CN103156599A - Detection method of electrocardiosignal R characteristic waves - Google Patents

Abstract

The invention discloses a detection method of electrocardiosignal R characteristic waves. A wavelet decomposition method for improving threshold value is carried out on collected electrocardiosignals to conduct filtering, then wavelet reconstruction protruding out of a QRS wave group is carried out, energy window conversion from time domain analysis to energy domain analysis is carried out, then a maximum value point is selected and optimized, and finally an wave selection is carried out according to R selection logics. The detection method uses multiresolution characteristics of wavelet analysis to decompose signals on different scales according to frequency to accordingly process the signals on the different scales in a targeted mode and increase algorithmic flexibility. The energy window conversion method is adopted to convert the signals from a time domain to an energy domain, effectively restrains interference of high-frequency noise and improves algorithmic stability. By means of wavelet reconstruction, the QRS wave groups are extracted, P waves and T waves serve as noise to be rejected, false detection caused by huge P waves and T waves in detection can be effectively avoided, and detection accuracy is improved.

Description

A kind of electrocardiosignal R characteristic wave detection method

Technical field

The present invention relates to electrocardiosignal detects and analysis technical field, particularly a kind of electrocardiosignal R characteristic wave detection method automatically.

Background technology

Heart disease has disguise and latency, is difficult to when not falling ill show on electrocardiogram, is again of short duration during morbidity, has little time to observe electrocardiogram.Need to carry 24 hours Holter to patient for this reason, carry out 24 hours ecg signal acquirings.But will cause like this data volume huge, the doctor needs the plenty of time to check one by one electrocardiogram, seeks abnormity point, has increased greatly doctor's burden.Simultaneously, can be subject to the impact of individual's cognition and emotion in diagnostic procedure, make heart disease diagnosis have subjectivity.Use signal of telecommunication automatic analysis technology and can correct this deviation.Existing parser is for the parser comparatively perfect of R ripple, but the treatment effect on high-frequency noise and QRS wave group form are changed is unsatisfactory.

Summary of the invention

The purpose of this invention is to provide a kind of electrocardiosignal R characteristic wave detection method, to solve the processing relatively poor problem of existing parser in reply high-frequency noise and the change of QRS wave group form.

The object of the present invention is achieved like this: electrocardiosignal R characteristic wave detection method provided by the present invention comprises the following steps:

A) signals collecting: the frequency acquisition with 250Hz gathers human ecg signal, and is stored as the data mode of TXT document, then reads in computer with the electrocardio primary signal data of Matlab software with described TXT document storage.

B) described electrocardio primary signal data are carried out Filtering Processing:

B-1) described electrocardio primary signal is carried out wavelet decomposition: select the DB6 small echo, signal is carried out 8 layers of decomposition, obtain the wavelet coefficient d on each yardstick _i

B-2) adopt improved calculated threshold method, ask for the threshold value of each yardstick, wavelet coefficient is carried out thresholding process:

$T_i=\frac{{\mathrm{\σ}}_i\sqrt{2\log \left(n\right)}}{e^{i-1}},i=1,...,8$

Wherein, T _iBe improved threshold value, i represents the wavelet decomposition number of plies, and e is natural constant, and n represents sampling number, σ _iAverage for the wavelet coefficient absolute value:

B-3) adopting the soft-threshold method to carry out thresholding to signal processes: choose different threshold values at different scale and carry out the thresholding processing, obtain filtered electrocardiosignal.

C) according to the frequency distribution scope feature of QRS wave group, select the 3rd, 4 yardsticks to carry out wavelet reconstruction, obtain the electrocardiosignal S' after reconstruct;

When the yardstick of wavelet reconstruction was carried out in selection, it mainly considered the different of QRS wave group and P ripple, T wave frequency distribution, and comprehensive QRS group frequency distributions is at yardstick, and P ripple, the T wave frequency less yardstick that distributes carries out wavelet reconstruction.

D) electrocardiosignal through wavelet reconstruction is carried out the energy window conversion, and chooses maximum point:

D-1) energy window conversion: press following formula, will transform to the energy domain analysis by time domain analysis through the electrocardiosignal S' of wavelet reconstruction, obtain the electrocardiosignal energy curve:

$E_n=\underset{n-\frac{N}{2}+1}{\overset{n+\frac{N}{2}}{\mathrm{\&Sigma;}}}{S^{{\&prime;}^2}}_n,n=\frac{N}{2},...,M-\frac{\mathrm{<figure-callout\; id="2"\; label="N"\; filenames="130402160628.png,130402160633.png"\; state="\{\{state\}\}">N</figure-callout>}}{2}$

Wherein, E _nThe energy value that represents n sampled point; N is selected length of window; M is total sampling number; S' _nN data that represent the electrocardiosignal S' after described wavelet reconstruction;

Described length of window is determined: according to the size of sample frequency, select the number of the minimum even number of samples point that the QRS wave group can be covered long as window as follows;

D-2) choose maximum point: resulting electrocardiosignal energy curve is carried out the hard-threshold processing, that is:

$Q_n=\left\{\begin{array}{c}{E}_n,E_n\&GreaterEqual;{\mathrm{<figure-callout\; id="0"\; label="T\; h"\; filenames="130402160628.png,130402160638.png"\; state="\{\{state\}\}">T</figure-callout>}}_h\\ 0,E_n<T_h\end{array}\right\}$

Wherein, T _hFor selected threshold value, get T _h=0.3*median (E _n),

The crest location of the electrocardiosignal energy curve after the processing of then selection process hard-threshold is as maximum point.

E) optimize maximum point: set 2 time threshold t ₁And t ₂, and t ₁＜t ₂When the interval of any two maximum points less than t ₁The time, just remove less that of amplitude between these two maximum points; When any two maximum points interval greater than t ₂The time, just seek another unrecognized extreme point between these two maximum points; When the interval of any two maximum points both more than or equal to t ₁, again less than or equal to t ₂, these two maximum points all keep, the maximum point through optimizing that finally obtains, and the corresponding QRS wave group of each described maximum point through optimizing.

F) according to the time point at each maximum point place in step e), the point of search signal amplitude maximum in the scope of each 7 sampled points of corresponding time point left and right on filtered electrocardiosignal described in step b) is as the R ripple that detects.

Step b-3) in, the expression formula of soft-threshold processing is:

${\hat{d}}_j=\left\{\begin{array}{c}\mathrm{sgn}\left(d_j\right)\left(\right|d_j|-T_j),\left|d_j\right|\&GreaterEqual;{\mathrm{<figure-callout\; id="0"\; label="T\; j"\; filenames="130402160628.png,130402160638.png"\; state="\{\{state\}\}">T</figure-callout>}}_j\\ 0,\left|d_j\right|<T_j\end{array}\right.,$ J=i wherein;

In step c), the expression formula of wavelet reconstruction is:

$S^{\&prime;}={\hat{d}}_3+{\hat{d}}_4,$

Wherein said

Be the signal after the thresholding processing.

Steps d-1) length of window described in is 26.

T in step e) ₁And t ₂Determine as follows:

With E _tAs steps d-2) meansigma methods of the interval of resulting all maximum points, regulation t ₁=0.5 * E _t, t ₂=1.5 * E _t

The present invention carries out choosing in the wavelet decomposition process specific wavelet basis function and the wavelet decomposition number of plies to signal, in being carried out the thresholding processing procedure, adopts signal simultaneously improved threshold method, make the myoelectricity that filtered signal is mingled with in the filtering electrocardiosignal disturb, when baseline drift and power frequency are disturbed, the Useful Information that kept as much as possible has improved the phenomenon of generic threshold value excess smoothness.The present invention extracts the QRS wave group, and P ripple, T ripple is used as noise eliminating by carrying out wavelet reconstruction, has effectively avoided tall and big P ripple, the flase drop that the T ripple causes in detection, has improved the precision that detects.In addition, the present invention has adopted the energy window alternative approach, goes analysis to solve in time-domain analysis thereby convert the signal into energy domain, and signal is subject to the impact of high-frequency noise, and can not be by the problem of whole filterings in filtering.In the energy window conversion, the present invention has taken into full account the leap time of temporal signatures and the QRS wave group of electrocardiosignal, carries out the long selection of window.The results show only fenestrate length is 26 o'clock, and the undetected phenomenon that many inspections that the spurious peaks of noise produces and the QRS ripple of low amplitude value cause could the most effectively be avoided.

Method of the present invention has solved in prior art existing problem aspect reply high-frequency noise and the change of QRS wave group form, can realize to electrocardiosignal R characteristic wave fast, accurately detect.

Description of drawings

Fig. 1 is flow chart of the present invention.

Fig. 2 is the electrocardio primary signal.

Fig. 3 is for carrying out filtered electrocardiosignal.

Fig. 4 is the electrocardiosignal of carrying out after wavelet reconstruction.

Fig. 5 is the electrocardiosignal of choosing maximum point of carrying out after the hard-threshold processing.

The flow chart of Fig. 6 for maximum point is optimized.

Fig. 7 is detected R ripple.

The specific embodiment

Embodiment 1:

The present embodiment is at Intel Pentium Dual E2200 2.20GHz, in save as 3.00GB, operating system is to realize in the computer of Window XP, whole R ripple detection algorithm adopts the Matlab language compilation.

Implementation process of the present invention is as shown in Figure 1:

A) ecg signal acquiring and output: (seven of Beijing logical first Electron equipment Co., Ltd of wound advanced in years leads electrocardiogram module (ECG CB) to utilize ecg signal acquiring equipment, its sample frequency is 250HZ, and gap length analysis time of electrocardiosignal is set as 10s) gather the electrocardiosignal of human body and store with the form of TXT.In Matlab software, the electrocardiogram (ECG) data in the TXT document is read in environment, as Fig. 2, obtain the electrocardio primary signal.

B) utilize the wavelet-decomposing method of improving threshold value to carry out Filtering Processing to the electrocardio primary signal:

B-1) select the DB6 small echo of Daubechies small echo series, carry out 8 layers of decomposition, as shown in table 1:

Table 1: under the sample frequency of 250Hz, the DB6 small echo is carried out 8 layers of decomposition

Scale	250HZ
			1	62.5-125
2	31.25-62.5
		3	15.625-31.25
4	7.8125-15.625
		5	3.90625-7.8125
6	1.953125-3.90625
		7	0.9765625-1.953125
8	0.48828125-0.9765625

Then extract the wavelet coefficient d on each yardstick _i, carry out the calculating of threshold value, to obtain through improved threshold value, that is:

$T_i=\frac{{\mathrm{\&sigma;}}_i\sqrt{2\log \left(n\right)}}{e^{i-1}},$ I=1 ..., 8 formulas (1)

Wherein i represents the wavelet decomposition number of plies, T _iBe improved threshold value, e is natural constant, and n represents sampling number, σ _iBe the average of wavelet coefficient absolute value, its expression formula is ${\mathrm{\&sigma;}}_i=\frac{\mathrm{median}\left|d_i\right|}{0.6745}.$

Compare existing fixed threshold method and minimax threshold method, after by formula (1), the algorithm of threshold value being improved, the threshold value after improvement has the sideband adaptivity, and has kept good denoising reconstruction property.

B-2) adopt the soft-threshold method, choose correspondingly to improved threshold value on each yardstick, by formula (2), electrocardiosignal is carried out the thresholding processing, that is:

${\hat{d}}_j=\left\{\begin{array}{c}\mathrm{sgn}\left(d_j\right)\left(\right|d_j|-T_j),\left|d_j\right|\&GreaterEqual;{\mathrm{<figure-callout\; id="0"\; label="T\; j"\; filenames="130402160628.png,130402160638.png"\; state="\{\{state\}\}">T</figure-callout>}}_j\\ 0,\left|d_j\right|<T_j\end{array}\right.$ Formula (2)

J=i wherein;

Thereby obtain filtered electrocardiosignal, as shown in Figure 3.

After the wavelet-decomposing method of improving threshold value by employing is carried out filtering, make filtered electrocardiosignal, the Useful Information that kept as much as possible has improved the phenomenon of generic threshold value excess smoothness, makes filter effect more stable.

C) give prominence to the wavelet reconstruction of QRS wave group: the frequency distribution scope of normal electrocardiosignal QRS wave group is 5-45Hz, as can be seen from Table 1, it mainly concentrates on 3,4 yardsticks, and P ripple and T wave frequency distribution are 0.05 to 10Hz, there is no or only has a small amount of distribution on 3,4 yardsticks, therefore, difference according to QRS wave group and P ripple, T wave frequency distribution, select 3,4 minimum yardsticks of QRS wave group and P ripple, T wavelength-division cloth frequency overlap to carrying out wavelet reconstruction through filtered electrocardiosignal, that is:

$S^{\&prime;}={\hat{d}}_3+{\hat{d}}_4$ Formula (3)

Wherein,

With

Be respectively by step b-2) to the result of the electrocardiosignal on 3,4 yardsticks after thresholding is processed.

Through after wavelet reconstruction, resulting electrocardiosignal is mainly the information of QRS wave group, has played the effect that highlights the QRS wave group, as shown in Figure 4.

D) electrocardiosignal after wavelet reconstruction is carried out by the energy window conversion of time domain analysis to the energy domain analysis, and chooses maximum point:

D-1) energy window conversion: by as shown in the formula (4), will transform to the energy domain analysis by time domain analysis through the electrocardiosignal of wavelet reconstruction, and obtain the electrocardiosignal energy curve:

$E_n=\underset{n-\frac{\mathrm{<figure-callout\; id="2"\; label="N"\; filenames="130402160628.png,130402160633.png"\; state="\{\{state\}\}">N</figure-callout>}}{2}+1}{\overset{n+\frac{N}{2}}{\mathrm{\&Sigma;}}}{S^{{\&prime;}^2}}_n,n=\frac{\mathrm{<figure-callout\; id="2"\; label="N"\; filenames="130402160628.png,130402160633.png"\; state="\{\{state\}\}">N</figure-callout>}}{2},...,M-\frac{\mathrm{<figure-callout\; id="2"\; label="N"\; filenames="130402160628.png,130402160633.png"\; state="\{\{state\}\}">N</figure-callout>}}{2}$ Formula (4)

Wherein, N is selected length of window (N=26), and M is total sampling number, S' _nN the data of electrocardiosignal S' after expression step c) wavelet reconstruction.

In the energy window conversion, choosing of window length is a key, and it determines directly whether R ripple detection algorithm is effective.Select the window long time, according to the size of sample frequency, select the number of the minimum even number of samples point that just in time the QRS wave group can be covered long as window.In the present embodiment, taken into full account the leap time of temporal signatures and the QRS wave group of electrocardiosignal, the selection of its N value is determined as follows: the sample frequency of electrocardiosignal is 250Hz, and normal QRS wave group generally is no more than 0.1s, be 25 sampled points, it is long for even number that we choose window, is 26.The results show only fenestrate length is 26 o'clock, and the undetected phenomenon that many inspections that the spurious peaks of noise produces and the QRS ripple of low amplitude value cause could the most effectively be avoided.

In time-domain analysis, signal is subject to the impact of high-frequency noise, and can not by whole filterings, for this problem, adopt the method for energy bed conversion that time domain analysis is transformed to the energy domain analysis in filtering.Energy domain has better robustness than time domain analysis to noise.As Fig. 5, after the energy window conversion, the QRS wave group location point of signal becomes more outstanding, and the interval between heart bat and the heart are clapped is more obvious, and the impact of high-frequency noise also dies down accordingly.

D-2) by setting threshold to choose maximum point: the signal energy curve that obtains is carried out the hard-threshold processing:

$Q_n=\left\{\begin{array}{c}{E}_n,E_n\&GreaterEqual;{\mathrm{<figure-callout\; id="0"\; label="T\; h"\; filenames="130402160628.png,130402160638.png"\; state="\{\{state\}\}">T</figure-callout>}}_h\\ 0,E_n<T_h\end{array}\right\}$ Formula (5)

Wherein, T _hFor selected threshold value, get T _h=0.3*median (E _n);

Then choose crest location through the electrocardiosignal energy curve after the hard-threshold processing as maximum point, as shown in Figure 5.

E) optimize maximum point: flow chart as given in Fig. 6, set 2 time threshold t ₁And t ₂, and t ₁＜t ₂When the interval of any two maximum points less than t ₁The time, just remove less that of amplitude between these two maximum points; When any two maximum points interval greater than t ₂The time, just seek another unrecognized extreme point between these two maximum points; As the interval of two maximum points both more than or equal to t ₁, again less than or equal to t ₂, these two maximum points all keep, the corresponding QRS wave group of each maximum point through optimizing that so finally obtains.

In Fig. 6, E _tExpression steps d-2) meansigma methods of the interval of resulting all maximum points, t ₁=0.5 * E _t, t ₂=1.5 * E _t

F) the R ripple is chosen: according to the time point at determined each maximum point place in step e), the point of search signal amplitude maximum in the scope of each 7 sampled points of corresponding time point left and right on electrocardiosignal after filtering, be the R ripple (Fig. 7) that detects in step b).

Claims (5)

1. an electrocardiosignal R characteristic wave detection method, is characterized in that, comprises the following steps:

A) signals collecting: the frequency acquisition with 250Hz gathers human ecg signal, and is stored as the data mode of TXT document, then reads in computer with the electrocardio primary signal data of Matlab software with described TXT document storage;

B) described electrocardio primary signal data are carried out Filtering Processing:

B-1) described electrocardio primary signal is carried out wavelet decomposition: select the DB6 small echo, signal is carried out 8 layers of decomposition, obtain the wavelet coefficient d on each yardstick _i

B-2) adopt improved calculated threshold method, ask for the threshold value of each yardstick, wavelet coefficient is carried out thresholding process:

$T_i=\frac{{\mathrm{\&sigma;}}_i\sqrt{2\log \left(n\right)}}{e^{i-1}},i=1,...,8$

Wherein, T _iBe improved threshold value, i represents the wavelet decomposition number of plies, and e is natural constant, and n represents sampling number, σ _iAverage for the wavelet coefficient absolute value:

B-3) adopting the soft-threshold method to carry out thresholding to signal processes: choose different threshold values at different scale and carry out the thresholding processing, obtain filtered electrocardiosignal;

C) according to the frequency distribution scope feature of QRS wave group, select the 3rd, 4 yardsticks to carry out wavelet reconstruction, obtain the electrocardiosignal S' after reconstruct;

D) electrocardiosignal through wavelet reconstruction is carried out the energy window conversion, and chooses maximum point:

D-1) energy window conversion: press following formula, will transform to the energy domain analysis by time domain analysis through the electrocardiosignal S' of wavelet reconstruction, obtain the electrocardiosignal energy curve:

$E_n=\underset{n-\frac{N}{2}+1}{\overset{n+\frac{N}{2}}{\mathrm{\&Sigma;}}}{S^{{\&prime;}^2}}_n,n=\frac{N}{2},...,M-\frac{N}{2}$

Wherein, E _nThe energy value that represents n sampled point; N is selected length of window; M is total sampling number; S' _nN data that represent the electrocardiosignal S' after described wavelet reconstruction;

Described length of window is determined: according to the size of sample frequency, select the number of the minimum even number of samples point that the QRS wave group can be covered long as window as follows;

D-2) choose maximum point: resulting electrocardiosignal energy curve is carried out the hard-threshold processing, that is:

$Q_n=\left\{\begin{array}{c}{E}_n,E_n\&GreaterEqual;T_h\\ 0,E_n<T_h\end{array}\right\}$

Wherein, T _hFor selected threshold value, get T _h=0.3*median (E _n),

The crest location of the electrocardiosignal energy curve after the processing of then selection process hard-threshold is as maximum point;

E) optimize maximum point: set 2 time threshold t ₁And t ₂, and t ₁＜t ₂, when the interval of any two maximum points less than t ₁The time, just remove less that of amplitude between these two maximum points; When any two maximum points interval greater than t ₂The time, just seek another unrecognized extreme point between these two maximum points; When the interval of any two maximum points both more than or equal to t ₁, again less than or equal to t ₂, these two maximum points all keep, the maximum point through optimizing that finally obtains, and the corresponding QRS wave group of each described maximum point through optimizing;

F) according to the time point at each maximum point place in step e), the point of search signal amplitude maximum in the scope of each 7 sampled points of corresponding time point left and right on filtered electrocardiosignal described in step b) is as the R ripple that detects.

2. a kind of electrocardiosignal R characteristic wave detection method according to claim 1, is characterized in that,

Step b-3) in, the expression formula of soft-threshold processing is:

${\hat{d}}_j=\left\{\begin{array}{c}\mathrm{sgn}\left(d_j\right)\left(\right|d_j|-T_j),\left|d_j\right|\&GreaterEqual;T_j\\ 0,\left|d_j\right|<T_j\end{array}\right.,$ J=i wherein.

3. a kind of electrocardiosignal R characteristic wave detection method according to claim 1, is characterized in that, in step c), the expression formula of wavelet reconstruction is:

$S^{\&prime;}={\hat{d}}_3+{\hat{d}}_4,$

Wherein said

Be the signal after the thresholding processing.

4. a kind of electrocardiosignal R characteristic wave detection method according to claim 1, is characterized in that steps d-1) described in length of window be 26.

5. a kind of electrocardiosignal R characteristic wave detection method according to claim 1, is characterized in that t in step e) ₁And t ₂Determine as follows:

With E _tAs steps d-2) meansigma methods of the interval of resulting all maximum points, regulation t ₁=0.5 * E _t, t ₂=1.5 * E _t

Images