JP2012019967A - Noise removing method and pulse photometer - Google Patents

Abstract

Since there are artifacts due to body movement, veins, tissue, etc., a noise removal method that rotates by φ based on each unique artifact and minimizes each unique artifact, and A pulse photometer to which the noise removal method is applied is realized.
A step of irradiating a living tissue with light of two different wavelengths, a step of converting and receiving light of each wavelength transmitted or reflected through the living tissue into an electric signal, and an electric signal of each wavelength A step of rotating the obtained discrete time series pulse wave data around an average value of each pulse wave data at an angle peculiar to a specific artifact by using a rotation matrix. A method for removing noise contained in pulse wave data.
[Selection] Figure 3

Description

  The present invention relates to signal processing for extracting two common signals extracted from a single medium at the same time to extract a common signal component, and more particularly, a pulse used in the medical field, particularly for diagnosis of the circulatory system. The present invention relates to improvement of signal processing in a photometer.

Various methods have been proposed for separating a signal component and a noise component from two signals extracted almost simultaneously from one medium.
In general, they are processed in the frequency domain or the time domain.
Even in the medical field, a device that measures the pulse waveform and pulse rate, which is called a photoelectric pulse wave meter, a concentration measurement of light-absorbing substances in blood, a device for measuring oxygen saturation SpO2, carbon monoxide hemoglobin, Met hemoglobin, etc. A device for measuring the concentration of special hemoglobin, a device for measuring the concentration of injected dye, and the like are known as pulse photometers.
In particular, the oxygen saturation SpO2 measuring device is called a pulse oximeter.

The principle of the pulse photometer is that pulse wave data obtained by transmitting or reflecting light of multiple wavelengths with different absorption to the target substance to the living tissue and continuously measuring the amount of transmitted or reflected light. The concentration of the target substance is obtained from the signal.
If noise is mixed in the pulse wave data, the correct concentration cannot be calculated, and there is a risk of mishandling.
Conventionally, in a pulse photometer, in order to reduce noise, a method of dividing a frequency band and paying attention to a signal component or taking a correlation between two signals has been proposed.
However, these methods have a problem that analysis takes time.

Therefore, in the patent No. 32701717 (Patent Document 1), the applicant of the present invention irradiates the biological tissue with light of two different wavelengths and expresses the magnitudes of the two pulse wave signals obtained from the transmitted light on the vertical axis, It has been proposed that a graph is drawn on the horizontal axis, the regression line is obtained, and the oxygen saturation or light-absorbing substance concentration in the arterial blood is obtained based on the slope of the regression line.
According to the present invention, measurement accuracy can be improved and power consumption can be reduced.
However, in order to obtain the regression line or its inclination using a large amount of sampling data for pulse wave signals of each wavelength, a lot of calculation processing is still required.

  In addition, although the present applicant uses frequency analysis in Japanese Patent No. 3924636 (Patent Document 2), the pulse wave signal itself is not extracted in the analysis instead of extracting the pulse wave signal itself as in the prior art. In order to obtain the fundamental frequency of a signal and further improve accuracy, a method of filtering a pulse wave signal using a filter using the harmonic frequency is proposed.

  Further, as a further improvement measure regarding the point of obtaining the fundamental frequency, the present applicant processes two common signals extracted from the same medium almost simultaneously in Japanese Patent No. 4352315 (Patent Document 3). The signal processing method which reduced the calculation processing burden which extracts the signal component of is provided.

Next, the contents of Patent Document 3 will be described.
In describing the embodiment of the invention described in Patent Document 3, the principle will be described using a pulse oximeter that measures arterial blood oxygen saturation as an example.
The technique of the invention described in Patent Document 3 is not limited to a pulse oximeter, and pulse photometry is performed on a blood absorption substance such as special hemoglobin (carbon monoxide hemoglobin, Met hemoglobin, etc.) or a dye injected into blood. It can be applied to a device (pulse photometer) that uses the principle of

The configuration of the pulse oximeter for measuring the arterial oxygen saturation is as shown in FIG. 8, which is a schematic configuration block diagram.
The light-emitting elements 1 and 2 that emit light of different wavelengths are driven by the drive circuit 3 so as to emit light alternately.
Light used for the light emitting elements 1 and 2 is infrared light (eg, 940 [nm]) that is less affected by arterial oxygen saturation, and red light (eg, 660 [nm]) that is highly sensitive to changes in arterial oxygen saturation. Good.

Light emitted from the light emitting elements 1 and 2 is transmitted through the living tissue 4, received by the photodiode 5, and converted into an electrical signal.
The reflected light may be received.
These converted signals are amplified by the amplifier 6 and distributed by the multiplexer 7 to the filters 8-1 and 8-2 corresponding to the respective optical wavelengths.
The signals distributed to the filters are filtered by the filters 8-1 and 8-2 to reduce noise components, and are digitized by the A / D converter 9.

Each signal sequence corresponding to digitized infrared light and red light forms a respective pulse wave signal.
Each digitized signal sequence is input to the processing unit 10, processed by a program stored in the ROM 12, the oxygen saturation SpO 2 is measured, and the value is displayed on the display unit 11.

<Noise reduction by rotation matrix and calculation of fundamental frequency of pulse wave>
First, measurement of the change in absorbance (light attenuation) of a light-absorbing substance in blood will be described.
14A and 14B show pulse wave data in which the light emitted from the light emitting elements 1 and 2 is transmitted through the living tissue 4 and received by the photodiode 5 and converted into an electrical signal. a) shows red light, and (b) shows infrared light.
In FIG. 14A, when the horizontal axis is time and the vertical axis is the received light output, the received light output from the photodiode 5 is superimposed with the direct current component (R ′) and the pulsating component (ΔR ′) of red light. It has a waveform.
Further, in FIG. 14B, when the horizontal axis is time and the vertical axis is the light reception output, the light reception output by the photodiode 5 includes the direct current component (IR ′) and the pulsation component (ΔIR ′) of infrared light. The waveform is superimposed.
FIG. 9 shows the ratio (IR = ΔIR ′ / IR ′) of the pulsating components (ΔR ′, ΔIR ′) to the DC components (R ′, IR ′) for 8 seconds in the pulse wave as shown in FIG. In addition, the average value of the data for 8 seconds is set to zero.
Note that the calculation is possible without performing the process of setting the average value to zero as shown in FIG.
FIG. 10 is a graph in which the infrared light IR data shown in FIG. 14 is plotted on the horizontal axis and the red light R data is plotted on the vertical axis.

Next, calculation processing for reducing noise using two rotation wave data signals of two wavelengths digitized by the A / D converter 9 will be described.
In addition, since infrared light and red light are alternately emitted, they are not emitted at the same time. However, adjacent infrared light reception values and red light reception values obtained at the same time were obtained. The pulse wave signal of infrared light and the pulse wave signal of red light for a predetermined time are developed on a two-dimensional orthogonal coordinate.
That is, the graph of FIG. 10 is created.
Further, by taking the ratio of the pulsating component to the DC component of the pulsating wave, the absorbance pulsation due to the pulse is approximated.
The transition seen in the graph of FIG. 10 is not 45 degrees because the amplitude of the pulsating component of the infrared light pulse wave is different from the amplitude of the pulsating component of the red light pulse wave, and noise. This is because is superimposed.

Next, a rotation calculation is performed on the developed pulse wave data using a rotation matrix.
A data string of the ratio (IR) of the pulsating component to the direct current component of the infrared light pulse wave,

  A data string of the ratio (R) of the pulsating component to the direct current component of the red light pulse wave is

And
Data of IR and R obtained at the same time ti is defined by a matrix as follows.
That is,

  Further, if the rotation matrix for rotating θ [rad] is A, A can be expressed as follows.

  Then, the following X is obtained by rotating S by θ [rad] by A.

  In addition to the above, the rotation matrix A is

May be used.
Here, a graph obtained by rotating the pulse wave data S by π / 30 [rad] from 0 to 9π / 30 [rad] is shown in FIG.
As can be seen from FIG. 11, the image is rotated around a point with zero horizontal axis and zero vertical axis (a point where both the red light pulse wave and the infrared light pulse wave are average), and θ is 9π / 30 [ rad], the region projected onto the horizontal axis (X1) is the smallest, and the region projected onto the vertical axis (X2) is the largest.
When θ is further rotated by π / 2 [rad] from 9π / 30 [rad] and rotated by 24π / 30 [rad] (= 12π / 15 [rad]), the area projected onto the horizontal axis (X1) is the maximum. It is clear that the region projected onto the vertical axis (X2) is the smallest.

Next, the measured pulse wave data S is processed into X by the rotation matrix A when θ is 9π / 30 [rad] and 24π / 30 [rad]. Explain how.
FIG. 12 shows a waveform of X obtained by processing the pulse wave data S shown in FIG. 9 by the rotation matrix A with θ being 9π / 30 [rad].
X1 (ti) where the area projected onto the horizontal axis is minimized is

  On the other hand, X2 (t i) where the area projected onto the horizontal axis is maximized is

Is calculated by
From the waveform of X1 in FIG. 12, it can be seen that noise has been removed.
When the pulse wave data S is processed by the rotation matrix A with θ being 24π / 30 [rad], the waveform of X2 is a waveform from which noise is removed.
X1 (ti) that maximizes the area projected on the horizontal axis is

  On the other hand, X2 (t i) that minimizes the area projected onto the horizontal axis is

Is calculated by
If the rotation angle θ is set so that the region projected onto the horizontal axis in this way is minimized and the pulse wave data S is processed, a pulse wave main component waveform with suppressed noise can be obtained.

Next, calculation of the fundamental frequency of the pulse wave will be described.
FIG. 13 shows the spectrum obtained by frequency analysis of the pulse wave signal shown in FIG. 9 before the noise is reduced and the pulse wave main component waveform whose noise is reduced using the rotation matrix.
The horizontal axis is frequency and the vertical axis is spectrum.
The spectrum of the pulse wave (Before-rotation) signal before the noise is reduced has a strong spectrum in the frequency band fn of the noise, and the spectrum of the fundamental frequency fs of the pulse wave signal hardly appears.
On the other hand, in the spectrum obtained by frequency analysis of a pulse wave main component waveform (after-rotation) in which noise is reduced using a rotation matrix, the spectrum of the fundamental frequency fs of the pulse wave signal is the spectrum of the noise frequency band fn. The fundamental frequency fs of the pulse wave signal can be obtained.
If the fundamental frequency fs [Hz] of the pulse wave signal is obtained, the pulse rate fs × 60 [times / min] can be easily obtained.

Thus, by using a rotation matrix of a predetermined angle, a pulse wave main component waveform with reduced noise can be obtained, and the fundamental frequency or pulse rate of the pulse wave signal can be obtained.
Here, the predetermined angle may be determined in advance or may be changed adaptively during the measurement period.

<Calculation of oxygen saturation>
FIG. 10 is a graph in which the infrared light IR data shown in FIG. 14 is plotted on the horizontal axis and the red light R data is plotted on the vertical axis, as described above. The slope of this graph is calculated using the norm ratio. Ask.
First, the L2 norm of the infrared light pulse wave data IR is obtained.
Since the infrared light pulse wave data string is IR = [IR (ti): ti = 0, 1, 2, 3,...] (1), the L2 norm can be expressed by the following equation.

Next, the L2 norm of the red light pulse wave data R is obtained.
Since the red light pulse wave data string is R = [R (ti): ti = 0, 1, 2, 3,...] (2), the L2 norm can be expressed by the following equation.

Therefore,

  Then, Φ correlates with the oxygen saturation SpO2, so if the function representing the correlation is f,

The oxygen saturation SpO2 can be obtained.
Φ is a ratio as described in Equation 12, but the actual rotation angle φ takes an arc tangent (ATN).

Japanese Patent No. 32701717 Japanese Patent No. 3924636 Japanese Patent No. 4352315

A pulse photometer to which the signal processing method described in Patent Document 3 is applied,
“Pulse photometer comprising light emitting means for irradiating a living tissue with light of two different wavelengths and light receiving means for converting light of each wavelength generated from the light emitting means and transmitted or reflected by the living tissue into an electrical signal The discrete time-series pulse wave data obtained from the electrical signals of the respective wavelengths are rotated around a mean value of each pulse wave data at a predetermined angle or an angle determined based on a predetermined condition. A pulse photometer comprising: a first processing unit that removes noise included in the discrete time-series pulse wave data using a rotation matrix. ”, Wherein artifacts are reduced most. It is rotated at a rotation angle.

  The problem of the present invention is that there are artifacts due to body movement, veins, tissue, etc., so that each artifact can be selectively rotated by selectively rotating by φ based on each unique artifact. And a pulse photometer to which the noise removal method is applied.

  In order to solve the above-described problems, a noise removal method of the present invention includes a step of irradiating a living tissue with light of two different wavelengths, and receiving light by converting each wavelength of light transmitted or reflected through the living tissue into an electrical signal. And rotating the discrete time-series pulse wave data obtained from the electrical signals of the respective wavelengths around the average value of each pulse wave data at an angle peculiar to a specific artifact using a rotation matrix; It is characterized by including.

In addition, the angle unique to the specific artifact is different from each other, and includes a step of rotating around the average value of each pulse wave data over a plurality of times at different angles. .
In addition, the angles unique to the specific artifact are respectively angles specific to body motion, vein, and tissue.

In order to solve the problem, the pulse photometer of the present invention includes a light emitting means for irradiating a living tissue with light of two different wavelengths, and light of each wavelength generated from the light emitting means and transmitted or reflected by the living tissue. In a pulse photometer equipped with a light receiving means for converting the signal into an electrical signal,
The discrete time series pulse wave data obtained from the electrical signal of each wavelength is rotated around the average value of each pulse wave data at an angle peculiar to a specific artifact using a rotation matrix and the discrete time series And noise processing means for removing a specific noise component included in the pulse wave data.

  According to the present invention, there are artifacts due to body movements, veins, tissue, etc., so that each artifact can be selectively rotated by selectively rotating by φ based on each unique artifact. And a pulse photometer to which the noise removal method is applied can be realized.

4 is a graph of discrete time-series pulse wave data in which infrared light IR and R data, which is an example for explaining the present invention, is represented by time on the horizontal axis and magnitude on the vertical axis. It is a spectrum figure which shows the result of having carried out FFT processing of the discrete time series pulse wave data of FIG. FIG. 2 is a spectrum diagram of a result obtained by rotating the discrete time-series pulse wave data (IR) of FIG. 1 around an average value of each pulse wave data at an angle φ = 0.6 using a rotation matrix. FIG. 2 is a spectrum diagram of a result obtained by rotating the discrete time-series pulse wave data (IR) of FIG. 1 around an average value of each pulse wave data at an angle φ = 0.7 using a rotation matrix. FIG. 2 is a spectrum diagram of a result obtained by rotating the discrete time-series pulse wave data (IR) of FIG. 1 around an average value of each pulse wave data at an angle φ = 0.8 using a rotation matrix. FIG. 2 is a spectrum diagram of a result obtained by rotating the discrete time-series pulse wave data (IR) of FIG. 1 around an average value of each pulse wave data at an angle φ = 0.9 using a rotation matrix. FIG. 2 is a spectrum diagram of a result obtained by rotating the discrete time series pulse wave data (IR) of FIG. 1 around an average value of each pulse wave data at an angle φ = 1.0 using a rotation matrix. It is a block diagram which shows schematic structure of a pulse photometer. It is a figure which shows the detected pulse wave. FIG. 15 is a graph in which infrared light IR data shown in FIG. 14 is plotted on the horizontal axis and red light R data is plotted on the vertical axis. It is the figure which rotated the graph of FIG. 10 every (pi) / 30 [rad]. It is a figure which shows the waveform of the pulse wave processed by the rotation matrix of rotation angle 9π / 30 [rad]. It is a figure which shows the spectrum of the waveform of X1 shown in FIG. It is a wave form diagram explaining the measurement principle of the fluctuation | variation of the light absorbency of the light absorption substance in the blood.

  Next, referring to FIG. 1 to FIG. 7, the artifacts of the present invention include those due to body movement, those due to veins, those due to tissues, and so on. A noise removal method for minimizing inherent artifacts will be described.

FIG. 1 is a discrete time-series pulse wave in which data of infrared light IR (solid line) and R (marked by *), which is an example for explaining the present invention, is represented with time on the horizontal axis and magnitude on the vertical axis. It is a graph of data.
This graph corresponds to the graph of FIG. 9 although the magnitude of noise (artifact) contained in the pulse wave is different.

FIG. 2 is a spectrum diagram showing the result of performing FFT processing on the discrete time-series pulse wave data of FIG. 1, and the solid line shows the IR spectrum and the broken line shows the R spectrum.
In FIG. 2, in order to understand the operation of the present invention more clearly, the pulse wave has φ = 0.6, φ = 0.7, φ = 0.8, φ = 0.9, and φ = 1.0. Artifacts equivalent to are included.
From FIG. 2, the presence of large artifact components in both IR and R can be seen near the frequencies corresponding to φ = 0.6, φ = 0.7, φ = 0.8, φ = 0.9, and φ = 1.0. .

FIG. 3 is a spectrum diagram of a result obtained by rotating the discrete time-series pulse wave data (IR) of FIG. 1 around the average value of each pulse wave data at an angle φ = 0.6 using a rotation matrix.
From FIG. 3, it can be understood that the artifacts in the vicinity of 5 Hz are almost removed, and other artifacts remain larger as the frequency is lower.

FIG. 4 is a spectrum diagram of the result of rotating the discrete time-series pulse wave data (IR) of FIG. 1 around the average value of each pulse wave data at an angle φ = 0.7 using a rotation matrix.
It can be understood from FIG. 4 that the artifacts near 4 Hz are almost eliminated, and that other artifacts remain larger as the frequency decreases at both sides of 4 Hz.

FIG. 5 is a spectrum diagram of the result of rotating the discrete time-series pulse wave data (IR) of FIG. 1 around the average value of each pulse wave data at an angle φ = 0.8 using a rotation matrix.
From FIG. 5, it can be understood that the artifacts in the vicinity of 3 Hz are almost removed, and the other artifacts remain large as they move away above and below 3 Hz.

FIG. 6 is a spectrum diagram of a result obtained by rotating the discrete time-series pulse wave data (IR) of FIG. 1 around the average value of each pulse wave data at an angle φ = 0.9 using a rotation matrix.
From FIG. 6, it can be understood that the artifacts near 2.5 Hz are almost removed, and other artifacts are reduced above and below 2.5 Hz, and the larger the distance remains.

FIG. 7 is a spectrum diagram of the result of rotating the discrete time-series pulse wave data (IR) of FIG. 1 around the average value of each pulse wave data at an angle φ = 1.0 using a rotation matrix.
From FIG. 7, it can be understood that artifacts in the vicinity of 2 Hz are almost removed, and other artifacts remain large as they move away from 2 Hz.

  In the description of FIGS. 3 to 7, an example in which φ = 0.6, φ = 0.7, φ = 0.8, φ = 0.9, and φ = 1.0 are separately rotated as the rotation angle φ. Although it is shown, it is possible to remove noise corresponding to multiple artifacts by rotating around the average value of each pulse wave data at different angles with respect to the result of rotating once. It is.

DESCRIPTION OF SYMBOLS 1 Light emitting element 2 Light emitting element 3 Drive circuit 4 Living body tissue 5 Photodiode 6 Converter 7 Multiplexer 8 Filter 9 A / D converter 10 Processing part 11 Display part 12 ROM
13 RAM

Claims (4)

Irradiating a living tissue with light of two different wavelengths;
Converting the light of each wavelength transmitted or reflected through the living tissue into an electrical signal and receiving it;
Rotating the discrete time-series pulse wave data obtained from the electrical signal of each wavelength around an average value of each pulse wave data at an angle specific to a specific artifact using a rotation matrix;
A method for removing noise contained in discrete time-series pulse wave data. The angle unique to the specific artifact is different from each other, and is rotated around the average value of each pulse wave data over a plurality of times at different angles.
The method for removing noise contained in discrete time-series pulse wave data according to claim 1, wherein: The angles specific to the specific artifact are angles specific to body movement, vein, and tissue, respectively.
The method for removing noise contained in discrete time-series pulse wave data according to claim 1 or 2. A pulse photometer comprising: a light emitting means for irradiating a living tissue with light of two different wavelengths; and a light receiving means for converting light of each wavelength generated from the light emitting means and transmitted or reflected by the living tissue into an electrical signal. ,
The discrete time series pulse wave data obtained from the electrical signal of each wavelength is rotated around the average value of each pulse wave data at an angle peculiar to a specific artifact using a rotation matrix and the discrete time series A noise processing means for removing a specific noise component included in the pulse wave data;
A pulse photometer characterized by comprising: