak spacing sequence can be calculated using formula (4) and formula (5). In addition, several first J peak spacings can be obtained based on the first J peak sequence, and the first J peak spacing sequence can be obtained based on several first J peak spacings, and the probability of the first J peak spacing sequence can be calculated using formula (4) and formula (5). More specifically, when the difference between the probability of the candidate J peak spacing sequence and the probability of the first J peak spacing sequence is within a preset range, the candidate J peaks involved can be selected from the candidate J peak sequence according to the candidate J peak spacing sequence to form the second J peak sequence; if the difference between the probability of the candidate J peak spacing sequence and the probability of the first J peak spacing sequence exceeds the preset range, any number of the candidate J peak spacings can be deleted to form a new candidate J peak spacing sequence, and the probability of the new candidate J peak spacing sequence can be calculated using formula (4) and formula (5) until the difference between the probability of the new candidate J peak spacing sequence and the probability of the first J peak spacing sequence is within the preset range, and the candidate J peaks involved are selected from the candidate J peak sequence according to the new candidate J peak spacing sequence to form the second J peak sequence, but the present invention is not limited to this.

[0094] Please also refer to Figure 2 and Figure 8 , the step S160 includes: obtaining a final J peak sequence of the cardiac impact signal according to the first J peak sequence and the second J peak sequence; and obtaining the heart rate estimation value according to the number of final J peaks in the final J peak sequence of the cardiac impact signal within a preset time.

[0095] Specifically, in this embodiment, the first J peak sequence and the second J peak sequence are combined to obtain the final J peak sequence of the cardiac impact signal, and the heart rate estimation value can be obtained by counting the number of final J peaks in the final J peak sequence of the cardiac impact signal within the preset time, but the present invention is not limited to this.

[0096] In summary, this embodiment provides a heart rate estimation method, which can divide the cardiac impact signal into a non-motion interference segment and a motion interference segment; a first J peak sequence can be obtained by performing J peak detection on the non-motion interference segment; a J peak detection model can be constructed based on the first J peak sequence and the point process model; a second J peak sequence can be obtained by performing J peak detection on the motion interference segment through the J peak detection model, and a heart rate estimation value can be obtained based on the first J peak sequence and the second J peak sequence. After the cardiac impact signal is divided into several evaluation signals, the evaluation function value corresponding to each evaluation signal can be obtained through the interference evaluation function, thereby obtaining the motion interference threshold, and then effectively identifying the non-motion interference segment and the motion interference segment in the cardiac impact signal according to the motion interference threshold. This embodiment can also filter out respiratory interference in the cardiac impact signal based on empirical wavelet transform and obtain a signal without respiratory interference; by performing a length transformation on the signal without respiratory interference, the initial J peak sequence of the cardiac impact signal can be obtained, and then the first J peak sequence corresponding to the segment without motion interference can be obtained; and a J peak detection model can be constructed based on the first J peak sequence and the point process model, and the second J peak sequence can be accurately detected from the segment with motion interference through the probability density function of the J peak detection model and the candidate J peak spacing.

[0097] Although the content of the present invention has been described in detail through the above preferred embodiments, it should be appreciated that the above description should not be considered as a limitation of the present invention. After reading the above content, it will be apparent to those skilled in the art that various modifications and substitutions of the present invention will occur. Therefore, the protection scope of the present invention should be limited by the appended claims.

Claims

1. A heart rate estimation method, characterized in that: include: Obtain cardiac shock signal; Dividing the heartbeat signal into a segment without motion interference and a segment with motion interference; Performing J peak detection on the non-action interference segment to obtain a first J peak sequence; Constructing a J peak detection model according to the first J peak sequence and the point process model; Using the J-peak detection model to perform J-peak detection on the segment with motion interference to obtain a second J-peak sequence; and Obtaining a heart rate estimation value according to the first J peak sequence and the second J peak sequence; The step of dividing the cardiac impulse signal into a segment without motion interference and a segment with motion interference comprises: Dividing the cardiac impulse signal into a plurality of evaluation signals according to a preset evaluation window length; Using an interference evaluation function to calculate each of the evaluation signals to obtain an evaluation function value corresponding to each of the evaluation signals; Dividing all the evaluation signals into a plurality of evaluation groups according to a preset cardiac window length; Selecting a maximum value from the evaluation function values ​​corresponding to all the evaluation signals in each evaluation group as the maximum value of each evaluation group; Obtaining an action interference threshold according to the maximum values ​​of all the evaluation groups; According to the motion interference threshold, segmentally process the evaluation function values ​​corresponding to all the evaluation signals in the heartbeat signal to obtain a first threshold segment and a second threshold segment; and Mapping the first threshold segment and the second threshold segment to the heart ballistic signal respectively to obtain the motion-free interference segment and the motion-interference segment; The expression of the interference evaluation function is as follows: Wherein, Q(i) represents the evaluation function value corresponding to the i-th evaluation signal; n represents the number of sampling points in the i-th evaluation signal; D(k) represents the signal value corresponding to the k-th sampling point in the i-th evaluation signal; The motion interference threshold is calculated using the following formula: threshold=3×min{Y′(1), Y′(2),…,Y′(h),…} Wherein, threshold represents the action interference threshold; {Y'(1), Y'(2), ..., Y'(h), ...} represents the set of maximum values ​​of all the evaluation groups; Y'(h) represents the maximum value of the hth evaluation group; The step of obtaining a heart rate estimation value according to the first J peak sequence and the second J peak sequence comprises: Obtaining a final J peak sequence of the ballistocardi signal according to the first J peak sequence and the second J peak sequence; and The heart rate estimation value is obtained according to the number of final J peaks in the final J peak sequence of the cardiac impulse signal within a preset time.

2. The heart rate estimation method according to claim 1, characterized in that: Before executing the step of performing J peak detection on the non-action interference segment to obtain a first J peak sequence, the method further includes: The heartbeat signal is processed to remove respiratory interference by using empirical wavelet transform to obtain a signal without respiratory interference.

3. The heart rate estimation method according to claim 2, characterized in that: The step of performing J peak detection on the non-action interference segment to obtain a first J peak sequence comprises: Dividing the de-respiratory interference signal into a plurality of transformed signals according to a window length of a preset length transformation window; Performing length transformation on each of the transformation signals to obtain a length transformation result of each of the transformation signals; Obtaining a length transformation result of the de-respiratory interference signal according to the length transformation results of all the transformation signals; Mapping all peaks in the length transformation result of the de-respiratory interference signal to the heartbeat signal to obtain an initial J peak sequence of the heartbeat signal; and An initial J peak sequence corresponding to the motion-free interference segment in the cardiac impulse signal is selected as the first J peak sequence.

4. The heart rate estimation method according to claim 3, characterized in that: The length of each of the transformed signals is transformed using the following formula: Among them, L(t) represents the length transformation result of the t-th transformed signal; w represents the window length of the preset length transformation window; m represents the number of sampling points in the t-th transformed signal; Z(b) represents the signal value corresponding to the b-th sampling point in the t-th transformed signal.

5. The heart rate estimation method according to claim 1, characterized in that: The J peak detection model is a normal probability model, and the probability density function of the J peak detection model is: Wherein, μ is the mean of the J-peak detection model, and σ is the variance of the J-peak detection model.

6. The heart rate estimation method according to claim 3, characterized in that: The step of using the J peak detection model to perform J peak detection on the segment with motion interference to obtain a second J peak sequence includes: Selecting an initial J peak sequence corresponding to the motion interference segment in the heartbeat signal as a candidate J peak sequence; Obtaining a plurality of candidate J peak distances according to the candidate J peak sequence, and all the candidate J peak distances constitute a candidate J peak distance sequence; Obtaining the probability of the candidate J peak spacing sequence according to the probability density function of the J peak detection model; and The second J peak sequence is selected from the candidate J peak sequences according to the probability of the candidate J peak interval sequences.

7. The heart rate estimation method according to claim 6, characterized in that: The probability of the candidate J peak spacing sequence is calculated using the following formula: Wherein, {JJ} represents the candidate J peak spacing sequence; P({JJ}) represents the probability of the candidate J peak spacing sequence; JJ r represents the rth candidate J peak distance; f(JJ r ) represents the probability density function value of the rth candidate J peak distance; s represents the number of the candidate J peak distances; δ=1.

Images