CN114795168B

CN114795168B - Method and system for calculating heart rate of vital sign parameter

Info

Publication number: CN114795168B
Application number: CN202210720787.3A
Authority: CN
Inventors: 陈煜�; 安俊华
Original assignee: Ocamar Information Technology Shanghai Co ltd
Current assignee: Ocamar Information Technology Shanghai Co ltd
Priority date: 2022-06-24
Filing date: 2022-06-24
Publication date: 2022-09-30
Anticipated expiration: 2042-06-24
Also published as: CN114795168A

Abstract

The invention provides a vital sign parameter heart rate calculation method and a vital sign parameter heart rate calculation system, which relate to the technical field of vital sign signal processing, and the method comprises the following steps: step S1: collecting data, and carrying out filtering analysis processing on the data; step S2: carrying out time domain peak point searching on the filtered data to obtain a time domain peak point interval; step S3: performing histogram statistics on the time domain peak point interval to obtain a time domain heart rate value; step S4: carrying out Fourier transform on the filtered data to obtain frequency spectrum data; step S5: searching the frequency spectrum data for a peak point to obtain a frequency domain peak point interval; step S6: counting the peak point interval of the frequency domain to obtain a theoretical heart rate value of the frequency domain; step S7: and comparing the time domain heart rate value with the frequency domain theoretical heart rate value to determine an actual heart rate value. The invention can improve the accuracy of vital sign data measurement.

Description

Method and system for calculating heart rate of vital sign parameter

Technical Field

The invention relates to the technical field of vital sign signal processing, in particular to a vital sign parameter heart rate calculation method and a vital sign parameter heart rate calculation system.

Background

The Ballistocardiography (BCG) technology can non-invasively measure the effect of blood ejected by a human body due to each beat of the heart on the motion of the human body, and obtain a corresponding BCG waveform signal. The BCG signal acquisition technology can non-invasively measure the body tiny vibration signals of the human body caused by heartbeat and respiration, so that the breathing rate and the heart rate of a patient can be monitored in a non-contact manner.

At present, most of calculations of the heart rate of the physical sign signals and the BCG signals are performed on a single time domain calculation method or a single frequency domain calculation method, but the data content of time domain processing is more, and the result cannot be quickly calculated in the frequency domain processing due to the constraint of acquisition time and sampling points.

In the prior art, the invention patent with publication number CN108056769B discloses a method and a device for analyzing and processing vital sign signals, and a vital sign monitoring device, including: acquiring an original signal acquired by a sensor; generating a vital sign time domain signal based on the original signal; calculating to obtain a first vital sign parameter based on the vital sign time domain signal; carrying out time-frequency transformation on the vital sign time domain signal with preset duration to obtain a vital sign frequency domain signal, and calculating to obtain a second vital sign parameter based on the vital sign frequency domain signal; and calculating to obtain final vital sign parameters based on the first vital sign parameters and the second vital sign parameters, wherein the final vital sign parameters comprise a final heart rate and/or a final breathing rate. The invention discloses a method for directly calculating a frequency domain heart rate value by using frequency doubling, and even if signal quality is increased as part of accuracy rate judgment, the frequency domain result has larger error due to lower frequency spectrum resolution.

Disclosure of Invention

Aiming at the defects in the prior art, the invention provides a vital sign parameter heart rate calculation method and a vital sign parameter heart rate calculation system.

According to the method and the system for calculating the heart rate of the vital sign parameters, provided by the invention, the scheme is as follows:

in a first aspect, a vital sign parameter heart rate calculation method is provided, the method including:

step S1: collecting data, and carrying out filtering analysis processing on the data;

step S2: performing time domain peak point searching on the filtered data to obtain a time domain peak point interval;

step S3: performing histogram statistics on the time domain peak point interval to obtain a time domain heart rate value;

step S4: carrying out Fourier transform on the filtered data to obtain frequency spectrum data;

step S5: searching the frequency spectrum data to obtain a frequency domain peak point interval;

step S6: counting the peak point interval of the frequency domain to obtain a theoretical heart rate value of the frequency domain;

step S7: and comparing the time domain heart rate value with the frequency domain theoretical heart rate value to determine an actual heart rate value.

