CN103630584B - Time domain inflection point extraction method for pulse relaxation signal - Google Patents
Time domain inflection point extraction method for pulse relaxation signal Download PDFInfo
- Publication number
- CN103630584B CN103630584B CN201310486085.4A CN201310486085A CN103630584B CN 103630584 B CN103630584 B CN 103630584B CN 201310486085 A CN201310486085 A CN 201310486085A CN 103630584 B CN103630584 B CN 103630584B
- Authority
- CN
- China
- Prior art keywords
- signal
- charging
- value
- pulse
- discharging
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Investigating Or Analyzing Materials By The Use Of Electric Means (AREA)
Abstract
The invention belongs to the technical field of food measurement and discloses a time domain inflection point extraction method for a pulse relaxation signal. The time domain inflection point extraction method comprises the following steps: carrying out segmentation and smoothing processing on an acquired pulse relaxation signal; removing noises and reestablishing a signal to obtain a smoothed pulse relaxation signal; carrying out normalization on the smoothed pulse relaxation signal, and pre-processing the pulse relaxation signal; rotating the pre-processed pulse relaxation signal at a certain angle to obtain a rotated pulse relaxation signal; finding out the maximum value of the rotated pulse relaxation signal and recording the position of the maximum value of the pulse relaxation signal; carrying out counter rotation processing on the rotated pulse relaxation signal to obtain a new rotated pulse relaxation signal; and obtaining an inflection point I according to a position point of the recorded maximum value. According to the method provided by the invention, the extraction efficiency and the accuracy of pulse time domain characteristics can be improved effectively.
Description
Technical field
The invention belongs to eatable measuration technical field, relate to the analyzing and processing technology of signal, specifically a kind of pulse relaxation signals Inflexion extracting method.
Background technology
In modern food science, food sense organ science is more and more subject to people's attention, although the development of human food's sense organ is swift and violent all the more and ripe now, and most organoleptic test result or the serious impact being subject to subjective factor.And the pulse relaxation signals in the present invention is measured by using the electronic tongues equipment of simulating human gustatory organ, by detecting the redox ability of food solution, electrode adsorption ability reacts the characteristic of this food solution, for the reach of science of food sense organ adds powerful power.
At present, the Method and Technology mainly applied in the Inflexion extracting of the process of pulse relaxation signals and signal has: method of derivation, reversion method and smothing filtering etc., wherein smothing filtering is that paired pulses relaxation signals does pretreated technical method, and method of derivation and reversion method extract the common method of pulse relaxation signals flex point.
Method of derivation, reversion method are all the spans that the value of the pulse relaxation signals obtained according to segmentation determines to extract pulse flex point, wherein method of derivation obtains extreme value according to secondary differentiate to determine pulse flex point, reversion rule asks the difference of adjacent signaling point continuously, obtains pulse signal flex point according to the reversion of difference.Although these two kinds of methods can determine the flex point of pulse relaxation signals substantially, pulse relaxation signals data volume is larger, when extracting flex point by these two kinds of methods, can affect the speed of extraction and the efficiency of software.
Smothing filtering utilizes the method for moving average to do smoothing processing to signal exactly, and but, selected level and smooth threshold value may not necessarily be suitable for all pulse relaxation signals, if threshold value is less than normal, the effect of denoising is just not obvious, if threshold value is bigger than normal, then can lose part useful information.So, look for optimal level and smooth threshold value, use smothing filtering to carry out denoising to signal, make Inflexion extracting reach better effect.
Summary of the invention
The present invention seeks to, for above-mentioned deficiency existing in prior art, to provide a kind of method being applicable to pulse relaxation signals Inflexion extracting, to reach better Inflexion extracting effect.For this reason, the present invention adopts following technical scheme.
A kind of time domain inflection point extraction method for pulse relaxation signal, paired pulses relaxation signals, at different frequencies, determines the pulse signal of one-period according to stimulating frequency and exciting voltage stepping setpoint frequency coefficient, periodic coefficient.Then, find out crest value M1 and the trough value M2 of periodic pulse signal respectively, then the signaling point back getting some (N) according to M1 and M2 is respectively as between charging zone and discharge range.Then from charging and discharging wayside signaling, flex point is extracted.
Specifically comprise the following steps:
1) by different sample frequency, segmentation is carried out to the pulse relaxation signals collected;
2) Inflexion extracting is carried out to each periodic signal of the signal of different frequency section; In this step, the method single periodic signal being extracted to flex point is as follows:
A. by sample frequency setting electric discharge charging coefficient;
B. on charging signals and discharge signal, the signal spacing of extracting flex point is obtained, i.e. charging wayside signaling and discharge range signal according to value;
C. to the smoothing filtering process of the charging and discharging wayside signaling obtained, remove noise, obtain the charging and discharging wayside signaling smoothly;
D. the charging and discharging wayside signaling after level and smooth is normalized, realizes the pre-service of charging and discharging wayside signaling;
E. to through pretreated charging and discharging wayside signaling coordinate axis, centered by initial point, according to the rotation of certain angle, obtain postrotational charging and discharging wayside signaling, the accuracy that flex point I extracts can be increased;
F. find out the most value of charging and discharging wayside signaling after over-rotation, record the ordinate size that charging and discharging wayside signaling is worth most;
G. carrying out despining process by recording the ordinate size be worth most, in the interval normalized signal of initial charging and discharging, obtaining the position (i.e. which put be flex point) of flex point according to the value ordinate size after despining;
H. according to the corner position marked in charging and discharging normalized signal, in the interval smooth signal of initial charging and discharging, the flex point value I needing to extract is obtained.
In the inventive solutions, pulse relaxation signals collection be the electrochemical signals detected under the different electromotive force of food solution.In gatherer process, pulse relaxation signals can be subject to the interference of the factors such as food solution composition, electrode, dissimilar noise can be mixed into, therefore, be suitable for the smoothing process of noise before paired pulses relaxation signals carries out Inflexion extracting, remove certain noise, then extract the flex point I of pulse relaxation signals.
The disposal of gentle filter described in step c, preferably adopts the method for moving average, the smoothing filtering of paired pulses relaxation signals, rebuilds pulse relaxation signals.The Global Information of smothing filtering not lossing signal again while suitably removing certain signal noise, is beneficial to the extraction to flex point more,
Method of moving average smothing filtering specifically carries out according to following method:
C-1. terminate to last pulse current value from first pulse current value.
C-2. each to traveling through n pulse current value averaged backward, be used as new pulse current value with this.
C-3. the pulse current value is on average used to be combined into new pulse relaxation signals.
Wherein n is threshold value, is preferably 5.
Normalized described in steps d, detailed process can refer to carries out as follows:
D-1. the maximal value Q of signal spacing is obtained
maxwith minimum value Q
min;
D-2. the pulse current value Q in ergodic signals interval
value, according to Q
maxand Q
min, obtain the pulse current value Q after normalization
value1, computing formula is as follows:
Q
value1=(Q
value-Q
min)/(Q
max-Q
min)
D-3. the charging and discharging wayside signaling after normalized, makes transverse and longitudinal coordinate values become a kind of relative value relation from the absolute value of physical system numerical value, for follow-up rotation is ready.
This invents said pulse relaxation signals Inflexion extracting method, smothing filtering paired pulses relaxation signals is used to carry out pre-service, and the scope of discrete method determination flex point is employed according to the feature of pulse relaxation signals, combine again signal be normalized, rotate, despun process, not only increase Inflexion extracting efficiency, more considerably increase accuracy.
Accompanying drawing explanation
Fig. 1 is a kind of original signal of pulse relaxation signals.
Fig. 2 is the frequency signal of a kind of frequency in the relaxation signals of pulse shown in Fig. 1.
Fig. 3 is the periodic signal in the single cycle of the relaxation signals of pulse shown in Fig. 2.
Fig. 4 is in the pulse relaxation signals single cycle between charging zone, discharge range signal, and wherein (a) is charging wayside signaling, and (b) is discharge range signal.
Fig. 5 is between charging zone, the smooth signal of discharge range signal after the disposal of gentle filter, and wherein (a) is smooth signal between charging zone, and (b) is discharge range smooth signal.
Fig. 6 is between the charging zone after normalized, discharge range signal, and wherein (a) is for being normalized signal between charging zone, and (b) is discharge range normalized signal.
Fig. 7 be through rotate charging zone between, discharge range signal, wherein (a) is rotating signal between charging zone, and (b) is discharge range rotating signal.
Fig. 8 for indicate value (" o ") charging zone between and discharge range rotating signal, wherein (a) for indicate value (" o ") charging zone between rotating signal, (b) is for indicating the discharge range rotating signal of value (" o ").
Fig. 9 is for marked flex point I(" o ") charging zone between and discharge range normalized signal, wherein (a) marked flex point I(" o ") charging zone between normalized signal, (b) is for marked flex point I(" o ") discharge range normalized signal.
Figure 10 is for marked flex point I(" o ") charging zone between and discharge range smooth signal, wherein (a) is flex point I(" o ") charging zone between smooth signal, (b) is flex point I(" o ") discharge range smooth signal.
Figure 11 is for marked flex point I(" o ") frequency signal.
Figure 12 is for marked flex point I(" o ") pulse relaxation signals.
Embodiment
For the ease of the understanding of the present invention, first introduce various pulse relaxation parameter implication, be mainly divided into signal parameter and discharge range signal parameter between charging zone, these parameters are as shown in Fig. 1 (c):
1. signal parameter between charging zone:
1) wave inception: the signal segment started in pulse relaxation signals between rising suddenly to signal is exactly wave inception, and now potential pulse signal maintains 0V, without any electric current in food solution.
2) charge point: it is exactly charge point that pulse relaxation signals is transitioned into the flex point between rising suddenly from wave inception, and now potential pulse signal sports nV(0<n<10 from 0V).
3) ripple is risen: the rising signals between charge point and pulse relaxation signals peak is called and rises ripple, that in food solution, two electrodes form double electrical layers, start charging, naturally under the effect of the component gathered at electrode surface in food solution at current potential, there is redox reaction, produce redox current.Rising ripple is exactly being formed by charging current and redox current.
4) crest: the peak of pulse relaxation signals is called as crest, now, charging current and redox current reach maximal value.
5) paddy falls: crest starts curved decline and arrives stable signal segment and be called and fall paddy, and now, along with the consumption gradually completing the redox materials being adsorbed on electrode surface with food of charging, redox current reduces gradually.
2. discharge range signal parameter:
1) point of discharge: falling paddy, to be transitioned into the flex point that signal declines suddenly be exactly point of discharge, and now potential pulse signal returns back to 0V.
2) ripple falls: the dropping signal between point of discharge and pulse relaxation signals minimum point is called and falls ripple, that the electrostatic double layer equilibrium system originally set up due to electrode surface is destroyed, electric double layer capacitance starts electric discharge, the charged ion of electrode surface accumulation is rapidly to food solution bulk and moves, and therefore produces the pulsed current signal reverse with last process.
3) trough: the minimum point of pulse relaxation signals is called as trough, now, reverse pulse current signal reaches maximal value.
4) peak is risen: trough starts curved rising and arrives stable signal segment and be called and rise peak, and now, along with potential pulse signal is continually and steadily in zero potential, the continuous discharge of electric double layer capacitance is until electrode surface replys charge balance, and inverse current continues exponentially and weakens.
5) Gui Bo: rise peak and start stable signal segment between signal ended and return ripple exactly in pulse relaxation signals, now inverse current is 0.
3. time index
1) t
frequency: frequency time, the pulse relaxation signals in the present invention is the pulse relaxation signals of the multi-frequency section of the different sample frequency of a kind of Different periods, so, according to sample frequency, signal can be divided into the pulse relaxation signals of different frequency section.
2) t
cycle: cycle length, the pulse relaxation signals of each frequency band has ten cycles, and the time that each cycle continues is exactly t
cycle.
3) t
discharge and recharge: the discharge and recharge time, the pulse relaxation signals in each cycle has between charging zone and discharge range signal respectively, the process of potential pulse signal from 0V to nV between charging zone, discharge range is then that potential pulse signal is from nV(0<n<10) be returned to the process of 0V, the time of these two wayside signalings is equal, is namely t
discharge and recharge.
4. waveform index
1) H1: rise wave height, namely crest is to wave inception length of perpendicular.
2) H2: fall wave height, namely trough is to returning ripple vertical length.
3) C1: the curvature of falling paddy.
4) C2: the curvature rising peak.
H1 and H2 represents the ion concentration in food solution, for assessment of the ionic strength in food.C1 and C2 represents the locomotivity of Active Ingredient in Food in charging process and in discharge process respectively.
5. unique point index
(1) M1: crest value, the maximal value of pulse relaxation signals.
(2) M2: trough value, the minimum value of pulse relaxation signals.
(3) I1: charge value (flex point), the current value of charge point.
(4) I2: place value (flex point), the current value of point of discharge.
M1 and M2 represents the charging and discharging rate travel of food effects of ion respectively, and I1 has reacted the redox ability of Active Ingredient in Food, and I2 has then reacted the adsorptive power of Active Ingredient in Food and electrode, simultaneously for assessment of food quality stability.
Example:
Be the KCl of 0.01mol/L below with concentration be example, illustrate the process extracting pulse relaxation signals flex point:
As shown in Figure 1, as can be seen from Figure, original signal has the signal composition of three frequency bands (20000Hz respectively, 2000Hz, 200Hz) to the original pulse relaxation signals of 1.KCl.
2., because electronic tongues equipment is with the frequency excitation pulse signal of 20000Hz, respectively with 20000Hz, 2000Hz, 200Hz tri-frequency collection pulse relaxation signals, each frequency band gathers the stand-by period of ten cycles and 3S.Therefore, original signal is divided into single frequency signal according to following coefficient of frequency:
A.t
frequency1=460000 (2 × 10 × 20000+3 × 20000);
B.t
frequency2=46000 (2 × 10 × 2000+3 × 2000);
C.t
frequency2=4600 (2 × 10 × 200+3 × 200).
The single frequency signal of KCl20000Hz as shown in Figure 2, for it, will further illustrate the process of pulse relaxation signals Inflexion extracting below.
3. when frequency acquisition is 20000Hz, periodic coefficient t
cyclebe then 40000, according to periodic coefficient, the frequency signal of 20000Hz be divided into periodic signal, KCl20000Hz periodic signal as shown in Figure 3.
4. discharge and recharge coefficient t
discharge and rechargebe 500, then according to discharge and recharge coefficient, periodic signal be divided into charging and discharging wayside signaling, concrete fragmentation procedure is as follows:
(1) first find maximal value Max1 and the Max2 in charge or discharge interval, record the position (namely at which point of charging and discharging signal) of maximal value.
(2) again according to discharge and recharge coefficient 500, travel through 500 forward from the position being worth most place, the signal after intercepting is exactly charging and discharging wayside signaling.
Wherein charge wayside signaling as shown in Figure 4 (a), puts a wayside signaling as shown in Figure 4 (b).
5., respectively to the smoothing filtering process of the transverse and longitudinal coordinate of charging and discharging wayside signaling, the result after smothing filtering as shown in Figure 5.
6. be normalized the transverse and longitudinal coordinate of charging and discharging wayside signaling respectively, normalized result as shown in Figure 6.
7. pair charging and discharging normalized signal is centered by initial point, carry out 45 ° clockwise and be rotated counterclockwise, postrotational result is as shown in Figure 7.
8. find out the most value of charging and discharging wayside signaling after over-rotation, record the ordinate size that charging and discharging wayside signaling is worth most, specifically, ask for minimum value Dmin and the maximal value Dmin of the interval rotating signal of charging and discharging respectively, and record minimum value Dmin and maximal value Dmax.Between the charging zone indicating value (representing with " o " in figure) and discharge range rotating signal as shown in Figure 8.
9. carry out despining process by recording the ordinate size be worth most, in normalized charging and discharging wayside signaling, obtain the position (namely which point is flex point) of flex point according to the value ordinate size after despining, between the charging zone having indicated flex point (" o ") and discharge range normalized signal as shown in Figure 9.
10., according to the corner position that marks in charging and discharging normalized signal, in the interval smooth signal of initial charging and discharging, obtain the flex point value I needing to extract, between the charging zone having indicated flex point (" o ") and discharge range smooth signal as shown in Figure 10.
11. pairs of each pulse relaxaton period signals carry out the Inflexion extracting of as above step, marked flex point I(" o ") 20000Hz single frequency signal and complete pulse relaxation signals respectively as shown in Figure 11,12.
Claims (2)
1. a time domain inflection point extraction method for pulse relaxation signal, is characterized in that, specifically comprises the following steps:
1) by different sample frequency, segmentation is carried out to the pulse relaxation signals collected;
2) Inflexion extracting is carried out to each periodic signal of the signal of different frequency section; As follows to the method for single periodic signal extraction flex point in this step:
A. by sample frequency setting electric discharge charging coefficient;
B. on charging signals and discharge signal, the signal spacing of extracting flex point is obtained, i.e. charging wayside signaling and discharge range signal according to value;
C. to the smoothing filtering process of the charging and discharging wayside signaling obtained, remove noise, obtain the charging and discharging wayside signaling smoothly;
D. the charging and discharging wayside signaling after level and smooth is normalized, realizes the pre-service of charging and discharging wayside signaling;
E. to through pretreated charging and discharging wayside signaling coordinate axis, centered by initial point, according to the rotation of certain angle, postrotational charging and discharging wayside signaling is obtained;
F. find out the most value of charging and discharging wayside signaling after over-rotation, record the ordinate size that charging and discharging wayside signaling is worth most;
G. carrying out despining process by recording the ordinate size be worth most, in the interval normalized signal of initial charging and discharging, obtaining the position of flex point according to the value ordinate size after despining;
H. according to the corner position marked in charging and discharging normalized signal, in the interval smooth signal of initial charging and discharging, the flex point value I needing to extract is obtained;
Wherein, the disposal of gentle filter described in step c, specifically carry out according to following method:
C-1. terminate to last pulse current value from first pulse current value;
C-2. each to traveling through n pulse current value averaged backward, be used as new pulse current value with this;
C-3. the pulse current value is on average used to be combined into new pulse relaxation signals;
Normalized described in steps d, detailed process is as follows:
D-1. the maximal value Q of signal spacing is obtained
maxwith minimum value Q
min;
D-2. the pulse current value Q in ergodic signals interval
value, according to Q
maxand Q
min, obtain the pulse current value Q after normalization
value1, computing formula is as follows:
Q
value1=(Q
value-Q
min)/(Q
max-Q
min)
Charging and discharging wayside signaling after normalization, makes transverse and longitudinal coordinate values become certain relative value relation from the absolute value of physical system numerical value, for follow-up rotation is ready.
2. according to the time domain inflection point extraction method for pulse relaxation signal described in claim 1, it is characterized in that, threshold value n is preferably 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310486085.4A CN103630584B (en) | 2013-10-16 | 2013-10-16 | Time domain inflection point extraction method for pulse relaxation signal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310486085.4A CN103630584B (en) | 2013-10-16 | 2013-10-16 | Time domain inflection point extraction method for pulse relaxation signal |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103630584A CN103630584A (en) | 2014-03-12 |
| CN103630584B true CN103630584B (en) | 2015-07-15 |
Family
ID=50211857
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310486085.4A Active CN103630584B (en) | 2013-10-16 | 2013-10-16 | Time domain inflection point extraction method for pulse relaxation signal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103630584B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6507181B1 (en) * | 1999-10-19 | 2003-01-14 | Abb Substation Automation Oy | Arrangement and method for finding out the number of sources of partial discharges |
| CN101911401A (en) * | 2008-01-04 | 2010-12-08 | 曼德斯必德技术公司 | Method and apparatus for reducing optical signal speckle |
-
2013
- 2013-10-16 CN CN201310486085.4A patent/CN103630584B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6507181B1 (en) * | 1999-10-19 | 2003-01-14 | Abb Substation Automation Oy | Arrangement and method for finding out the number of sources of partial discharges |
| CN101911401A (en) * | 2008-01-04 | 2010-12-08 | 曼德斯必德技术公司 | Method and apparatus for reducing optical signal speckle |
Non-Patent Citations (2)
| Title |
|---|
| 《电子舌多频大幅脉冲伏安法弛豫响应信号的电化学机理研究》;田师一;《中国博士论文全文数据库》;20130815(第08期);1-132 * |
| 张夏宾 等.《新型伏安型多频脉冲电子舌及其应用》.《浙江大学学报(工学版)》.2008,第42卷(第10期), * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103630584A (en) | 2014-03-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102680860B (en) | Automatic fault-point locating method for traveling-wave based fault location of high-voltage electric power lines | |
| CN101828918B (en) | Electrocardiosignal R peak detection method based on waveform characteristic matching | |
| CN101797156B (en) | Decision method of respiratory wave of human body respiration monitoring system | |
| CN103989462B (en) | The extracting method of a kind of pulse wave fisrt feature point and second feature point | |
| US11337638B2 (en) | ECG signal processing method and apparatus | |
| US9204816B2 (en) | Method and system for determining body impedance | |
| CN107688553B (en) | Method for detecting electrocardiographic waveform characteristics based on wavelet transform and logistic regression algorithm | |
| CN103405227A (en) | Double-layer morphological filter based electrocardiosignal preprocessing method | |
| CN105030233B (en) | A kind of electrocardiosignal ST sections of recognition methods | |
| CN103230267B (en) | Anti-movement-interference extraction method for pulse rates | |
| CN103006207A (en) | Electrocardiosignal analysis method of T wave alternans scatter diagram method based on morphology | |
| CN101251445B (en) | Method for analysis of fractal characteristic of rotating machinery bump-scrape acoustic emission signal | |
| CN112168152A (en) | Method and device for detecting respiration and heartbeat and computer readable storage medium | |
| RU2015146506A (en) | DEVICE AND METHOD FOR REMOVING MOTION ARTHFACTS DURING ECG | |
| CN109633266A (en) | A kind of frequency measurement method, system, device and computer readable storage medium | |
| CN106214143B (en) | Transient motion interference identification method and heart rate confidence calculations method | |
| CN103083011B (en) | Electrocardio R wave crest real-time locating method assisted by chest impedance second order difference image | |
| CN103630584B (en) | Time domain inflection point extraction method for pulse relaxation signal | |
| CN104808055B (en) | A kind of digital measuring method of signal frequency | |
| CN103549947A (en) | Real-time and accurate electrocardiographic wave drawing method of smartphone platforms | |
| CN106333678B (en) | Method and system for detecting brain waves in sleep state in electroencephalogram signals | |
| CN103027692B (en) | Dynamic spectrum data processing method based on uncertainty | |
| CN110063726B (en) | Electrocardiosignal single-lead f-wave extraction method and device | |
| CN106175755B (en) | Sleep state detector for sleep state analysis | |
| Escalona et al. | Real-time system for high-resolution ECG diagnosis based on 3D late potential fractal dimension estimation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |