A multi-channel BCG signal detection and analysis method, a heart rate detection method and a device

Abstract

The present invention discloses a multi-channel BCG signal detection and analysis method, comprising the following steps: (1) collecting at least four juxtaposed BCG signals; (2) selecting one of the optimal signal data for processing as the basis for subsequent heart rate detection. The present invention also provides corresponding heart rate detection methods and devices respectively. Among them, the heart rate detection method comprises the following steps: (a) performing EMD decomposition on the optimal signal to obtain a plurality of intrinsic mode components; (b) obtaining the spectrum of each independent component; (c) calculating the signal power ratio of 1-10 Hz based on the spectrum; (d) taking the signal greater than the signal power ratio threshold as the starting point, and simultaneously taking the amplitude value corresponding to the starting point as the threshold; (e) collecting the starting point positions of S1 and S2; (f) determining S1; (g) obtaining heart rate data based on S1. The present invention has well improved and enhanced the acquisition and calculation accuracy of heart sound and heart rate, providing an effective means and guarantee for non-contact and long-term heart sound signal monitoring.

Description

Technical Field

[0001] The present invention relates to the field of medical technology, and in particular to a multi-channel BCG signal detection and analysis method and a heart rate detection method and device. Background Art

[0002] As one of the most important physiological signals of the human body, heart sound can reflect the mechanical activity of the heart and has irreplaceable clinical value in ECG monitoring. Heart sound usually consists of four parts, namely: the first heart sound (S1), the second heart sound (S2), the third heart sound (S3), and the fourth heart sound (S4). S3 and S4 are relatively weak and can only be heard under special circumstances, so the main purpose of heart sound segmentation is to find S1 and S2.

[0003] Heart rate is one of the most important parameters of heart sounds and is the basis for subsequent related analysis. A cardiac cycle refers to the interval between two consecutive S1s, and the calculation of heart rate is based on the positioning of the cardiac cycle. Therefore, positioning S1 becomes the key to calculating heart rate. At present, many solutions for monitoring heart rate based on a single PVDF sensor have limited acquisition range and more interference during actual application because the acquired signal is affected by the sensitivity and placement of a single PVDF. It is impossible to obtain a BCG signal with a high signal-to-noise ratio, resulting in low accuracy in acquiring heart sound signals, large errors in calculating heart rate, and difficulty in accurately reflecting the contraction and relaxation time of the heart. Summary of the invention

[0004] In view of the shortcomings of the above-mentioned prior art, the present invention provides a multi-channel BCG signal detection and analysis method and a heart rate detection method and device, which can not only obtain high signal-to-noise ratio signals, but also obtain the complete structure of heart sounds, realize non-contact heart rate calculation and provide a powerful means for subsequent heart disease monitoring.

[0005] To achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0006] A multi-channel BCG signal detection and analysis method comprises the following steps:

[0007] (1) Collect at least four parallel BCG signals;

[0008] (2) Select the best signal data from one channel for processing as the basis for subsequent heart rate detection. In this step, the process of selecting the best signal data is as follows:

[0009] (21) Filtering all raw BCG signal data;

[0010] (22) Perform first-order differential signal analysis on the filtered data calculate;

[0011] (23) Calculate the standard deviation of the first-order difference signal in sequence std and the short-time energy En ;

[0012] (24) Determine the optimal signal path according to the maximum standard deviation value and the short-time energy value.

[0013] In the above step (24), the present invention may also select the signal channels corresponding to the two maximum values according to the standard deviation values as the alternative channels, and then continue the subsequent processing to determine the optimal signal path. The processing process is as follows:

[0014] (24) Select the signal channels corresponding to the two maximum values according to the standard deviation values as the alternative channels;

[0015] (25) Calculate the autocorrelation coefficients of the two alternative channels respectively: ACF(j) = C j / C0, where C0 is the signal variance, , is the signal at the time point, is the signal data length, is the signal mean value, is the lag coefficient, is the autocorrelation coefficient with the lag coefficient being ;

[0016] (26) Based on the autocorrelation coefficient ACF , extract the peak points in the coefficient ACF (200:900) to obtain ACFmax , and then calculate the energy proportion based on 10 points before and after and 40 points before and after the peak points. The calculation formula is as follows:

[0017] ;

[0018] (27) Perform smoothing processing on ACF max and Pacf , perform smoothing filtering based on 50 points before and after the data points to obtain the smoothed characteristic parameter ACF max _100, Pacf_100 ;

[0019] (28) Perform comprehensive judgment on the alternative channels based on ACF max and Pacf to finally determine the optimal signal channel for the subsequent heart rate detection. The judgment criterion is: if ACF max (1) > ACF max (2), andPacf (1) > Pacf If (2), then select channel 1; otherwise, select channel 2.

[0020] Furthermore, in the step (21), the filtering range is 1 - 20 Hz.

[0021] Specifically, in the step (22), the first-order difference signal is calculated using the following formula:

[0022] ;

[0023] In the formula, diff_data(i) is the first-order difference signal of the i-th data point; filter_data(i) is the filtered data of the i-th data point.

[0024] Specifically, it is characterized in that in the step (23), the standard deviation of the first-order difference signal is calculated using the following formula:

[0025] ;

[0026] In the formula, is the mean value of the first-order difference signal .

[0027] Specifically, in the step (23), the short-time energy of the first-order difference signal is calculated using the following formula: .

[0028] Based on the above method, the present invention also provides a corresponding heart rate detection method, including the following steps:

[0029] (a) Decompose the determined optimal signal to obtain multiple intrinsic mode components ; ;

[0030] (b) Perform a fast Fourier transform on to obtain the spectrum of each independent component;

[0031] (c) Calculate the signal power ratio of the 1 - 10 Hz signal based on the spectrum Pr ;

[0032] (d) Set the signal power ratio threshold thr , and use the signal greater than the signal power ratio threshold as the starting point, and at the same time use the amplitude value corresponding to the starting point as the threshold;

[0033] (e) After determining the starting point, when the subsequent data is less than the threshold, it is determined as the end point, and the starting point positions of the first heart sound S1 and the second heart sound S2 are respectively collected;

[0034] (f)According to the durations of S1 and S2, when T S1 > T S2 , T S1S2 < T S2S1 , the peak points of the envelope amplitudes S1 and S2 En max (S1) > En max (S2), determine S1;

[0035] (g)Extract the peak points within the starting point of S1 based on adjacent S1s, subtract the peak points to obtain heart rate data, and achieve heart rate detection.

[0036] Specifically, in the step (a), The decomposition process is as follows:

[0037] A) Determine all the extreme points of the differential signal , and then use cubic spline interpolation to obtain the upper envelope composed of all the maximum and minimum values and the lower envelope ;

[0038] B) Denote the average value of the upper and lower envelope lines of the differential signal as , remove the instantaneous average from the signal to obtain the instantaneous high-frequency component and calculate to obtain the intrinsic mode component ;

[0039] C) Remove from the original data , and repeat the above process until is finally decomposed into N pieces of components, that is, obtain .

[0040] Preferably, in the step (d), set the signal power ratio threshold thr to 0.2 - 0.6.

[0041] In addition, based on a general inventive concept, the present invention also provides a device for implementing the above multi-channel BCG signal detection and analysis method and heart rate detection method, including a plurality of PVDF sensors arranged on the mattress in a full-bridge manner for collecting human BCG signals, a signal processing device for filtering the BCG signals, a transmitting module for sending the processed data, and a central processing terminal connected to the transmitting module for selecting the optimal one of the BCG signals and implementing heart rate detection; the signal processing device includes a preamplification module, a high-pass filtering module, a first-stage amplification module, a 50Hz filtering module, and a low-pass filtering module connected in sequence; the preamplification module is connected to all the PVDF sensors at the same time; the low-pass filtering module is connected to the transmitting module.

[0042] Furthermore, the outside of the signal processing device is also covered with a shielding layer.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] (1) By collecting the BCG signals of four channels of the human body, the present invention has a relatively wide collection range, and the lying position of the human body is not restricted too much. Then, through signal quality evaluation, the optimal signal channel is determined, and subsequent heart rate calculation is performed based on this. In this way, the present invention can not only obtain more accurate heart rate data, but also obtain a relatively accurate heart sound signal structure, accurately reflecting the heart contraction and relaxation time.

[0045] (2) The present invention EMD decomposes the signal, calculates through spectrum analysis, combines the signal power ratio threshold to determine the starting point and ending point of S1 and S2 collection, and finally determines S1 according to the judgment condition to obtain the heart rate data. By adopting this method, the present invention can effectively avoid the interference of noise on inaccurate heart rate calculation in the non-contact heart rate calculation process and obtain an obvious heartbeat pattern.

[0046] (3) Through the effective combination of software and hardware, each link is closely linked and complements each other, which well improves the acquisition and calculation accuracy of heart sound and heart rate, providing an effective means and strong guarantee for non-contact and long-term heart sound signal monitoring. Therefore, the present invention is very suitable for popularization and application. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a schematic diagram of the layout of PVDF sensors in an embodiment of the present invention.

