CN111759292B - Device and method for comprehensive measurement of human heart rate, respiration and blood oxygen - Google Patents

Abstract

The invention relates to a device and method for measuring human physiological health parameters, in particular to a device and method for comprehensively measuring human heart rate, respiration and blood oxygen. The purpose of the present invention is to solve the problems existing in the prior art, such as large size, complicated operation, affected test signal authenticity, large error, single function, and no display of blood volume pulse waveform, and provides a human body heart rate, respiration and blood oxygen Comprehensive measuring device and method. The device is provided with a display screen installation slot on the top of the housing on the base, and a fixing piece installation slot on the side; the capacitive display screen is installed in the display screen installation slot, the light source camera fixing piece is installed in the fixing piece installation slot; the light source camera fixing piece is installed in the fixing piece installation slot; There are light source installation holes at the bottom and camera installation holes at the top; the LED polychromatic light source is installed at the light source installation hole, and the color camera is installed at the camera installation hole; the development board is located inside the housing; the development board is connected to an external power supply through a power adapter. The method is carried out using this device.

Description

Device and method for comprehensive measurement of human heart rate, respiration and blood oxygen

technical field

The invention relates to a human body physiological health parameter measuring device and method, in particular to a human body heart rate, respiration and blood oxygen comprehensive measuring device and method.

Background technique

Heart rate, respiration and blood oxygen are important human physiological health parameters, which have important reference value in clinical diagnosis, and all require professional testing equipment for testing. In addition, most of the health testing equipment on the market today is large in size, complicated in operation and single in function, and requires professional medical staff to complete the corresponding parameter measurement, which is not conducive to health monitoring and medical diagnosis of families and individuals.

According to the different signal acquisition methods, the physiological health parameter detection equipment can be divided into two types: contact type and non-contact type.

Most contact measurement devices rely on accelerometers, and such devices generally need to fit the corresponding parts of the subject to be measured. Therefore, contact measurement equipment has great limitations in use. For example, the patient to be measured has trauma and cannot fit the sensor. In this case, measurement cannot be achieved.

Non-contact measurement equipment mainly uses optical or ultrasonic methods to obtain signals, which can overcome the problem of limited use of contact measurement equipment. At present, the mainstream non-contact physiological health parameter detection method is to use the interaction between light and biological tissue to extract physiological information, and photoplethysmography (PPG technology) is the most widely used method to obtain the photoplethysmography of the subject. The basic principle of the signal (PPG signal) technology is to irradiate the light source on the densely distributed human blood vessels, such as fingertips or earlobes. Due to the periodic circulation of the human heart to eject blood, the blood vessels are also periodically expanded and contracted. The transmitted or reflected light intensity also changes periodically, and the photodetector reflects the blood volume change of the corresponding part of the subject by recording the intensity information of the transmitted light in a period of time.

The existing physiological health parameter detection equipment generally uses a single-point photoelectric detector, which is easily interfered by noise, and the measured signal signal-to-noise ratio is low, resulting in a large error in the measurement result.

Taking the commercially available finger-clip pulse oximeter based on the PPG principle as an example, this pulse oximeter mainly has the following four shortcomings:

First, it will put pressure on the test subject's fingertips, which will affect the authenticity of the test signal;

Second, single-point photoelectric detectors are generally used, and the signal-to-noise ratio of the measured signal is low, resulting in a large error in the measurement result;

Third, generally only the detection of heart rate and blood oxygen can be realized, and the function is single;

Fourth, the blood volume pulse waveform of the tester is generally not displayed.

SUMMARY OF THE INVENTION

The purpose of the present invention is to solve the technical problems of the prior art, such as large body size, complicated operation, affected test signal authenticity, large error, single function, and no display of blood volume pulse waveform, and provides a human body heart rate, respiration and blood pressure Oxygen comprehensive measuring device and method.

For solving the above-mentioned technical problems, the technical solutions provided by the present invention are as follows:

The invention provides a comprehensive measurement device for human heart rate, respiration and blood oxygen, which is special in that it includes a casing, a base, a capacitive display screen, an LED polychromatic light source, a color camera, a light source camera fixture, a development board and a power adapter ;

The housing is arranged on the base, the top of the housing is provided with a display screen installation slot, and the side of the housing is provided with a fixing piece installation slot;

The capacitive display screen is installed in the display screen installation groove, and the light source camera fixing member is installed in the fixing member installation groove;

The bottom of the light source camera fixing member is provided with a light source installation hole, and the top of the light source camera fixing member is provided with a camera installation hole;

The LED polychromatic light source is installed at the light source installation hole, the color camera is installed at the camera installation hole, and the color camera receives the outgoing light of the LED polychromatic light source;

The development board is located inside the casing, the development board is connected to the LED polychromatic light source through wires, and the color camera and the capacitive display screen are respectively connected through the data cable, and the development board sends out control signals to control the LED polychromatic light source, the color camera and the capacitive display screen respectively. ;

The development board is connected to an external power supply through a power adapter.

Further, the LED polychromatic light source adopts a white light source.

Further, the color camera adopts an area array CMOS sensor.

Further, the area array CMOS sensor is an area array OV5647-CMOS sensor.

Further, the development board is a Raspberry Pi 4B development board, and the power adapter is a Raspberry Pi 4B power adapter.

Further, the development board is provided with a band-pass filter, and the band-pass filter is a third-order Butterworth band-pass filter.

Further, the capacitive display screen adopts a TFT capacitive display screen.

Further, the overall size formed by the casing mounted on the base is 135mm×82mm×60mm.

The present invention also provides a method for comprehensively measuring human heart rate, respiration and blood oxygen using the above-mentioned comprehensive measuring device for human heart rate, respiration and blood oxygen, which is special in that it includes the following steps:

1) The development board collects video data from the color camera, identifies the target area for each frame of image in the video data, calculates the average gray value of the target area, and obtains a gray-frame number matrix, that is, a one-dimensional PPG signal; The target area is the area covered by the finger in each frame of image;

2) Through the baseline drift correction algorithm based on the single-layer morphological opening and closing operation, the baseline drift correction of the signal is performed to obtain a one-dimensional relative grayscale-frame number matrix, the relative grayscale of the ordinate, and the number of frames of the abscissa, that is, after the baseline correction. signal of;

3) Determine the highest signal peak and lowest signal valley in each signal cycle of the baseline-corrected signal

3.1) Obtain the sampling frame rate f _s of the color camera, and use the sampling frame rate f _s to calculate the moving step step by the following formula

in,

step is the moving step, the unit is frame/second;

f _s is the sampling frame rate, in Hz;

3.2) Set a one-dimensional signal sequence f(n ₀ , n ₁ , n ₂ ... n _k ), where n _k is the number of frames, f(n _k ) is the grayscale value, traverse all elements of the signal sequence, and pass The following relationship is used to obtain the coordinate values of all signal peaks and signal valleys in each signal period of the baseline-corrected signal, and then calculate the frame number difference between adjacent signal peaks or adjacent signal valleys in each signal period;

Signal peak relationship:

f(n _k-step )<f(n _k )>f(n _k+step );

Signal valley relation:

f(n _k-step )>f(n _k )<f(n _k+step );

3.3) Calculate the minimum interval delta using the following formula

Among them, delta is the minimum interval, in frames per second;

3.4) Compare the frame number difference between adjacent signal peaks or adjacent signal valleys in each signal cycle obtained in step 3.2) with the minimum interval delta obtained in step 3.3), if the frame number of adjacent signal peaks or adjacent signal valleys is The difference is less than the minimum interval delta, indicating that the adjacent signal peaks or adjacent signal valleys are in the same signal cycle, then the signal peak with the largest gray value in the adjacent signal peaks is taken as the highest signal peak in the cycle, and the adjacent signal peaks are taken as the highest signal peak in the cycle. The signal valley with the smallest gray value in the valley is regarded as the lowest signal valley of the cycle; if the frame number difference between adjacent signal peaks or adjacent signal valleys is greater than or equal to the minimum interval delta, it indicates that the adjacent signal peaks or adjacent signal valleys are located in different positions. In the signal period, the two adjacent signal peaks are respectively regarded as the highest signal peaks of their respective signal periods, and the two adjacent signal valleys are respectively regarded as the lowest signal valleys of their respective signal periods;

4) Calculate heart rate

Using the abscissa corresponding to the highest signal peak of all signal cycles obtained in step 3.4), calculate the heart rate according to the following formula

in,

HR is heart rate, the unit is times/min;

n is the signal period, the value is 1,2,3...,n;

loc _n is the abscissa (that is, the number of frames) corresponding to the highest signal peak of the nth signal period;

f _s is the sampling frame rate, in Hz;

5) Calculate the breathing rate

Use a bandpass filter with a passband frequency of 0.1Hz-0.4Hz to perform bandpass filtering on the baseline-corrected signal obtained in step 2), and then perform Fourier transform on the filtered signal to obtain the amplitude of the bandpass filtered signal. - Frequency information and search for the frequency corresponding to the maximum amplitude, this frequency is the breathing frequency, and the breathing rate is calculated by the following formula

Res=60* _ploc

in,

Res is the respiratory rate, in times/min;

p _loc is the respiratory frequency, in Hz;

6) Calculate blood oxygen saturation based on color video images

6.1) Utilize the described human heart rate, respiration and blood oxygen comprehensive measurement device, adopt the calculation method of the average gray value in step 1), calculate the average gray value of each frame image R channel and G channel in the video respectively, Obtain the signals of the R channel and the G channel, calculate the AC volume after the logarithm of the R channel and the G channel signal, and then calculate the ratio R' of the AC volume of the R channel and the G channel, where the AC volume is defined as the highest in each signal cycle. The difference between the relative gray value of the signal peak and the lowest signal valley;

6.2) Using the ratio R' obtained in step 6.1) and the following non-linear blood oxygen saturation calculation formula that conforms to the change trend of human blood oxygen saturation, calculate the blood oxygen saturation

in,

SPO ₂ is blood oxygen saturation;

R' is the AC volume ratio after the logarithm of the signals obtained from the dual channels of R and G of the video image;

A, B, C and D are constants.

Further, in step 6.2), the calibration method of the nonlinear blood oxygen saturation calculation formula is as follows:

Use the existing oximeter to detect the multiple sets of blood oxygen saturation SPO ₂ of the test subject, and then use the multiple sets of blood oxygen saturation SPO ₂ and the ratio R' obtained in step 6.1) to minimize the nonlinear blood oxygen saturation calculation formula. A quadratic fit determines the values of parameters A, B, C, and D.

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

1. Compared with the existing finger-clip pulse oximeter, it will put pressure on the tester's fingertips and affect the authenticity of the signal. The single-point photodetector used has a low signal-to-noise ratio and can only achieve The detection of heart rate and blood oxygen has a single function, and does not display the blood volume pulse waveform of the tester. The device and method for comprehensive measurement of human heart rate, respiration and blood oxygen provided by the present invention use an LED polychromatic light source as the light source (such as red, The white light source formed by the combination of green and blue monochromatic light), the subject's fingertip is placed between the light source and the color camera, the color camera is used to collect the video signal, and the image processing algorithm is used to obtain the sense of each frame of the video signal. The region of interest (Region Of Interest ROI, that is, the target region is identified), and then the grayscale average value of the region is obtained to reflect the change of the subject's blood volume during this time period, and to improve the signal-to-noise of the obtained signal. There is no need to put pressure on the tester's fingertips, which can ensure the authenticity of the signal, and the function of respiration detection and the real blood volume pulse wave waveform display function of the test subject is also added, that is, the heart rate, respiration, blood oxygen and blood volume pulse wave can be realized at the same time. Comprehensive measurement of health parameters such as waveforms.

2. The present invention proposes a nonlinear measurement model for calculating blood oxygen saturation based on color video images, which reduces the requirement for the monochromaticity of the light source, and is more in line with the changing trend of human blood oxygen saturation compared with the existing linear models. , effectively improve the blood oxygen saturation measurement accuracy.

3. The size of the comprehensive measuring device for human heart rate, respiration and blood oxygen provided by the present invention (that is, the overall size formed by the casing being installed on the base) is 135mm×82mm×60mm, with a concise operation interface, easy to use and small in size, It can be widely used in home personal health monitoring and medical diagnosis.

4. The real blood volume pulse waveform of the subject displayed by the comprehensive measuring device for human heart rate, respiration and blood oxygen provided by the present invention can be used as a tool for Chinese medicine to observe the patient's pulse. Using the device of the present invention, the traditional "feeling the pulse" (the doctor places the index finger, middle finger and ring finger on the tester's radial artery to feel the pulse) can be transformed into a measurable "seeing the pulse", such as: the blood volume and the pulse wave drop in the middle The height of the gorge is closely related to the blood pressure of the test subject, and this change will help the development of traditional Chinese medicine and go to the world.

Description of drawings

1 is a cross-sectional view of a comprehensive measuring device for human heart rate, respiration and blood oxygen according to the present invention;

FIG. 2 is a schematic structural diagram of a light source camera fixture in FIG. 1;

Fig. 3 is the flow chart of single-layer morphological baseline drift correction in step 3) of the comprehensive measurement method of human heart rate, respiration and blood oxygen of the present invention;

4 is a schematic diagram of structural elements in step 3) of the comprehensive measurement method for human heart rate, respiration and blood oxygen of the present invention in the correction of the baseline drift of the monolayer morphology;

Fig. 5 is the comparison diagram of the signal curve before and after the correction of the baseline drift of the monolayer morphology in step 3) of the comprehensive measurement method of human heart rate, respiration and blood oxygen of the present invention, a is before correction (with baseline drift), b is correction back;

Fig. 6 is the initial startup system operation interface of the capacitive display screen of the comprehensive measuring device for human heart rate, respiration and blood oxygen according to the present invention;

FIG. 7 is the measurement result displayed on the capacitive display screen after one measurement by the comprehensive measuring device for human heart rate, respiration and blood oxygen according to the present invention.

Description of reference numbers:

1-shell, 11-display installation slot, 12-fixture installation slot, 13-camera installation screw hole, 14-light source outlet hole, 15-fixture installation screw hole, 2-base, 3-capacitive display screen, 4-LED polychromatic light source, 5-color camera, 6-light source camera fixture, 61-light source mounting hole, 62-camera mounting hole, 7-development board, 8-finger.

Detailed ways

The present invention will be further described below with reference to the accompanying drawings.

The present invention provides a comprehensive measurement device for human heart rate, respiration and blood oxygen, as shown in Figures 1 and 2, comprising a casing 1 (that is, a structure for supporting and protecting internal electronic devices), a base 2, a capacitor Display screen 3, LED polychromatic light source (LED composite color light source, that is, a variety of monochromatic light composite light sources), color camera 5 (for recording color video), light source camera fixture 6, development board 7 and power adapter. The LED polychromatic light source 4 can use a white LED light source (that is, white light formed by the composite of three monochromatic lights of red, green and blue); the capacitive display screen 3 uses a 5-inch TFT capacitive display screen; the color camera 5. An area array CMOS sensor (also called area array CMOS, area array CMOS image sensor, and CMOS camera) is used. The area array CMOS sensor is preferably an area array OV5647-CMOS sensor; the development board 7 is a Raspberry Pi 4B development board. There is a band-pass filter on the development board 7, and the band-pass filter can be a 3rd-order Butterworth band-pass filter; the power adapter can be a Raspberry Pi 4B power adapter.

The housing 1 is arranged on the base 2, the top of the housing 1 is provided with a display screen installation slot 11, and the side of the housing 1 is provided with a fixing member installation slot 12; the capacitive display screen 3 is installed in the display screen installation slot 11, so The light source camera fixing member 6 is installed in the fixing member installation groove 12, or a fixing member mounting screw hole 15 can be opened on the light source camera fixing member 6, and the light source camera fixing member 6 is fixed in the housing 1 by screws; The bottom of the camera fixing member 6 is provided with a light source mounting hole 61, and the top of the light source camera fixing member 6 is provided with a camera mounting hole 62; the LED polychromatic light source 4 is installed at the light source mounting hole 61, and the top of the LED polychromatic light source 4 is connected to the light source mounting hole. 61 holes are flush; the color camera 5 (through the camera mounting screw hole 13, the aperture is Φ2.5mm) is installed at the camera mounting hole 62; the color camera 5 receives the outgoing light of the LED polychromatic light source 4; the development board 7 is located inside the housing 1; the light source camera fixing member 6 is provided with a light source outlet hole 14, and the development board 7 is connected to the LED polychromatic light source 4 through a wire (the wire passes through the light source outlet hole 14, and the positive pole of the LED polychromatic light source 4 is connected to the raspberry. The (BCM) GPIO17 pin of the I/O pin is connected to the GPIO17 pin, and the negative pole is connected to 0V), and the color camera 5 and the capacitive display screen 3 are respectively connected through the OV5647 data cable (that is, the capacitive display screen 3 is connected to the raspberry through the video data cable. The development board 7 sends out control signals to control the LED polychromatic light source 4, the color camera 5 and the capacitive display screen 3 respectively; the development board 7 is connected to an external power supply through a power adapter. The connection of the components in the device can be connected by screws (the diameter of the threaded hole is Φ3mm). The overall size formed by the housing 1 installed on the base 2 is 135mm×82mm×60mm.

The dynamic process when the device is running:

When the above device is powered on, the system automatically loads and starts the software program, and displays the operation interface of the software on the capacitive display screen 3. After the start, the operation interface is shown in Figure 6. The subject placed his finger 8 on the upper end of the light source hole on one side of the device, and completely covered the light source hole with his fingertip, and then clicked the "Start" button on the operation interface of the capacitive display screen. At this time, the LED polychromatic light source 4 was clicked. light, the color camera 5 (CMOS camera) starts to collect video signals, and the system starts to measure the user's health parameters. During the measurement process, the user needs to keep his finger 8 still as much as possible, and the algorithm program on the development board 7 starts to process the video signal in real time. The default acquisition time of the set program (for example, 10 seconds), after the acquisition time ends, the LED polychromatic light source 4 is automatically turned off, the color camera 5 signal acquisition stops (that is, the video recording is stopped), and the capacitive display screen operation interface displays the corresponding measurement results ( Including heart rate, respiration, blood oxygen and blood volume pulse wave waveform), as shown in Figure 7 is the operation interface after one measurement, the person to be tested can click the "Continue" button again to carry out the second test (the operation interface is simple, use The user only needs to click the "Start" button when the limb is in a stable state to complete the measurement operation).

When healthy adults are not exercising and are emotionally stable, their heart rate data is generally 60-90 beats/min, and their breathing is 16-20 beats/min. The baby's heart rate will be higher than ordinary people, at 120-160 beats/min. min, breathing 40-60 breaths/min. The signal is collected by a CMOS camera, and the sampling frequency is fixed at about 30Hz, which satisfies the Nyquist sampling law.

The method for comprehensively measuring human heart rate, respiration and blood oxygen using the above-mentioned comprehensive measuring device for human heart rate, respiration and blood oxygen includes the following steps:

1) The device starts to work, and the CMOS camera obtains an instruction to start recording video, and previews the picture in real time (that is, the recorded video picture of finger 8 is displayed on the upper right corner of the operation interface of the capacitive display screen 3); Collect (read) video data, identify (calibrate) the target area for each frame of image in the video data, calculate the average gray value of the target area, and obtain a gray-frame number matrix, that is, a one-dimensional PPG signal (one of which is one-dimensional PPG signal). The frame image identifies a target area and calculates an average gray value. All the video frames collected by the CMOS camera together form the signal sequence during this period. For example: the sampling frame rate of the CMOS camera is 30Hz, and the sampling rate is 10 seconds, then A total of 300 frames of images are obtained, and each frame of image obtains an average gray value, then a one-dimensional grayscale-frame number matrix is obtained in this 10s time period, that is, a one-dimensional signal sequence. The target area is the image in each frame. The area covered by finger 8); the target area is the area covered by finger 8; the average gray value is defined as the sum of all pixel gray values in the target area divided by the total number of pixels in the target area;

2) Since there is an obvious baseline drift in the collected PPG original signal, as shown in Figure 3, here, through the existing baseline drift correction algorithm based on the single-layer morphological opening and closing operation, the original signal is opened and closed once. The baseline of the signal is obtained by the closing operation, and the baseline drift correction is performed on the signal by subtracting the baseline from the original signal to obtain a one-dimensional relative grayscale-frame number matrix, the relative grayscale of the ordinate, and the number of frames of the abscissa, that is, the signal after baseline correction;

The morphological open operation means that the signal is first eroded and then expanded, and the closed operation means that the signal is first expanded and then eroded.

Let the one-dimensional signal sequence be f(n)={0,1,2,3...N-1}, where f(n) refers to the one-dimensional signal sequence of length N, and the value of n ranges from 0 to N- 1. Example: f(n)={0,1,2,3,4} represents a discrete digital signal of length 5, then f(0)=0, f(1)=1, f(2)= 2, f(2)=3, f(4)=4; the structural element is k(m)={0,1,2,3...M-1}, k(m) refers to the structure of width M Element, the value of m ranges from 0 to M-1, and M is generally slightly less than the length of the signal to be reserved. In the present invention, the length of the signal to be reserved is 30, so M=28, where {0, 1, 2, 3 ...M-1} represents the height and shape of the structural element. For example, the structural element of the present invention is a linear structural element with a width of 28 and a height of 1, which can be expressed as k(m)={1,1,1,1, 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1}, where k(0)=1, k(1)=1, k(2)=1, k(3)=1...k(27)=1, as shown in FIG. 4 .

Then the expansion operation of the signal f(n) with respect to the structuring element k(m) is defined as:

Then the corrosion of the signal f(n) with respect to the structuring element k(m) is defined as:

闭操作定义为

基线校正前、后的效果对比如图5所示；The opening operation is defined as eroding and then dilating a one-dimensional signal:

The closing operation is defined as

The comparison of the effect before and after baseline correction is shown in Figure 5;

3) Determine the highest signal peak and lowest signal valley in each signal cycle of the baseline-corrected signal

3.1) Obtain the sampling frame rate f _s of the color camera 5, and use the sampling frame rate f _s to calculate the moving step step by the following formula

in,

step is the moving step, the unit is frame/second;

f _s is the sampling frame rate, in Hz;

3.2) Set a one-dimensional signal sequence f(n ₀ , n ₁ , n ₂ ... n _k ), where n _k is the number of frames, f(n _k ) is the grayscale value, traverse all elements of the signal sequence, and pass The following relationship is used to obtain the coordinate values of all signal peaks and signal valleys in each signal period of the baseline-corrected signal, and then calculate the frame number difference between adjacent signal peaks or adjacent signal valleys in each signal period;

Signal peak relationship:

f(n _k-step )<f(n _k )>f(n _k+step );

Signal valley relation:

f(n _k-step )>f(n _k )<f(n _k+step );

3.3) Calculate the minimum interval delta using the following formula

Among them, delta is the minimum interval, in frames per second;

3.4) Compare the frame number difference between adjacent signal peaks or adjacent signal valleys in each signal cycle obtained in step 3.2) with the minimum interval delta obtained in step 3.3), if the frame number of adjacent signal peaks or adjacent signal valleys is The difference is less than the minimum interval delta, indicating that the adjacent signal peaks or adjacent signal valleys are in the same signal cycle, then the signal peak with the largest gray value in the adjacent signal peaks is taken as the highest signal peak in the cycle, and the adjacent signal peaks are taken as the highest signal peak in the cycle. The signal valley with the smallest gray value in the valley is regarded as the lowest signal valley of the cycle; if the frame number difference between adjacent signal peaks or adjacent signal valleys is greater than or equal to the minimum interval delta, it indicates that the adjacent signal peaks or adjacent signal valleys are located in different positions. In the signal period, the two adjacent signal peaks are respectively regarded as the highest signal peaks of their respective signal periods, and the two adjacent signal valleys are respectively regarded as the lowest signal valleys of their respective signal periods;

4) Calculate heart rate

Using the abscissa corresponding to the highest signal peak of all signal cycles obtained in step 3.4), since the time interval between two peaks can be equivalent to the cycle of a single PPG signal, the heart rate is calculated according to the following formula

in,

HR is heart rate, the unit is times/min;

n is the signal period, the value is 1,2,3...,n;

loc _n is the abscissa (that is, the number of frames) corresponding to the highest signal peak of the nth signal period;

f _s is the sampling frame rate, in Hz;

5) Calculate the breathing rate

Use a bandpass filter with a passband frequency of 0.1Hz-0.4Hz (preferably a 3rd-order Butterworth bandpass filter) to perform bandpass filtering on the baseline-corrected signal obtained in step 2), and then perform Fourier transform on the filtered matrix. Lie transform obtains the amplitude-frequency information of the band-pass filtered signal and searches for the frequency corresponding to the maximum amplitude, which is the breathing frequency, and the breathing rate is calculated by the following formula

Res=60* _ploc

in,

Res is the respiratory rate, in times/min;

p _loc is the respiratory frequency, in Hz;

6) Calculate blood oxygen saturation based on color video images

Current (i.e. prior art) pulse oximeters generally use a dual-wavelength linear model to calculate blood oxygen saturation. This dual-wavelength linear model is a simplified model, and it has been proved that the error is large when the blood oxygen is low, and its The monochromaticity of the incident light with two wavelengths is required to be high. The commonly used dual-wavelength combination is 660nm and 940nm two wavelengths. After using these two wavelengths of light for time-division driving to obtain dual-channel pulse signals, extract the AC value AC and DC value DC respectively, and obtain the ratio R,

R=(AC _λ660 /DC _λ660 )/(AC _λ940 /DC _λ940 )

The formula for calculating blood oxygen saturation is

SPO ₂ =k ₁ R+k ₂

k ₁ and k ₂ are constants obtained by performing least squares fitting on the actual blood oxygen value and its corresponding R value.

The present invention proposes a nonlinear blood oxygen saturation calibration method, the details are as follows

6.1) Since the signal discrimination of the image in the R channel and the G channel is better, using the comprehensive measurement device for human heart rate, respiration and blood oxygen, using the calculation method of the average gray value in step 1), respectively Calculate the average gray value of the R channel and G channel of the frame image, obtain the signals of the R channel and the G channel, calculate the logarithm of the signal of the R channel and the G channel, and then calculate the R channel and the G channel. Ratio R', where the amount of AC is defined as the difference between the relative gray values of the highest signal peak and the lowest signal valley in each signal cycle;

6.2) Using the ratio R' obtained in step 6.1) and the following non-linear blood oxygen saturation calculation formula that conforms to the change trend of human blood oxygen saturation, calculate the blood oxygen saturation

in,

SPO ₂ is blood oxygen saturation;

R' is the AC volume ratio after the logarithm of the signals obtained from the dual channels of R and G of the video image;

A, B, C and D are constants.

In step 6.2), the calibration method of the nonlinear blood oxygen saturation calculation formula is as follows:

Use the existing oximeter to detect the multiple sets of blood oxygen saturation SPO ₂ of the test subject, and then use the multiple sets of blood oxygen saturation SPO ₂ and the ratio R' obtained in step 6.1) to minimize the nonlinear blood oxygen saturation calculation formula. A quadratic fit determines the values of parameters A, B, C, and D.

The derivation process of the above formula is as follows:

According to Lamber-Beer (Lamber-Beer law), we get

in,

I ₀ is the incident light intensity, the unit is lux (lux or lx);

ε ₀ is the total molar extinction coefficient of the non-pulsatile components and venous blood components in the tissue, the unit is L·mol ^-1 ·cm ^-1 , representing the absorption coefficient when the concentration is 1 mol/L;

C ₀ is the total concentration of non-pulsatile components and venous blood components in the tissue, in mol/L;

l is the optical path of light in non-pulsating tissue, in meters (m);

I is the intensity of light emitted through the tissue, in lux (lux or lx);

为氧合血红蛋白的摩尔消光系数，单位为L·mol^-1·cm^-1；

is the molar extinction coefficient of oxyhemoglobin, in L·mol ^-1 ·cm ^-1 ;

ε _Hb is the molar extinction coefficient of deoxyhemoglobin, in L·mol ^-1 ·cm ^-1 ;

为脉动血液中氧合血红蛋白的浓度，单位为mol/L；

is the concentration of oxyhemoglobin in pulsatile blood, in mol/L;

C _Hb is the concentration of deoxyhemoglobin in pulsatile blood, in mol/L;

L _AC is the optical path of light in the tissue pulsation component, in meters (m);

Take the logarithm of both sides of equation (1):

为交流成分,则单色光入射时出射光的交流成分可以表示为：In the above formula, In(I ₀ ) and -ε ₀ C ₀ l are DC components,

is the AC component, then the AC component of the outgoing light when the monochromatic light is incident can be expressed as:

The AC component obtained when irradiated with dual wavelengths is expressed as:

Among them, λ ₁ and λ ₂ represent different wavelengths;

Compare the AC signals at their respective wavelengths:

Further derivation can be obtained:

Then according to the definition formula of blood oxygen saturation:

Further derivation can be obtained:

Substituting Equation (7) into Equation (9), the formula for calculating blood oxygen saturation under dual wavelengths can be obtained:

令

得：Divide the right side of the above equation by

make

have to:

Since a color camera (color camera) generally contains three channels of information RGB, and each channel has a certain spectral response width, the signal obtained in each channel is essentially the result of the superposition of multiple wavelengths. It is impossible to obtain a signal of a certain wavelength by using a color camera, but since the tissue structure of each person's finger is basically similar, and its optical properties are basically the same, the equivalent extinction coefficient under complex light can be obtained by least squares fitting;

则得到复色光下血氧饱和度计算的经验公式：make

Then the empirical formula for calculating blood oxygen saturation under complex light is obtained:

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, but not to limit them. For those of ordinary skill in the art, the specific technical solutions recorded in the foregoing embodiments can be modified. , or equivalently replace some of the technical features, and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions protected by the present invention.

Claims (2)

1. A method for comprehensively measuring human heart rate, respiration and blood oxygen using a human heart rate, respiration and blood oxygen comprehensive measuring device, the human heart rate, respiration and blood oxygen comprehensive measuring device comprising a casing (1), a base (2) ), capacitive display screen (3), LED polychromatic light source (4), color camera (5), light source camera fixture (6), development board (7) and power adapter; The casing (1) is arranged on the base (2), a display screen installation slot (11) is formed on the top of the casing (1), and a fixing member installation groove (12) is formed on the side of the casing (1); The capacitive display screen (3) is installed in the display screen installation groove (11), and the light source camera fixing member (6) is installed in the fixing member installation groove (12); A light source mounting hole (61) is provided at the bottom of the light source camera fixture (6), and a camera mounting hole (62) is provided on the top of the light source camera fixture (6); The LED polychromatic light source (4) is installed at the light source installation hole (61), the color camera (5) is installed at the camera installation hole (62), and the color camera (5) receives the LED polychromatic light source (4). outgoing light; The development board (7) is located inside the housing (1), and the development board (7) is connected to the LED polychromatic light source (4) through wires, and is respectively connected to the color camera (5) and the capacitive display screen (3) through data cables, The development board (7) sends out control signals to respectively control the LED polychromatic light source (4), the color camera (5) and the capacitive display screen (3); The development board (7) is connected to an external power supply through a power adapter; It is characterized in that, comprises the following steps: 1) The development board (7) collects video data from the color camera (5), identifies the target area for each frame of the video data, calculates the average gray value of the target area, and obtains a gray-frame number matrix, namely One-dimensional PPG signal; the target area is the area covered by the finger (8) in each frame of image; 2) Through the baseline drift correction algorithm based on the single-layer morphological opening and closing operation, the baseline drift correction of the signal is performed to obtain a one-dimensional relative grayscale-frame number matrix, the relative grayscale of the ordinate, and the number of frames of the abscissa, that is, after the baseline correction. signal of; 3) Determine the highest signal peak and lowest signal valley in each signal cycle of the baseline-corrected signal 3.1) Obtain the sampling frame rate f _s of the color camera (5), and use the sampling frame rate f _s to calculate the moving step step by the following formula

in, step is the moving step, the unit is frame/second; f _s is the sampling frame rate, in Hz; 3.2) Set a one-dimensional signal sequence f(n ₀ , n ₁ , n ₂ ... n _k ), where n _k is the number of frames, f(n _k ) is the grayscale value, traverse all elements of the signal sequence, and pass The following relationship is used to obtain the coordinate values of all signal peaks and signal valleys in each signal period of the baseline-corrected signal, and then calculate the frame number difference between adjacent signal peaks or adjacent signal valleys in each signal period; Signal peak relationship: f(n _k-step )<f(n _k )>f(n _k+step ); Signal valley relation: f(n _k-step )>f(n _k )<f(n _k+step ); 3.3) Calculate the minimum interval delta using the following formula

Among them, delta is the minimum interval, in frames per second; 3.4) Compare the frame number difference between adjacent signal peaks or adjacent signal valleys in each signal cycle obtained in step 3.2) with the minimum interval delta obtained in step 3.3), if the frame number of adjacent signal peaks or adjacent signal valleys is The difference is less than the minimum interval delta, indicating that the adjacent signal peaks or adjacent signal valleys are in the same signal cycle, then the signal peak with the largest gray value in the adjacent signal peaks is taken as the highest signal peak in the cycle, and the adjacent signal peaks are taken as the highest signal peak in the cycle. The signal valley with the smallest gray value in the valley is regarded as the lowest signal valley of the cycle; if the frame number difference between adjacent signal peaks or adjacent signal valleys is greater than or equal to the minimum interval delta, it indicates that the adjacent signal peaks or adjacent signal valleys are located in different positions. In the signal period, the two adjacent signal peaks are respectively regarded as the highest signal peaks of their respective signal periods, and the two adjacent signal valleys are respectively regarded as the lowest signal valleys of their respective signal periods; 4) Calculate heart rate Using the abscissa corresponding to the highest signal peak of all signal cycles obtained in step 3.4), calculate the heart rate according to the following formula

in, HR is heart rate, the unit is times/min; n is the signal period, the value is 1,2,3...,n; loc _n is the abscissa (that is, the number of frames) corresponding to the highest signal peak of the nth signal period; f _s is the sampling frame rate, in Hz; 5) Calculate the breathing rate Use a bandpass filter with a passband frequency of 0.1Hz-0.4Hz to perform bandpass filtering on the baseline-corrected signal obtained in step 2), and then perform Fourier transform on the filtered signal to obtain the amplitude of the bandpass filtered signal. - Frequency information and search for the frequency corresponding to the maximum amplitude, this frequency is the breathing frequency, and the breathing rate is calculated by the following formula Res=60* _ploc in, Res is the respiratory rate, in times/min; p _loc is the respiratory frequency, in Hz; 6) Calculate blood oxygen saturation based on color video images 6.1) Utilize the described human heart rate, respiration and blood oxygen comprehensive measurement device, adopt the calculation method of the average gray value in step 1), calculate the average gray value of each frame image R channel and G channel in the video respectively, Obtain the signals of the R channel and the G channel, calculate the AC volume after the logarithm of the R channel and the G channel signal, and then calculate the ratio R' of the AC volume of the R channel and the G channel, where the AC volume is defined as the highest in each signal cycle. The difference between the relative gray value of the signal peak and the lowest signal valley; 6.2) Using the ratio R' obtained in step 6.1) and the following non-linear blood oxygen saturation calculation formula that conforms to the change trend of human blood oxygen saturation, calculate the blood oxygen saturation

in, SPO ₂ is blood oxygen saturation; R' is the AC volume ratio after the logarithm of the signals obtained from the dual channels of R and G of the video image; A, B, C and D are constants. 2. the method for comprehensive measurement of human heart rate, respiration and blood oxygen according to claim 1, is characterized in that, in step 6.2), the calibration method of described non-linear blood oxygen saturation calculation formula is as follows: Use the existing oximeter to detect the multiple sets of blood oxygen saturation SPO ₂ of the test subject, and then use the multiple sets of blood oxygen saturation SPO ₂ and the ratio R' obtained in step 6.1) to minimize the nonlinear blood oxygen saturation calculation formula. A quadratic fit determines the values of parameters A, B, C, and D.

Images