CN101556724B - Safety management system of optical fiber perimeter and pattern recognition method thereof - Google Patents
Safety management system of optical fiber perimeter and pattern recognition method thereof Download PDFInfo
- Publication number
- CN101556724B CN101556724B CN2008101036518A CN200810103651A CN101556724B CN 101556724 B CN101556724 B CN 101556724B CN 2008101036518 A CN2008101036518 A CN 2008101036518A CN 200810103651 A CN200810103651 A CN 200810103651A CN 101556724 B CN101556724 B CN 101556724B
- Authority
- CN
- China
- Prior art keywords
- signal
- optical fiber
- management system
- intrusion
- safety management
- 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.)
- Expired - Fee Related
Links
Images
Landscapes
- Alarm Systems (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
The invention relates to a safety management system of optical fiber perimeter, comprising a light source, a multiplexing system, an optical fiber sensor array, a demodulating system, a signal processing and pattern recognition system and an alarm management system. The input end of the multiplexing system is connected with the light source by adopting optical fiber, and the multiplexing system plays the role of optical fiber channel multiplexing; the input end of the optical fiber sensor array is connected with the output end of the multiplexing system by adopting optical fiber to sense the change of external physical quantity; the input end of the demodulating system is connected with the output end of the multiplexing system to convert optical signals into electric signals and restore the change of the outside physical quantity which is sensitive to the optical fiber sensor array; the input end of the signal processing and pattern recognition system is connected with the output end of the demodulating system to carry out processing and recognition to the signals output by the demodulating system; and the input end of the alarm management system is connected with the output end of the signal processing and pattern recognition system to carry out alarming according to the output result of the signal processing and pattern recognition system and manage the communication with the outside.
Description
Technical field
The present invention relates to a kind of safety management system of optical fiber perimeter, comprise fibre optic sensor arra, signal Processing and demodulating system and pattern recognition system.This invention can be applied to the monitoring of border security, oil pipeline safety and the occasions such as protection of important foundation facility.Not only comprise optical fiber sensing system itself, and have powerful communication, control and data managing capacity.
Background technology
Circumference safety is one of of paramount importance aspect in the safety-security area.From the boundary line safety of long distance, the remote monitoring of oil pipeline, the protection to short-range important foundation facility all has great significance to national security and national economy.In recent years, along with the continuous progress of computer technology, the communication technology, sensing technology, various new security protection means have appearred.Compare with means such as traditional high-tension cable, recording monitors, the distribution type fiber-optic monitoring scheme has many advantages, by Optical Fiber Transmission, is not subjected to the interference of external electromagnetic signal as signal; Compare with radio sensing network, be not subjected to the restriction of power consumption condition, can realize real-time remote monitoring.
Fiber optic sensor technology has obtained in multiple field to use widely, because high sensitivity, radioresistance and electromagnetic interference (EMI) are arranged and be easy to advantages such as large-scale network-estabilishing, optical fiber acceleration transducer and network thereof have incomparable technical advantage in circumference safety, petroleum pipe line monitoring field.
Existing optical fiber sensing system as patent CN1862239A, CN1995935A, adopts ordinary optic fibre to constitute the scheme of interferometer, and two significant disadvantages are arranged: the first, and vibration detecting sensitivity is not high; The second, bearing accuracy is not high.Above two kinds of schemes just experienced simple pressure signal, quantity of information is limited, utilizing these signals is far from being enough to the variety of event type.
Optical fiber acceleration transducer is based on the energy transducer of the ground vibration acceleration of optical fiber technology, the earthquake motion signal on ground is transformed into the variation of light signal phase place, utilize detection means to restore actual signal, the ground vibration signal is done real-time monitoring.In the security protection application scenario, system can carry out early warning for possible intrusion event, and can differentiate, classify the type of incident, and directive function is played in the quick enforcement of counte-rplan.
Come from ground unmanned sensor (UGS) system of battlefield in using the earliest based on the Target Recognition of earthquake motion signal.Generally comprising multiple sensors such as sound, vibrations, infrared, image in the UGS system forms.The earthquake motion sensor is a kind of effective detection mode, and when target was invaded, target travel produced the stimulation to ground, thereby produces the earthquake motion signal to propagating at a distance.Carry out perception by the earthquake motion sensor,, can make the judgement of event type by analysis to various earthquake motion signals.According to document, because the earthquake motion signal is typical nonstationary random signal, so the identification of all kinds of incidents is needed well to extract the feature of signal.Simultaneously, also the recognizer of high-accuracy is claimed.
Technology contents
The objective of the invention is to, a kind of safety management system of optical fiber perimeter and mode identification method thereof are provided, this system and method thereof can be applied to the application of boundary line safety, oil pipeline safety and other circumference safety occasion.
The invention provides a kind of safety management system of optical fiber perimeter, it is characterized in that, this system comprises:
One light source;
One multiplex system, the input end of this multiplex system adopt optical fiber to be connected with light source, and this multiplex system plays the multiplexing effect of optical-fibre channel;
One fibre optic sensor arra, the input end of this fibre optic sensor arra adopts optical fiber to be connected with the output terminal of multiplex system, this fibre optic sensor arra comprises the earthquake motion sensor that a plurality of polyphones connect, and the caused earthquake motion signal of responsive extraneous intrusion incident can be realized remote multiple spot monitoring;
One demodulating system, the input end of this demodulating system is connected with the output terminal of multiplex system, and light signal is changed into electric signal, and recover fibre optic sensor arra the variation of responsive external physical quantity;
One signal Processing and pattern recognition system, this signal Processing is connected with the output terminal of demodulating system with the input end of pattern recognition system, and the signal of demodulating system output is handled and discerned;
One alarm management system, the input end of this alarm management system is connected with the output terminal of pattern recognition system with signal Processing, reports to the police according to the output result of signal Processing and pattern recognition system, and the management and the external world communicate by letter.
The present invention also provides a kind of mode identification method of safety management system of optical fiber perimeter, and this method adopts the described system of claim 1, it is characterized in that this method comprises the steps:
Step 1: signal input step, this step are the intrusion signals that the receiving demodulation system is recovered;
Step 2: Signal Pretreatment step, this step are carried out the pre-service of denoising to the signal of signal input step;
Step 3: characteristic extraction step, this step extracts the essential attribute of different intrusion incidents to by conversion and statistics to signal, for further identification is ready;
Step 4: the design procedure of sorter, this step structure is to the intrusion signal employed sorter model of classifying;
Step 5: learning procedure, this learning procedure is included between characteristic extraction step and the classifier design step, characteristic parameter to different event in the feature database, constantly be input in the sorter that has designed, when classification performance reaches preferably, learning procedure is finished, and at this moment, preserves the classifier parameters of having succeeded in school;
Step 6: the output result, this step provides the type that the intrusion incident is taken place.
Description of drawings
For further specifying concrete technology contents of the present invention, below in conjunction with embodiment and accompanying drawing describes in detail as after, wherein:
Fig. 1 is a safety management system of optical fiber perimeter block diagram of the present invention;
Fig. 2 is an expression sef-adapting filter subsystem;
Fig. 3 is a representative ring context self-adapting constant false alarm rate algorithm flow chart;
Fig. 4 is that representation feature extracts process flow diagram;
Fig. 5 is an expression pattern-recognition process flow diagram;
Fig. 6 is a presentation of events investigative range synoptic diagram;
Fig. 7 is that the expression personnel invade institute and cause the earthquake motion signal;
Fig. 8 is that expression vehicle intrusion institute causes the earthquake motion signal;
Fig. 9 is an expression safety management system of optical fiber perimeter functional structure chart.
Embodiment
See also shown in Figure 1, a kind of safety management system of optical fiber perimeter of the present invention, this system comprises:
One light source 101;
One multiplex system 102, the input end of this multiplex system 102 adopt optical fiber to be connected with light source 101, and this multiplex system plays the multiplexing effect of optical-fibre channel;
One fibre optic sensor arra 103, the input end of this fibre optic sensor arra 103 adopt optical fiber to be connected the variation of responsive external physical quantity with the output terminal 102 of multiplex system; This fiber array 103 comprises: the sensor 1031 that a plurality of polyphones connect, can realize remote multiple spot monitoring; These a plurality of sensors 1031 are shock sensors, the caused earthquake motion signal of responsive extraneous intrusion incident;
One demodulating system 104, the input end of this demodulating system 104 is connected with the output terminal of multiplex system 102, the electric signal that changes into of light signal, and recover fibre optic sensor arra 103 the variation of responsive external physical quantity;
One signal Processing and pattern recognition system 105, this signal Processing is connected with the output terminal of demodulating system 104 with the input end of pattern recognition system 105, and the signal of demodulating system 104 outputs is handled and discerned; This signal Processing and pattern recognition system 105 comprise:
One photoelectric conversion module 1051 (in conjunction with consulting Fig. 9), this module input is connected by the output terminal of optical fiber with multiplex system 102, and light signal is changed into electric signal;
One D/A converter module 1052, this module input is connected with photoelectric conversion module 1051 output terminals, and analog electrical signal is changed into digital electric signal;
One multiple digital signal Processing plate module 1053, this module input is connected with D/A converter module 1052 output terminals, reception from the signal of D/A converter module the row operation of going forward side by side handle, recover the earthquake motion signal that is detected by a plurality of Fibre Optical Sensors 1031, and vibration signal is handled and is discerned over the ground.
One alarm management system 106, the input end of this alarm management system 106 is connected with the output terminal of signal Processing with pattern recognition system 105, reports to the police according to the output result of signal Processing and pattern recognition system 105, and the management and the external world communicate by letter;
Described alarm management system 106 comprises:
One bus 1061 (in conjunction with consulting Fig. 9) is responsible for communicating by letter between multiple digital signal-processing board 1053 and the main frame 1062;
One main frame 1062 is as the arithmetic core of safety management system of optical fiber perimeter;
One invades log database 1063, communicates by user interface 1066 and main frame, invades daily record in the record circumference; Described intrusion log database 1063 comprises: intrusion time, the type of invading incident, intrusion sum of events are invaded locale;
One property data base 1064 communicates by user interface 1066 and main frame, various intrusion event features in the record circumference safety; This property data base comprises: the caused earthquake motion original signal of intrusion incident; Invade the characteristic quantity of incident through the difference after the feature extraction;
One communication module 1065, this module manages by main frame 1062, is responsible for and extraneous communicating by letter; This communication module 1065 comprises following feature: a serial ports can be realized the transmission of alarm log, with other security protection mode, utilizes video, the infrared fusion, communicates with the result and merges; Directly link to each other by Ethernet interface, have Telnet and management function with LAN (Local Area Network); To realize invading the transmission of event vibration signal remote live;
One user interface 1066, this user interface is based on the human-computer interaction interface on the main frame, and the system manager can be configured safety management system of optical fiber perimeter by this interface.
See also shown in Figure 5ly, the invention provides a kind of mode identification method of safety management system of optical fiber perimeter, this method adopts foregoing system, it is characterized in that, this method comprises the steps:
Step 1: signal input step 501, this step are the intrusion signals that receiving demodulation system 104 is recovered;
Step 2: Signal Pretreatment step 502, this step are carried out the pre-service of denoising to the signal of signal input step 501; Wherein said Signal Pretreatment step 502 comprises following steps: the branch frame of signal, windowing step, this step are responsible for continuous intrusion signal is divided into several Frames, to satisfy further the needs of handling, discerning; The denoising step of signal, this step is responsible for removing the influence of outside noise, improves the signal to noise ratio (S/N ratio) of intrusion signal, improves the detection and the recognition accuracy of system; Whether the detection steps of intrusion signal, this step signal after to denoising is surveyed, represent to have had the intrusion incident to take place to determine this signal;
The detection steps of described intrusion signal is a self-adaption constant false alarm rate intrusion detection method, and the method includes the steps of: branch frame signal input step 301 (in conjunction with consulting Fig. 3), and what this step received is the signal of intrusion signal after undue frame, denoising; Counting statistics running mean energy step 302, this step is calculated the average energy in a period of time; Threshold value step of updating 303, this step is upgraded according to following formula according to statistics running mean energy step 302 calculated energy, obtains adaptive threshold; The interior data point number 304 of calculation window greater than adaptive threshold; Determining step 305, in the window that this step determining step draws greater than the data point number of adaptive threshold whether greater than a predefined value, if outer signals is detected so, if not, will rotate back into branch frame signal input step 301 so proceeds; The output result step;
The denoising step of described signal comprises: an adaptive noise filter, and this adaptive noise filter is responsible for the removal of extraneous periodic noise or interference, improves the signal to noise ratio (S/N ratio) of input signal;
Described sef-adapting filter comprises:
One transmission system 201 (in conjunction with consulting Fig. 2), this transmission system represent that extraneous periodic noise signal is along the caused change of face of land transmission transmission;
One adaptive algorithm 202, this adaptive algorithm step constitute the feedback between transmission system 201 and the signal output, and according to the feedback signal of output, the parameter of the adjusting transmission system 201 that adaptive algorithm is real-time makes real-time being removed of periodic noise;
Step 3: characteristic extraction step 503, this step extracts the essential attribute of different intrusion incidents to by conversion and statistics to signal, for further identification is ready;
Described characteristic extraction step 503 comprises following steps: the signal receiving step, and the thing that this step receives is through the intrusion signal after the Signal Pretreatment 502; The power spectrum shift step 5031 of signal (in conjunction with consulting Fig. 4), this step calculates the power Spectral Estimation of signal, prepares for extracting power spectrum aspect signal; The wavelet transformation step 5032 of signal, this step are to wavelet transformation, for the feature of extracting the small echo aspect is prepared; The power spectrum resonance peak selects step 5033, this step to search out the position of the formant frequency in the power spectrum of signal, as characteristic quantity; Power spectrum shape Statistics amount calculation procedure 5034, this step are calculated the shape Statistics amount of the power spectrum that obtains from power spectrum shift step 5031, as characteristic quantity; Wavelet energy calculation procedure 5035, this step calculates wavelet energy according to the wavelet coefficient that step 5032 obtains, as characteristic quantity; Normalization associating feature 5036, resonance peak characteristic quantity, power spectrum shape Statistics characteristic quantity, the wavelet energy characteristic quantity that this step selects the power spectrum resonance peak step 5033 to obtain united one group of new characteristic quantity of formation, for unitized unit, does normalized;
Step 4: the design procedure 504 of sorter, this step structure is to the intrusion signal employed sorter model of classifying;
The design procedure 504 used sorters of described sorter are the sorters of a support vector machine, and support vector machine is that the primitive characteristics amount is mapped in the higher dimensional space, asks for the step of optimal classification face in higher dimensional space;
Step 5: learning procedure 505, this learning procedure is included between characteristic extraction step 503 and the classifier design step 504, characteristic parameter to different event in the feature database, constantly be input in the sorter that has designed, when classification performance reaches preferably, learning procedure is finished, and at this moment, preserves the classifier parameters of having succeeded in school;
Described learning procedure 505 comprises:
According to invading affair character extracting method structural attitude sample set, set up training set and test set respectively; By a radially basic kernel function, characteristic quantity from the status spatial mappings to higher dimensional space, in higher dimensional space, ask for the optimal classification face; One can be used for polytypic support vector machine, and many classification problems are changed into the pairwise classification problem earlier;
Step 6: output result 506, this step provides the type that the intrusion incident is sent out.
Please consult Fig. 1 again, Fig. 1 has illustrated the basic composition of optical fiber perimeter security system, mainly comprises light source 101, multiplex system 102, fibre optic sensor arra 103, demodulating system 104, signal Processing and pattern recognition system 105 and alarm management system 106 compositions.
Wherein light source 101 is wideband light sources, for example can be the ASE light source of a wavelength coverage from 1530nm to 1560nm, or a super-radiance light emitting diode (SLD).What light source sent couples light in the optical fiber, through multiplex system 102, arrives fibre optic sensor arra.Multiplex technique in many optical fiber communications such as wavelength-division multiplex (WDM), Time Division Multiplexing and space division multiplexing (SDM) technology all have been applied to optical fiber sensing system.Be without loss of generality, in this enforcement, illustrate with wavelength-division multiplex technique.Fibre optic sensor arra 103 is in series by a plurality of sensor nodes, below being that example illustrates based on fiber grating (FBG), in array each fiber grating corresponding to a specific centre wavelength, i.e. Prague (Bragg) wavelength.The light that sends when light source arrives each fiber grating, will reflect in centre wavelength so.When intrusion incident in the external world's took place, the vibration of the caused soil of incident (acceleration signal) can cause the stretching of grating, and this stretching causes the drift of centre wavelength, and is as follows:
Δλ
B＝λ
BG
eε
In the following formula, λ
BBe the bragg wavelength of grating, G
eBe a coefficient, ε is the grating strain amount.Like this, the variation of the acceleration of ground vibration signal just is converted into the variation of the centre wavelength of wavelength.Multiplex system 102 is responsible in the sensor array, and different fiber gratings decomposes corresponding operation wavelength, can separate wavelength division multiplexer with one and realize.Demodulating system 104 changes into wavelength change can be for the digital electric signal of handling.The scheme of demodulation has multiple, and it is method comparatively commonly used that phase place produces carrier wave (PGC) method, and the detail of its computing method can be consulted relevant references.
Fibre optic sensor arra 103 is laid according to circumference planning, and the distance between two sensors is determining the precision that total system is surveyed.According to the measuring and calculating of sensitivity, the personnel that single-sensor can detect beyond the 60m walk about, and can reach more than the 120m the investigative range of vehicle.In order to be unlikely to have the blind area to exist, the distance between two sensors is made as 50m, and to the circumference of a 2km, required number of sensors is 40.Signal demodulating system 104 recovers the earthquake motion signal of all passages, and entering signal is handled and identification process then.
Signal Processing and pattern recognition system as shown in Figure 1 mainly comprise the flow process of the pre-service and the pattern-recognition of signal.As shown in Figure 5, the pre-service 501 main effects of signal are the denoisings of signal, and useful signal is enhanced.The method of signal denoising commonly used has traditional filtering method, methods such as sef-adapting filter.Because the earthquake motion signal that the intrusion incident causes is mainly based on low-frequency component, so adopt the bandpass filter of a bandwidth 1Hz～200Hz, the high frequency noise filtering that the external world is possible.But simultaneously, if there is the one-period noise in applied environment, and amplitude is bigger, so just can utilize a sef-adapting filter that it is removed, and the structure of sef-adapting filter as shown in Figure 2.
As shown in Figure 2, sef-adapting filter is made up of a transmission system 201 and adaptive algorithm module 202 and an arithmetic element 203 is formed.An input that passage is wanted signal d and additive noise n of sef-adapting filter, another one then is noise signal n ', they are from same noise source, but by extraneous environment change, by adaptive step, the variation of real-time tracking outside noise, through after the arithmetical unit 203, noise is by real-time balancing out.In the present invention, use to be adjacent to the input of the resulting signal of shock sensor as two passages, Wai Jie periodic noise so will be by real-time balancing out, and useful signal is enhanced.Lowest mean square (LMS) algorithm is adopted in the realization of sef-adapting filter, and the update algorithm of sef-adapting filter is as follows:
w
k(n+1)＝w
k(n)+2βe(n)x(n-k)，k＝0，1，L，N-1
Signal is through after the pre-treatment step, and useful signal is enhanced, and enters constant false alarm rate target detection algorithm unit, as shown in Figure 3.The constant false alarm rate algorithm divides in frame 301, slip statistical average unit 302, threshold value updating block 303, the window threshold value counting unit 304 and identifying unit 305 to form by signal.As previously mentioned, the branch frame signal enters running mean unit 302 after by windowing process, and this unit calculates statistical average in time span L square with energy, and is as follows:
Threshold value updating block 303 is responsible for calculating adaptive threshold according to top result of calculation, and is as follows:
β[m]＝α*LMeanE[m-1]
α represents the parameter that can regulate to obtain the compromise between detectivity and the false alarm rate in the following formula.At last by absolute value in the identifying unit statistics sequence greater than the counting of adaptive threshold, if greater than a threshold value, just be judged to be the intrusion incident and take place, otherwise continue to upgrade threshold value.
When an intrusion incident is detected, enter the pattern-recognition step.Pattern-recognition comprises feature extraction, and the design procedure 504 of sorter and learning procedure 505 are formed.Characteristic extraction step as shown in Figure 4, comprise power Spectral Estimation 5031 and wavelet transformation step 5032, then from the power Spectral Estimation result, calculate resonance peak 5033 and shape Statistics amount 5034, calculate wavelet energy 5035, finally obtain normalized associating feature 5036 by wavelet coefficient.This method to provide actual outfield actual signal result in order better illustrating, to have provided the signal when personnel walk about with wheeled vehicle process single-sensor device respectively.
Power Spectral Estimation comprises classical way and modern spectral estimation method.For better description, adopt this periodogram analysis that advances that the power spectrum of signal is estimated.Fig. 9 shows the power Spectral Estimation of the earthquake motion signal of personnel's signal and wheeled vehicle amount.
Sorting technique based on SVM is to ask for the step of optimal classification face in higher dimensional space, by one " kernel function ", original feature vector is mapped to a higher dimensional space.Adopt the libSVM software package to test in testing procedure, its basic decision function is as follows:
In following formula,
Be the optimum solution corresponding with support vector, x
iRepresent each support vector, b
*Be the threshold value of classification, can obtain by any a pair of support vector.
Training set is to obtain by top characteristic extraction step, as the input of support vector machine, adopts the optimized parameter of the method searching support vector machine of cross validation in training step.After training was finished, the model that trains kept as recognition mode, used down in order to actual environment.
The demodulating system 104 of Fibre Optical Sensor network and signal Processing 105 steps are all finished on digital signal processing (DSP) plate.Shown in 9, light signal enters digital-to-analog conversion 1052 steps through behind the photoelectric conversion module 1051, enters the DSP module.The polylith dsp board links to each other with main frame 1062 by bus 1053.Main frame 1062 can be realized operations such as DSP module initialization, parameter configuration.Polylith DSP disposable plates constitutes parallel organization, can finish real-time demodulation of 16 road Fibre Optical Sensors and pattern recognition process for every, and the main frame 1062 that sends to that recognition result is real-time is handled.
Management software system can obtain original signal, characteristic and the recognition result etc. of process dsp board by bus 1061.Software comprises a user interface 1066, shows the signal of raw data, the daily record of characteristic quantity and intrusion thereof (time, place, event type etc.); By user interface, the user can manage and invade the affair character storehouse, and different intrusion incidents is carried out signal analysis, new intrusion incident is carried out the renewal of database.Be responsible for real-time Transmission, long-range login and the function of management of data with the communication interface 1065 in the external world.Communication interface comprises a serial port, a network interface (LAN (Local Area Network) LAN or wide area network WLAN), has not only realized the function of telecommunication, and with other security protection mode such as image, collaborative work such as infrared.
As an embodiment, in perimeter application, the earthquake motion signal that the optical fiber earthquake motion sensor personnel of recording walk about and vehicle causes when the sensor is the invasion personnel signals of walking about as shown in Figure 7; Fig. 8 shows and invades the earthquake motion signal that vehicle causes.Fig. 6 shows the incident detection scope.Personnel's signal of walking about is regular ground shock signal, and signals of vehicles then to be a variable signal along with distance constantly strengthen the step of decay gradually.Gather some dissimilar intrusion incidents in advance, set up the feature samples collection, supporting vector machine model is trained, the model that storage has been built up; When practical application, the model that trains is read in, live signal is through pre-service and feature extraction step then, and as the input of sorter, output is the result who obtains.
The characteristics of a kind of safety management system of optical fiber perimeter of the present invention:
1, multisensor node, remote Real-Time Monitoring ability;
2, the detectivity of the high detectivity of earthquake motion time, low false alarm rate;
3, the high-accuracy of multiple event identification, classification capacity;
4, accurate location and the follow-up control thereof of multiple affair;
5, the adaptive ability of applied environment;
6, powerful system alarm management, communication capacity.
Above-mentioned title of the present invention and content are decided by convenient description technology contents of the present invention, but not in order to limit interest field of the present invention. And equivalent application or element conversion, alternative that every foundation invention spirit of the present invention is done all should be encompassed in the claim protection domain of the present invention.
Claims (13)
1. a safety management system of optical fiber perimeter is characterized in that, this system comprises:
One light source;
One multiplex system, the input end of this multiplex system adopt optical fiber to be connected with light source, and this multiplex system plays the multiplexing effect of optical-fibre channel;
One fibre optic sensor arra, the input end of this fibre optic sensor arra adopts optical fiber to be connected with the output terminal of multiplex system, this fibre optic sensor arra comprises the earthquake motion sensor that a plurality of polyphones connect, and the caused earthquake motion signal of responsive extraneous intrusion incident can be realized remote multiple spot monitoring;
One demodulating system, the input end of this demodulating system is connected with the output terminal of multiplex system, and light signal is changed into electric signal, and recover fibre optic sensor arra the variation of responsive external physical quantity;
One signal Processing and pattern recognition system, this signal Processing is connected with the output terminal of demodulating system with the input end of pattern recognition system, and the signal of demodulating system output is handled and discerned;
One alarm management system, the input end of this alarm management system is connected with the output terminal of pattern recognition system with signal Processing, reports to the police according to the output result of signal Processing and pattern recognition system, and the management and the external world communicate by letter.
2. safety management system of optical fiber perimeter as claimed in claim 1 is characterized in that, wherein said signal Processing and pattern recognition system comprise:
One photoelectric conversion module, this module input is connected with the output terminal of multiplex system by optical fiber, and light signal is changed into electric signal;
One D/A converter module, this module input is connected with the photoelectric conversion module output terminal, and analog electrical signal is changed into digital electric signal;
One multiple digital signal Processing plate module, this module input is connected with the D/A converter module output terminal, reception from the signal of D/A converter module the row operation of going forward side by side handle, recover the earthquake motion signal that detects by a plurality of Fibre Optical Sensor, and vibration signal is handled and is discerned over the ground.
3. safety management system of optical fiber perimeter as claimed in claim 1 is characterized in that, wherein said alarm management system comprises:
One bus is responsible for communicating by letter between multiple digital signal-processing board and the main frame;
One main frame is as the arithmetic core of safety management system of optical fiber perimeter;
One invades log database, communicates by user interface and main frame, invades daily record in the record circumference;
One property data base communicates by user interface and main frame, various intrusion event features in the record circumference safety;
One communication module, this module manages by main frame, is responsible for and extraneous communicating by letter;
One user interface, this user interface is based on the human-computer interaction interface on the main frame, and the system manager can be configured safety management system of optical fiber perimeter by this interface.
4. safety management system of optical fiber perimeter as claimed in claim 3 is characterized in that, wherein said intrusion log database comprises: intrusion time, the type of invading incident, intrusion locale.
5. safety management system of optical fiber perimeter as claimed in claim 3 is characterized in that, wherein said property data base comprises:
The caused earthquake motion original signal of intrusion incident;
Invade the characteristic quantity of incident through the difference after the feature extraction.
6. safety management system of optical fiber perimeter as claimed in claim 3 is characterized in that, wherein said communication module comprises following feature:
A serial ports can be realized the transmission of alarm log, with other security protection mode, utilizes video, the infrared fusion, communicates with the result and merges;
Directly link to each other by Ethernet interface, have Telnet and management function with LAN (Local Area Network);
To realize invading the transmission of event vibration signal remote live.
7. the mode identification method of a safety management system of optical fiber perimeter, this method adopts the described system of claim 1, it is characterized in that this method comprises the steps:
Step 1: signal input step, this step are the intrusion signals that the receiving demodulation system is recovered;
Step 2: Signal Pretreatment step, this step are carried out the pre-service of denoising to the signal of signal input step;
Step 3: characteristic extraction step, this step extract the essential attribute of different intrusion incidents by conversion and statistics to signal, for further identification is ready;
Step 4: the design procedure of sorter, this step structure is to the intrusion signal employed sorter model of classifying;
Step 5: learning procedure, this learning procedure is included between characteristic extraction step and the classifier design step, characteristic parameter to different event in the feature database, constantly be input in the sorter that has designed, when classification performance reaches preferably, learning procedure is finished, and at this moment, preserves the classifier parameters of having succeeded in school;
Step 6: the output result, this step provides the type that the intrusion incident is taken place.
8. the mode identification method of safety management system of optical fiber perimeter as claimed in claim 7 is characterized in that, wherein said Signal Pretreatment step comprises following steps:
The branch frame of signal, windowing step, this step are responsible for continuous intrusion signal is divided into several Frames, to satisfy further the needs of handling, discerning;
The denoising step of signal, this step is responsible for removing the influence of outside noise, improves the signal to noise ratio (S/N ratio) of intrusion signal, improves the detection and the recognition accuracy of system;
Whether the detection steps of intrusion signal, this step signal after to denoising is surveyed, represent to have had the intrusion incident to take place to determine this signal.
9. the mode identification method of safety management system of optical fiber perimeter as claimed in claim 8 is characterized in that, the denoising step of wherein said signal comprises:
An adaptive noise filter, this adaptive noise filter is responsible for the removal of extraneous periodic noise or interference, improves the signal to noise ratio (S/N ratio) of input signal.
10. the mode identification method of safety management system of optical fiber perimeter as claimed in claim 9 is characterized in that, wherein said sef-adapting filter comprises:
One transmission system, this transmission system represent that extraneous periodic noise signal transmits caused change along the face of land;
One adaptive algorithm, this adaptive algorithm step constitute the feedback between transmission system and the signal output, and according to the feedback signal of output, the parameter of the adjusting transmission system that adaptive algorithm is real-time makes real-time being removed of periodic noise.
11. the mode identification method of safety management system of optical fiber perimeter as claimed in claim 7 is characterized in that, wherein said characteristic extraction step comprises following steps:
The signal receiving step, what this step received is through the intrusion signal after the Signal Pretreatment;
The power spectrum shift step of signal, this step calculates the power Spectral Estimation of signal, prepares for extracting power spectrum aspect signal;
The wavelet transformation step of signal, this step are to wavelet transformation, for the feature of extracting the small echo aspect is prepared;
The power spectrum resonance peak selects step, this step to search out the position of the formant frequency in the power spectrum of signal, as characteristic quantity;
Power spectrum shape Statistics amount calculation procedure, this step are calculated the shape Statistics amount of the power spectrum that obtains from the power spectrum shift step, as characteristic quantity;
The wavelet energy calculation procedure, this step calculates wavelet energy according to the wavelet coefficient that the wavelet transformation step of signal obtains, as characteristic quantity;
Normalization associating feature, resonance peak characteristic quantity, power spectrum shape Statistics characteristic quantity, the wavelet energy characteristic quantity that this step selects the power spectrum resonance peak step to obtain united one group of new characteristic quantity of formation, for unitized unit, does normalized.
12. the mode identification method of safety management system of optical fiber perimeter as claimed in claim 7, it is characterized in that, the used sorter of the design procedure of wherein said sorter is the sorter of a support vector machine, support vector machine is that the primitive characteristics amount is mapped in the higher dimensional space, asks for the step of optimal classification face in higher dimensional space.
13. the mode identification method of safety management system of optical fiber perimeter as claimed in claim 7 is characterized in that, wherein said learning procedure comprises:
According to invading affair character extracting method structural attitude sample set, set up training set and test set respectively;
By a radially basic kernel function, characteristic quantity from the low level spatial mappings to higher dimensional space, in higher dimensional space, ask for the optimal classification face;
One can be used for polytypic support vector machine, and many classification problems are changed into the pairwise classification problem earlier.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2008101036518A CN101556724B (en) | 2008-04-09 | 2008-04-09 | Safety management system of optical fiber perimeter and pattern recognition method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2008101036518A CN101556724B (en) | 2008-04-09 | 2008-04-09 | Safety management system of optical fiber perimeter and pattern recognition method thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101556724A CN101556724A (en) | 2009-10-14 |
| CN101556724B true CN101556724B (en) | 2011-06-01 |
Family
ID=41174827
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2008101036518A Expired - Fee Related CN101556724B (en) | 2008-04-09 | 2008-04-09 | Safety management system of optical fiber perimeter and pattern recognition method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101556724B (en) |
Families Citing this family (39)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101840615B (en) * | 2010-04-09 | 2012-01-11 | 武汉理工大学 | Self-adaptive warning system for intelligently addressing circumferential intrusions of fiber Bragging grating |
| CN101871809B (en) * | 2010-06-25 | 2012-09-12 | 北京奥普科达科技有限公司 | Defense area optical fiber distribution type vibration sensor and perimeter alarm system adopting same |
| CN101916492A (en) * | 2010-08-10 | 2010-12-15 | 卢军翔 | Optical fiber intelligent sensor used in security system |
| CN101968550B (en) * | 2010-09-29 | 2013-03-06 | 山东大学 | Rock stratum recognizing device and method based on array optical fiber sensor |
| CN102663032B (en) * | 2012-03-26 | 2014-03-05 | 电子科技大学 | Fiber grating fence invasion event mode recognition method |
| CN103544791B (en) * | 2012-07-10 | 2016-01-20 | 中国矿业大学(北京) | Based on the underground system for monitoring intrusion of seismic event |
| CN102902954A (en) * | 2012-08-29 | 2013-01-30 | 无锡泛太科技有限公司 | Pedestrian detection system and method based on internet-of-things intelligent household security gateway |
| CN103123739B (en) * | 2012-12-05 | 2016-01-20 | 北京航天易联科技发展有限公司 | Optical fiber perimeter intrusion alarm system signal processing unit proving installation |
| CN103236127B (en) * | 2013-05-06 | 2016-06-22 | 无锡成电光纤传感科技有限公司 | A kind of fiber fence system for monitoring intrusion and mode identification method thereof |
| CN103913645B (en) * | 2014-03-31 | 2016-08-17 | 南京航空航天大学 | Fibre optic sensor arra and antenna pattern measurement device, measuring method |
| CN104240455B (en) * | 2014-08-07 | 2016-08-17 | 北京航天控制仪器研究所 | A kind of disturbance event recognition methods in distribution type fiber-optic pipeline safety early warning system |
| CN104791612B (en) * | 2015-03-23 | 2017-12-05 | 中国石油天然气集团公司 | A kind of adaptive noise cancelling arrangement being applied in optical fiber early warning system |
| CN104916059B (en) * | 2015-05-18 | 2017-06-16 | 中国民航大学 | M Z interferometer types airport circumference early warning system mode identification method based on EEMD |
| CN105046859B (en) * | 2015-06-02 | 2017-07-07 | 北方工业大学 | Optical fiber intrusion detection method based on vibration signal space-time two-dimensional sparse representation K-S detection |
| CN105222885B (en) * | 2015-06-19 | 2018-07-24 | 北方工业大学 | Optical fiber vibration detection method and device |
| CN104966076A (en) * | 2015-07-21 | 2015-10-07 | 北方工业大学 | Optical fiber intrusion signal classification and identification method based on support vector machine |
| CN105067101A (en) * | 2015-08-05 | 2015-11-18 | 北方工业大学 | Fundamental tone frequency characteristic extraction method based on vibration signal for vibration source identification |
| CN105182435A (en) * | 2015-08-31 | 2015-12-23 | 中国科学院半导体研究所 | Distributed multipoint invasion detection system based on polarization maintaining fiber |
| CN105204084B (en) * | 2015-09-10 | 2018-02-09 | 北方工业大学 | Optical fiber intrusion signal identification method based on L DA algorithm model |
| CN105608822B (en) * | 2015-12-25 | 2021-09-14 | 中国铁道科学研究院集团有限公司电子计算技术研究所 | Railway perimeter intrusion vibration optical fiber alarm method and device |
| CN105841793A (en) * | 2016-04-15 | 2016-08-10 | 深圳艾瑞斯通技术有限公司 | Optical fiber vibration source identification method, device and system |
| CN106067244B (en) * | 2016-06-29 | 2018-08-17 | 中国科学院上海微系统与信息技术研究所 | A kind of sensor threshold value adaptive regulation method and system for perimeter detection system |
| CN106644031B (en) * | 2016-10-08 | 2019-07-26 | 天津理工大学 | A kind of high stability fibre-optical sensing device and demodulation method for eliminating light intensity disturbance |
| CN106504451B (en) * | 2016-11-14 | 2018-08-14 | 中国人民解放军国防科学技术大学 | A kind of demodulating algorithm of the optical fiber perimeter safety signal based on matrix theory |
| CN106408822A (en) * | 2016-11-25 | 2017-02-15 | 同方威视技术股份有限公司 | Vibration sensing optical cable used for perimeter surface and perimeter security system |
| CN107356962B (en) * | 2017-07-14 | 2019-03-08 | 北京知觉科技有限公司 | Micro-seismic Signals localization method and device based on fibre optical sensor |
| CN107170166A (en) * | 2017-07-14 | 2017-09-15 | 四川中科通服智能安防科技有限公司 | Defence area supervising device and intelligent security guard cloud platform |
| 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 |
| CN109374116B (en) * | 2018-12-07 | 2020-12-29 | 武汉理工光科股份有限公司 | Excavation behavior identification method of buried optical fiber sensing vibration detection system |
| CN111458059B (en) * | 2019-01-21 | 2022-01-28 | 中国科学院上海光学精密机械研究所 | Continuous vehicle collision detection system and method thereof |
| CN110631682A (en) * | 2019-09-26 | 2019-12-31 | 华北水利水电大学 | Strain safety monitoring method for cable tunnel bearing body |
| CN110631683A (en) * | 2019-09-26 | 2019-12-31 | 华北水利水电大学 | Building rigid part strain safety monitoring method |
| CN110567571A (en) * | 2019-09-26 | 2019-12-13 | 华北水利水电大学 | oil pipeline vibration detection method based on optical fiber distributed monitoring |
| CN111766636B (en) * | 2020-06-29 | 2023-05-26 | 中国人民解放军63983部队 | Military vehicle intrusion recognition device based on multi-source optical fiber sensing |
| CN114200875A (en) * | 2020-07-02 | 2022-03-18 | 广东电网有限责任公司江门供电局 | Super-radiation double-light-source external hidden danger inspection system |
| CN114059474A (en) * | 2021-10-24 | 2022-02-18 | 廖永宏 | Intelligent expressway accident sensing and intelligent induction security device and method |
| CN114268365B (en) * | 2021-12-02 | 2023-07-11 | 国网甘肃省电力公司酒泉供电公司 | Communication optical cable intelligent early warning method and system based on visualization technology |
| CN115394024B (en) * | 2022-08-10 | 2024-02-23 | 武汉烽理光电技术有限公司 | Walking monitoring and predicting method and device based on grating array |
| CN115798131B (en) * | 2023-02-13 | 2023-04-28 | 成都陆迪盛华科技有限公司 | Multi-dimensional characteristic intrusion detection method based on distributed optical fibers |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1862239A (en) * | 2006-06-15 | 2006-11-15 | 华中科技大学 | Distributed optical fiber vibration sensing method and apparatus thereof |
| CN101021961A (en) * | 2007-03-15 | 2007-08-22 | 中国科学院上海光学精密机械研究所 | Single-mode optical fibre circumferential protective sensor |
| CN101025845A (en) * | 2007-03-21 | 2007-08-29 | 中国科学院上海光学精密机械研究所 | Perimeter defence optical fiber sensor |
-
2008
- 2008-04-09 CN CN2008101036518A patent/CN101556724B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1862239A (en) * | 2006-06-15 | 2006-11-15 | 华中科技大学 | Distributed optical fiber vibration sensing method and apparatus thereof |
| CN101021961A (en) * | 2007-03-15 | 2007-08-22 | 中国科学院上海光学精密机械研究所 | Single-mode optical fibre circumferential protective sensor |
| CN101025845A (en) * | 2007-03-21 | 2007-08-29 | 中国科学院上海光学精密机械研究所 | Perimeter defence optical fiber sensor |
Non-Patent Citations (2)
| Title |
|---|
| JP特开平9-152126A 1997.06.10 |
| 蓝天，张春熹等.全光纤周界安全防范系统.光学技术.2008,34(2),259. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101556724A (en) | 2009-10-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101556724B (en) | Safety management system of optical fiber perimeter and pattern recognition method thereof | |
| CN108932480B (en) | Distributed optical fiber sensing signal feature learning and classifying method based on 1D-CNN | |
| CN105931402B (en) | Optical fiber perimeter intrusion detection method based on image identification | |
| CN104240455B (en) | A kind of disturbance event recognition methods in distribution type fiber-optic pipeline safety early warning system | |
| CN107256635B (en) | Vehicle identification method based on distributed optical fiber sensing in intelligent traffic | |
| CN102045120B (en) | Vibration signal identification method for optical fiber perimeter system | |
| CN102401667B (en) | There is optical fiber distributed perturbation method for sensing and the system of disturbance character recognition function | |
| CN103236127B (en) | A kind of fiber fence system for monitoring intrusion and mode identification method thereof | |
| CN110570613A (en) | Fence vibration intrusion positioning and mode identification method based on distributed optical fiber system | |
| CN111649817B (en) | Distributed optical fiber vibration sensor system and mode identification method thereof | |
| CN111222743B (en) | Method for judging vertical offset distance and threat level of optical fiber sensing event | |
| CN103196465B (en) | Phase sensitive optical time-domain reflectometer (phi-OTDR) sensing signal noise separation and signal extraction method | |
| CN102663032B (en) | Fiber grating fence invasion event mode recognition method | |
| CN103116957B (en) | A kind of method of optical fiber perimeter security and protection system shielding climate impact | |
| Shi et al. | Multi-event classification for Φ-OTDR distributed optical fiber sensing system using deep learning and support vector machine | |
| CN103617684A (en) | Interference type optical fiber perimeter vibration intrusion recognition algorithm | |
| CN111537056A (en) | Pipeline along-line third-party construction dynamic early warning method based on SVM and time-frequency domain characteristics | |
| CN113670430A (en) | Distributed optical fiber vibration sensing intelligent disturbance identification method | |
| CN116912586A (en) | Substation intrusion event identification method, system, equipment and medium | |
| CN104833378B (en) | Method for identifying interference signal of optical fiber perimeter system | |
| Khan et al. | Prior recognition of flash floods: Concrete optimal neural network configuration analysis for multi-resolution sensing | |
| CN111951505B (en) | Fence vibration intrusion positioning and mode identification method based on distributed optical fiber system | |
| CN101626270B (en) | Event pre-warning and classifying method by external safety pre-warning and positioning system of photoelectric composite cables | |
| CN113627470B (en) | Zero-order learning-based unknown event classification method for optical fiber early warning system | |
| CN103903270A (en) | Regularized valid characteristic section selecting method of optical fiber link monitoring signals |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20110601 Termination date: 20150409 |
|
| EXPY | Termination of patent right or utility model |