[0048] Figure 2 It is a block diagram of the device structure in an embodiment of the present invention.

[0049] Figure 3 It is a schematic diagram of the first-order difference signal and the original signal in an embodiment of the present invention.

[0050] Figure 4 This is a comparison schematic diagram of four signal channels in an embodiment of the present invention.

[0051] Figure 5 This is a short-time energy schematic diagram of four signal channels in an embodiment of the present invention.

[0052] Figure 6 This is a schematic diagram of the intrinsic mode components in an embodiment of the present invention.

[0053] Figure 7 This is a schematic diagram during S1 and S2 heartbeats in an embodiment of the present invention.

[0054] Figure 8 This is a schematic diagram of the S1 and S2 start point detection in an embodiment of the present invention. Detailed implementation manners

[0055] The present invention will be further described below in conjunction with the accompanying drawings and embodiments. The implementation manners of the present invention include but are not limited to the following embodiments. Embodiment

[0056] This embodiment provides a multi-channel BCG signal detection and analysis method, which acquires BCG signals of four channels, combines a four-channel signal quality evaluation algorithm to obtain an optimal signal channel for subsequent heart sound and heart rate calculation. The implementation process of this embodiment is as follows:

[0057] First, four-channel BCG signals are acquired; in this embodiment, the acquisition of the BCG signals is realized by multiple PVDF sensors arranged in a full-bridge manner on the mattress for acquiring human BCG signals, and the layout structure is as Figure 1 shown, Figure 1 wherein, reference numeral 1 is a PVDF sensor, and reference numeral 2 is a mattress.

[0058] Then, one of the optimal signal data is selected for processing. In this embodiment, the signal processing is realized by a signal processing device, and then the signal data is transmitted to a central processing terminal for optimal signal selection processing through a transmitting module (in this embodiment, a radio frequency module for wireless transmission), and its device structure is as Figure 2 shown. The signal processing device includes a preamplification module, a high-pass filtering module, a first-stage amplification module, a 50Hz filtering module, and a low-pass filtering module connected in sequence; the preamplification module is simultaneously connected to all PVDF sensors; the low-pass filtering module is connected to the transmitting module. The signal processing device in this embodiment can effectively obtain BCG signals with high signal-to-noise ratio through a processing method of two-stage amplification and multi-stage filtering in different ranges. And in order to further avoid signal interference, the outside of the signal processing device is also coated with a shielding layer.

[0059] The specific process of selecting the optimal signal data in this embodiment is as follows:

[0060] (1) Filter all the original BCG signal data, and the low-pass filtering range is selected as 20 Hz;

[0061] (2) Perform first-order difference signal calculation on the filtered data, and its calculation formula is as follows:

[0062] ;

[0063] In the formula, diff_data(i) is the first-order difference signal of the i-th data point; filter_data(i) is the filtered data of the i-th data point; in this embodiment, the first-order difference signal and the original signal are as Figure 3 shown;

[0064] (3) Calculate the standard deviation std and short-time energy En of the first-order difference signal in sequence. Among them, the standard deviation is calculated using the following formula:

[0065] ;

[0066] In the formula, is the mean value of the first-order difference signal , in this embodiment, the comparison of the four signal channels is as Figure 4 shown;

[0067] The short-time energy is calculated using the following formula:

[0068] ;

[0069] In this embodiment, the short-time energy of the four signal channels is as Figure 5 shown;

[0070] (4) Determine the optimal signal according to the maximum standard deviation value and short-time energy value. The selected optimal signal channel can be used as the basis for subsequent heart rate detection.

[0071] Embodiment 2

[0072] The difference from Embodiment 1 is that in this embodiment, two signal channels with the two maximum standard deviation values are selected as alternative channels according to the standard deviation value, and then subsequent processing is continued to finally determine the optimal signal. The processing process is as follows (the previous process is the same as that in Embodiment 1, and here only how to finally determine the optimal signal after selecting the alternative channels is introduced):

[0073] (1)Calculate the autocorrelation coefficients of the two alternative channels respectively: ACF(j) = C j / C0, where C0 is the signal variance, ;

[0074] (2)Based on the autocorrelation coefficients ACF , extract the peak points in the coefficients ACF (200:900) to obtain ACFmax , and then calculate the energy proportion Pacf based on 10 points before and after the peak points and 40 points before and after. The calculation formula is as follows:

[0075] ;

[0076] (3)Perform smoothing processing on ACF max and Pacf , and perform smoothing filtering based on 50 points before and after the data points to obtain the smoothed characteristic parameters ACF max _100, Pacf_100 ;

[0077] (4)Make a comprehensive judgment on the alternative channels based on ACF max and Pacf to finally determine the optimal signal path for the subsequent heart rate detection; the main judgment criteria are: if ACF max (1) > ACF max (2), and Pacf (1) > Pacf (2), then select Channel 1; otherwise, select Channel 2. The method of calculating and processing the autocorrelation coefficient adopted in this embodiment determines the optimal signal path, with better channel quality, and can more effectively locate whether the signal is a heartbeat or noise signal in the subsequent process, making the heart rate detection more accurate.

[0078] Example 3

[0079] Based on the optimal signal path selected and processed in Example 1 or 2, this example obtains the heart sounds S1 and S2, and finally calculates the heart rate to determine the heartbeat mode. The main process is as follows:

[0080] The first step is to perform decomposition on the optimal signal to obtain multiple intrinsic mode components ; the decomposition process is as follows:

[0081] (1)Determine the differential signal All extreme points, and then use cubic spline interpolation to obtain the upper envelope composed of all maxima and minima and the lower envelope ;

[0082] (2) Denote the average value of the upper and lower envelope lines of the differential signal as , and remove the instantaneous average from the signal to obtain the instantaneous high-frequency component Calculate to obtain the intrinsic mode components ;

[0083] (3) Remove from the original data , and repeat the above process until is finally decomposed into N individual components, that is, obtain . In this embodiment, the intrinsic mode components are as shown in Figure 6 .

[0084] Second step: Perform a fast Fourier transform on to obtain the spectrum of each independent component .

[0085] Third step: Calculate the signal power ratio of the 1 - 10 Hz signal based on the spectrum Pr , as follows:

[0086] .

[0087] Fourth step: Set the signal power ratio threshold thr (set to 0.2 in this embodiment), and use the signal greater than the signal power ratio threshold as the starting point, and at the same time use the amplitude value corresponding to the starting point as the threshold;

[0088] After determining the starting point, when the subsequent data is less than the threshold, it is determined as the end point, and the starting point positions of the first heart sound S1 and the second heart sound S2 are collected respectively;

[0089] Then, according to the durations of S1 and S2, when T S1 > T S2 , T S1S2 < T S2S1 , and the peak points of the envelope amplitudes S1 and S2 En max (S1)> En max (S2), determine S1; in this embodiment, during the S1 and S2 heartbeats, as shown in Figure 7 , the detection of the S1 and S2 starting points is as shown inFigure 8 as shown

[0090] Finally, based on adjacent S1, extract the peak points within the starting point of S1, subtract the peak points to obtain the heart rate data, and achieve heart rate detection.

[0091] Through the effective combination of software and hardware, the present invention effectively solves the uncertainty of heart rate calculation for heartbeat patterns, and at the same time provides a powerful means and guarantee for long-term, continuous, and non-contact heart sound signal monitoring. Compared with the prior art, the present invention has prominent substantive features and significant progress.

[0092] The above embodiments are only one of the preferred embodiments of the present invention and should not be used to limit the protection scope of the present invention. Any meaningless changes or polish made in the main design concept and spirit of the present invention, as long as the technical problems solved are still consistent with the present invention, should be included in the protection scope of the present invention.

Claims

1. A multi-channel BCG signal detection and analysis method, characterized in that It includes the following steps: (1) Collect at least four parallel BCG signals; (2) Select one of the optimal signal data for processing as the basis for subsequent heart rate detection; In this step, the process of selecting the optimal signal data is as follows: (21) Filter all the original BCG signal data; (22) Perform a first-order difference signal on the filtered data Calculation; (23) Calculate the standard deviation of the first-order difference signal and the short-time energy in sequence std and the short-time energy En ; (24) Select two signal channels with the two maximum values as alternative channels according to the standard deviation value; (25) Calculate the autocorrelation coefficients of the two alternative channels respectively: ACF(j) = C j / C0, where C0 is the signal variance, , is the signal at the time point, is the signal data length, is the signal mean, is the lag coefficient, is the autocorrelation coefficient when the lag coefficient is ; (26) Based on the autocorrelation coefficient ACF , extract the peak points in the coefficient ACF (200:900) to obtain ACFmax , and then calculate the energy proportion based on 10 points before and after the peak point and 40 points before and after Pacf . The calculation formula is as follows: ； (27) Pair ACF max and Pacf are smoothed. Smoothing filtering is performed based on 50 points before and after the data points to obtain the smoothed characteristic parameters ACF max _100, Pacf_100 ; (28) Based on the alternative channels ACF max _100 and Pacf_100 make a comprehensive judgment, and finally determine the optimal signal channel for the subsequent heart rate detection; the judgment criterion is: if ACF max _100 (1) > ACF max _100 (2), and Pacf_ 100 (1) > Pacf_100 (2), then select Channel 1, otherwise select Channel 2.

2. The multi-channel BCG signal detection and analysis method according to claim 1, wherein In the step (21), the filtering range is 1 - 20 Hz.

3. A multi-channel BCG signal detection and analysis method according to claim 1, characterized in that, In the step (22), the following formula is used to calculate the first-order difference signal: ； In the formula, diff_data(i) is the first-order difference signal of the i-th data point; filter_data(i) is the filtered data of the i-th data point.

4. A multi-channel BCG signal detection and analysis method according to claim 1, characterized in that In the step (23), the following formula is used to calculate the standard deviation of the first-order difference signal: ； In the formula, is the mean value of the first-order difference signal ; is the first-order difference signal of the th data point.

5. A multi-channel BCG signal detection and analysis method according to claim 1, characterized in that In the step (23), the following formula is used to calculate the short-time energy of the first-order difference signal: 。 6. An apparatus for implementing the multi-channel BCG signal detection and analysis method according to any one of claims 1 to 5, characterized in that, It includes a plurality of PVDF sensors arranged in a full-bridge manner on the mattress for collecting human BCG signals, a signal processing device for filtering the BCG signals, a transmitting module for sending the processed data, and a central processing terminal connected to the transmitting module for selecting one of the optimal BCG signals and realizing heart rate detection; The signal processing device includes a pre-amplification module, a high-pass filtering module, a first-stage amplification module, a 50 Hz filtering module, and a low-pass filtering module connected in sequence; The pre-amplification module is simultaneously connected to all the PVDF sensors; The low-pass filtering module is connected to the transmitting module.

7. A heart rate detection method, characterized in that, Based on the multi-channel BCG signal detection and analysis method according to any one of claims 1 to 5, it includes the following steps: (a) For the optimal signal data perform decomposition to obtain a plurality of intrinsic mode components ; (b) Perform a fast Fourier transform on to obtain the spectrum of each independent component ; (c) Calculate the proportion of the signal power in the range of 1 - 10 Hz based on the spectrum Pr ; (d) Set the signal power ratio threshold thr , and use the signal greater than the signal power ratio threshold as the starting point, and at the same time use the amplitude value corresponding to the starting point as the threshold; the signal power ratio threshold thr is 0.2 - 0.6; (e) After determining the starting point, when the subsequent data is less than the threshold value, it is determined as the end point, and the starting point positions of the first heart sound S1 and the second heart sound S2 are respectively collected; (f) According to the durations of S1 and S2, when T S1 > T S2 , T S1S2 < T S2S1 , the peak points of the envelope amplitudes of S1 and S2 En max (S1) > En max (S2), determine S1; (g) Based on adjacent S1s, extract the peak points within the S1 starting point, subtract the peak points to obtain heart rate data, and realize heart rate detection.

8. A heart rate detection method according to claim 7, characterized in that, In the step (a), The decomposition process is as follows: A) Determine the differential signal of all the extreme points, and then use cubic spline interpolation to obtain the upper envelope composed of all the maxima and minima, and the lower envelope ; B) Denote the average value of the upper and lower envelopes of the differential signal as , and remove the instantaneous average from the signal to obtain the instantaneous high-frequency component Calculate to obtain the intrinsic mode component i mf ; C) In the original data remove , repeat the above process until is finally decomposed into N i mf components, that is, obtain .

9. An apparatus for implementing the heart rate detection method according to any one of claims 7 to 8, characterized in that, It includes a plurality of PVDF sensors arranged in a full-bridge manner on the mattress for collecting human BCG signals, a signal processing device for filtering the BCG signals, a transmitting module for sending the processed data, and a central processing terminal connected to the transmitting module for selecting one of the optimal BCG signals and realizing heart rate detection; The signal processing device includes a pre-amplification module, a high-pass filtering module, a first-stage amplification module, a 50 Hz filtering module, and a low-pass filtering module connected in sequence; The pre-amplification module is simultaneously connected to all the PVDF sensors; The low-pass filtering module is connected to the transmitting module.

10. The device according to claim 9, characterized in that, The outside of the signal processing device is also coated with a shielding layer.

Images