CN104240455B - A kind of disturbance event recognition methods in distribution type fiber-optic pipeline safety early warning system - Google Patents
A kind of disturbance event recognition methods in distribution type fiber-optic pipeline safety early warning system Download PDFInfo
- Publication number
- CN104240455B CN104240455B CN201410385097.2A CN201410385097A CN104240455B CN 104240455 B CN104240455 B CN 104240455B CN 201410385097 A CN201410385097 A CN 201410385097A CN 104240455 B CN104240455 B CN 104240455B
- Authority
- CN
- China
- Prior art keywords
- disturbance event
- wavelet
- disturbance
- decision tree
- tree classifier
- 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
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000003066 decision tree Methods 0.000 claims abstract description 35
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 18
- 238000012549 training Methods 0.000 claims abstract description 18
- 230000008569 process Effects 0.000 claims abstract description 15
- 230000003993 interaction Effects 0.000 claims abstract description 5
- 239000000284 extract Substances 0.000 claims abstract description 4
- 239000000835 fiber Substances 0.000 claims description 10
- 230000007547 defect Effects 0.000 claims description 9
- 239000013307 optical fiber Substances 0.000 claims description 9
- 230000007613 environmental effect Effects 0.000 claims description 8
- 238000001228 spectrum Methods 0.000 claims description 6
- 238000004422 calculation algorithm Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 4
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 239000012634 fragment Substances 0.000 claims 1
- 235000014676 Phragmites communis Nutrition 0.000 abstract description 7
- 238000012360 testing method Methods 0.000 description 7
- 238000013461 design Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 6
- 230000003287 optical effect Effects 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 5
- 230000008859 change Effects 0.000 description 5
- 238000003909 pattern recognition Methods 0.000 description 5
- 238000004891 communication Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 230000002085 persistent effect Effects 0.000 description 3
- 238000010276 construction Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000001149 cognitive effect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000006698 induction Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000011946 reduction process Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B29/00—Checking or monitoring of signalling or alarm systems; Prevention or correction of operating errors, e.g. preventing unauthorised operation
- G08B29/18—Prevention or correction of operating errors
- G08B29/185—Signal analysis techniques for reducing or preventing false alarms or for enhancing the reliability of the system
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B13/00—Burglar, theft or intruder alarms
- G08B13/18—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength
- G08B13/181—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength using active radiation detection systems
- G08B13/183—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength using active radiation detection systems by interruption of a radiation beam or barrier
- G08B13/186—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength using active radiation detection systems by interruption of a radiation beam or barrier using light guides, e.g. optical fibres
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Computer Security & Cryptography (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
The invention discloses the disturbance event recognition methods in a kind of distribution type fiber-optic pipeline safety early warning system, when there is disturbance event, two-way sampled signal being carried out wavelet de-noising process；Wavelet de-noising being processed Hou mono-tunnel sampled signal and extracts eigenvalue, eigenvalue includes that vibrating reed segment length, time domain energy, k rank initial point are away from, each layer low-frequency wavelet coefficients ENERGY E after k rank centre-to-centre spacing, the degree of bias, kurtosis and wavelet decompositionj, j=1 to 7；13 eigenvalues extracted are sent into decision tree classifier, utilizes decision tree classifier to draw disturbance event classification；In the case of occurring that mistake occur in new disturbance event classification or the disturbance event classification drawn at decision tree classifier, by the disturbance event classification of storage in data base is modified, realize man-machine interaction incremental learning, according to amended disturbance event classification, decision tree classifier is carried out on-line training.The present invention can accurately draw disturbance event type.
Description
Technical field
The present invention relates to the disturbance event recognition methods in a kind of distribution type fiber-optic pipeline safety early warning system, belong to technical field of optical fiber sensing.
Background technology
Development, the progress of science and technology along with society; protection for various social propertys and communal facility gradually comes into one's own; fix particularly with traffic route, energy pipeline, these positions of communication line and for the wider defence target of geographic range; its urgent demand is in the case of safety pre-warning system has extra long distance early warning; Intelligent Recognition ability to disturbance event pattern, the difficult point meeting this type of demand is: 1. have extra long distance real-time early warning and disturbance event stationkeeping ability；2. there is the Intelligent Recognition ability to disturbance event pattern, the i.e. disturbance information of intelligence process and identify；3. there is the change for environment and new type disturbance event occurs and carries out incremental learning ability.
Distribution type fiber-optic pipeline safety early warning system is based on Sagnac principle of interference, and its sensing principle is as it is shown in figure 1, position disturbance location by measuring the phase place change of the two ways of optical signals that disturbance causes along pipeline sensing optic cable.Two-beam meets in fiber optic loop, and owing to sharing same bonder, without optical path difference, if not having disturbance event to act on sensing fiber ring, then the two-beam exported does not has phase contrast, will not interfere phenomenon when bonder exports.If have a disturbance event to act on sensing fiber ring certain some, the phase place of two-beam changes generation interference at this point, and the two-way interference signal that photodetector receives can reflect the phase contrast of two-beam, i.e. contains the positional information of disturbance point.
Existing similar distributed optical fiber vibration sensing system, ensureing under disturbance event warning, location premise, for the accuracy improving disturbance event type identification, the demands such as the false positive event that the noise caused of reduction environment noise and human activity causes, use more and vibration signal is carried out Time-domain Statistics characteristic acquisition characteristic vector, again by statistical analysis, the method that characteristic component arranges threshold value distinguishes all kinds of disturbance events.Such method disadvantageously, complicated for characteristic component com-parison and analysis, the disturbance event needing the combination of multiple eigenvalue to distinguish cannot effectively be distinguished, and classification accuracy cannot ensure；In the environment of can be only applied to specify, the disturbance event classification to specifying is distinguished, it is impossible to carries out the incremental learning change with response environment and the appearance of new type disturbance event again, lacks versatility.Disturbance event recognition methods in existing distribution type fiber-optic pipeline safety early warning system also has the grader using support vector machine, the grader using support vector machine becomes more dependent on the order of accuarcy of training sample data, when certain sample error is bigger, overall training effect can be produced a very large impact, affect classification accuracy further.
Summary of the invention
The technical problem to be solved is: provide the disturbance event recognition methods in a kind of distribution type fiber-optic pipeline safety early warning system, can accurately draw disturbance event type in the case of all types of disturbance event characteristics of signals are more close.
The present invention includes following technical scheme:
Disturbance event recognition methods in a kind of distribution type fiber-optic pipeline safety early warning system, comprises the following steps:
S1, when the amplitude of two-way sampled signal exceedes setting value, it is judged that there is disturbance event, then carries out wavelet de-noising process to two-way sampled signal；
S2, processes Hou mono-tunnel sampled signal to wavelet de-noising and extracts eigenvalue, and eigenvalue includes that vibrating reed segment length, time domain energy, k rank initial point are away from, each layer low-frequency wavelet coefficients ENERGY E after k rank centre-to-centre spacing, the degree of bias, kurtosis and wavelet decompositionj, j=1 to 7；Use DB4 wavelet basis that wavelet de-noising processes Hou mono-tunnel sampled signal and be decomposed into 7 floor, each layer low-frequency wavelet coefficients ENERGY E after wavelet decompositionjComputing formula as follows:
In formula, ajL () is the l low-frequency wavelet coefficients value on wavelet decomposition jth layer yardstick, j=1 to 7, L are the low frequency coefficient length on signal wavelet decomposition jth layer yardstick；
13 eigenvalues extracted are sent into decision tree classifier, are utilized decision tree classifier to draw disturbance event classification by S3；If disturbance event classification is environmental background noise, do not report to the police, the eigenvalue of extraction is sent directly into database purchase；If disturbance event classification is intrusion event, report to the police, two-way sampled signal after processing wavelet de-noising carries out time-frequency conversion and obtains Pwelch power spectrum, thus obtain the defect spectrum of disturbance location, in composing according to defect, intrusion event in sensor fibre is positioned by the frequency values of trap wave point, finally the eigenvalue of extraction is sent into data base and stores；
In the case of occurring that mistake occur in new disturbance event classification or the disturbance event classification drawn at decision tree classifier, by the disturbance event classification of storage in data base is modified, realize man-machine interaction incremental learning, according to amended disturbance event classification, decision tree classifier is carried out on-line training.
Decision tree classifier arranges training sample space size according to the number of eigenvalue, and the size of every class disturbance event training sample is at least 5 to 10 times of the number of eigenvalue.
The method that in composing according to defect, intrusion event in sensor fibre is positioned by the frequency values of trap wave point is the position R according to equation below calculation perturbation point distance far-end reflecting module:
K is selected trap wave point sequence number, K=1,3,5；C is the light velocity, and n is refractive index, fs,nullFrequency values for selected trap wave point.
Decision tree classifier uses ID3 algorithm.
The present invention compared with prior art has the advantage that
(1) present invention combines Time-Frequency Analysis and statistical analysis means, utilize the distribution of the vibrating reed segment length of noise reduction process post-sampling signal, time domain energy, frequency domain energy, k rank initial point away from, energy corresponding to k rank centre-to-centre spacing, the degree of bias, kurtosis, signal wavelet decomposition each layer low frequency component as characteristic vector, effectively can classify for disturbing signal characteristic, improve the accuracy of perturbation mode identification.
(2) compared with other sorting techniques existing, the present invention design by decision tree classifier, it is possible to effectively complicated to characteristic component com-parison and analysis, the disturbance event needing the combination of multiple eigenvalue to differentiate carries out pattern recognition.Decision tree classifier classification speed is fast, and amount of calculation is relatively small, is easily converted to classifying rules；Classification accuracy is high, and the regular accuracy excavated from decision tree classifier is high and readily appreciates.
(3) by the design of decision tree classifier, make the incremental learning function with man-machine interaction, not only can be during the use of system, revise existing tagsort system, improve constantly the accuracy to existing perturbation mode identification, and cope with change or the appearance of new disturbance event of environment, improve the versatility of system.
Accompanying drawing explanation
Fig. 1 is to interfere distributed vibrating sensing schematic diagram based on Sagnac type；
Fig. 2 is disturbance location and algorithm for pattern recognition flow chart；
Fig. 3 is decision tree classifier algorithm effect figure of the present invention.
Detailed description of the invention
The most just combine accompanying drawing the present invention is described further.
As it is shown in figure 1, the distribution type fiber-optic pipeline safety early warning system of the present invention, including light source, the first bonder, time delay optical fiber, the second bonder, sensor fibre, far-end reflecting module, photodetector and computer.The two paths of signals of photodetector output is sent into computer and is carried out sampling acquisition two-way sampled signal, utilizes computer to carry out disturbance event pattern recognition, carries out disturbance event recognition methods in a computer as in figure 2 it is shown, specifically include following steps:
S1, when the amplitude of two-way sampled signal exceedes setting value, it is judged that there is disturbance event, then two-way sampled signal carried out wavelet de-noising process.Basic skills is that the Noise signal collected is carried out multi-scale wavelet transformation, extracts the wavelet coefficient of signal as far as possible, suppress simultaneously or even remove the wavelet coefficient belonging to noise, then use inverse wavelet transform reconstruction signal under each yardstick.After utilizing wavelet de-noising that sampled signal is processed, the interference of random noise in system be can effectively reduce, precision and the accuracy rate of disturbance event type identification of disturbance event location are effectively improved.
S2, wavelet de-noising processes Hou mono-tunnel sampled signal extracting eigenvalue, eigenvalue includes that vibrating reed segment length, time domain energy, k rank initial point are away from, each layer low frequency component is corresponding after k rank centre-to-centre spacing, the degree of bias, kurtosis, wavelet decomposition energy.By disturbing signal is carried out time-frequency domain specificity analysis and statistical characteristic analysis obtain vibrating reed segment length, time domain energy, k rank initial point away from, k rank centre-to-centre spacing, skewness and kurtosis；By each layer low-frequency wavelet coefficients energy after Multiscale Wavelet Decomposition acquisition wavelet decomposition.Vibrating reed segment length is to characterize the characteristic quantity of disturbing signal persistent period length, can arrange threshold value by the amplitude of sampled signal and obtain the initial time of disturbance event.Time domain energy is to be characterized in the characteristic quantity of the intensity of signal in the disturbing signal persistent period, can obtain by the absolute value of the amplitude of signal in the persistent period is carried out summation.
The square of data divides can be initial point away from centre-to-centre spacing:
K rank initial point away from for:
K rank centre-to-centre spacing is:
In above formula, k be initial point away from the exponent number with centre-to-centre spacing, N is number of samples, xiFor ith sample data value,For sample mean.K value, away from, k rank centre-to-centre spacing, can be adjusted by k rank initial point according to perturbation mode classifying quality.K typically takes 3-5.
The degree of bias and kurtosis are to portray the skewness of sampled data, the index of cabrage degree, relevant with the square of data.
Degree of bias g1Computing formula be:
In above formula, s is that data standard is poor, u3It is 3 rank centre-to-centre spacing.The degree of bias is the index portraying data symmetry, and its degree of bias of data that average is symmetrical is 0, and the more dispersed data degree of bias in right side is just, the more dispersed data degree of bias in left side is negative.
Kurtosis g2Computing formula be:
When the population distribution of data is normal distribution, kurtosis is approximately 0；When being distributed more dispersed compared with the afterbody of normal distribution, kurtosis is just, otherwise kurtosis is negative.When kurtosis is timing.Both sides Outliers is more, when kurtosis is for time negative, and both sides Outliers is less.
The present invention uses DB4 wavelet basis that wavelet de-noising processes Hou mono-tunnel sampled signal and is decomposed into 7 floor, by each for signal layer low-frequency wavelet coefficients ENERGY EjAs signal characteristic value, formula is as follows:
In formula, ajL () is the l low-frequency wavelet coefficients value on signal wavelet decomposition jth layer yardstick, j=1 to 7, L are the low frequency coefficient length on signal wavelet decomposition jth layer yardstick.Through field test, finding to utilize DB4 wavelet basis to extract eigenvalue after signal carries out 7 layers of wavelet decomposition, perturbation mode classifying quality is best.
S3, eigenvalue step S2 obtained send into decision tree classifier, draw disturbance event classification.
Decision tree classifier uses top-down recursive fashion, according to eigenvalue, disturbance event is classified, in categorizing process, internal node decision tree carries out the comparison of eigenvalue, and judge from the branch that this internal node is downward according to different characteristic value, until reaching certain leaf node, thus find the classification belonging to this disturbance event.Each internal node of decision tree represents the once test to a certain eigenvalue, and each edge represents a test process, and each leaf node represents the distribution of certain classification.
If disturbance event classification is environmental background noise, then system is not reported to the police, and only the eigenvalue of signal is sent directly into database purchase.
If disturbance event classification is intrusion event, then report to the police, two-way sampled signal after processing wavelet de-noising carries out time-frequency conversion and obtains Pwelch power spectrum, thus obtain the defect spectrum of disturbance location, in composing according to defect, intrusion event in sensor fibre is positioned by the frequency values of trap wave point, finally the eigenvalue of extraction is sent into data base and stores.The method that in composing according to defect, intrusion event in sensor fibre is positioned by the frequency values of trap wave point is: according to the position R of equation below calculation perturbation point distance far-end reflecting module:
In above formula, K is selected trap wave point sequence number, K=1,3,5；C is the light velocity, and n is refractive index, fs,nullFrequency values for selected trap wave point.
If disturbance event classification is not one of known intrusion event classification, while to intrusion event locating alarming, primary signal and characteristic vector after treatment are sent directly into data base, after the new classification disturbance event labelling that it has been stored in data base by domain expert or duty personnel, re-start the design of decision tree classifier again, it is achieved man-machine interaction incremental learning.
In the incremental learning link of grader of the present invention, mainly include the incremental learning to old classification new samples and two processes of incremental learning to the new samples belonging to new classification.Owing to the collection of sample is constantly accumulation, an abundant and perfect process.In order to obtain higher nicety of grading, fashionable when there being new samples to add, newly-increased sample and old sample are re-started training altogether.For decision tree classifier, owing to its construction algorithm is carried out on the basis of grasping global information, and after newly increasing a class, relation between every class may all there occurs change, decision tree classifier may and non-optimal solution be redesigned to reach optimal solution for new global information in the way of incremental learning by old tree construction in the case.
Decision tree classifier arranges training sample space size according to the number of eigenvalue, and for ensureing classifier design effect, the size of general every class disturbance event training sample is at least 5 to 10 times of the number of eigenvalue.
Native system combines data base and disturbance event is carried out category label, thus obtain the disturbance event record with time order and function order, system provides security alarm information according to different field, different grades of warning demand, thus obtains about applied environment the most cognitive；While utilizing known signal characteristic and environmental information to classify, existing tagsort system can be revised automatically according to domain expert or duty personnel to the demarcation of safety early warning signal and decision-making again.
Example
In instances, data acquisition was tested from Shandong CHINAUNICOM Weifang to Anqiu section communications optical cable second phase, test optical fibre cables length 45 kilometers, and wherein in section communication optical cable laying underground piping, optic cable is maked somebody a mere figurehead.Native system sample frequency is 1.5MHz, and number of samples N is 50000.
According to use demand, disturbance event is divided into tactile cable, environmental disturbances, environment noise three types, in system operation, finds that environmental disturbances, essentially from the one section of aerial optical cable walked through the woods, produces disturbance trigger warning due to branch touching optical cable.Staff combines the information such as disturbing signal waveform and disturbance location, and the disturbing signal data that recent months picks up from test optical fibre cables are analyzed, and in system database, disturbance event are carried out category label.Owing to system utilizes 13 eigenvalues to classify disturbing signal, for ensureing classifier design effect, every kind of disturbance event classification number of training is set to 150, carries out decision tree classifier training, demarcated classification training sample eigenvalue as shown in table 1.Feature 1, feature 2 to feature 13 are followed successively by vibrating reed segment length, time domain energy, k rank initial point away from, each layer low frequency component is corresponding after k rank centre-to-centre spacing, the degree of bias, kurtosis, wavelet decomposition energy.
Classification training sample eigenvalue demarcated by table 1
Utilization has been demarcated classification training sample and can have been obtained decision tree classifier as shown in Figure 3, as shown in Figure 3, participate in characteristic of division and be chosen as feature 2, feature 4, feature 5, feature 6, feature 7, feature 9, feature 11, feature 12, feature 13, respectively to induction signal time domain energy, k rank initial point away from, k rank centre-to-centre spacing, the degree of bias, kurtosis and corresponding 3rd, 5,6, the 7 layers of low-frequency wavelet coefficients energy of signal wavelet decomposition.
For verifying the effectiveness of system disturbance pattern recognition function, select the signal of three class disturbance trigger events respectively, and calculate characteristic of division value and send in the decision tree classifier shown in Fig. 3, the eigenvalue of signal is judged by each node of decision tree classifier, judge from the branch that this internal node is downward according to different characteristic value, until reaching certain leaf node, thus find the classification belonging to this disturbance event, decision tree classifier output result is respectively " touching cable ", " environmental disturbances ", " environment noise ", demonstrate the effectiveness of system disturbance event schema recognition methods.Wherein, the characteristic of division value of three class disturbance trigger events is as shown in table 2.
Table 2 three class disturbance trigger event characteristic of division value
| Classification | Feature 2 | Feature 4 | Feature 5 | Feature 6 | Feature 7 | Feature 9 | Feature 11 | Feature 12 | Feature 13 |
| Touch cable | 1012 | 25 | -0.00137607 | 1.36537 | 0.147763 | 0.147714 | 0.145013 | 0.140563 | 0.124105 |
| Environmental disturbances | 396 | 2 | 0.00308111 | 12.5099 | 0.143769 | 0.143766 | 0.143689 | 0.143289 | 0.137979 |
| Environment noise | 87 | 9 | 0.00104344 | 0.0260809 | 0.142924 | 0.142922 | 0.142861 | 0.142744 | 0.142719 |
The noisy data training complete decision tree classifier that recent months picks up from test optical fibre cables is finally utilized to carry out pattern recognition checking, and carry out the on-the-spot test of perturbation mode identification in real time, during final system runs, perturbation mode discrimination is more than 95%, reach desired design effect, meet actual application demand.Thus also demonstrate reliability and the effectiveness of the disturbance event recognition methods in distribution type fiber-optic pipeline safety early warning system proposed by the invention.
The unspecified part of the present invention belongs to general knowledge as well known to those skilled in the art.
Claims (4)
1. the disturbance event recognition methods in distribution type fiber-optic pipeline safety early warning system, described distributed
Optical fiber pipe safety pre-warning system includes light source, the first bonder, time delay optical fiber, the second bonder, sensing
Optical fiber, far-end reflecting module, photodetector and computer, the two paths of signals of photodetector output is sent into
Computer carries out sampling and obtains two-way sampled signal, it is characterised in that carry out disturbance event knowledge in a computer
Method for distinguishing comprises the following steps:
S1, when the amplitude of two-way sampled signal exceedes setting value, it is judged that there is disturbance event, then to two
Road sampled signal carries out wavelet de-noising process；
S2, processes Hou mono-tunnel sampled signal to wavelet de-noising and extracts eigenvalue, and eigenvalue includes vibrating fragment
Length, time domain energy, k rank initial point are away from, each layer low frequency after k rank centre-to-centre spacing, the degree of bias, kurtosis and wavelet decomposition
Wavelet coefficient ENERGY Ej, j=1 to 7；Use DB4 wavelet basis that wavelet de-noising processes Hou mono-tunnel sampling letter
Number it is decomposed into 7 floor, each layer low-frequency wavelet coefficients ENERGY E after wavelet decompositionjComputing formula as follows:
In formula, ajL () is the l low-frequency wavelet coefficients value on wavelet decomposition jth layer yardstick, j=1 to 7, L
For the low frequency coefficient length on signal wavelet decomposition jth layer yardstick；
13 eigenvalues extracted are sent into decision tree classifier, are utilized decision tree classifier to draw disturbance by S3
Event category；If disturbance event classification is environmental background noise, do not report to the police, by direct for the eigenvalue extracted
Send into database purchase；If disturbance event classification is intrusion event, report to the police, after wavelet de-noising is processed
Two-way sampled signal carry out time-frequency conversion obtain Pwelch power spectrum, thus obtain disturbance location defect
Spectrum, in composing according to defect, intrusion event in sensor fibre is positioned by the frequency values of trap wave point, finally will carry
The eigenvalue taken is sent into data base and is stored；
Occurring that mistake occur in new disturbance event classification or the disturbance event classification drawn at decision tree classifier
In the case of Wu, by the disturbance event classification of storage in data base is modified, realize man-machine interaction
Incremental learning, carries out on-line training according to amended disturbance event classification to decision tree classifier.
2. disturbance event recognition methods as claimed in claim 1, it is characterised in that decision tree classifier root
Arranging training sample space size according to the number of eigenvalue, the size of every class disturbance event training sample is at least
5 to 10 times of the number of eigenvalue.
3. disturbance event recognition methods as claimed in claim 1, it is characterised in that fall in composing according to defect
The method that intrusion event in sensor fibre is positioned by the frequency values of wave point is for according to equation below calculation perturbation
The position R of some distance far-end reflecting module:
K is selected trap wave point sequence number, K=1,3,5；C is the light velocity, and n is refractive index, fs,nullFall into for selected
The frequency values of wave point.
4. disturbance event recognition methods as claimed in claim 1, it is characterised in that decision tree classifier is adopted
Use ID3 algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410385097.2A CN104240455B (en) | 2014-08-07 | 2014-08-07 | A kind of disturbance event recognition methods in distribution type fiber-optic pipeline safety early warning system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410385097.2A CN104240455B (en) | 2014-08-07 | 2014-08-07 | A kind of disturbance event recognition methods in distribution type fiber-optic pipeline safety early warning system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104240455A CN104240455A (en) | 2014-12-24 |
| CN104240455B true CN104240455B (en) | 2016-08-17 |
Family
ID=52228444
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410385097.2A Active CN104240455B (en) | 2014-08-07 | 2014-08-07 | A kind of disturbance event recognition methods in distribution type fiber-optic pipeline safety early warning system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104240455B (en) |
Families Citing this family (41)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104568122B (en) * | 2015-01-21 | 2017-10-31 | 东南大学 | A kind of distributed optical fiber vibration sensing system disturbance event identification and localization method |
| CN104729667B (en) * | 2015-03-25 | 2017-11-07 | 北京航天控制仪器研究所 | A kind of disturbance kind identification method in distributed optical fiber vibration sensing system |
| CN105184319B (en) * | 2015-09-02 | 2019-07-30 | 同方威视技术股份有限公司 | Recognition methods, device and the circumference intrusion alarm system of optical fiber perimeter invasion signal |
| CN105204084B (en) * | 2015-09-10 | 2018-02-09 | 北方工业大学 | Optical Fiber Intrusion Signal Recognition Method Based on LDA Algorithm Model |
| CN105654645B (en) * | 2016-01-22 | 2018-04-10 | 北京北邮国安技术股份有限公司 | A kind of optical fiber security signal processing method and system |
| CN105703832B (en) * | 2016-03-01 | 2018-08-31 | 深圳艾瑞斯通技术有限公司 | A kind of fiber-optic signal processing method and processing device |
| CN105841793A (en) * | 2016-04-15 | 2016-08-10 | 深圳艾瑞斯通技术有限公司 | Optical fiber vibration source identification method, device and system |
| CN106504451B (en) * | 2016-11-14 | 2018-08-14 | 中国人民解放军国防科学技术大学 | A kind of demodulating algorithm of the optical fiber perimeter safety signal based on matrix theory |
| CN106706109B (en) * | 2016-12-15 | 2019-10-01 | 北方工业大学 | A vibration source identification method and system based on time-domain two-dimensional characteristics |
| CN106683305A (en) * | 2017-01-22 | 2017-05-17 | 唐山市镜圆科技有限公司 | Anti-eavesdrop optical-fiber alarm system |
| CN107270952B (en) * | 2017-07-27 | 2020-03-31 | 天津求实飞博科技有限公司 | Long-distance optical fiber distributed disturbance sensing signal processing method based on optical frequency domain reflection |
| CN107590516B (en) * | 2017-09-16 | 2020-09-22 | 电子科技大学 | Leak detection and identification method of gas pipeline based on optical fiber sensing data mining |
| CN107862821A (en) * | 2017-12-12 | 2018-03-30 | 上海师范大学 | A kind of perimeter security system |
| CN108197646A (en) * | 2017-12-28 | 2018-06-22 | 中国电子科技集团公司第五十四研究所 | A kind of target classification identification method for distributed optical fiber sensing system |
| CN108961641B (en) * | 2018-07-24 | 2021-03-19 | 民航成都电子技术有限责任公司 | Method for reducing false alarm of capacitance enclosure alarm system based on classification tree |
| CN109374116B (en) * | 2018-12-07 | 2020-12-29 | 武汉理工光科股份有限公司 | Excavation behavior identification method of buried optical fiber sensing vibration detection system |
| CN109374115B (en) * | 2018-12-07 | 2021-08-03 | 武汉理工光科股份有限公司 | Oil-gas pipeline external damage vibration monitoring and identifying method based on phi-OTDR |
| CN109489800A (en) * | 2018-12-14 | 2019-03-19 | 广东世港信息科技有限公司 | A kind of disturbance event recognition methods in distribution optic cable vibration safety pre-warning system |
| CN110275896B (en) * | 2019-05-28 | 2021-07-20 | 深圳供电局有限公司 | Optical cable intrusion construction event identification method, device, device and readable storage medium |
| CN110146116B (en) * | 2019-06-19 | 2021-04-27 | 南昌航空大学 | Positioning method for Sagnac optical fiber sensing under multipoint disturbance |
| CN110837856B (en) * | 2019-10-31 | 2023-05-30 | 深圳市商汤科技有限公司 | Neural network training and target detection method, device, equipment and storage medium |
| CN111121946B (en) * | 2019-12-18 | 2021-09-28 | 东南大学 | Method for accurately determining abnormal value at multiple points in large dynamic range and large discrete single area |
| CN113283455B (en) * | 2020-02-20 | 2024-09-06 | 中国石油天然气股份有限公司 | Working condition identification method, device, computer equipment and storage medium |
| CN111911818A (en) * | 2020-07-10 | 2020-11-10 | 中海广东天然气有限责任公司 | A sonic leak monitoring method based on skewness and kurtosis signal features |
| CN112179475B (en) * | 2020-08-20 | 2021-09-07 | 电子科技大学 | A Multi-source Aliasing Distributed Optical Fiber Vibration Sensing Signal Separation Method |
| CN111951505B (en) * | 2020-08-25 | 2022-02-08 | 青岛大学 | Fence vibration intrusion positioning and mode identification method based on distributed optical fiber system |
| US11501616B2 (en) * | 2020-09-11 | 2022-11-15 | Network Integrity Systems, Inc. | Test device for verifying operation of an optical fiber monitoring system utilizing a far end optical shutter |
| US11502751B2 (en) | 2020-09-11 | 2022-11-15 | Network Integrity Systems, Inc. | Test device for verifying operation of an optical fiber monitoring system organized in a common housing |
| US11631308B2 (en) | 2020-09-11 | 2023-04-18 | Network Integrity Systems, Inc. | Test device for verifying operation of an optical fiber monitoring system by injection of longitudinal strain |
| US11501617B2 (en) | 2020-09-11 | 2022-11-15 | Network Integrity Systems, Inc. | Test device for verifying operation of an optical fiber monitoring system |
| US11515940B2 (en) | 2020-09-11 | 2022-11-29 | Network Integrity Systems, Inc. | Test device for verifying operation of an optical fiber monitoring system |
| US11502748B2 (en) | 2020-09-11 | 2022-11-15 | Network Integrity Systems, Inc. | Test device for verifying operation of an optical fiber monitoring system utilizing complex test signal |
| US11502750B2 (en) | 2020-09-11 | 2022-11-15 | Network Integrity Systems, Inc. | Test device for verifying operation of an optical fiber monitoring system protecting multiple fiber runs |
| US11502749B2 (en) | 2020-09-11 | 2022-11-15 | Network Integrity Systems, Inc. | Test device for verifying operation of an optical fiber monitoring system utilizing outgoing communication |
| CN113390447B (en) * | 2021-06-17 | 2022-03-29 | 复旦大学 | Frequency response compensation method of optical fiber differential interference sensing system based on frequency response compensation |
| CN113988170B (en) * | 2021-10-26 | 2022-12-23 | 国家石油天然气管网集团有限公司 | Event classification machine learning algorithm of optical fiber early warning system |
| CN114268365B (en) * | 2021-12-02 | 2023-07-11 | 国网甘肃省电力公司酒泉供电公司 | Communication optical cable intelligent early warning method and system based on visualization technology |
| CN115060184B (en) * | 2022-05-18 | 2024-07-16 | 武汉迪信达科技有限公司 | Optical fiber perimeter intrusion detection method and system based on recursion diagram |
| CN114943257B (en) * | 2022-06-02 | 2024-09-20 | 常熟理工学院 | Synchronous noise reduction and disturbance classification identification method for distributed optical fiber sensing system |
| CN114910156B (en) * | 2022-06-08 | 2024-11-08 | 上海市信息管线有限公司 | A method, device, equipment and storage medium for monitoring disturbance of underground pipeline |
| CN116091897A (en) * | 2023-03-14 | 2023-05-09 | 山东省计算中心（国家超级计算济南中心） | Distributed optical fiber sensing event identification method and system based on light weight |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101556724B (en) * | 2008-04-09 | 2011-06-01 | 中国科学院半导体研究所 | Optical Fiber Perimeter Security Management System and Its Pattern Recognition Method |
| US8705020B2 (en) * | 2010-01-30 | 2014-04-22 | Optellios, Inc. | Method and apparatus for disturbance detection |
| CN102280001B (en) * | 2011-07-29 | 2013-01-16 | 电子科技大学 | Distributed optical fiber fence intrusion detection and location method based on phi-OTDR (Optical Time Domain Reflectometer) |
| CN202404694U (en) * | 2011-09-23 | 2012-08-29 | 无锡科晟光子科技有限公司 | Adaptive disturbance signal identification module of distributing type optical fiber sensing application system |
| CN102401667B (en) * | 2011-09-29 | 2016-03-30 | 北京航空航天大学 | There is optical fiber distributed perturbation method for sensing and the system of disturbance character recognition function |
-
2014
- 2014-08-07 CN CN201410385097.2A patent/CN104240455B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN104240455A (en) | 2014-12-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104240455B (en) | A kind of disturbance event recognition methods in distribution type fiber-optic pipeline safety early warning system | |
| CN109489800A (en) | A kind of disturbance event recognition methods in distribution optic cable vibration safety pre-warning system | |
| CN104729667B (en) | A kind of disturbance kind identification method in distributed optical fiber vibration sensing system | |
| CN103617684B (en) | Interference-type optical fiber circumference vibrating intruding recognizer | |
| CN103968933B (en) | Fiber-optic vibration signal recognition method based on fuzzy field feature | |
| Zhang et al. | Time series prediction of microseismic multi-parameter related to rockburst based on deep learning | |
| Lin et al. | Power quality disturbance feature selection and pattern recognition based on image enhancement techniques | |
| CN108932480A (en) | The study of distributing optical fiber sensing signal characteristic and classification method based on 1D-CNN | |
| CN105606198A (en) | Fiber vibration sensing system two-order signal feature extraction and determining method | |
| CN109357747B (en) | An Online Train Identification and Speed Estimation Method Based on Optical Fiber Vibration Signal | |
| CN116805061A (en) | Leakage event judging method based on optical fiber sensing | |
| CN110533007A (en) | A kind of vehicle-mounted strain of bridge influences the identification of line feature intelligent and extracting method | |
| Alsina-Pagès et al. | Anomalous events removal for automated traffic noise maps generation | |
| CN117195750B (en) | A method for constructing a landslide disaster sensitivity model with reference to time series deformation | |
| CN111539393A (en) | Oil-gas pipeline third-party construction early warning method based on EMD decomposition and LSTM | |
| Cipollone et al. | Benchmarking the mesoscale variability in global ocean eddy-permitting numerical systems | |
| Zhong et al. | Recognition and prediction of ground vibration signal based on machine learning algorithm | |
| Zhang et al. | Unsupervised learning method for events identification in φ-OTDR | |
| CN104964736B (en) | Identification method of optical fiber intrusion vibration source based on maximum expectation classification of time-frequency characteristics | |
| CN107576380A (en) | A kind of three-dimensional vibrating Modulation recognition method towards Φ OTDR techniques | |
| Jia et al. | Rapid classification of local seismic events using machine learning | |
| Lyu et al. | Two-stage intrusion events recognition for vibration signals from distributed optical fiber sensors | |
| Chen et al. | Coal mine microseismic identification and first-arrival picking based on Conv-LSTM-Unet | |
| CN110020190B (en) | A method and system for verifying suspicious threat indicators based on multi-example learning | |
| Qiu et al. | A new detection method based on CFAR and DE for OFPS |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | 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 |