CN108399696A - Intrusion behavior recognition methods and device - Google Patents
Intrusion behavior recognition methods and device Download PDFInfo
- Publication number
- CN108399696A CN108399696A CN201810243918.7A CN201810243918A CN108399696A CN 108399696 A CN108399696 A CN 108399696A CN 201810243918 A CN201810243918 A CN 201810243918A CN 108399696 A CN108399696 A CN 108399696A
- Authority
- CN
- China
- Prior art keywords
- energy
- gradient
- tested
- time series
- series signal
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B13/00—Burglar, theft or intruder alarms
- G08B13/02—Mechanical actuation
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E40/00—Technologies for an efficient electrical power generation, transmission or distribution
- Y02E40/70—Smart grids as climate change mitigation technology in the energy generation sector
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The present invention provides a kind of intrusion behavior recognition methods and devices, are related to the technical field of security protection, which includes：Obtain time series signal, wherein time series signal is sequence signal made of the tested amplitude for enclosing boundary's vibration arranges at any time；Characteristic quantity is extracted from time series signal, wherein characteristic quantity includes：The maximum value of the tested peak swing for enclosing boundary's vibration, energy, spectrum energy breadth coefficient, energy gradient and energy gradient gradient；Intrusion behavior identification is carried out using characteristic quantity as the input quantity of machine learning algorithm, to determine the tested intrusion behavior type for enclosing boundary by the output quantity of machine learning algorithm.The present invention alleviates conventional machines study identification intrusion behavior technical problem of high cost and poor for applicability.
Description
Technical field
The present invention relates to the technical fields of security protection, more particularly, to a kind of intrusion behavior recognition methods and device.
Background technology
Intrusion behavior identifies, that is, is detected invasion vibration intrusion behavior to be identified, belongs to passive discerning skill
Art.Vibration detecting is invaded, i.e., lays the acquisition of vibration detecting node in metal fence or solid wall and encloses on boundary's entity protective device
Vibration signal and this is analyzed, to generate intrusion alarm.Currently used Detection Techniques include individual axis acceleration, three axis
The MEMS such as acceleration, geophone (Micro-Electro-Mechanical System, abbreviation, MEMS), vibration
Optical fiber, vibration wireline etc. require Detection Techniques to can recognize that percussion, rock and climb three kinds of main intrusion behaviors in business；
Also, invasive biology algorithm want can to handle wind, vehicle by etc. ambient noises, with meet simultaneously higher recognition accuracy and
Two key indexes of lower rate of false alarm, and handle delay and should be less than 1~2 second.
The invasive biology algorithm of early stage is realized on hardware device, specifically, algorithm is run simultaneously in hardware node equipment
Switching value alarm signal is directly exported, the adjusting of algorithm is realized by the regulation button in equipment.Such realization method
Advantage is that algorithmic technique framework is simple, and system installation and debugging are convenient, fast response time；The disadvantage is that since hardware computation ability limits
It makes, relative complex calculating can not be loaded and run on node, the algorithm that can be realized is relatively easy, can not accurately go to invasion
To classify, rate of false alarm and rate of failing to report are higher.
With networking and digitlization that security and guard technology develops, mainstream vibrating intruding Detection Techniques realize perception data
Acquisition digitlization and transmission network, hardware node equipment are merely responsible for acquisition vibration signal or carry out simple data processing, enters
It invades recognizer and realizes that invasive biology algorithm realizes networking on central server in the background.Under this realization method, algorithm
It is not limited by hardware node equipment computing capability, and algorithm can be trained, optimizes and dispose as needed, thus, it is possible to real
Existing algorithm is complex, and can export the alarm signal for including structuring intrusion behavior.
Specifically, networking invasive biology algorithm is basically divided into both direction：
One direction is parameter configuration, that is, amplitude, wavelength, wave number, zero-crossing rate are extracted from the time series signal of vibration
Equal characteristic quantities, for different types of intrusion behavior, based on features described above amount configuration determination rule, and according to invasive biology business
Type and operating status, operating parameter etc. sound state adjustment is carried out to parameter of regularity.Such regular drive based on business is calculated
Method it can be readily appreciated that technology realize it is upper simple and feasible, the disadvantage is that rule configuration and adjust it is comparatively laborious, it is more dependent on practice warp
It tests.
Another direction is machine learning, i.e., designs a kind of invasive biology machine learning according to invasive biology business scenario
Then algorithm exercises supervision training to algorithm model by loading the largely data manually demarcated, will after evaluating and optimizing
Algorithm model is deployed in actual items and is applied.Common machine learning algorithm includes EM, SVM, neural network, decision tree
Deng.Such invasive biology algorithm for meeting Artificial Intelligence Development trend has output accurate after big data training and optimization
The high advantage of rate.But machine learning at present identifies intrusion behavior, is arranged according to specific invasive biology business scenario
Corresponding machine recognition algorithm, poor universality, to there are of high cost, algorithm environment adaptability it is not strong and in practical effect
The shortcomings of not good enough.
In conclusion current networking invasive biology algorithm, although relative complex algorithm can be realized, parameter is matched
It is relatively complicated to set rule configuration；And although machine learning does not configure relatively complicated disadvantage, it is of high cost and poor for applicability.
Invention content
In view of this, the purpose of the present invention is to provide a kind of intrusion behavior recognition methods and device, to alleviate current machine
Device study identification intrusion behavior technical problem of high cost and poor for applicability.
In a first aspect, an embodiment of the present invention provides a kind of intrusion behavior recognition methods, including：
Obtain time series signal, wherein the time series signal is that the tested amplitude for enclosing boundary's vibration arranges at any time
Made of sequence signal；
Characteristic quantity is extracted from the time series signal, wherein the characteristic quantity includes：Described be tested encloses boundary's vibration
The maximum value of peak swing, energy, spectrum energy breadth coefficient, energy gradient and energy gradient gradient；
Intrusion behavior identification is carried out using the characteristic quantity as the input quantity of machine learning algorithm, to pass through the engineering
The output quantity for practising algorithm determines the tested intrusion behavior type for enclosing boundary.
With reference to first aspect, an embodiment of the present invention provides the first possible embodiments of first aspect, wherein from
Characteristic quantity is extracted in the time series signal, including：
Amplitude maximum in the time series signal is determined as the tested peak swing for enclosing boundary's vibration；
Discrete Fourier transform is carried out to the time series signal, obtains the tested spectrogram for enclosing boundary's vibration, and
Target characteristic amount is extracted from the spectrogram, wherein the target characteristic amount includes：The tested energy for enclosing boundary's vibration,
The maximum value of spectrum energy breadth coefficient, energy gradient and energy gradient gradient.
The possible embodiment of with reference to first aspect the first, an embodiment of the present invention provides second of first aspect
Possible embodiment, wherein discrete Fourier transform is carried out to the time series signal, the tested boundary that encloses is obtained and vibrates
Spectrogram, and extract target characteristic amount from the spectrogram, including：
Window is handled by preset signals to be split the time series signal, obtains multiple Time Sub-series letters
Number；
Discrete Fourier transform is carried out to the Time Sub-series signal in the signal processing window, obtains multiple sons
Spectrogram；
The target characteristic amount is extracted from multiple sub- spectrograms.
Second of possible embodiment with reference to first aspect, an embodiment of the present invention provides the third of first aspect
Possible embodiment, wherein the target characteristic amount is extracted from multiple sub- spectrograms, including：
Multiple default division ranges of spectrum distribution are obtained, and the tested boundary that encloses is calculated based on the sub- spectrogram and is vibrated
In multiple default spectrum energy breadth coefficients divided in range；
The sum of the vibrational energy in the sub- spectrogram is calculated, multiple energy values are obtained, by the collection of multiple energy values
It closes and is determined as the tested energy for enclosing boundary's vibration；
The ratio of any two adjacent energies in multiple energy values is determined as the energy gradient；
By in multiple energy gradients, the ratio of any two energy gradients is determined as the energy gradient
Gradient, to search the maximum value of the energy gradient gradient from the energy gradient gradient.
With reference to first aspect, an embodiment of the present invention provides the 4th kind of possible embodiments of first aspect, wherein will
The characteristic quantity carries out intrusion behavior identification as the input quantity of machine learning algorithm, including：
Obtain trained neural network, wherein the trained neural network include an input layer, three
Hidden layer and an output layer；
The characteristic quantity is input to the neural network in the input layer；
The recognition result of the neural network is obtained from the output layer, and the recognition result is determined as described be tested
Enclose the intrusion behavior type on boundary.
The 4th kind of possible embodiment with reference to first aspect, an embodiment of the present invention provides the 5th kind of first aspect
Possible embodiment, wherein the spectrum energy breadth coefficient includes low frequency energy breadth coefficient, intermediate frequency Energy distribution coefficient
With high-frequency energy breadth coefficient, the intrusion behavior type include it is following any one：No intrusion behavior is tapped, is rocked, climbing
It climbs；
The input layer includes seven parameter input ends, and seven parameter input ends are respectively used to input：It is described tested
Enclose peak swing, energy, low frequency energy breadth coefficient, intermediate frequency Energy distribution coefficient, high-frequency energy breadth coefficient, the energy of boundary's vibration
The maximum value of quantitative change rate and energy gradient gradient；
And the output layer includes four result output ends, four result output ends are respectively used to export：Without invasion
Behavior is tapped, is rocked, climbing.
Second aspect, the embodiment of the present invention also provide a kind of intrusion behavior identification device, including：
Acquisition module, for obtaining time series signal, wherein the time series signal is to be tested to enclose shaking for boundary's vibration
Sequence signal made of width arranges at any time；
Extraction module, for extracting characteristic quantity from the time series signal, wherein the characteristic quantity includes：It is described
The maximum of the tested peak swing for enclosing boundary's vibration, energy, spectrum energy breadth coefficient, energy gradient and energy gradient gradient
Value；
Identification module, for carrying out intrusion behavior identification using the characteristic quantity as the input quantity of machine learning algorithm, with
The tested intrusion behavior type for enclosing boundary is determined by the output quantity of the machine learning algorithm.In conjunction with second aspect, this hair
Bright embodiment provides the first possible embodiment of second aspect, wherein
The embodiment of the present invention brings following advantageous effect：The intrusion behavior recognition methods includes：Obtain time series letter
Number, wherein time series signal is sequence signal made of the tested amplitude for enclosing boundary's vibration arranges at any time；Believe from time series
Characteristic quantity is extracted in number, wherein characteristic quantity includes：The tested peak swing for enclosing boundary's vibration, energy, spectrum energy breadth coefficient,
The maximum value of energy gradient and energy gradient gradient；Invasion row is carried out using characteristic quantity as the input quantity of machine learning algorithm
For identification, the intrusion behavior type for enclosing boundary is tested to be determined by the output quantity of machine learning algorithm.The present invention passes through above-mentioned spy
The exact classification to intrusion behavior can be realized in sign amount, versatile, alleviates conventional machines study identification intrusion behavior cost
High and poor for applicability technical problem.
Other features and advantages of the present invention will illustrate in the following description, also, partly become from specification
It obtains it is clear that understand through the implementation of the invention.The purpose of the present invention and other advantages are in specification and attached drawing
Specifically noted structure is realized and is obtained.
To enable the above objects, features and advantages of the present invention to be clearer and more comprehensible, preferred embodiment cited below particularly, and coordinate
Appended attached drawing, is described in detail below.
Description of the drawings
It, below will be to specific in order to illustrate more clearly of the specific embodiment of the invention or technical solution in the prior art
Embodiment or attached drawing needed to be used in the description of the prior art are briefly described, it should be apparent that, in being described below
Attached drawing is some embodiments of the present invention, for those of ordinary skill in the art, before not making the creative labor
It puts, other drawings may also be obtained based on these drawings.
Fig. 1 is a kind of flow chart for intrusion behavior recognition methods that the embodiment of the present invention one provides；
Fig. 2 is a kind of tested coordinate schematic diagram for enclosing boundary that the embodiment of the present invention one provides；
Fig. 3 is the vibrational waveform figure in the case of a kind of no intrusion behavior that the embodiment of the present invention one provides；
Fig. 4 is the spectrogram in the case of a kind of no intrusion behavior that the embodiment of the present invention one provides；
Fig. 5 is the vibrational waveform figure in the case of a kind of percussion that the embodiment of the present invention one provides；
Fig. 6 is the spectrogram in the case of a kind of percussion that the embodiment of the present invention one provides；
Fig. 7 be the embodiment of the present invention one provide it is a kind of rock in the case of spectrogram；
Fig. 8 be the embodiment of the present invention one provide it is a kind of rock in the case of vibrational waveform figure；
Fig. 9 is the spectrogram in the case of a kind of climbing that the embodiment of the present invention one provides；
Figure 10 is the vibrational waveform figure in the case of a kind of climbing that the embodiment of the present invention one provides；
Figure 11 is the vibrational waveform figure in the case of a kind of climbing that the embodiment of the present invention one provides；
Figure 12 is the spectrogram in the case of a kind of wind noise that the embodiment of the present invention one provides；
Figure 13 is a kind of schematic diagram for neural network that the embodiment of the present invention one provides；
Figure 14 is a kind of structure diagram of intrusion behavior identification device provided by Embodiment 2 of the present invention.
Icon：100- acquisition modules；200- extraction modules；300- identification modules.
Specific implementation mode
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with attached drawing to the present invention
Technical solution be clearly and completely described, it is clear that described embodiments are some of the embodiments of the present invention, rather than
Whole embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art are not making creative work premise
Lower obtained every other embodiment, shall fall within the protection scope of the present invention.
For networking invasive biology algorithm, networking invasive biology algorithm is divided into the invasive biology algorithm generally used at present
Based on two kinds of the method for parameter configuration and machine learning method of business rule driving, the former is the problem is that rule configures more
Cumbersome, the shortcomings that the latter is of high cost and poor for applicability.Based on this, a kind of intrusion behavior recognition methods of present invention offer and dress
It sets, to alleviate cumbersome, of high cost, the poor for applicability technical problem of current invasive biology algorithm configuration.
Embodiment one
A kind of intrusion behavior recognition methods provided in an embodiment of the present invention, as shown in Figure 1, including：
Step S102 obtains time series signal, wherein time series signal be tested to enclose amplitude that boundary vibrates at any time
Sequence signal made of arrangement；
Step S104, extracts characteristic quantity from time series signal, wherein characteristic quantity includes：It is tested to enclose boundary's vibration most
The maximum value of large amplitude, energy, spectrum energy breadth coefficient, energy gradient and energy gradient gradient；
Step S106 carries out intrusion behavior identification, to pass through machine using characteristic quantity as the input quantity of machine learning algorithm
The output quantity of learning algorithm determines the tested intrusion behavior type for enclosing boundary.
It should be noted that the tested peak swing for enclosing boundary's vibration, that is, the tested maximum value for enclosing boundary's Oscillation Amplitude；Energy,
The sum of vibrational energy in a continuous time section may be used to characterize in the tested energy value for enclosing boundary's vibration；Spectrum energy point
Cloth coefficient is tested the distribution proportion for enclosing boundary's vibrational energy in each frequency；Energy gradient, that is, tested to enclose boundary's vibrational energy
To the first order derivative of time；Energy gradient gradient, that is, tested to enclose second derivative of boundary's vibrational energy to the time.
Intrusion behavior or noise jamming are presented as the variation of amplitude and frequency, and amplitude on tested fence vibrational waveform
Variation with frequency with external force size is applied to tested fence, action time, position of action point have relationship, tested fence is applied outer
The characterized Invasion type of tested fence of power size, action time, position of action point, thus, the variation of amplitude and frequency
The Invasion type for embodying tested fence extracts the characteristic quantity in relation to amplitude and frequency from time series signal：It is tested to enclose boundary
The peak swing of vibration, the maximum value of energy, spectrum energy breadth coefficient, energy gradient and energy gradient gradient, can be very
Intrusion behavior is identified well.
In embodiments of the present invention, the exact classification to intrusion behavior can be realized by features described above amount, it is versatile,
Alleviate conventional machines study identification intrusion behavior technical problem of high cost and poor for applicability.
Specifically, vibrating intruding Detection Techniques using the hardware nodes equipment detection such as MEMS sensor it is tested enclose in boundary due to
Vibration signal caused by intrusion behavior or noise, hardware device node output time series signal, the identification of all intrusion behaviors
It is based on time series signal.According to nyquist sampling law, hardware node equipment sample frequency is set as vibration letter to be analyzed
2 times of number maximum frequency.
Under normal condition, time series signal is stable；When encountering external interference, the waveform of time series signal is
It can change, output signal is presented as vibrational waveform.Due to tested fence physical arrangement limitation, it is assumed that residing for tested fence
Plane is X/Y plane as shown in Figure 2, then is tested the Z axis that the main deformation of fence is happened at vertical fence plane, therefore hardware
Node device emphasis acquires and analyzes the time series signal on Z axis.
First the vibrational waveform and spectrogram of different intrusion behaviors are introduced below, in the time sequence of description vibrational waveform
In row figure, horizontal axis parameter is the time, and longitudinal axis parameter is amplitude；In spectrogram, horizontal axis parameter is frequency, and longitudinal axis parameter is mould.
(1) under normal circumstances, i.e., no any intrusion behavior and noise jamming
From Fig. 3 and Fig. 4：A stable horizontal linear is substantially presented in vibrational waveform and frequency spectrum, vibrational waveform
Slight perturbations are the ambient noise under normal condition.
(2) behavior is tapped
As seen from Figure 5：Vibrational waveform is presented as a sharp pulse, and amplitude is relatively high and wavelength is relatively small, and amplitude is big
It is small to depend on percussion dynamics.
As seen from Figure 6：Signal on frequency spectrum is concentrated mainly on high-frequency region, and crest frequency is more than 40Hz, and signal
Attenuation process is usually very fast, and process was lasted generally within 1 second.
(3) behavior is rocked
As seen from Figure 7：Apparent sine wave is embodied on vibrational waveform, wavelength is larger, and amplitude depends on rocking dynamics, believes
Number attenuation process is slower, and waveform there may be aftershock, lasted generally within 1~2 second by process.
As seen from Figure 8：Signal on frequency spectrum is concentrated mainly on low frequency region, crest frequency 10Hz.
(4) behavior is climbed
As seen from Figure 9：Occurs the feature that an amplitude is gradually increased and then is gradually reduced on vibrational waveform, amplitude variations are slow
Slowly, local feature is similar to a series of percussion of different dynamics, and signal attenuation process is slower, and process is lasted generally on 3 seconds left sides
It is right.
As seen from Figure 10：Crest frequency is 30Hz, and the signal distributions on frequency spectrum are in low frequency and mid-frequency region, but energy is main
Concentrate on mid-frequency region.
(5) wind noise
From Figure 11：Wind noise is intermittent, duration average out to 2~5 seconds, vibration wave to the effect of fence
Shape lacks apparent feature.
From Figure 12：Energy concentrates on intermediate frequency and high-frequency region, but without apparent crest frequency, distribution is relatively
Uniformly.
In view of the difference of vibrational waveform and spectrogram in above-mentioned different intrusion behaviors, the characteristic quantity that the present invention passes through extraction：
The maximum of the tested peak swing for enclosing boundary's vibration, energy, spectrum energy breadth coefficient, energy gradient and energy gradient gradient
Value, is identified intrusion behavior.
First, the present invention extracts characteristic quantity from time series signal.
In one optional embodiment of the embodiment of the present invention, step S104 extracts characteristic quantity from time series signal,
Including：
By the amplitude maximum A in time series signalmaxIt is determined as the tested peak swing for enclosing boundary's vibration；
Discrete Fourier transform is carried out to time series signal, obtains the tested spectrogram for enclosing boundary's vibration, and from spectrogram
Middle extraction target characteristic amount, wherein target characteristic amount includes：The tested energy, spectrum energy breadth coefficient, energy for enclosing boundary's vibration
The maximum value of change rate and energy gradient gradient.
It should be noted that AmaxLess than preset value Alimit, then it is assumed that it is up state；Amax>Alimit, then
Think that there may be intrusion behavior or noise jammings.
In another optional embodiment of the embodiment of the present invention, discrete Fourier transform is carried out to time series signal,
The tested spectrogram for enclosing boundary and vibrating is obtained, and extracts target characteristic amount from spectrogram, including：
Window is handled by preset signals to be split time series signal, obtains multiple Time Sub-series signals；
Discrete Fourier transform is carried out to the Time Sub-series signal in signal processing window, obtains multiple sub- spectrograms；
Target characteristic amount is extracted from multiple sub- spectrograms.
Specifically, can preset signals be handled window time to be set as 2.5 seconds, the overlapping time of adjacent window apertures is 0.5
Second.Window is handled by preset signals to be split time series signal, that is, time series signal is divided into multiple durations
For 2.5 seconds Time Sub-series signals, and the overlapping time of two neighboring Time Sub-series signal was 0.5 second.
In another optional embodiment of the embodiment of the present invention, target characteristic amount is extracted from multiple sub- spectrograms, is wrapped
It includes：
Multiple default division ranges of spectrum distribution are obtained, and the tested boundary that encloses is calculated based on sub- spectrogram and is vibrated multiple pre-
If dividing the spectrum energy breadth coefficient in range；
The sum of the vibrational energy in sub- spectrogram is calculated, multiple energy values are obtained, the set of multiple energy values is determined as
The tested energy for enclosing boundary's vibration；
The ratio of any two adjacent energies in multiple energy values is determined as energy gradient；
By in multiple energy gradients, the ratio of any two energy gradients is determined as energy gradient gradient, with from energy
The maximum value of energy gradient gradient is searched in quantitative change rate gradient.
Specific process is as follows：
(1) the tested spectrum energies for enclosing boundary's vibration in multiple default division ranges are calculated based on sub- spectrogram and is distributed system
Number, process are as follows：
Multiple default division ranges can be classified as three low frequency, intermediate frequency and high frequency ranges according to the height of frequency, wherein low
Frequency ranging from 1Hz~20Hz, intermediate frequency range are 20Hz~35Hz, and high-frequency range is 35Hz~50Hz.Define spectrum energy distribution
Coefficient F=Es/Et, wherein Es be some frequency range energy and, Et be energy in all frequency ranges and, then low frequency energy
Measure breadth coefficient FlowFor low-frequency range energy and in all frequency ranges energy and the ratio between, intermediate frequency energy distribution system
FmidFor intermediate frequency range energy and in all frequency ranges energy and the ratio between, high-frequency energy compartment system FhighFor high frequency
The energy of range and in all frequency ranges energy and the ratio between.
It should be noted that low-frequency range 1Hz~20Hz, intermediate frequency range 20Hz~35Hz, high-frequency range 35Hz~50Hz,
It is corresponding in turn to frequency spectrum when rocking, climb and tapping, thus, the concentrated area of spectrum energy distribution is fallen in low frequency, intermediate frequency and height
A range in frequency, then intrusion behavior is corresponding type.In practical applications, it is contemplated that in tested fence physical characteristic
Difference, filter frequency range can be adjusted as needed, i.e., so that low frequency, intermediate frequency and high-frequency range repartition.
(2) the sum of the vibrational energy in sub- spectrogram is calculated, multiple energy values are obtained, the set of multiple energy values is determined
To be tested the energy for enclosing boundary's vibration, process is as follows：
Signal processing window is to define E in the case of 2.5 secondstEnergy in time range thus defines Et-1It is upper one
The energy accumulation value of a signal processing window, an energy aggregation S can be defined by thus promotingE={ Et-n,…Et-1,Et}。
It should be noted that EtMore than energy preset value ElimitAt the time of, it is carved at the beginning of identifying climbing behavior.
(3) ratio of any two adjacent energies in multiple energy values is determined as energy gradient, process is as follows：
Define energy ratio coefficients R=Et/Et-1, indicate the energy ratio of former and later two signal processing windows, thus obtain one
A energy ratio coefficient sets SR={ Rt-n-+1..., Rt-1, Rt}。
It should be noted that wind noise leads to the E at previous momentt-1It is larger, therefore energy ratio Et/Et-1Relative to normal
Under the conditions of value want low, while wind effect caused by energy increase gradient it is lower than intrusion behavior.In addition the energy spectrum of wind is concentrated
It in intermediate frequency and high-frequency region, and is uniformly distributed, this point and intrusion behavior different from.
(4) by multiple energy gradients, the ratio of any two energy gradients is determined as energy gradient gradient, with from
The maximum value of energy gradient gradient is searched in energy gradient gradient, process is as follows：
The maximum value in energy gradient gradient can be taken to be denoted as Gmax in the case where energy gradient is positive value,
Whether Gmax is to tap or rock the aftershock brought to distinguish.According to field test data, the energy variation that taps and rock
Rate gradient is usually larger, and the energy variation gradient climbed is relatively small.
In features described above amount extraction process, according to the vibrational waveform and spectrogram of different intrusion behaviors, to each feature
It measures and is described in detail with the correspondence of intrusion behavior type.Value range and intrusion behavior based on each characteristic quantity
Correspondence, one neural network of training, then by the Application of Neural Network in step S106, using characteristic quantity as machine learning
The input quantity of algorithm carries out intrusion behavior identification.Specifically, step S106, using characteristic quantity as the input quantity of machine learning algorithm
Intrusion behavior identification is carried out, including：
Obtain trained neural network, wherein trained neural network is hidden including an input layer, three
Layer and an output layer；
Characteristic quantity is input to neural network in input layer；
The recognition result of neural network is obtained from output layer, and recognition result is determined as the tested intrusion behavior class for enclosing boundary
Type.
In the embodiment of the present invention, by trained neural fusion machine learning algorithm, intrusion behavior is carried out
Intrusion behavior is identified in identification, the neural network for using three hidden layers of band, avoids the neural network institute of overcomplicated
The overfitting phenomenon of appearance, to also avoid adverse effect of the overfitting phenomenon to algorithm accuracy rate.
In another optional embodiment of the embodiment of the present invention, spectrum energy breadth coefficient includes low frequency energy distribution system
Number, intermediate frequency Energy distribution coefficient and high-frequency energy breadth coefficient, intrusion behavior type include it is following any one：It goes without invasion
For, tap, rock, climb；
Input layer includes seven parameter input ends, and seven parameter input ends are respectively used to input：It is tested to enclose boundary's vibration most
Large amplitude, energy, low frequency energy breadth coefficient, intermediate frequency Energy distribution coefficient, high-frequency energy breadth coefficient, energy gradient and energy
The maximum value of quantitative change rate gradient；
And output layer includes four result output ends, four result output ends are respectively used to export：No intrusion behavior strikes
It hits, rock, climb, neural network is as shown in figure 13.
The feature duration set that the present invention finally chooses is { Amax, Flow, Fmid, Fhigh, Et, Rt, Gmax }.The embodiment of the present invention
For the characteristic quantity negligible amounts of machine learning, therefore often occur in effectively avoiding machine learning algorithm from applying excessive quasi-
While conjunction problem, algorithm load and run time are short, and accuracy is high, with scale can be applied in practical application.
Embodiment two
A kind of intrusion behavior identification device provided in an embodiment of the present invention, as shown in figure 14, including：
Acquisition module 100, for obtaining time series signal, wherein time series signal is to be tested to enclose shaking for boundary's vibration
Sequence signal made of width arranges at any time；
Extraction module 200, for extracting characteristic quantity from time series signal, wherein characteristic quantity includes：It is tested to enclose boundary and shake
The maximum value of dynamic peak swing, energy, spectrum energy breadth coefficient, energy gradient and energy gradient gradient；
Identification module 300, for carrying out intrusion behavior identification using characteristic quantity as the input quantity of machine learning algorithm, with logical
The output quantity for crossing machine learning algorithm determines the tested intrusion behavior type for enclosing boundary.
In embodiments of the present invention, acquisition module 100 obtains time series signal, and extraction module 200 is believed from time series
Characteristic quantity is extracted in number, characteristic quantity is carried out intrusion behavior identification by identification module 300, with
The tested intrusion behavior type for enclosing boundary is determined by the output quantity of machine learning algorithm.Wherein, characteristic quantity includes：It is tested to enclose boundary and shake
The maximum value of dynamic peak swing, energy, spectrum energy breadth coefficient, energy gradient and energy gradient gradient, the present invention
The exact classification to intrusion behavior can be realized by features described above amount, it is versatile, alleviate conventional machines study identify into
Invade behavior technical problem of high cost and poor for applicability.
In one optional embodiment of the embodiment of the present invention, extraction module, including：
First extraction unit, for the amplitude maximum in time series signal to be determined as the tested maximum for enclosing boundary's vibration
Amplitude；
Second extraction unit obtains being tested and encloses boundary's vibration for carrying out discrete Fourier transform to time series signal
Spectrogram, and target characteristic amount is extracted from spectrogram, wherein target characteristic amount includes：The tested energy for enclosing boundary's vibration, frequency spectrum
The maximum value of Energy distribution coefficient, energy gradient and energy gradient gradient.
In another optional embodiment of the embodiment of the present invention, the second extraction unit, including：
Divide subelement, time series signal is split for handling window by preset signals, obtains multiple sons
Time series signal；
Subelement is converted, for carrying out discrete Fourier transform to the Time Sub-series signal in signal processing window, is obtained
To multiple sub- spectrograms；
Subelement is extracted, for extracting target characteristic amount from multiple sub- spectrograms.
In another optional embodiment of the embodiment of the present invention, extraction subelement is used for：
Multiple default division ranges of spectrum distribution are obtained, and the tested boundary that encloses is calculated based on sub- spectrogram and is vibrated multiple pre-
If dividing the spectrum energy breadth coefficient in range；
The sum of the vibrational energy in sub- spectrogram is calculated, multiple energy values are obtained, the set of multiple energy values is determined as
The tested energy for enclosing boundary's vibration；
The ratio of any two adjacent energies in multiple energy values is determined as energy gradient；
By in multiple energy gradients, the ratio of any two energy gradients is determined as energy gradient gradient, with from energy
The maximum value of energy gradient gradient is searched in quantitative change rate gradient.
In another optional embodiment of the embodiment of the present invention, identification module includes：
Acquiring unit, for obtaining trained neural network, wherein trained neural network includes one defeated
Enter layer, three hidden layers and an output layer；
Input unit, for characteristic quantity to be input to neural network in input layer；
Output unit, the recognition result for obtaining neural network from output layer, and recognition result is determined as tested enclose
The intrusion behavior type on boundary.
In another optional embodiment of the embodiment of the present invention, spectrum energy breadth coefficient includes：Low frequency energy is distributed
Coefficient, intermediate frequency Energy distribution coefficient and high-frequency energy breadth coefficient, intrusion behavior type include it is following any one：It goes without invasion
For, tap, rock, climb；
Input layer includes seven parameter input ends, and seven parameter input ends are respectively used to input：It is tested to enclose boundary's vibration most
Large amplitude, energy, low frequency energy breadth coefficient, intermediate frequency Energy distribution coefficient, high-frequency energy breadth coefficient, energy gradient and energy
The maximum value of quantitative change rate gradient；
And output layer includes four result output ends, four result output ends are respectively used to export：No intrusion behavior strikes
It hits, rock, climb.
The computer program product of intrusion behavior recognition methods and device that the embodiment of the present invention is provided, including store
The computer readable storage medium of program code, the instruction that said program code includes can be used for executing in previous methods embodiment
The method, specific implementation can be found in embodiment of the method, and details are not described herein.
It is apparent to those skilled in the art that for convenience and simplicity of description, the system of foregoing description
It with the specific work process of device, can refer to corresponding processes in the foregoing method embodiment, details are not described herein.
In addition, in the description of the embodiment of the present invention unless specifically defined or limited otherwise, term " installation ", " phase
Even ", " connection " shall be understood in a broad sense, for example, it may be being fixedly connected, may be a detachable connection, or be integrally connected；It can
Can also be electrical connection to be mechanical connection；It can be directly connected, can also indirectly connected through an intermediary, Ke Yishi
Connection inside two elements.For the ordinary skill in the art, above-mentioned term can be understood at this with concrete condition
Concrete meaning in invention.
It, can be with if the function is realized in the form of SFU software functional unit and when sold or used as an independent product
It is stored in a computer read/write memory medium.Based on this understanding, technical scheme of the present invention is substantially in other words
The part of the part that contributes to existing technology or the technical solution can be expressed in the form of software products, the meter
Calculation machine software product is stored in a storage medium, including some instructions are used so that a computer equipment (can be
People's computer, server or network equipment etc.) it performs all or part of the steps of the method described in the various embodiments of the present invention.
And storage medium above-mentioned includes：USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), arbitrary access are deposited
The various media that can store program code such as reservoir (RAM, Random Access Memory), magnetic disc or CD.
In the description of the present invention, it should be noted that term "center", "upper", "lower", "left", "right", "vertical",
The orientation or positional relationship of the instructions such as "horizontal", "inner", "outside" be based on the orientation or positional relationship shown in the drawings, merely to
Convenient for the description present invention and simplify description, do not indicate or imply the indicated device or element must have a particular orientation,
With specific azimuth configuration and operation, therefore it is not considered as limiting the invention.
In addition, term " first ", " second ", " third " are used for description purposes only, it is not understood to indicate or imply phase
To importance.
Finally it should be noted that：Embodiment described above, only specific implementation mode of the invention, to illustrate the present invention
Technical solution, rather than its limitations, scope of protection of the present invention is not limited thereto, although with reference to the foregoing embodiments to this hair
It is bright to be described in detail, it will be understood by those of ordinary skill in the art that：Any one skilled in the art
In the technical scope disclosed by the present invention, it can still modify to the technical solution recorded in previous embodiment or can be light
It is readily conceivable that variation or equivalent replacement of some of the technical features；And these modifications, variation or replacement, do not make
The essence of corresponding technical solution is detached from the spirit and scope of technical solution of the embodiment of the present invention, should all cover the protection in the present invention
Within the scope of.Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. a kind of intrusion behavior recognition methods, which is characterized in that including：
Obtain time series signal, wherein the time series signal is that the tested amplitude for enclosing boundary's vibration arranges at any time
Sequence signal；
Characteristic quantity is extracted from the time series signal, wherein the characteristic quantity includes：The tested maximum for enclosing boundary's vibration
The maximum value of amplitude, energy, spectrum energy breadth coefficient, energy gradient and energy gradient gradient；
Intrusion behavior identification is carried out using the characteristic quantity as the input quantity of machine learning algorithm, to be calculated by the machine learning
The output quantity of method determines the tested intrusion behavior type for enclosing boundary.
2. according to the method described in claim 1, it is characterized in that, extract characteristic quantity from the time series signal, including：
Amplitude maximum in the time series signal is determined as the tested peak swing for enclosing boundary's vibration；
Discrete Fourier transform is carried out to the time series signal, obtains the tested spectrogram for enclosing boundary's vibration, and from institute
It states and extracts target characteristic amount in spectrogram, wherein the target characteristic amount includes：The tested energy for enclosing boundary's vibration, frequency spectrum
The maximum value of Energy distribution coefficient, energy gradient and energy gradient gradient.
3. according to the method described in claim 2, it is characterized in that, carrying out discrete fourier change to the time series signal
It changes, obtains the tested spectrogram for enclosing boundary and vibrating, and target characteristic amount is extracted from the spectrogram, including：
Window is handled by preset signals to be split the time series signal, obtains multiple Time Sub-series signals；
Discrete Fourier transform is carried out to the Time Sub-series signal in the signal processing window, obtains multiple sub- frequency spectrums
Figure；
The target characteristic amount is extracted from multiple sub- spectrograms.
4. according to the method described in claim 3, it is characterized in that, extracting the target signature from multiple sub- spectrograms
Amount, including：
Multiple default division ranges of spectrum distribution are obtained, and the tested boundary that encloses is calculated based on the sub- spectrogram and is vibrated more
A default spectrum energy breadth coefficient divided in range；
The sum of the vibrational energy in the sub- spectrogram is calculated, multiple energy values are obtained, the set of multiple energy values is true
It is set to the tested energy for enclosing boundary's vibration；
The ratio of any two adjacent energies in multiple energy values is determined as the energy gradient；
By in multiple energy gradients, the ratio of any two energy gradients is determined as the energy gradient ladder
Degree, to search the maximum value of the energy gradient gradient from the energy gradient gradient.
5. according to the method described in claim 1, it is characterized in that, using the characteristic quantity as the input quantity of machine learning algorithm
Intrusion behavior identification is carried out, including：
Obtain trained neural network, wherein the trained neural network is hidden including an input layer, three
Layer and an output layer；
The characteristic quantity is input to the neural network in the input layer；
The recognition result of the neural network is obtained from the output layer, and the recognition result is determined as described be tested and encloses boundary
Intrusion behavior type.
6. according to the method described in claim 5, it is characterized in that,
The spectrum energy breadth coefficient includes low frequency energy breadth coefficient, intermediate frequency Energy distribution coefficient and high-frequency energy distribution system
Number, the intrusion behavior type include it is following any one：No intrusion behavior is tapped, is rocked, climbing；
The input layer includes seven parameter input ends, and seven parameter input ends are respectively used to input：Described be tested encloses boundary
The peak swing of vibration, energy, low frequency energy breadth coefficient, intermediate frequency Energy distribution coefficient, high-frequency energy breadth coefficient, energy quantitative change
The maximum value of rate and energy gradient gradient；
And the output layer includes four result output ends, four result output ends are respectively used to export：Without intrusion behavior,
It taps, rock, climb.
7. a kind of intrusion behavior identification device, which is characterized in that including：
Acquisition module, for obtaining time series signal, wherein the time series signal be the tested amplitude for enclosing boundary's vibration with
Sequence signal made of Time alignment；
Extraction module, for extracting characteristic quantity from the time series signal, wherein the characteristic quantity includes：It is described tested
Enclose the peak swing of boundary's vibration, the maximum value of energy, spectrum energy breadth coefficient, energy gradient and energy gradient gradient；
Identification module, for carrying out intrusion behavior identification using the characteristic quantity as the input quantity of machine learning algorithm, to pass through
The output quantity of the machine learning algorithm determines the tested intrusion behavior type for enclosing boundary.
8. device according to claim 7, which is characterized in that the extraction module, including：
First extraction unit encloses boundary's vibration for the amplitude maximum in the time series signal to be determined as described be tested
Peak swing；
Second extraction unit obtains described tested enclosing boundary and shaking for carrying out discrete Fourier transform to the time series signal
Dynamic spectrogram, and target characteristic amount is extracted from the spectrogram, wherein the target characteristic amount includes：Described be tested is enclosed
The energy of boundary's vibration, the maximum value of spectrum energy breadth coefficient, energy gradient and energy gradient gradient.
9. device according to claim 8, which is characterized in that second extraction unit, including：
Divide subelement, the time series signal is split for handling window by preset signals, obtains multiple sons
Time series signal；
Subelement is converted, for carrying out discrete fourier change to the Time Sub-series signal in the signal processing window
It changes, obtains multiple sub- spectrograms；
Subelement is extracted, for extracting the target characteristic amount from multiple sub- spectrograms.
10. device according to claim 9, which is characterized in that the extraction subelement is used for：
Multiple default division ranges of spectrum distribution are obtained, and the tested boundary that encloses is calculated based on the sub- spectrogram and is vibrated more
A default spectrum energy breadth coefficient divided in range；
The sum of the vibrational energy in the sub- spectrogram is calculated, multiple energy values are obtained, the set of multiple energy values is true
It is set to the tested energy for enclosing boundary's vibration；
The ratio of any two adjacent energies in multiple energy values is determined as the energy gradient；
By in multiple energy gradients, the ratio of any two energy gradients is determined as the energy gradient ladder
Degree, to search the maximum value of the energy gradient gradient from the energy gradient gradient.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810243918.7A CN108399696A (en) | 2018-03-22 | 2018-03-22 | Intrusion behavior recognition methods and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810243918.7A CN108399696A (en) | 2018-03-22 | 2018-03-22 | Intrusion behavior recognition methods and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108399696A true CN108399696A (en) | 2018-08-14 |
Family
ID=63092932
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810243918.7A Pending CN108399696A (en) | 2018-03-22 | 2018-03-22 | Intrusion behavior recognition methods and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108399696A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109064696A (en) * | 2018-08-17 | 2018-12-21 | 成都九洲电子信息系统股份有限公司 | The optical fiber perimeter security system realized based on deep learning |
| CN109377692A (en) * | 2018-11-13 | 2019-02-22 | 武汉研希科技有限公司 | A kind of Intelligent optical fiber vibration anti-intrusion method for early warning and system |
| CN110276921A (en) * | 2019-07-08 | 2019-09-24 | 重庆邮电大学 | Indoor invasion and space structure based on radio signal characteristics change recognition methods |
| CN112309063A (en) * | 2020-10-30 | 2021-02-02 | 魏运 | Method and device for extracting hybrid fiber intrusion signal feature spectrum |
| CN112567656A (en) * | 2018-08-15 | 2021-03-26 | 三菱电机株式会社 | Signal detection device and signal detection method |
| CN114323246A (en) * | 2021-12-17 | 2022-04-12 | 北京特里尼斯石油技术股份有限公司 | Pipeline safety monitoring method and device |
| CN114926947A (en) * | 2022-05-09 | 2022-08-19 | 成都理想科技开发有限公司 | Method for detecting vibration and alarming |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010026917A (en) * | 2008-07-23 | 2010-02-04 | Mitsubishi Electric Engineering Co Ltd | Intruder detection device |
| CN102968868A (en) * | 2012-10-31 | 2013-03-13 | 武汉理工光科股份有限公司 | Fiber bragg grating perimeter intrusion behavior identification method and system based on time domain envelopment analysis |
| CN106600869A (en) * | 2016-12-31 | 2017-04-26 | 上海华魏光纤传感技术有限公司 | Fence intrusion identification method for fiber fence security protection system |
| CN106600870A (en) * | 2016-12-31 | 2017-04-26 | 上海华魏光纤传感技术有限公司 | Distributed optical fiber fence vibration invasion identifying system |
| CN106874833A (en) * | 2016-12-26 | 2017-06-20 | 中国船舶重工集团公司第七0研究所 | A kind of mode identification method of vibration event |
| CN107180521A (en) * | 2017-04-19 | 2017-09-19 | 天津大学 | Optical fiber perimeter security protection intrusion event recognition methods and device based on comprehensive characteristics |
| CN107369282A (en) * | 2017-08-23 | 2017-11-21 | 武汉理工光科股份有限公司 | A kind of buried COTDR circumferences intrusion alarm system and signal processing method |
-
2018
- 2018-03-22 CN CN201810243918.7A patent/CN108399696A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010026917A (en) * | 2008-07-23 | 2010-02-04 | Mitsubishi Electric Engineering Co Ltd | Intruder detection device |
| CN102968868A (en) * | 2012-10-31 | 2013-03-13 | 武汉理工光科股份有限公司 | Fiber bragg grating perimeter intrusion behavior identification method and system based on time domain envelopment analysis |
| CN106874833A (en) * | 2016-12-26 | 2017-06-20 | 中国船舶重工集团公司第七0研究所 | A kind of mode identification method of vibration event |
| CN106600869A (en) * | 2016-12-31 | 2017-04-26 | 上海华魏光纤传感技术有限公司 | Fence intrusion identification method for fiber fence security protection system |
| CN106600870A (en) * | 2016-12-31 | 2017-04-26 | 上海华魏光纤传感技术有限公司 | Distributed optical fiber fence vibration invasion identifying system |
| CN107180521A (en) * | 2017-04-19 | 2017-09-19 | 天津大学 | Optical fiber perimeter security protection intrusion event recognition methods and device based on comprehensive characteristics |
| CN107369282A (en) * | 2017-08-23 | 2017-11-21 | 武汉理工光科股份有限公司 | A kind of buried COTDR circumferences intrusion alarm system and signal processing method |
Non-Patent Citations (3)
| Title |
|---|
| 朱程辉 等: "基于时频特征的光纤周界入侵振动信号识别与定位", 《中国激光》 * |
| 盛媛媛 等: "光纤传感器振动信号特征提取研究", 《光电技术应用》 * |
| 邹东伯 等: "分布式光纤振动传感信号识别的研究", 《激光技术》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112567656A (en) * | 2018-08-15 | 2021-03-26 | 三菱电机株式会社 | Signal detection device and signal detection method |
| CN112567656B (en) * | 2018-08-15 | 2022-08-19 | 三菱电机株式会社 | Signal detection device, signal detection method, control circuit, and storage medium |
| CN109064696A (en) * | 2018-08-17 | 2018-12-21 | 成都九洲电子信息系统股份有限公司 | The optical fiber perimeter security system realized based on deep learning |
| CN109377692A (en) * | 2018-11-13 | 2019-02-22 | 武汉研希科技有限公司 | A kind of Intelligent optical fiber vibration anti-intrusion method for early warning and system |
| CN109377692B (en) * | 2018-11-13 | 2024-03-12 | 武汉研希科技有限公司 | Intelligent optical fiber vibration anti-intrusion early warning method and system |
| CN110276921A (en) * | 2019-07-08 | 2019-09-24 | 重庆邮电大学 | Indoor invasion and space structure based on radio signal characteristics change recognition methods |
| CN112309063A (en) * | 2020-10-30 | 2021-02-02 | 魏运 | Method and device for extracting hybrid fiber intrusion signal feature spectrum |
| CN114323246A (en) * | 2021-12-17 | 2022-04-12 | 北京特里尼斯石油技术股份有限公司 | Pipeline safety monitoring method and device |
| CN114926947A (en) * | 2022-05-09 | 2022-08-19 | 成都理想科技开发有限公司 | Method for detecting vibration and alarming |
| CN114926947B (en) * | 2022-05-09 | 2024-04-26 | 成都理想科技开发有限公司 | Method for detecting vibration and alarming |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108399696A (en) | Intrusion behavior recognition methods and device | |
| CN104751580B (en) | The recognition methods of distributing optical fiber sensing signal mode and system | |
| Christie et al. | Acoustics based terrain classification for legged robots | |
| CN105095624B (en) | A kind of recognition methods of Fibre Optical Sensor vibration signal | |
| CN110570613A (en) | Fence vibration intrusion positioning and mode identification method based on distributed optical fiber system | |
| CN105931402A (en) | Optical fiber perimeter intrusion monitoring method based on image recognition | |
| Zhong et al. | A study on the stationarity and Gaussianity of the background noise in land-seismic prospecting | |
| CN106886043B (en) | Reservoir detection method based on seismic data deep learning | |
| CN103183263A (en) | Security system of structure and elevator provided with same | |
| CN111520192B (en) | Non-contact tunnel engineering construction rock burst real-time forecasting optimization method | |
| CN104766099B (en) | Distribution type fiber-optic circumference vibration signal processing and recognition methods based on image | |
| CN111160106A (en) | Method and system for extracting and classifying optical fiber vibration signal features based on GPU | |
| CN110138471A (en) | A kind of detection method and relevant apparatus | |
| Ghorbani et al. | A modern method to improve of detecting and categorizing mechanism for micro seismic events data using boost learning system | |
| CN114742096A (en) | Intrusion alarm method and system based on vibration optical fiber detection and complete action extraction | |
| Birnie et al. | Bidirectional recurrent neural networks for seismic event detection | |
| CN106197646A (en) | The detection of a kind of fiber-optic vibration reduces the method for error and fine vibration detection device | |
| CN114545504A (en) | Desert seismic exploration random noise modeling method based on deep learning | |
| Yi et al. | An intelligent crash recognition method based on 1DResNet-SVM with distributed vibration sensors | |
| CN117454187A (en) | Integrated model training method based on frequency domain limiting target attack | |
| Wang et al. | Identification and classification of water absorption profile of distributed optical fiber vibration signal based on XGBoost algorithm | |
| CN107688197B (en) | Thin bed prediction method and apparatus | |
| Peng et al. | Seismic signal recognition using improved BP neural network and combined feature extraction method | |
| CN115795353A (en) | Underground metal target classification method and system based on unbalanced data set | |
| CN109285598A (en) | The mobile phone projection technology for having color mood regulation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180814 |