Preferably, the step S1 includes:

step S1.1: the effectiveness judgment is carried out on the collected data, and abnormal points are removed to obtain effective data;

step S1.2: carrying out high-pass filtering on the effective data, and filtering out high-frequency signals uniformly distributed in a signal interval to obtain a preprocessed signal;

step S1.3: and filtering the preprocessing example data, extracting useless signals from the signals, and subtracting the useless signals from the original data to obtain new data.

Preferably, the step S2 includes:

step S2.1: searching peak points of the new data according to a set scale to find characteristic point data;

step S2.2: the effective points with the amplitude and the time meeting the requirements are called, the amplitude of points around each peak point is compared, and the invalid points which do not meet the requirements are removed; after the data are all subjected to secondary filtering, new processing data are obtained;

step S2.3: and calculating the difference of the points after the secondary peak point filtering, and subtracting the coordinate index of the previous peak point from the coordinate index corresponding to the next peak point to obtain the time domain peak point interval value.

Preferably, the step S3 includes:

step S3.1: performing histogram statistics on the time domain peak point interval value;

step S3.2: and obtaining the interval statistical peak points of the time domain, and calculating the heart rate value in the time domain.

Preferably, the step S5 includes:

step S5.1: searching the peak point of the frequency spectrum data according to a certain scale;

step S5.2: after traversing each frequency point of the frequency spectrum, calculating the interval between the next peak point and the previous peak point of the frequency spectrum, and according to the distribution of the characteristic peak points of the frequency spectrum under different heart rates under the condition of the interval value of the peak points, the lower the heart rate is, the smaller the interval of the peak points of the frequency spectrum is, and the higher the heart rate is, the larger the interval between the peak points of the frequency spectrum is;

the relationship between the interval and the corresponding heart rate is expressed by the formula:

heart rate = (spectral interval (sampling rate/number of fourier points) 60 times/min).

Preferably, the step S6 includes:

step S6.1: counting the interval values, and when the interval values meet the condition that the occurrence times are the most or a plurality of interval values are relatively close, taking the interval values as theoretical heart rate intervals obtained in a frequency spectrum;

step S6.2: and comparing the theoretical heart rate intervals with the theoretical heart rate table according to the corresponding frequency spectrum peak point intervals to find out corresponding theoretical heart rate values.

In a second aspect, a vital sign parameter heart rate calculation system is provided, the system comprising:

module M1: collecting data, and carrying out filtering analysis processing on the data;

module M2: performing time domain peak point searching on the filtered data to obtain a time domain peak point interval;

module M3: performing histogram statistics on the time domain peak point interval to obtain a time domain heart rate value;

module M4: carrying out Fourier transform on the filtered data to obtain frequency spectrum data;

module M5: searching the frequency spectrum data for a peak point to obtain a frequency domain peak point interval;

module M6: counting the peak point interval of the frequency domain to obtain a theoretical heart rate value of the frequency domain;

module M7: and comparing the time domain heart rate value with the frequency domain theoretical heart rate value to determine an actual heart rate value.

Preferably, said module M1 comprises:

module M1.1: the effectiveness judgment is carried out on the collected data, and abnormal points are removed to obtain effective data;

module M1.2: carrying out high-pass filtering on the effective data, and filtering out high-frequency signals uniformly distributed in a signal interval to obtain a preprocessed signal;

module M1.3: and filtering the preprocessing example data, extracting useless signals in the signals, and subtracting the useless signals from the original data to obtain new data.

Preferably, said module M2 comprises:

module M2.1: searching peak points of the new data according to a set scale to find characteristic point data;

module M2.2: the effective points with the amplitude and the time meeting the requirements are called, the amplitude of points around each peak point is compared, and the invalid points which do not meet the requirements are removed; after the data are all subjected to secondary filtering, new processing data are obtained;

module M2.3: calculating the difference of the points after the secondary peak point filtering, and subtracting the coordinate index of the previous peak point from the coordinate index corresponding to the next peak point to obtain a time domain peak point interval value;

the module M3 includes:

module M3.1: performing histogram statistics on the time domain peak point interval value;

module M3.2: and obtaining the interval statistical peak points of the time domain, and calculating the heart rate value in the time domain.

Preferably, said module M5 comprises:

module M5.1: searching the peak point of the frequency spectrum data according to a certain scale;

module M5.2: after traversing each frequency point of the frequency spectrum, calculating the interval between the next peak point and the previous peak point of the frequency spectrum, and according to the distribution of the characteristic peak points of the frequency spectrum under different heart rates under the condition of the interval value of the peak points, the interval of the frequency spectrum peak points is smaller when the heart rate is lower, and the interval between the frequency spectrum peak points is larger when the heart rate is higher;

heart rate = (spectral interval (sampling rate/number of fourier points) × 60 times/min);

the module M6 includes:

module M6.1: counting the interval values, and when the interval values meet the condition that the occurrence times are the most or a plurality of interval values are relatively close, taking the interval values as theoretical heart rate intervals obtained in a frequency spectrum;

module M6.2: and comparing the theoretical heart rate intervals with the theoretical heart rate table according to the corresponding frequency spectrum peak point intervals to find out corresponding theoretical heart rate values.

Compared with the prior art, the invention has the following beneficial effects:

1. the distribution characteristics of each group of heartbeat data are observed more visually by counting the time domain/frequency domain peak point intervals, and only one-dimensional array traversal is performed by using the statistics in the measurement range, so that the program complexity is reduced, and the execution speed of the controller is improved;

2. when Fourier change is carried out, time domain data are turned and combined, on the basis of not increasing sampling time, the resolution of a frequency spectrum is increased, and the accuracy of a frequency domain algorithm is improved;

3. by using the rule of peak point intervals in the spectrogram: the larger the heart rate is, the larger the interval is, the smaller the heart rate is, the smaller the interval is, and the practicability of Fourier transform is improved through the characteristic of multiple frequency;

4. and comparing the time domain calculation result with the frequency domain calculation result, and performing time-frequency interaction, thereby improving the accuracy of the final calculation result.

Drawings

Other features, objects and advantages of the invention will become more apparent upon reading of the detailed description of non-limiting embodiments with reference to the following drawings:

FIG. 1 is a schematic flow diagram of the present invention;

FIG. 2 is a data diagram of an example of preprocessing;

FIG. 3 is a graph comparing unwanted data with raw data;

FIG. 4 is the new data;

FIG. 5 is a peak finding graph;

FIG. 6 is a schematic diagram of invalid point culling;

FIG. 7 is new process data;

FIG. 8 is a graph of interval values;

FIG. 9 is a peak point diagram in the conventional case;

FIG. 10 is a peak point diagram in the case where two or more points are present in the same number;

FIG. 11 is a spectral fence effect;

FIG. 12 is data collected at 100 Hz;

FIG. 13 is the corresponding Fourier transform result;

FIG. 14 is a process of flip expansion;

FIG. 15 is the result of the inverse Fourier transform after the inverse expansion;

FIG. 16 is a frequency bin interval of 40 heart rates;

FIG. 17 is a frequency interval of 67 heart rates;

FIG. 18 shows frequency intervals for a heart rate of 93;

fig. 19 shows the frequency interval of 120 heart rates.

Detailed Description

The present invention will be described in detail with reference to specific examples. The following examples will assist those skilled in the art in further understanding the invention, but are not intended to limit the invention in any way. It should be noted that it would be obvious to those skilled in the art that various changes and modifications can be made without departing from the spirit of the invention. All falling within the scope of the present invention.

The embodiment of the invention provides a vital sign parameter heart rate calculation method, which specifically comprises the following steps of:

step S1: and collecting data and carrying out filtering analysis processing on the data.

The step S1 specifically includes:

step S1.1: the method comprises the steps of judging the effectiveness of collected data, removing abnormal points to obtain effective data, comparing original data input by a sensor one by one in the process, carrying out smooth filtering on elements smaller than or larger than a preset value, using an average value to fill in a group according to 5 numerical values, wherein the operation is essentially to check input parameters of the sensor; if there are a large number of numerical operations, the sensors are reconfigured on the hardware.

Step S1.2: and carrying out high-pass filtering on the effective data, and filtering out some high-frequency signals uniformly distributed in a signal interval to obtain a preprocessed signal.

Step S1.3: referring to fig. 2, the preprocessing example data is subjected to some filtering processes, which may be wavelet processing, IIR filtering, etc., and after useless signals (low frequency or high frequency) in the signals are extracted, the useless signals are subtracted from the original data to obtain new data. Wherein FIG. 3 is a comparison of extracted garbage data with raw data; fig. 4 is new data.

Step S2: and searching the peak point of the time domain of the filtered data to obtain the time domain peak point interval.

The step S2 specifically includes:

step S2.1: referring to fig. 5, searching peak points of new data according to a set scale to find feature point data; due to the particularity of the BCG signal, when new data is obtained in step S1.3, peaks of the upper half of the waveform are not found and retained, and peaks of the lower half of the waveform can also be found.

Step S2.2: the points with amplitude and time meeting the requirements are called effective points, because some time scales can be met by searching peak points according to a certain scale, but the points with amplitude not meeting can also be identified, but the points can only be called invalid points, and the correct interval can be interfered in the calculation of the interval between the subsequent peak points and the peak points; therefore, the amplitude values of points around each peak point need to be related for comparison, and invalid points which do not meet the requirements are removed; for example, as shown in fig. 6, the amplitude of point 2 does not satisfy 70% of the amplitude of point 1, and the time interval between the two points is small, so point 2 is definitely an invalid point, and can be eliminated; after the data are filtered for the second time according to the rule, new processing data shown in fig. 7 can be obtained.

Step S2.3: and (4) calculating the difference of the points after the secondary peak point filtering, and subtracting the coordinate index of the previous peak point from the coordinate index corresponding to the next peak point to obtain a time domain peak point interval value, which is shown in fig. 8. The upper part and the lower part are respectively calculated and then put into an array without the difference of the sequence.

Step S3: and carrying out histogram statistics on the time domain peak point interval to obtain a time domain heart rate value.

The step S3 includes:

step S3.1: performing histogram statistics on the time domain peak interval value group; for example, the interval value and the range (30-200 times/min, corresponding time domain interval range 2000-30 under 100Hz sampling rate) of the actual heart rate are compared in 2000-30 traversal ways, and the indexes meeting the conditions are accumulated under the array according to the current comparison interval; since the peak point is not as obvious as that of fig. 9 in some cases, but as that of fig. 10, two or more points with the same number are present and are far apart, which causes difficulty in finding the peak point by interval statistics, and the deviation of the final heart rate value is large; to solve this problem, the summed average is taken for two points around each point.

Step S3.2: and the heart rate value in the time domain can be directly calculated by obtaining the interval statistical peak point of the time domain. For example, at a sampling rate of 100Hz, the index of the coordinates of the interval statistical peak points is 115, and if the index starts from 30, the heart rate value of the time domain is 60/((115-30) × 0.01s) =70 times/min; the heart rate value in the time domain can be obtained through the steps.

Step S4: and carrying out Fourier transform on the filtered data to obtain frequency spectrum data.

Step S4 includes the following steps:

step S4.1: carrying out Fourier transform on the preprocessed data subjected to the high-pass filtering in the step S1.2; for a known sampling frequency Fs of the original signal, the number of sampling points is N, then the frequency that can be analyzed by the discrete fourier transform is Fs/N (i.e., the resolution of the frequency), and in some cases, the sampling frequency is 1024Hz, and if the number of sampling points is 1024, the resolution of the frequency is 1 Hz; in this case, due to the spectrum barrier effect fig. 11, the point between 1Hz and 2Hz cannot be represented, for example, the dotted line in fig. 11 is the frequency point between 2 and 3 of the spectrum; the method of making the frequency resolution smaller is only to decrease the sampling rate or to increase the number of sampling points.

The human heart rate range is lower within 0.5 Hz-3.3 Hz, so the sampling rate is not set to be too high, but the Nyquist sampling theorem can be satisfied; increasing the number of sampling points means that the sampling time needs to be increased, certain response time is needed in the use occasion of the vital sign monitoring mattress, and the increase of the number of sampling points can increase the calculation burden of a calculation element, so that the acquisition time cannot be excessively increased, and the calculation speed is obviously reduced when 4096-point discrete Fourier transform (FFT) is processed by 32-bit calculation equipment; therefore, if the FFT is used alone to calculate the heart rate result and not combined with the comparison of the time domain result, a large error exists; in addition, in some cases, when the multiple frequency in the frequency spectrum is not obvious, the range interval where the heart rate is located cannot be determined by the interval between the multiple frequency points, so that the FFT cannot obtain an accurate result.

The method is provided for turning over and expanding the original data, the amplitude of characteristic points in the frequency spectrum can be enhanced on the basis of ensuring the computing power, the originally protruded frequency peak point is more prominent, and meanwhile, the frequency spectrum resolution is doubled on the basis of not increasing the sampling time by expanding the data acquisition point; FIG. 12 is data collected at 100Hz, and FIG. 13 is the corresponding Fourier transform result; FIG. 14 is a process of flip expansion; FIG. 15 is the result of the inverse augmented Fourier transform; compared with the spectral resolution of 0.97 in fig. 13 and the spectral resolution of 0.49 in fig. 15, the precision is doubled.

Step S5: and searching the peak points of the frequency spectrum data to obtain the frequency domain peak point interval.

The step S5 includes: step S5.1: and searching the peak point of the expanded FFT frequency spectrum according to a certain scale.

Step S5.2: after traversing each frequency point of the frequency spectrum, calculating the interval between the next peak point and the previous peak point of the frequency spectrum, and according to the distribution of the characteristic peak points of the frequency spectrum under different heart rates under the condition of the interval value of the peak points, the interval of the frequency spectrum peak points is smaller when the heart rate is lower, and the interval between the frequency spectrum peak points is larger when the heart rate is higher;

heart rate = (spectral interval (sampling rate/number of fourier points) × 60 times/min).

The reason for this is that the heart rate at a high frequency corresponds to a higher fundamental frequency, so that the distance between points expressed in the frequency spectrum by the multiple frequencies in the fourier spectrum is larger; for example, the frequency doubling of 30 times/minute (0.5Hz frequency point) is (1Hz), and the frequency doubling of 60 times/minute (1Hz frequency point) is (2Hz), so that the interval between 0.5 and 1Hz is always much smaller than the interval between 1Hz and 2Hz (under the same Fourier transform spectrum resolution); for example, the frequency point interval of 40 heart rates (fig. 16), the frequency point interval of 67 heart rates (fig. 17), the frequency point interval of 93 heart rates (fig. 18), and the frequency point interval of 120 heart rates (fig. 19) are the spectrum intervals at different heart rates.

Step S6: and counting the peak point interval of the frequency domain to obtain a theoretical heart rate value of the frequency domain.

The step S6 includes: step S6.1: and (3) counting the interval values, finding out the interval value with the most occurrence times or a plurality of interval values which are relatively close (can be an average value or a mode), and taking the interval value as a theoretical heart rate interval obtained in the frequency spectrum.

Step S6.2: comparing the theoretical heart rate intervals with the theoretical heart rate table (2048 sampling points at 100 Hz) 1 according to the frequency spectrum peak point intervals, and finding out corresponding theoretical heart rate values; this value is compared to the time domain as the theoretical heart rate value in the frequency domain.

TABLE 1 theoretical heart rate table

Step S7: and comparing the time domain heart rate value with the frequency domain theoretical heart rate value to determine an actual heart rate value. Specifically, the method is performed according to the flowchart 1, and if the error between the time domain value and the frequency domain value is small, it is indicated that the heart rate value in the time domain is the actual heart rate value.

The embodiment of the invention provides a vital sign parameter heart rate calculation method and a vital sign parameter heart rate calculation system, wherein a result obtained by a time domain fast statistical algorithm is compared with a frequency domain multiple frequency point algorithm in a consolidation mode, so that the vital sign data measurement accuracy is improved.

Those skilled in the art will appreciate that, in addition to implementing the system and its various devices, modules, units provided by the present invention as pure computer readable program code, the system and its various devices, modules, units provided by the present invention can be fully implemented by logically programming method steps in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers and the like. Therefore, the system and various devices, modules and units thereof provided by the invention can be regarded as a hardware component, and the devices, modules and units included in the system for realizing various functions can also be regarded as structures in the hardware component; means, modules, units for performing the various functions may also be regarded as structures within both software modules and hardware components for performing the method.

The foregoing description of specific embodiments of the present invention has been presented. It is to be understood that the present invention is not limited to the specific embodiments described above, and that various changes or modifications may be made by one skilled in the art within the scope of the appended claims without departing from the spirit of the invention. The embodiments and features of the embodiments of the present application may be combined with each other arbitrarily without conflict.

Claims

1. A vital sign parameter heart rate calculation method is characterized by comprising the following steps:

step S2: carrying out time domain peak point searching on the filtered data to obtain a time domain peak point interval;

step S5: searching the frequency spectrum data for a peak point to obtain a frequency domain peak point interval;

step S7: comparing the time domain heart rate value with the frequency domain theoretical heart rate value to determine an actual heart rate value;

the step S6 includes:

2. Method for vital sign parameter heart rate calculation according to claim 1, wherein step S1 comprises:

step S1.3: and filtering the preprocessing example data, extracting useless signals in the signals, and subtracting the useless signals from the original data to obtain new data.

3. Method for vital sign parameter heart rate calculation according to claim 2, wherein step S2 comprises:

step S2.1: searching the peak points of the new data according to a set scale to find out the feature point data;

4. Method for vital sign parameter heart rate calculation according to claim 3, wherein step S3 comprises:

5. Method for vital sign parameter heart rate calculation according to claim 1, wherein step S5 comprises:

heart rate (spectrum interval (sampling rate/number of fourier points) 60 times/min).

6. A vital sign parameter heart rate computing system, comprising:

module M2: carrying out time domain peak point searching on the filtered data to obtain a time domain peak point interval;

module M5: searching the frequency spectrum data to obtain a frequency domain peak point interval;

module M7: comparing the time domain heart rate value with the frequency domain theoretical heart rate value to determine an actual heart rate value;

the module M6 includes:

7. Vital signs parameter heart rate calculation system according to claim 6, wherein the module M1 comprises:

8. Vital signs parameter heart rate calculation system according to claim 7, wherein the module M2 comprises:

module M2.1: searching the peak points of the new data according to a set scale to find out the feature point data;

module M2.2: the effective points with the amplitude and the time meeting the requirements are called, the amplitude values of the points around each peak point are compared, and the invalid points which do not meet the requirements are removed; after the data are all subjected to secondary filtering, new processing data are obtained;

the module M3 includes:

9. The vital sign parameter heart rate computing system of claim 6, wherein the module M5 comprises: