CN106580276A - Pulse wave conduction time acquisition method based on correlation - Google Patents

Abstract

The invention relates to a pulse wave conduction time acquisition method based on correlation. The pulse wave conduction time acquisition method based on correlation includes the steps: utilizing an electrode slice and a photoelectric pulse wave sensor to synchronously acquire a human body ECG (electrocardiograph) signal and a photoelectric pulse wave signal; determining the correlation coefficient formula of the data of the human body ECG signal and the data of the photoelectric pulse wave signal in a set length; in a time domain, moving the photoelectric pulse wave signal to the left with offset of one unit, and utilizing the correlation coefficient formula to compute the correlation coefficient of ECG and PPG after offset; repeatedly computing the correlation coefficient of the ECG and the PPG until obtaining m correlation coefficients; and selecting the maximum correlation coefficient among the m correlation coefficients, and using the position number I to represent the offset, moved by the photoelectric pulse wave, corresponding to the correlation coefficient so as to obtain the pulse wave conduction time PTT = i*0.5*103. The pulse wave conduction time acquisition method based on correlation selects the region of the ECG signal and the photoelectric pulse wave signal, and improves the accuracy for measurement of pulse wave conduction time, and can get better effect while gauss white noise and baseline drift are added, thus having greater potential application prospects.

Description

A kind of pulse wave translation time acquisition methods based on dependency

Technical field

The present invention relates to a kind of pulse wave translation time acquisition methods.More particularly to a kind of pulse wave based on dependency Conduction time acquisition methods.

Background technology

Pulse wave translation time is conduction time of the human body artery pulse pressure ripple in arteries.Pulse wave translation time Acquisition methods typically have two big class, and one kind is conventional method, calculate the two-way pulse wave signal of diverse location, calculate pulse wave and pass Lead the time, another kind of measuring method is using electrocardiosignal and pulse wave signal measurement pulse wave translation time.

General measuring method is exactly the arterial pulse that two diverse locations are synchronously measured in pulse wave communication process Ripple, obtains pulse wave translation time by measuring the propagation time of certain corresponding point, but due to pulse wave be one have it is many The complex wave that weight subharmonic is combined, its waveform is complicated and is ceaselessly being changed in communication process again, so often Rule algorithm does not ensure that the accuracy of measurement.

At present conventional pulse wave translation time measuring method by electrocardiosignal and Photoelectric Pulse Wave Signal Time difference is obtaining pulse wave translation time.Which mainly selects some characteristic points, such as selects the R crest values of human ecg signal Point, is passed so as to obtain pulse wave as the Origin And Destination of pulse wave translation time with some characteristic points of Photoelectric Pulse Wave Signal Lead the time.The method cannot be accurately identified because of individual characteristics point, cause the inaccuracy of end product.Based on correlation Property pulse wave translation time acquisition methods, then be the one piece of data for choosing electrocardiosignal and Photoelectric Pulse Wave Signal, it is to avoid Due to the error that indivedual points cause.Selection fragment is calculated based on it is critical only that for pulse wave translation time acquisition methods of dependency The correlation coefficient of interior electrocardiosignal and Photoelectric Pulse Wave Signal, the translation of Photoelectric Pulse Wave Signal when searching correlation coefficient maximum Amount.

The content of the invention

The technical problem to be solved be to provide it is a kind of accurately can obtain pulse wave translation time based on phase The pulse wave translation time acquisition methods of closing property.

The technical solution adopted in the present invention is：A kind of pulse wave translation time acquisition methods based on dependency, including Following steps：

1) using electrode slice and photoelectric sphyg wave sensor synchronous acquisition human ecg signal and Photoelectric Pulse Wave Signal；

2) determine the formula of correlation coefficient of human ecg signal data and Photoelectric Pulse Wave Signal data in preseting length：

In formula, R0 is correlation coefficient, and the starting point of preseting length is Loc_1, and the terminal of preseting length is Loc_2, and Xi is people Any one value of the body-centered signal of telecommunication between starting point Loc_1 and terminal Loc_2, EX are human ecg signal in starting point Loc_1 And the average between terminal Loc_2, Yi be Photoelectric Pulse Wave Signal any one between starting point Loc_1 and terminal Loc_2 Value, EY be average of the Photoelectric Pulse Wave Signal between starting point Loc_1 and terminal Loc_2, n=Loc_2-Loc_1；

3) in time domain, Photoelectric Pulse Wave Signal is moved to the left into the side-play amount of 1 unit, i.e. PPG (Loc_1+1, Loc_ 2+1), the phase relation of ECG (Loc_1, Loc_2) and PPG (Loc_1+1, Loc_2+1) after skew is calculated using formula of correlation coefficient Number, is designated as R1, wherein, ECG is human ecg signal, and PPG is Photoelectric Pulse Wave Signal；

4) 3) repeat step, calculates the correlation coefficient of ECG (Loc_1, Loc_2) and PPG (Loc_1+i, Loc_2+i), directly To obtaining m correlation coefficient：R1, R2 ... Rm, wherein m >=i；

5) coefficient R i of maximum in m coefficient R 1, R2 ... Rm is selected, by the light corresponding to coefficient R i The side-play amount of electric pulse wave movement is represented with i, so as to obtain pulse wave translation time PTT, PTT=i × 0.5 × 10³。

Step 2) include：A R ripple of human ecg signal is found, with R ripples as starting point Loc_1, one section of setting length is chosen The data of degree, final position are designated as Loc_2, and it is human ecg signal appointing between starting point Loc_1 and terminal Loc_2 to set Xi One value of meaning, EX is average of the human ecg signal between starting point Loc_1 and terminal Loc_2, and Yi is Photoelectric Pulse Wave Signal Any one value between starting point Loc_1 and terminal Loc_2, EY are Photoelectric Pulse Wave Signal in starting point Loc_1 and terminal Average between Loc_2, n=Loc_2-Loc_1, the correlation coefficient are designated as R0, then：

The R0=corrcoef used in matlab softwares (ECG (Loc_1, Loc_2), PPG (Loc_1, Loc_2)).

A kind of pulse wave translation time acquisition methods based on dependency of the present invention, have the advantages that：

1st, one section of region of electrocardiosignal and photoelectric sphyg ripple that the present invention chooses, rather than single characteristic point calculates pulse Ripple conduction time, improves the accuracy of pulse wave translation time measurement；

2nd, the present invention equally has preferable effect in the case where white Gaussian noise and baseline drift is added, with bigger Potential application foreground.

Description of the drawings

Fig. 1 is a kind of pulse wave translation time acquisition methods flow chart based on dependency of the present invention；

Fig. 2 be in the present invention pulse wave signal to left schematic diagram.

Specific embodiment

With reference to a kind of pulse wave translation time acquisition methods based on dependency of embodiment and accompanying drawing to the present invention It is described in detail.

A kind of pulse wave translation time acquisition methods based on dependency of the present invention, are that one kind is adopted by correlation analysiss The method that the human ecg signal of collection and Photoelectric Pulse Wave Signal obtain pulse wave translation time PTT.

As shown in figure 1, a kind of pulse wave translation time acquisition methods based on dependency of the present invention comprise the steps：

1) using electrode slice and photoelectric sphyg wave sensor synchronous acquisition human ecg signal (ECG) and photoelectric sphyg ripple letter Number (PPG)；

2) determine the formula of correlation coefficient of human ecg signal data and Photoelectric Pulse Wave Signal data in preseting length：

In formula, R0 is correlation coefficient, and the starting point of preseting length is Loc_1, and the terminal of preseting length is Loc_2, and Xi is people Any one value of the body-centered signal of telecommunication between starting point Loc_1 and terminal Loc_2, EX are human ecg signal in starting point Loc_1 And the average between terminal Loc_2, Yi be Photoelectric Pulse Wave Signal any one between starting point Loc_1 and terminal Loc_2 Value, EY be average of the Photoelectric Pulse Wave Signal between starting point Loc_1 and terminal Loc_2, n=Loc_2-Loc_1；

The determination of formula of correlation coefficient includes：A R ripple of human ecg signal is found, with R ripples as starting point Loc_1, choosing The data of one section of preseting length are taken, final position is designated as Loc_2, it is human ecg signal in starting point Loc_1 and terminal to set Xi Any one value between Loc_2, EX are average of the human ecg signal between starting point Loc_1 and terminal Loc_2, and Yi is light Any one value of electric pulse wave signal between starting point Loc_1 and terminal Loc_2, EY are Photoelectric Pulse Wave Signal in starting point Average between Loc_1 and terminal Loc_2, n=Loc_2-Loc_1, the correlation coefficient are designated as R0, then：

The R0=corrcoef used in matlab softwares (ECG (Loc_1, Loc_2), PPG (Loc_1, Loc_2)).

3) in time domain, Photoelectric Pulse Wave Signal is moved to the left into the side-play amount of 1 unit, i.e. PPG (Loc_1+1, Loc_ 2+1), the phase relation of ECG (Loc_1, Loc_2) and PPG (Loc_1+1, Loc_2+1) after skew is calculated using formula of correlation coefficient Number, is designated as R1, wherein, ECG is human ecg signal, and PPG is Photoelectric Pulse Wave Signal；

4) 3) repeat step, calculates the correlation coefficient of ECG (Loc_1, Loc_2) and PPG (Loc_1+i, Loc_2+i), directly To obtaining m correlation coefficient：R1, R2 ... Rm, wherein m >=i；

5) coefficient R i of maximum in m coefficient R 1, R2 ... Rm is selected, by the light corresponding to coefficient R i The side-play amount of electric pulse wave movement is represented with i, so as to obtain pulse wave translation time PTT, PTT=i × 0.5 × 10³。

Example is given below：

1st, using Analog Discovery capture cards synchronous acquisition electrocardiosignal (ECG) and Photoelectric Pulse Wave Signal (PPG), sample frequency is 2KHz, and data (length is 8000) are uploaded on PC；

2nd, a R ripple being found in ECG signal, then choosing one piece of data, same length is equally chosen in PPG signals Data, calculate the correlation coefficient of ECG signal and PPG signals；

3rd, Photoelectric Pulse Wave Signal is moved to the left into 1 unit (0.5 × 10^-3S), then calculate and above-mentioned electrocardiosignal Correlation coefficient, determines whether maximum；

4th, as shown in Fig. 2 being moved to the left 1,2,3 successively ... ..., m unit calculates electrocardiosignal and photoelectric sphyg respectively The correlation coefficient of ripple；

5th, as shown in Fig. 2 a value for selecting correlation coefficient maximum, records translational movement i (i≤m) now, by translational movement Calculate PTT now, PTT=i × 0.5 × 10³。

Claims (2)

1. a kind of pulse wave translation time acquisition methods based on dependency, it is characterised in that comprise the steps：

1) using electrode slice and photoelectric sphyg wave sensor synchronous acquisition human ecg signal and Photoelectric Pulse Wave Signal；

2) determine the formula of correlation coefficient of human ecg signal data and Photoelectric Pulse Wave Signal data in preseting length：

$R0=\frac{\underset{i=1}{\overset{n}{\Σ}}(Xi-EX)(Yi-EY)}{\sqrt{\underset{i=1}{\overset{n}{\Σ}}{(Xi-EX)}^2\·\underset{i=1}{\overset{n}{\Σ}}{(Yi-EY)}^2}}$

In formula, R0 is correlation coefficient, and the starting point of preseting length is Loc_1, and the terminal of preseting length is Loc_2, and Xi is people body-centered Any one value of the signal of telecommunication between starting point Loc_1 and terminal Loc_2, EX are human ecg signal at starting point Loc_1 and end Average between point Loc_2, Yi be any one value of Photoelectric Pulse Wave Signal between starting point Loc_1 and terminal Loc_2, EY The average for being Photoelectric Pulse Wave Signal between starting point Loc_1 and terminal Loc_2, n=Loc_2-Loc_1；

3) in time domain, Photoelectric Pulse Wave Signal is moved to the left into the side-play amount of 1 unit, i.e. PPG (Loc_1+1, Loc_2+ 1) phase relation of ECG (Loc_1, Loc_2) and PPG (Loc_1+1, Loc_2+1) after skew, is calculated using formula of correlation coefficient Number, is designated as R1, wherein, ECG is human ecg signal, and PPG is Photoelectric Pulse Wave Signal；

4) 3) repeat step, calculates the correlation coefficient of ECG (Loc_1, Loc_2) and PPG (Loc_1+i, Loc_2+i), until To m correlation coefficient：R1, R2 ... Rm, wherein m >=i；

5) coefficient R i of maximum in m coefficient R 1, R2 ... Rm is selected, by the photoelectricity arteries and veins corresponding to coefficient R i The side-play amount of ripple of fighting movement is represented with i, so as to obtain pulse wave translation time PTT, PTT=i × 0.5 × 10³。

2. a kind of pulse wave translation time acquisition methods based on dependency according to claim 1, it is characterised in that step It is rapid 2) to include：A R ripple of human ecg signal is found, the data of one section of preseting length with R ripples as starting point Loc_1, is chosen, eventually Point position is designated as Loc_2, sets any one value that Xi is human ecg signal between starting point Loc_1 and terminal Loc_2, EX The average for being human ecg signal between starting point Loc_1 and terminal Loc_2, Yi are Photoelectric Pulse Wave Signal in starting point Loc_1 Any one value and terminal Loc_2 between, EY are that Photoelectric Pulse Wave Signal is equal between starting point Loc_1 and terminal Loc_2 Value, n=Loc_2-Loc_1, the correlation coefficient are designated as R0, then：

$R0=\frac{\underset{i=1}{\overset{n}{\Σ}}(Xi-EX)(Yi-EY)}{\sqrt{\underset{i=1}{\overset{n}{\Σ}}{(Xi-EX)}^2\·\underset{i=1}{\overset{n}{\Σ}}{(Yi-EY)}^2}},$

The R0=corrcoef used in matlab softwares (ECG (Loc_1, Loc_2), PPG (Loc_1, Loc_2)).