CN107918162A - A kind of intelligent identification Method of main line of communication surrounding enviroment disturbing signal - Google Patents
A kind of intelligent identification Method of main line of communication surrounding enviroment disturbing signal Download PDFInfo
- Publication number
- CN107918162A CN107918162A CN201711134881.6A CN201711134881A CN107918162A CN 107918162 A CN107918162 A CN 107918162A CN 201711134881 A CN201711134881 A CN 201711134881A CN 107918162 A CN107918162 A CN 107918162A
- Authority
- CN
- China
- Prior art keywords
- main line
- feature
- identification method
- surrounding enviroment
- intelligent identification
- 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
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V8/00—Prospecting or detecting by optical means
- G01V8/10—Detecting, e.g. by using light barriers
Landscapes
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Geophysics (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a kind of intelligent identification Method of main line of communication surrounding enviroment disturbing signal, optical cable for sensing is laid on the monitoring objective of the main line of communication, monitoring objective and Analysis interference signal source are sensed by optical cable for sensing, the discriminance analysis step gathered after interference signal source is as follows：S1 establishes vectorial pond；S2 establishes vectorial pond, and by the way of Feature Selection, final combination of eigenvectors is filtered out from described eigenvector pond.Further segmentation attenuates the frequency window that wavelet packet broadens once big with scale, can find the time frequency optimum base for being most suitable for signal to be analyzed.
Description
Technical field
The present invention relates to a kind of signal monitoring to screen identification technology, more particularly to a kind of main line of communication surrounding enviroment disturbance letter
Number intelligent identification Method.
Background technology
In order to detect the attribute of event in disturbance event (harmful or harmless) and position exactly, done in traffic
In line surrounding enviroment disturbing signal intelligent identification technology, pattern-recognition is wherein the most key technology, is permitted identification methods
Due to its own unique advantage, become research hotspot.The weight of the 1980s Back-propagation (BP) algorithm
New discovery and successful application have promoted artificial neural network to study and using upsurge.Neural net method has compared with statistical method
It is good probabilistic model, Parameter Self-learning, Generalization Capability are not depended on, so far the extensive use still in pattern-recognition.So
And the design and realization of neutral net depend on experience, Generalization Capability cannot ensure optimal.
Since the 1990s, the proposition of support vector machines (SVM) has attracted pattern-recognition bound pair Statistical Learning Theory
With the great interest of kernel method (Kernel methods).Compared with neutral net, the advantages of support vector machines is to pass through optimization
One extensive bouds on error automatically determines an optimal grader structure, so as to have more preferable Generalization Capability.
The content of the invention
The present invention overcomes the deficiencies in the prior art, there is provided a kind of intelligent recognition of main line of communication surrounding enviroment disturbing signal
Method.
To reach above-mentioned purpose, the technical solution adopted by the present invention is：A kind of main line of communication surrounding enviroment disturbing signal
Intelligent identification Method, optical cable for sensing are laid on the monitoring objective of the main line of communication, are sensed monitoring objective by optical cable for sensing and are divided
Interference signal source is analysed, the discriminance analysis step gathered after interference signal source is as follows：
S1 establishes vectorial pond：The sample rate for drafting digital signal is 2f, and the sample rate is carried out k layers of WAVELET PACKET DECOMPOSITION,
And then obtain 2kA bandwidth isWide frequency band, wherein, the wavelet packet coefficient of kth layer isWherein i decomposes for respective layer
Node ID, i=0,1,2,2k- 1, m identify for wavelet packet locus；
From Pa Saiwaer energy integral formula, wavelet packet coefficient has energy dimension, for carrying out the frequency band of signal
The ENERGY E of energy spectrometer, then i-th of frequency band of kth layerk,iFor：
Normalized, obtain required frequency band energy percentage：
K layers of WAVELET PACKET DECOMPOSITION, obtain 2kA frequency band energy percentage parameter, for composition characteristic vector；
Obtain 2kTie up the feature vector pond of wavelet packet parameter composition；
S2 establishes vectorial pond：By the way of Feature Selection, filtered out from described eigenvector pond final feature to
Amount combination.
In a preferred embodiment of the present invention, the interference signal source include can interfere with detection target vehicle pass through,
Foreign body intrusion, machinery operation and falling rocks collision.
In a preferred embodiment of the present invention, in step s 2, remove the most weak feature of separating capacity, and remove and cause
Obscure the feature of generation, obtain the feature vector of a low-dimensional.
In a preferred embodiment of the present invention, in S2 steps, more class Support Vectors are built using man-to-man mode
Machine, for each signal to carrying out the screening of feature vector.
In a preferred embodiment of the present invention, using time domain parameter, power spectrum parameters, small echo bag parameter as order arrayed feature
Value.
In a preferred embodiment of the present invention, the optical cable for sensing is installed on detection host, and the detection host passes through
Data cable attended operation terminal.
In a preferred embodiment of the present invention, the monitoring objective is junction device or wheel track.
The present invention solves defect present in background technology, and the present invention possesses following beneficial effect：
(1) further segmentation attenuates the frequency window that wavelet packet broadens once big with scale, can find and be most suitable for treating point
Analyse the time frequency optimum base of signal.
(2) combined for different signals, it is special for order arrangement using time domain parameter, power spectrum parameters, small echo bag parameter
Value indicative, can so realize intelligent recognition, and discrimination ensures the intelligence of identification also due to take in the advantage of three concurrently.
Embodiment
Presently in connection with embodiment, the present invention is described in further detail.
A kind of intelligent identification Method of main line of communication surrounding enviroment disturbing signal, disturbing signal is mainly by personnel, passing car
And falling rocks mud-rock flow etc. form, be highly suitable for the detection of this kind of disturbance object based on optical fiber inductive signal target acquisition.Separately
Outside, these vibration signals usually contain some distinctive characteristic attributes, this just provides method for further identification.From letter
Number processing from the viewpoint of, feature extraction can be carried out from time-frequency domain.
Since optical cable for sensing is laid on the specific monitoring objective of the main line of communication, as roadbed equipment, wheel track, optical cable for sensing are several
Always carve and be subject to the disturbance of various signals and can be gathered by system, therefore the importance of signal identification is self-evident.
The basic thought of time frequency analysis is the Copula of design time and frequency, and signal is described at the same time not with the function
With the energy density and intensity in time and frequency time, it is intended to disclose how many frequency component in signal, and each component
As how the time becomes.For time-frequency domain, using wavelet analysis.Wavelet packet is wavelet transformation, multiresolution analysis and son
With the popularization decomposed, there is any multi-resolution decomposition characteristic.Wavelet packet is by with scale, once frequency window that is big and broadening is further
Segmentation attenuates, and can find the time frequency optimum base for being most suitable for signal to be analyzed.
Assuming that the sample rate of gained digital signal is 2f, if k layers of WAVELET PACKET DECOMPOSITION are done, with regard to that can obtain 2kA bandwidth is
Wide frequency band, wherein, the wavelet packet coefficient of kth layer isWherein i for the layer decompose node ID, i=0,1,
2···,2k- 1, m are then wavelet packet locus marks.
Contact Parseval energy (Parseval's theorem) integral formula and understand that wavelet packet coefficient has energy dimension, can
For carrying out the Frequency Band Energy Analysis Using of signal.The then ENERGY E of i-th of frequency band of kth layerk,iFor：
Normalized, obtain required frequency band energy percentage：
In the present system, 3 layers of WAVELET PACKET DECOMPOSITION are done, obtain 8 frequency band energy percentage parameters, for composition characteristic to
Amount.
The feature vector pond of 8 dimension wavelet packet parameter compositions is obtained.On this basis, it is necessary to use Feature Selection skill
Art, from this feature vector pond, filters out final combination of eigenvectors.It is weaker very that this process can remove separating capacity
To the feature for the generation that causes confusion, the feature vector of a more low-dimensional is obtained, improves system identification power.Be employed herein support to
Amount machine (SVM) works to complete pattern-recognition, and the evaluation index of pattern-recognition effect is classification error rate (Classification
Error Rate)。
More classification SVM are built due to the use of the mode of " one-to-one ", therefore each signal can be directed to (Signal
Pair the screening of feature vector) is carried out, it is real using time domain parameter, power spectrum parameters, small echo bag parameter as order arrayed feature value
Test that the results are shown in Table 1.
1 Feature Selection result of table
| Signal combines | Screening obtains characteristic value | Eigenvalue rate |
| Vehicle passes through vs foreign body intrusions | 5,8,13,14,15,17,18,19 | 84.3% |
| Vehicle passes through vs machinery operations | 13,14 | 100% |
| Vehicle is collided by vs falling rocks | 1,3,5,11,16,17 | 29.4% |
| Foreign body intrusion vs machinery operations | 6,8,15,16,20 | 65.6% |
| Foreign body intrusion vs falling rocks collides | 1,2,6,8,10,13,14,16,20 | 40.8% |
| Machinery operation vs falling rocks collides | 3,5,7,8,10,12,13 | 6.8% |
From table 1 it follows that power spectrum parameters recognition effect is generally best, but it is directed to different signal groups
Close, the discrimination of three kinds of characteristic parameters is each advantageous, such as small echo bag parameter is passed through for vehicle and two class signal of machinery operation
Differentiation rate, and time domain parameter to vehicle pass through and falling rocks collision alarm differentiation rate.
The desirable embodiment according to the present invention is enlightenment above, and by above-mentioned description, related personnel completely can be with
Without departing from the scope of the technological thought of the present invention', various changes and amendments are carried out.The technical scope of this invention
The content being not limited on specification, it is necessary to determine the technical scope according to the scope of the claims.
Claims (7)
1. a kind of intelligent identification Method of main line of communication surrounding enviroment disturbing signal, it is characterised in that optical cable for sensing is laid in friendship
On the monitoring objective of logical main line, monitoring objective and Analysis interference signal source are sensed by optical cable for sensing, after gathering interference signal source
Discriminance analysis step it is as follows：
S1 establishes vectorial pond：The sample rate for drafting digital signal is 2f, and the sample rate is carried out k layers of WAVELET PACKET DECOMPOSITION, and then
Obtain 2kA bandwidth isWide frequency band, wherein, the wavelet packet coefficient of kth layer isWherein i decomposes node for respective layer
Sequence number, i=0,1,2,2k- 1, m identify for wavelet packet locus；
From Pa Saiwaer energy integral formula, wavelet packet coefficient has energy dimension, for carrying out the frequency band energy of signal
Analyze, then the ENERGY E of i-th of frequency band of kth layerk,iFor：
Normalized, obtain required frequency band energy percentage：
K layers of WAVELET PACKET DECOMPOSITION, obtain 2kA frequency band energy percentage parameter, for composition characteristic vector；
Obtain 2kTie up the feature vector pond of wavelet packet parameter composition；
S2 establishes vectorial pond：By the way of Feature Selection, final feature vector group is filtered out from described eigenvector pond
Close.
2. a kind of intelligent identification Method of main line of communication surrounding enviroment disturbing signal according to claim 1, its feature exist
In：The interference signal source includes can interfere with the vehicle process of detection target, foreign body intrusion, machinery operation and falling rocks collision.
3. a kind of intelligent identification Method of main line of communication surrounding enviroment disturbing signal according to claim 1, its feature exist
In：In step s 2, remove the most weak feature of separating capacity, and remove the feature for the generation that causes confusion, obtain a low-dimensional
Feature vector.
4. a kind of intelligent identification Method of main line of communication surrounding enviroment disturbing signal according to claim 1, its feature exist
In：In S2 steps, multi-category support vector machines are built using man-to-man mode, for each signal to carrying out feature
The screening of vector.
5. a kind of intelligent identification Method of main line of communication surrounding enviroment disturbing signal according to claim 4, its feature exist
In：Using time domain parameter, power spectrum parameters, small echo bag parameter as order arrayed feature value.
6. a kind of intelligent identification Method of main line of communication surrounding enviroment disturbing signal according to claim 1, its feature exist
In：The optical cable for sensing is installed on detection host, and the detection host passes through data cable attended operation terminal.
7. a kind of intelligent identification Method of main line of communication surrounding enviroment disturbing signal according to claim 1, its feature exist
In：The monitoring objective is junction device or wheel track.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711134881.6A CN107918162A (en) | 2017-11-16 | 2017-11-16 | A kind of intelligent identification Method of main line of communication surrounding enviroment disturbing signal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711134881.6A CN107918162A (en) | 2017-11-16 | 2017-11-16 | A kind of intelligent identification Method of main line of communication surrounding enviroment disturbing signal |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107918162A true CN107918162A (en) | 2018-04-17 |
Family
ID=61896334
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711134881.6A Pending CN107918162A (en) | 2017-11-16 | 2017-11-16 | A kind of intelligent identification Method of main line of communication surrounding enviroment disturbing signal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107918162A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6728645B1 (en) * | 2003-01-07 | 2004-04-27 | Electro-Optics Research & Development Ltd. | Method and system for automatic identification of objects type according to their characteristic spectrum of vibration frequencies |
| CN103136587A (en) * | 2013-03-07 | 2013-06-05 | 武汉大学 | Power distribution network operating state classification recognition method based on support vector machine |
| CN103345641A (en) * | 2013-07-16 | 2013-10-09 | 杭州电子科技大学 | Hand electromyographic signal motion recognition method based on wavelet entropy and support vector machine |
| US20150051083A1 (en) * | 2012-02-15 | 2015-02-19 | Battelle Memorial Institute | Methods and compositions for identifying repeating sequences in nucleic acids |
-
2017
- 2017-11-16 CN CN201711134881.6A patent/CN107918162A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6728645B1 (en) * | 2003-01-07 | 2004-04-27 | Electro-Optics Research & Development Ltd. | Method and system for automatic identification of objects type according to their characteristic spectrum of vibration frequencies |
| US20150051083A1 (en) * | 2012-02-15 | 2015-02-19 | Battelle Memorial Institute | Methods and compositions for identifying repeating sequences in nucleic acids |
| CN103136587A (en) * | 2013-03-07 | 2013-06-05 | 武汉大学 | Power distribution network operating state classification recognition method based on support vector machine |
| CN103345641A (en) * | 2013-07-16 | 2013-10-09 | 杭州电子科技大学 | Hand electromyographic signal motion recognition method based on wavelet entropy and support vector machine |
Non-Patent Citations (1)
| Title |
|---|
| 张嵩: "光纤周界安防系统的振动信号识别研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109768985B (en) | Intrusion detection method based on flow visualization and machine learning algorithm | |
| CN102779281B (en) | Vehicle type identification method based on support vector machine and used for earth inductor | |
| CN106443310B (en) | A kind of transformer fault detection method based on SOM neural network | |
| CN106022229B (en) | The abnormal behaviour recognition methods with the Back propagation neural networks of self-adaptive enhancement algorithm is extracted based on video motion information characteristics | |
| CN105184319B (en) | Recognition methods, device and the circumference intrusion alarm system of optical fiber perimeter invasion signal | |
| CN104021570B (en) | Directive in a kind of video cross the border and mix line detecting method | |
| CN103832433B (en) | Deviation and front truck anti-collision alarm system and its implementation | |
| CN110570613A (en) | Fence vibration intrusion positioning and mode identification method based on distributed optical fiber system | |
| CN106371144B (en) | A kind of device of unmanned plane discovery and early warning | |
| CN110084094B (en) | Unmanned aerial vehicle target identification and classification method based on deep learning | |
| CN104483326A (en) | High-voltage wire insulator defect detection method and high-voltage wire insulator defect detection system based on deep belief network | |
| CN104240455A (en) | Method for identifying disturbance event in distributed type optical fiber pipeline security early-warning system | |
| CN107622678A (en) | A kind of intelligent traffic control system and its method based on image procossing | |
| CN103323749A (en) | Multi-classifier information fusion partial discharge diagnostic method | |
| CN109405960B (en) | Aviation noise identification method | |
| CN108519149B (en) | Tunnel accident monitoring and alarming system and method based on sound time-frequency domain analysis | |
| CN108445868A (en) | A kind of automobile intelligent fault diagnosis system and method based on modern signal processing technology | |
| CN108280409A (en) | Large-space video smoke detection method based on multi-feature fusion | |
| CN108506736B (en) | A kind of invasion signal diagnostic system and method | |
| CN108664875A (en) | Underground belt-conveying monitoring method based on image recognition | |
| CN109120592A (en) | A kind of Web abnormality detection system based on user behavior | |
| CN106960153A (en) | The kind identification method and device of virus | |
| CN106292282A (en) | Reading intelligent agriculture environmental monitoring systems based on big data | |
| CN106644035B (en) | Vibration source identification method and system based on time-frequency transformation characteristics | |
| CN107918162A (en) | A kind of intelligent identification Method of main line of communication surrounding enviroment disturbing signal |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180417 |