CN106503642B - A kind of model of vibration method for building up applied to optical fiber sensing system - Google Patents
A kind of model of vibration method for building up applied to optical fiber sensing system Download PDFInfo
- Publication number
- CN106503642B CN106503642B CN201610907115.8A CN201610907115A CN106503642B CN 106503642 B CN106503642 B CN 106503642B CN 201610907115 A CN201610907115 A CN 201610907115A CN 106503642 B CN106503642 B CN 106503642B
- Authority
- CN
- China
- Prior art keywords
- optical fiber
- signal
- vibration
- fiber
- model
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Abstract
The present invention provides a kind of method for building up of model of vibration applied to optical fiber sensing system, the following steps are included: vibration signal is adjusted to optical fiber interference signal output by fiber-optic vibration device, it is that the second reference axis establishes signal condition figure that the distance for combining the time for generating signal using the signal strength of output, signal period, generating signal, which is the first reference axis, time at a distance from optical fiber starting point with fiber position,;Acquire the signal condition figure under different Vibration Conditions, artificial mark label is carried out to these data, the data information of artificial mark label and signal condition figure is input to convolutional neural networks model and carries out image recognition and CNN training, obtains the model of vibration for being applied to optical fiber sensing system;Image recognition and the CNN training of the convolutional neural networks model include: the first stage, forward propagation stage;Second stage, back-propagation stage.Technical solution of the present invention has the characteristics that discrimination is high, identification accuracy is good, recognition capability is strong.
Description
Technical field
The invention belongs to technical field of optical fiber sensing more particularly to a kind of models of vibration applied to optical fiber sensing system
Method for building up.
Background technique
With the rapid development of society, important has more and more been shown to the long-range monitoring and protection of some important areas, such as
Airport, petroleum pipeline, military base, nuclear power plant, prison, bank etc..It is heavy that fiber-optic vibration safety pre-warning system can acquire these
The various vibration signals for wanting area peripheral edge obtain vibration source type, if detecting to region by analyzing periphery vibration signal characteristics
Harmful vibration source occurs, and can carry out early warning in time, and report the specific location of hazardous events, reach to important area or region
The real-time guard on periphery, the purpose for reducing property loss.Optical fiber is to very sensitive from extraneous vibration signal, if not to adopting
The vibration signal collected is handled, and be will lead to optical fiber and is carried out indiscriminate alarm to different classes of vibration event, bring compared with
High false alarm rate reduces the practicability of optical fiber safety early-warning system, and carrying out identification to vibration signal is a kind of feasible processing
Means.Currently, the research in relation to vibration signal processing and identification mainly has the blind recognition skill of blind source separate technology, vibration signal
Art, wavelet analysis and transformation, linear classifier support vector machines etc..Applied to optical fiber early warning identifying system mainly using
DTW algorithm.Due to the continuous development of system, recognition result is required constantly to be promoted, DTW algorithm is no longer satisfied system and needs
It asks, existing main problem is the following:
1, current recognizer discrimination is lower, not can be carried out effective identification.
2, under the big environment of noise, discrimination is remarkably decreased, and not can be carried out accurate identification.
3, fiber distance is remoter, and the oscillation intensity for collecting data is lower, and recognition result is remarkably decreased.
4, model is fairly simple, cannot handle more complicated identification problem.
Summary of the invention
Against the above technical problems, the invention discloses a kind of foundation sides of model of vibration applied to optical fiber sensing system
Method carries out attributive character analysis according to the collected different vibration signal of fiber-optic vibration safety pre-warning system, and establishes corresponding
Characteristic model Recognition of Vibration Sources is carried out to collected vibration signal, substantially increases identification by neural network recognization algorithm
Accuracy, have higher discrimination.
In this regard, the technical solution of the present invention is as follows:
A kind of method for building up of the model of vibration applied to optical fiber sensing system, comprising the following steps:
Step S1: vibration signal is adjusted to by optical fiber interference signal output by fiber-optic vibration device, utilizes the signal of output
The distance that intensity, signal period combine the time for generating signal, generate signal, with fiber position for the with optical fiber starting point at a distance from
One reference axis, time are that the second reference axis establishes signal condition figure;Wherein, if the signal condition figure is for characterizing the optical fiber
Information of the dry fiber position in several temporal signal values;
Step S2: acquiring the signal condition figure under different Vibration Conditions, artificial mark label is carried out to these data, by people
The data information of work mark label and signal condition figure is input to convolutional neural networks model and carries out image recognition and CNN training,
Obtain the model of vibration for being applied to optical fiber sensing system;
Wherein, the image recognition of the convolutional neural networks model and CNN training the following steps are included:
First stage, propagation stage forward, detailed process is as follows:
Multi-class problem, total c class, total N number of training sample, error term are discussed using square error cost function are as follows:
In formula (1),Indicate the vector of the kth dimension of the corresponding label of n-th of sample,Indicate that n-th of sample is corresponding
Network output k-th output;C is total class of representative sample;For multi-class problem, exporting general tissue is " one-of-c "
Form, that is, the output node output of the only corresponding class of the input is positive, and the position of other classes or node are 0 or negative
Number, the activation primitive depending on your output layer;Sigmoid is exactly that 0, tanh is exactly -1;
It is the summation of the error of each training sample based on the error on whole training sets, for n-th sample
Error indicates are as follows:
Then partial derivative of the cost function E about each weight of network is calculated according to BP rule;It is indicated with l current
Layer, the output of current layer can indicate are as follows:
Xl=f (ul),with,ul=Wlxl-1+bl (3)
In formula, X represents output, and u represents input, and w represents weight, and b represents biasing, and wherein which of network l represent
Layer;
Second stage, in the back-propagation stage, detailed process is as follows:
The error that backpropagation is returned, i.e., the sensitivity of the base of each neuron are as follows:
The sensitivity that l layers of backpropagation are as follows:
δl=(Wl+1)Tδl+1οf'(ul) (5)
In formula (5), T represents matrix transposition, and W represents weight, and u represents input, is neuron signal in the technical program
Input, " ο " indicate each element multiplication;
The sensitivity of the neuron of output layer are as follows:
δL=f ' (uL)ο(yn-tn). (6)
Finally, carrying out right value update with delta rule to each neuron;It is stated with the form of vector, for l
Layer, the derivative that error is combined into matrix for this layer each weight, that is, group are the input and the sensitivity of this layer of this layer are as follows:
In formula (7), wherein l indicates the number of plies of neural network, and representative is which layer X is vector;W is l layers of weight,
η is learning rate, and δ indicates sensitivity;There is a specific learning rate for each weight.
Wherein, the convolutional neural networks model includes convolutional layer and pond layer, wherein C layers are characterized extract layer, each
The input of neuron is connected with the local receptor field of preceding layer, and extracts the feature of the part, once the local feature is extracted
Afterwards, its positional relationship between other features is also decided therewith;S layers are Feature Mapping layers, each computation layer of network by
Multiple Feature Mapping compositions, each Feature Mapping are a plane, and the weight of all neurons is equal in plane.Feature Mapping knot
Activation primitive of the structure using the small sigmoid function of influence function core as convolutional network, so that Feature Mapping has displacement not
Denaturation.
Input picture by and three trainable filters and can biasing set carry out convolution, in C1 layer generation three after convolution
A Feature Mapping figure, then every group of four pixels are summed again in Feature Mapping figure, and weighted value, biasing is set, and passes through one
Sigmoid function obtains three S2 layers of Feature Mapping figure.These mapping graphs obtain C3 layers into filtering excessively again.This hierarchical structure
S4 is generated as S2 again.Finally, these pixel values are rasterized, and are connected into a vector and be input to traditional nerve net
Network is exported.
Further, since the neuron on a mapping face shares weight, thus reduce the number of network freedom parameter, drops
The complexity of low network parameter selection.Each of convolutional neural networks feature extraction layer (C- layers) all followed by use
Seek the computation layer (S- layers) of local average and second extraction, this distinctive structure of feature extraction twice makes network in identification
There is higher distortion tolerance to input sample.
There is convolutional neural networks model powerful learning ability and efficient feature representation ability, more important advantage to be
Information is successively extracted from Pixel-level initial data to abstract semantic concept, this makes it extract the global characteristics of image and upper
There is advantage outstanding in terms of context information.
This technical solution, using based on convolutional neural networks (Convolutional Neural Networks, CNN)
Image recognition algorithm carries out model foundation, and the recognizer with higher efficiency improves discrimination.Firstly, being shaken using optical fiber
The effective informations such as dynamic signal, fiber-optic vibration intensity, distance, time, generate fiber-optic signal state diagram, preferably thermodynamic chart, finally
By the image recognition algorithm based on CNN, the training of model is carried out to the picture of a large amount of thermodynamic chart, obtains being applied to optical fiber biography
The model of vibration of sensing system.Identification classification, preferably thermodynamic chart are carried out just to new signal condition figure using the model that training obtains
Available oscillatory type.The technical program proposes that the optical fiber early warning based on optical fiber vibration sensing signal thermodynamic chart identifies and calculates
Method carries out attributive character analysis according to the collected different vibration signal of fiber-optic vibration safety pre-warning system, and establishes corresponding
Vibration performance model Recognition of Vibration Sources is carried out to collected vibration signal, is substantially increased by neural network recognization algorithm
The accuracy of identification.
As a further improvement of the present invention, it is described with fiber position at a distance from optical fiber starting point for the first reference axis, when
Between establish signal condition figure for the second reference axis, comprising:
If being at a distance from optical fiber starting point with fiber position the first reference axis, time is that the second reference axis, fiber position exist
Dry temporal signal value is that third reference axis establishes signal condition figure.
As a further improvement of the present invention, the signal condition figure is thermodynamic chart, wherein the pixel in the thermodynamic chart
Value is for characterizing several fiber positions of the optical fiber in the information of several temporal signal values.
As a further improvement of the present invention, to thermodynamic chart by color from shallowly to deep come indicate shockproofness by it is small to
Greatly.
The invention also discloses a kind of determination methods of the model of vibration of optical fiber sensing system, comprising the following steps:
Step A: reception optical fiber electric signal obtains and corresponds to several fiber positions in the fiber optic telecommunications number at several
Between signal value;Wherein, the fiber optic telecommunications number are converted to by the optical signal of fiber reflection;
Step B: being permanent reference axis, time as axis of ordinates to establish thermodynamic chart at a distance from optical fiber starting point using fiber position;
Pixel value in the thermodynamic chart is for characterizing several fiber positions of the optical fiber in the letter of several temporal signal values
Breath;
Step C: pre-processing thermodynamic chart, then according to claim 1 obtain described in be applied to Fibre Optical Sensor system
The model of vibration of system carries out control identification, obtains current model of vibration classification.Wherein, the thermodynamic chart of generation is located in advance
Reason, filters out a part of noise, is conducive to identify.
As a further improvement of the present invention, in step A, fiber-optic vibration threshold values is set, if fiber optic telecommunications number exceed light
Fibre vibration threshold values, then without subsequent step.
As a further improvement of the present invention, the fiber position and the optical fiber starting point are calculated using L=(c*T/n)/2
The distance between;Wherein, c is the speed of light in a vacuum, and T is that time of the fiber position in the fiber optic telecommunications number is long
Degree, n are the refractive index of the optical fiber.
Compared with prior art, the invention has the benefit that
Using technical solution of the present invention, by utilizing the suitable convolutional neural networks mould of convolutional neural networks algorithms selection
Type, the signal condition figure generated to the fiber-optic vibration signal of acquisition identify, establish model of vibration, and produce using historical data
Raw thermodynamic chart is trained model, ginseng is adjusted, and identified to new thermodynamic chart with trained model, compared to biography
The DTW algorithm of system increases significantly on discrimination, and identification accuracy is good, and recognition capability is strong and applicability is wide.Experimental result
Show that this method can be very good identification vibration, fully meets safety-security area and demand is identified to intrusion behavior.
Detailed description of the invention
Fig. 1 is a kind of flow chart of the method for building up of the model of vibration applied to optical fiber sensing system of the present invention.
Fig. 2 is the structural schematic diagram of convolutional neural networks model of the present invention.
Specific embodiment
With reference to the accompanying drawing, preferably embodiment of the invention is described in further detail.
A kind of method for building up of the model of vibration applied to optical fiber sensing system, comprising the following steps:
Step S1: being adjusted to optical fiber interference signal for vibration signal by fiber-optic vibration device and export, reception optical fiber electric signal,
If being not above preset vibration threshold values, using output signal strength, the signal period combine generate signal when
Between, generate signal distance, be that axis of ordinates establishes heating power as axis of abscissas, time at a distance from optical fiber starting point using fiber position
Figure;Pixel value in the thermodynamic chart is for characterizing the optical fiber several fiber positions in several temporal signal values
Information, thermodynamic chart is by color from shallowly indicating that shockproofness is ascending to deep.
Step S2: acquiring the thermodynamic chart under different Vibration Conditions, carries out artificial mark label to these data, will manually mark
The data information of note label and signal condition figure is input to convolutional neural networks model and carries out image recognition and CNN training, obtains
The model of vibration applied to optical fiber sensing system.Wherein, the label manually marked is preferably hand digging, mechanical digging
Pick, animal are passed through, machinery passes through.
Wherein, the image recognition of the convolutional neural networks model and CNN training the following steps are included:
First stage, propagation stage forward, detailed process is as follows:
Multi-class problem, total c class, total N number of training sample, error term are discussed using square error cost function are as follows:
In formula (1),Indicate the vector of the kth dimension of the corresponding label of n-th of sample,Indicate that n-th of sample is corresponding
K-th of output of network output;C is total class of representative sample;For multi-class problem, exporting general tissue is " one-of-c "
Form, that is, the output node output of the only corresponding class of the input are positive, and the position of other classes or node are 0 or bear
Number, the activation primitive depending on your output layer;Sigmoid is exactly that 0, tanh is exactly -1;
Because the error on whole training sets is the summation of the error of each training sample, we are first examined here
Consider the BP for a sample.For the error of n-th of sample, indicate are as follows:
Then partial derivative of the cost function E about each weight of network is calculated according to BP rule;It is indicated with l current
Layer, the output of current layer can indicate are as follows:
Xl=f (ul),with,ul=Wlxl-1+bl (3)
In formula (3), X represents output, and u represents input, and w represents weight, and b represents biasing, and wherein which of network l represent
Layer;
Export activation primitive f () can there are many kinds of, usually sigmoid function or hyperbolic tangent function.
Output is compressed to [0,1] by sigmoid, so last output average value generally tends to 0.So if by our training number
It is 1 according to zero-mean and variance is normalized to, convergence can be increased during gradient declines.For normalized data set
For, hyperbolic tangent function is also good selection.
Second stage, in the back-propagation stage, detailed process is as follows:
The error that backpropagation is returned can regard the sensitivity sensitivities of the base of each neuron as, wherein
The meaning of sensitivity be exactly we base b variation how much, error can change how much, that is, error is to the change rate of base, also
It is derivative, the error that backpropagation is returned, i.e., the sensitivity of the base of each neuron is as shown in following formula (4).Wherein, second
A equal sign is obtained according to the chain rule of derivation.
BecauseSoThat is the sensitivity of bias baseWith error E pair
One node fully enters the derivative of uIt is equal.This derivative is exactly to allow high-rise error back propagation to the mind of bottom
The pen come.
The sensitivity that l layers of backpropagation are as follows:
δl=(Wl+1)Tδl+1οf'(ul) (5)
In formula (5), T represents matrix transposition, and W represents weight, and u represents neuron signal input, and " ο " indicates each element
It is multiplied;
The sensitivity of the neuron of output layer are as follows:
δL=f ' (uL)ο(yn-tn). (6)
Finally, carrying out right value update with delta (i.e. δ) rule to each neuron.It is specifically exactly to be given to one
Fixed neuron obtains its input, is then zoomed in and out with the delta of this neuron (i.e. δ).With the form table of vector
It states and is exactly, for l layers, error is the input of this layer (equal to upper for the derivative of each weight (group is combined into matrix) of the layer
One layer of output) multiplication cross with the sensitivity (form that the δ of this layer of each neuron is combined into a vector) of this layer.Then
To partial derivative be exactly that the weight of neuron of this layer has updated multiplied by a negative learning rate:
In formula (7), wherein l indicates the number of plies of neural network, and representative is which layer X is vector, and w is l layers of weight,
η is learning rate, and the sensitivity before δ expression is similar for the more new-standard cement of bias base;In fact, for each weight
(W)ijThere is a specific learning rate ηIj。
Specific implementation process is as shown in Figure 1.In Fig. 1, vibration signal is adjusted to by fiber optic interferometric by fiber-optic vibration device
Signal output;Using the signal strength of output, the signal period combines the time for generating signal, and the distance for generating signal generates to obtain
Thermodynamic chart acquires heating power diagram data under different situations, manually to this by color from shallowly shockproofness size is indicated to deep
A little data are labeled, and are input to designed convolutional neural networks model using these data and are trained, and finally obtaining needs
The identification model wanted.
Depth network has powerful learning ability and efficient feature representation ability, and more important advantage is from Pixel-level
Initial data successively extracts information to abstract semantic concept, this makes it in the global characteristics and contextual information for extracting image
Aspect has advantage outstanding, and wherein convolutional layer and pond layer are the important components of convolutional neural networks.
Input picture by and three trainable filters and can biasing set carry out convolution, in C1 layer generation three after convolution
A Feature Mapping figure, then every group of four pixels are summed again in Feature Mapping figure, and weighted value, biasing is set, and passes through one
Sigmoid function obtains three S2 layers of Feature Mapping figure.These mapping graphs obtain C3 layers into filtering excessively again.This hierarchical structure
S4 is generated as S2 again.Finally, these pixel values are rasterized, and are connected into a vector and be input to traditional nerve net
Network is exported, and network structure Fig. 2 is as follows.
In Fig. 2, generally, C layers are characterized extract layer, the input and the local receptor field phase of preceding layer of each neuron
Even, and the feature of the part is extracted, after the local feature is extracted, its positional relationship between other features is also true therewith
It decides;S layers are Feature Mapping layers, and each computation layer of network is made of multiple Feature Mappings, and each Feature Mapping is one
Plane, the weight of all neurons is equal in plane.The sigmoid function conduct small using influence function core of Feature Mapping structure
The activation primitive of convolutional network, so that Feature Mapping has shift invariant.
Further, since the neuron on a mapping face shares weight, thus reduce the number of network freedom parameter, drops
The complexity of low network parameter selection.Each of convolutional neural networks feature extraction layer (C- layers) all followed by use
Seek the computation layer (S- layers) of local average and second extraction, this distinctive structure of feature extraction twice makes network in identification
There is higher distortion tolerance to input sample.
For the performance of authentication image recognizer, after the completion of image identification system design, generated using acquisition data
Thermodynamic chart is tested, and the model of vibration for being applied to optical fiber sensing system is established using the above method, there is vibration signal again
When, to the determination method of the model of vibration of the vibration signal, comprising the following steps:
Step A: reception optical fiber electric signal obtains and corresponds to several fiber positions in the fiber optic telecommunications number at several
Between signal value;Fiber-optic vibration threshold values is set, if fiber optic telecommunications number exceed fiber-optic vibration threshold values, without subsequent step.
Wherein, the fiber optic telecommunications number are converted to by the optical signal of fiber reflection;
Step B: being permanent reference axis, time as axis of ordinates to establish thermodynamic chart at a distance from optical fiber starting point using fiber position;
Pixel value in the thermodynamic chart is for characterizing several fiber positions of the optical fiber in the letter of several temporal signal values
Breath;Wherein, the distance between the fiber position and the optical fiber starting point are calculated using L=(c*T/n)/2;Wherein, c is that light exists
Speed in vacuum, T are time span of the fiber position in the fiber optic telecommunications number, and n is the refractive index of the optical fiber.
Step C: pre-processing thermodynamic chart, then according to claim 1 obtain described in be applied to Fibre Optical Sensor system
The model of vibration of system carries out control identification, obtains current model of vibration classification.
It is used to detect top vibration source classification since our fibre optical sensor is arranged on soil depths, we select
Big machinery excavation, hand digging, automobile pass through, strike four kinds of well typical vibration source classifications and mankind or animal by this normal
The violate-action signal seen has carried out neural network filter test.Difference is vibrated caused by different behaviors, generation
Interference signal is also different, and after FPGA is handled, the frequency variation signal waveform and amplitude of generation also have apparent difference.Therefore,
We carry out mode knowledge to FPGA treated frequency variation signal using the convolutional neural networks image recognition algorithm of host computer
It does not model, and has carried out prognostic experiment, the experimental results showed that 95% or more has all been reached to the predictablity rate of vibration behavior, and
Traditional DTW algorithm is compared, experimental result such as table 1:
1 experimental result comparative analysis of table
The foregoing is only a preferred embodiment of the present invention, is not intended to limit the present invention in any form;It is all
The those of ordinary skill of the industry can be described in by specification and described above and swimmingly implement the present invention;But it is all familiar
Professional and technical personnel without departing from the scope of the present invention, makes using disclosed above technology contents
A little variation, modification and evolution equivalent variations, be equivalent embodiment of the invention;Meanwhile all realities according to the present invention
The variation, modification and evolution etc. of matter technology any equivalent variations to the above embodiments, still fall within technology of the invention
Within the protection scope of scheme.
Claims (7)
1. a kind of method for building up of the model of vibration applied to optical fiber sensing system, which comprises the following steps:
Step S1: by fiber-optic vibration device by vibration signal be adjusted to optical fiber interference signal output, using the signal strength of output,
The distance that signal period combines the time for generating signal, generates signal, is sat with fiber position at a distance from optical fiber starting point for first
Parameter, time are that the second reference axis establishes signal condition figure;Wherein, the signal condition figure for characterize the optical fiber several
Information of the fiber position in several temporal signal values;
Step S2: acquiring the signal condition figure under different Vibration Conditions, carries out artificial mark label to these data, will manually mark
The data information of note label and signal condition figure is input to convolutional neural networks model and carries out image recognition and CNN training, obtains
The model of vibration applied to optical fiber sensing system;
Wherein, the image recognition of the convolutional neural networks model and CNN training the following steps are included:
First stage, propagation stage forward, detailed process is as follows:
Multi-class problem, total c class, total N number of training sample, error term are discussed using square error cost function are as follows:
In formula (1),Indicate the vector of the kth dimension of the corresponding label of n-th of sample,Indicate the corresponding network of n-th of sample
K-th of output of output;C is total class of representative sample;For multi-class problem, the shape that general tissue is " one-of-c " is exported
Formula, that is, the output node output of the only corresponding class of the input are positive, and the position of other classes or node are 0 or negative,
Activation primitive depending on output layer;Sigmoid is exactly that 0, tanh is exactly -1;
It is the summation of the error of each training sample based on the error on whole training sets, for the error of n-th of sample,
The summation of the error of each training sample indicates are as follows:
Then cost function E is calculated according to BP ruleNPartial derivative about each weight of network;Current layer is indicated with l, when
The output of front layer can indicate are as follows:
Xl=f (ul),with,ul=Wlxl-1+bl (3)
In formula, X represents the output for using vector form, and u represents input, and W represents weight, and b represents biasing, and wherein l represents net
Which layer of network;
Second stage, in the back-propagation stage, detailed process is as follows:
The error that backpropagation is returned, i.e., the sensitivity of the base of each neuron are as follows:
The sensitivity that l layers of backpropagation are as follows:
In formula (5), T represents matrix transposition, and W represents weight, and u represents input,Indicate each element multiplication;
The sensitivity of the neuron of output layer are as follows:
Finally, carrying out right value update with delta rule to each neuron;It is stated with the form of vector, for l layers, accidentally
Derivative of the difference for each weight of this layer are as follows:
In formula (7), wherein l indicates the number of plies of neural network, which layer representative is, X is the output with vector form;WlIt is l
The weight of layer, η is learning rate,Indicate l layers of backpropagation of sensitivity;For each weight have one it is specific
Learning rate.
2. the method for building up of the model of vibration according to claim 1 applied to optical fiber sensing system, it is characterised in that: institute
State be the first reference axis, time at a distance from optical fiber starting point with fiber position is that the second reference axis establishes signal condition figure, comprising:
It is the first reference axis, time at a distance from optical fiber starting point with fiber position is the second reference axis, fiber position at several
Temporal signal value is that third reference axis establishes signal condition figure.
3. the method for building up of the model of vibration according to claim 1 applied to optical fiber sensing system, it is characterised in that: institute
Stating signal condition figure is thermodynamic chart, wherein the pixel value in the thermodynamic chart is for characterizing several fiber positions of the optical fiber
In the information of several temporal signal values.
4. the method for building up of the model of vibration according to claim 3 applied to optical fiber sensing system, it is characterised in that: heat
Try hard to through color from shallowly indicating that oscillation intensity is ascending to deep.
5. a kind of determination method of the model of vibration of optical fiber sensing system, which comprises the following steps:
Step A: reception optical fiber electric signal obtains and corresponds to several fiber positions in the fiber optic telecommunications number in several times
Signal value;Wherein, the fiber optic telecommunications number are converted to by the optical signal of fiber reflection;
Step B: being that axis of ordinates establishes thermodynamic chart as axis of abscissas, time at a distance from optical fiber starting point using fiber position;It is described
Pixel value in thermodynamic chart is for characterizing several fiber positions of the optical fiber in the information of several temporal signal values;
Step C: pre-processing thermodynamic chart, then according to claim 1 obtain described in be applied to optical fiber sensing system
Model of vibration carries out control identification, obtains current model of vibration classification.
6. the determination method of the model of vibration of optical fiber sensing system according to claim 5, it is characterised in that: in step A,
Fiber-optic vibration threshold values is set, if fiber optic telecommunications number exceed fiber-optic vibration threshold values, without subsequent step.
7. the determination method of the model of vibration of optical fiber sensing system according to claim 5, it is characterised in that: use L=
(cLight*M/nOptical fiber)/2 calculate the distance between the fiber position and the optical fiber starting point;Wherein, cLightFor the speed of light in a vacuum
Degree, M are time span of the fiber position in the fiber optic telecommunications number, nOptical fiberFor the refractive index of the optical fiber.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610907115.8A CN106503642B (en) | 2016-10-18 | 2016-10-18 | A kind of model of vibration method for building up applied to optical fiber sensing system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610907115.8A CN106503642B (en) | 2016-10-18 | 2016-10-18 | A kind of model of vibration method for building up applied to optical fiber sensing system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106503642A CN106503642A (en) | 2017-03-15 |
CN106503642B true CN106503642B (en) | 2019-09-20 |
Family
ID=58293680
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610907115.8A Active CN106503642B (en) | 2016-10-18 | 2016-10-18 | A kind of model of vibration method for building up applied to optical fiber sensing system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106503642B (en) |
Families Citing this family (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108871760B (en) * | 2018-06-07 | 2020-07-17 | 广东石油化工学院 | Efficient gear fault mode identification method |
CN108932480B (en) * | 2018-06-08 | 2022-03-15 | 电子科技大学 | Distributed optical fiber sensing signal feature learning and classifying method based on 1D-CNN |
CN109598342B (en) * | 2018-11-23 | 2021-07-13 | 中国运载火箭技术研究院 | Decision network model self-game training method and system |
CN110135283A (en) * | 2019-04-25 | 2019-08-16 | 上海大学 | The signal recognition method of optical fiber perimeter defence system based on FastDTW algorithm |
CN110084986B (en) * | 2019-04-29 | 2021-11-09 | 西人马联合测控(泉州)科技有限公司 | Perimeter security method and device |
CN110363216A (en) * | 2019-05-31 | 2019-10-22 | 上海波汇软件有限公司 | A kind of implementation method applied to on-line training DNN model in DAS system |
CN110852187B (en) * | 2019-10-22 | 2023-04-07 | 华侨大学 | Method and system for identifying perimeter intrusion event |
CN111160106B (en) * | 2019-12-03 | 2023-12-12 | 上海微波技术研究所(中国电子科技集团公司第五十研究所) | GPU-based optical fiber vibration signal feature extraction and classification method and system |
CN111538246B (en) * | 2020-07-08 | 2020-10-09 | 浙江浙能天然气运行有限公司 | System and method for estimating interference of mechanical equipment on distributed optical fiber sensor |
CN112539772B (en) * | 2020-11-02 | 2023-04-07 | 上海大学 | Positioning method of Sagnac distributed optical fiber sensing system based on convolutional neural network integrated learning |
CN112364836A (en) * | 2020-12-07 | 2021-02-12 | 无锡科晟光子科技有限公司 | Vibration optical fiber signal classification method based on full convolution neural network |
CN114415018A (en) * | 2022-01-19 | 2022-04-29 | 山东超晟光电科技有限公司 | Self-learning grating interference spectrum analysis technology for motor fault early warning |
CN114684217B (en) * | 2022-03-16 | 2024-03-01 | 武汉理工大学 | Rail transit health monitoring system and method |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105698916A (en) * | 2016-03-01 | 2016-06-22 | 深圳艾瑞斯通技术有限公司 | Optical fiber vibration model determination method and optical fiber early warning device and system |
CN105787439A (en) * | 2016-02-04 | 2016-07-20 | 广州新节奏智能科技有限公司 | Depth image human body joint positioning method based on convolution nerve network |
CN105933063A (en) * | 2016-06-20 | 2016-09-07 | 深圳艾瑞斯通技术有限公司 | Optical fiber signal processing method and device, and optical fiber sensing system |
CN105973451A (en) * | 2016-05-09 | 2016-09-28 | 深圳艾瑞斯通技术有限公司 | Optical fiber vibration model determination method and device |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20110307935A1 (en) * | 2010-06-09 | 2011-12-15 | Verizon Patent And Licensing Inc. | Video content delivery optimization over mobile wireless networks |
-
2016
- 2016-10-18 CN CN201610907115.8A patent/CN106503642B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105787439A (en) * | 2016-02-04 | 2016-07-20 | 广州新节奏智能科技有限公司 | Depth image human body joint positioning method based on convolution nerve network |
CN105698916A (en) * | 2016-03-01 | 2016-06-22 | 深圳艾瑞斯通技术有限公司 | Optical fiber vibration model determination method and optical fiber early warning device and system |
CN105973451A (en) * | 2016-05-09 | 2016-09-28 | 深圳艾瑞斯通技术有限公司 | Optical fiber vibration model determination method and device |
CN105933063A (en) * | 2016-06-20 | 2016-09-07 | 深圳艾瑞斯通技术有限公司 | Optical fiber signal processing method and device, and optical fiber sensing system |
Non-Patent Citations (1)
Title |
---|
光纤光栅振动传感器的振动理论分析;陈帮 等;《武汉理工大学学报(信息与管理工程版)》;20120430;第34卷(第2期);第143-146页 * |
Also Published As
Publication number | Publication date |
---|---|
CN106503642A (en) | 2017-03-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106503642B (en) | A kind of model of vibration method for building up applied to optical fiber sensing system | |
Singh et al. | Transforming sensor data to the image domain for deep learning—An application to footstep detection | |
CN109034044B (en) | Pedestrian re-identification method based on fusion convolutional neural network | |
CN106874840B (en) | Vehicle information recognition method and device | |
CN103390164B (en) | Method for checking object based on depth image and its realize device | |
CN112926405B (en) | Method, system, equipment and storage medium for detecting wearing of safety helmet | |
CN109214349A (en) | A kind of object detecting method based on semantic segmentation enhancing | |
CN103971106B (en) | Various visual angles facial image gender identification method and device | |
CN104269006B (en) | Mode identification method for optical fiber early warning system | |
CN109886242A (en) | A kind of method and system that pedestrian identifies again | |
Shamrat et al. | A deep learning approach for face detection using max pooling | |
CN102136024A (en) | Biometric feature identification performance assessment and diagnosis optimizing system | |
CN109827652A (en) | One kind being directed to Fibre Optical Sensor vibration signal recognition and system | |
Islam et al. | Performance prediction of tomato leaf disease by a series of parallel convolutional neural networks | |
CN113807314A (en) | Millimeter wave radar video fusion method based on micro-Doppler effect | |
Buchanan et al. | Deep convolutional neural networks for detecting dolphin echolocation clicks | |
Ge et al. | Coarse-to-fine foraminifera image segmentation through 3D and deep features | |
Monigari et al. | Plant leaf disease prediction | |
CN105930793A (en) | Human body detection method based on SAE characteristic visual learning | |
CN110321867A (en) | Shelter target detection method based on part constraint network | |
CN110390949A (en) | Acoustic Object intelligent identification Method based on big data | |
CN106682604B (en) | Blurred image detection method based on deep learning | |
Reyes et al. | Safety gear compliance detection using data augmentation-assisted transfer learning in construction work environment | |
Shishkin et al. | Analysis of image clusterization methods for oceanographical equipment | |
Gill et al. | Face Mask Detection Using ResNet50 Model and fine tuning it on various hyperactive parameters |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |