CN111222461A - Method for identifying invasion signal of optical fiber vibration detection system - Google Patents
Method for identifying invasion signal of optical fiber vibration detection system Download PDFInfo
- Publication number
- CN111222461A CN111222461A CN202010011195.5A CN202010011195A CN111222461A CN 111222461 A CN111222461 A CN 111222461A CN 202010011195 A CN202010011195 A CN 202010011195A CN 111222461 A CN111222461 A CN 111222461A
- Authority
- CN
- China
- Prior art keywords
- signal
- event
- vibration
- real
- image
- 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
- 238000000034 method Methods 0.000 title claims abstract description 37
- 239000013307 optical fiber Substances 0.000 title claims abstract description 29
- 238000001514 detection method Methods 0.000 title claims description 15
- 230000009545 invasion Effects 0.000 title abstract description 7
- 238000013145 classification model Methods 0.000 claims abstract description 29
- 238000012544 monitoring process Methods 0.000 claims abstract description 24
- 238000012706 support-vector machine Methods 0.000 claims abstract description 14
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 12
- 238000012549 training Methods 0.000 claims description 18
- 238000009412 basement excavation Methods 0.000 claims description 14
- 238000004364 calculation method Methods 0.000 claims description 13
- 238000010586 diagram Methods 0.000 claims description 10
- 239000000835 fiber Substances 0.000 claims description 10
- 230000006870 function Effects 0.000 claims description 8
- 238000005070 sampling Methods 0.000 claims description 7
- 238000001914 filtration Methods 0.000 claims description 6
- 238000002372 labelling Methods 0.000 claims description 5
- 230000001066 destructive effect Effects 0.000 claims description 3
- 238000010801 machine learning Methods 0.000 description 5
- 230000003287 optical effect Effects 0.000 description 5
- 238000000354 decomposition reaction Methods 0.000 description 4
- 238000005259 measurement Methods 0.000 description 3
- 238000005065 mining Methods 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000008054 signal transmission Effects 0.000 description 2
- 238000001228 spectrum Methods 0.000 description 2
- 241000282414 Homo sapiens Species 0.000 description 1
- 108010076504 Protein Sorting Signals Proteins 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 239000013598 vector Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01H—MEASUREMENT OF MECHANICAL VIBRATIONS OR ULTRASONIC, SONIC OR INFRASONIC WAVES
- G01H17/00—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves, not provided for in the preceding groups
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B13/00—Burglar, theft or intruder alarms
- G08B13/18—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength
- G08B13/181—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength using active radiation detection systems
- G08B13/183—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength using active radiation detection systems by interruption of a radiation beam or barrier
- G08B13/186—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength using active radiation detection systems by interruption of a radiation beam or barrier using light guides, e.g. optical fibres
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
- G06F2218/04—Denoising
- G06F2218/06—Denoising by applying a scale-space analysis, e.g. using wavelet analysis
Abstract
The invention discloses a method for identifying a long-distance intelligent monitoring optical fiber vibration invasion signal. The identification method of the invention is to obtain real-time signals from the monitoring signals of the monitored optical fiber perimeter system, process the real-time signals and judge the category of the real-time signals. According to labels marked with different vibration signal event types in advance, a zero-crossing rate and energy characteristics of the signals are extracted to train and recognize classification models of the different vibration event types, initial event types of the signals are monitored in real time through the models, and finally the results are compared with recognition results matched with a graph to determine the final intrusion signal event types. The invention not only utilizes the image characteristic of the vibration signal, but also utilizes the zero-crossing and energy characteristic of the vibration signal, and the false recognition rate is greatly reduced by the method of combining the support vector machine and the image matching algorithm.
Description
Technical Field
The invention relates to a long-distance intelligent monitoring method, in particular to a method for identifying an optical fiber vibration invasion signal.
Background
Optical fiber is a common communication conducting medium, and in the era of rapid development of 5G technology, optical fiber is becoming more and more important in internet of things technology and social security. The impact of the fiber operating environment may directly affect the implementation of the fiber function. The use of optical fibers for intrusion signal alarms is therefore essential for the detection of fiber vibration, particularly in the event of a destructive event. However, the similarity and complexity of various intrusion signals make detection and classification alarm of vibration signals quite difficult, and how to accurately identify various intrusion signals and reduce the false alarm rate is a technical problem to be solved first in intrusion signal detection of an optical fiber vibration detection system.
Chinese patent 2013106132601 discloses a device and method for expanding the monitoring area of an optical cable vibration intrusion monitoring system. The device comprises a passive infrared detector, a vibration signal transmission medium and a vibration signal decoder, wherein the passive infrared detector and the vibration signal decoder are connected through the vibration signal transmission medium. And (4) utilizing a passive infrared detector to sense intrusion behavior. This patent is assisted optic fibre vibration intrusion monitoring system and is carried out the invasion and take precautions against, has expanded optical cable vibration intrusion monitoring system's monitoring area, reduces the invasion monitoring blind spot, reduces the requirement of sensing optical cable to the rail, but can't distinguish and cause the optical cable vibration classification, and its misstatement rate is great consequently. Similar patents such as ZL 2016486278, ZL2016104767007, etc. suffer from the same disadvantages.
To solve the defects of the prior art, the chinese invention patent 2010105235522 discloses a vibration signal identification method of an optical fiber perimeter system, which comprises the following steps: signal acquisition, windowing, band-pass filtering, wavelet denoising, vibration event detection, characteristic parameter extraction, mode matching and classification, and the beneficial effects are that: more characteristic parameters such as short-time energy E, short-time average amplitude M, short-time average zero-crossing rate Z and detail signal energy E of each scale of wavelet decomposition are introducedwAnd a vibration signal power spectrum P, so that the category of the external vibration signal can be more accurately judged. A similar patent is also 2015105567662. However, due to the factors causing the fiber vibration, the complexity of the signal and the signal characteristics makes the characteristic parameters introduced at the early stage rather difficult, and it is difficult to judge the exact class of the intrusion signalOtherwise, the prior art makes the optical fiber damage prevention work difficult to be put to practical use.
Disclosure of Invention
The invention provides a method for identifying an intrusion signal of an optical fiber vibration detection system, which can solve the problem of the prior art.
The invention discloses a method for identifying an intrusion signal of an optical fiber vibration detection system, which comprises the following steps: the method comprises the steps of obtaining a real-time signal from a monitoring signal of a monitored optical fiber perimeter monitoring system, processing the real-time signal, judging the category of the real-time signal, obtaining labels of characteristic samples according to vibration signals of different vibration events obtained in advance, establishing a classification model for identifying the different vibration events through training and learning of computer software, obtaining an initial event type of the real-time monitoring signal according to the classification model, and determining a final intrusion signal event type through a graph matching result by combining a vibration signal waterfall graph obtained by the real-time signal.
Preferably, the method for identifying the intrusion signal of the fiber vibration detection system according to the present invention comprises the steps of firstly inputting previously obtained different vibration events (such as damage events of excavator excavation events and manual excavation events, and normal events of vehicle walking events and noise events) to respectively perform event type labeling on vibration signal feature samples, extracting zero crossing rate and energy features of the vibration signals for the vibration signals with fixed fiber lengths and sampling frequencies, generating the feature samples, labeling the previously obtained training data feature samples according to the event types of the vibration signals for the damage events and the normal events, training a classification model by using a supervised learning algorithm SVM (support vector machine) to classify and identify the damage events or the normal events of the real-time signals, and respectively labeling the excavator excavation events of the upper-stage training data on the basis of the model identification real-time signals for classifying the two types of events, The method comprises the following steps of artificially excavating labels of four different vibration signal characteristic samples of an event, a vehicle walking event and a noise event, repeatedly training a model, classifying and identifying the type of an intrusion event, and performing similarity matching with an event image library established in advance according to a corresponding vibration signal waterfall graph obtained by combining a signal to be detected on the result of classification model identification, wherein the matching process is as follows: and binarizing all images, carrying out median filtering to obtain the fingerprint of each image, calculating the Hamming distance between the image fingerprint in an image library and the waterfall fingerprint of the real-time monitoring signal obtained from the monitoring signal of the monitored optical fiber perimeter monitoring system, determining the similarity between the two images according to the calculated Hamming distance value, obtaining the similarity between the image of the real-time monitoring signal and the image library according to the result, and finally determining the final intrusion signal event type through a judgment rule.
Preferably, when the supervised learning algorithm svm classification model is used for training, the kernel function used by svm selects a Radial Basis Function (RBF), the kernel parameter is 100, the penalty factor is 10, and the optimal parameters of the classification algorithm svm are used, so that the effect is better.
Preferably, the method for identifying the intrusion signal of the optical fiber vibration detection system according to the present invention obtains the corresponding waterfall graph image according to the optical fiber vibration signals with different fixed lengths of the periods, and then performs similarity calculation on the image, wherein the calculation process is as follows: carrying out binarization on a waterfall graph of a real-time signal and a waterfall graph in an event library, wherein a foreground pixel value is 0, a background pixel value is 255, then carrying out median filtering, and then uniformly scaling pictures to 9 x 8 to obtain pictures with 72 pixels; converting the zoomed picture into a 256-level gray scale image; subtracting adjacent pixels, for example, subtracting the left pixel value from the right pixel value, and generating 8 different differences among 9 pixels in each row, so as to obtain 64 difference values; and when the difference result is positive number or zero, recording as 1, otherwise, recording as 0, thus obtaining the fingerprints of the images, calculating to obtain the Hamming distance according to the fingerprints of the two images, and finally judging that the two images are similar when the Hamming distance is less than 5. By adopting the image similarity calculation method, the accuracy is higher in the signal waterfall chart similarity calculation.
Preferably, in the method for identifying an intrusion signal of an optical fiber vibration detection system according to the present invention, based on the calculation result that the decision rule is that the waterfall graph of the real-time signal and the similarity of the image library corresponding to the classification model result of the signal are based on, if the number of the similar images is more than two thirds of the image library, the intrusion signal and the event library are considered to have high similarity, the classification result has high accuracy, at this time, the final intrusion event type is the event type predicted by the classification model, when the similarity result is less than two thirds of the image library, the similarity with other libraries is checked, if the image library type with the largest number of similar images is not consistent with the event type predicted by the classification model, the number of the similar images calculated by the event library corresponding to the prediction event of the classification model is subtracted, when the value is greater than 1/4 image libraries, and the final intrusion signal type at the moment is the event type corresponding to the image library with the most similar pictures, otherwise, the final intrusion signal type is the event type predicted by the classification model. The judgment rule can well combine the machine learning algorithm svm with the pattern matching method, so that the optical fiber intrusion signal identification method with higher accuracy is realized, and the problem of higher false alarm rate is effectively solved.
From the middle of the 60 s to the middle of the 70 s in the 20 th century, machine learning methods are gradually developed and mature and applied to various fields, the machine tries to solve the problem of traditional mode identification by simulating the learning process of human beings, and at the present stage, the machine learning methods are continuously expanded, the application range is continuously expanded, and the problems in various fields are already solved. The method of the invention fully utilizes machine learning to solve the identification problem of the intrusion signal of the optical fiber vibration detection system.
The technical scheme of the method is as follows: firstly, obtaining a vibration signal with fixed optical fiber length and sampling frequency, simultaneously extracting the zero crossing rate and energy characteristics of the vibration signal, then carrying out event type marking on a vibration signal characteristic sample, and training a classification model by utilizing an svm algorithm. And on the basis of the signals identified by the classification model, similarity calculation is carried out on the vibration waterfall graph and a vibration waterfall graph obtained by corresponding actual measurement through an event image library established in advance, and finally the final event type is judged according to two comparative similarity results.
The difference of the invention from the prior art is that the extracted characteristics are different, the invention not only utilizes the energy characteristics of the vibration signal, but also utilizes the image characteristics of the vibration signal, the characteristics of the intrusion signal can be fully utilized by adopting the technical means, and simultaneously, the method of combining the support vector machine algorithm and the image matching algorithm is utilized, so the false recognition rate is greatly reduced.
Detailed Description
The invention is illustrated below with reference to one embodiment.
The first step is as follows: firstly, acquiring optical fiber vibration signals (x (i)) i ═ 1, 2., n with fixed optical fiber length and sampling frequency, and manually marking event types of various vibration signals, wherein the marking types comprise excavator excavation, manual excavation, vehicle walking and noise.
The second step is that: the signal that the invasion signal causes the fiber vibration has the following main characteristics: the short duration and the sudden nature of the vibration, these characteristics can be reacted to the energy characteristics by the zero crossing rate of the optical signal. The zero-crossing rate calculation formula is as follows:
where x (i) is the signal sequence and N is the window function length. In order to obtain energy characteristics, wavelet decomposition is carried out on the vibration signal, the wavelet decomposition adopts Mallat algorithm, Daubechies is used as a wavelet base, Db5 wavelet is selected, and 7-level decomposition is carried out on the vibration signal to obtain 1 low-frequency energy a7And 7 high frequency energies b1、b2、b3、b4、b5、b6、b7The energy eigenvector of the constructed vibration signal is as follows:
ξ＝[a7,b1,b2,b3,b4,b5,b6,b7]
then, the characteristic vectors are normalized to obtain a zero-crossing rate signal diagram and an energy signal diagram respectively, wherein the abscissa in the diagrams is a sampling physical position point, and the ordinate is the sampling times. Experiments show that when the outside does not disturb the signal, the zero crossing rate and the energy characteristics of the vibration signal are not changed; when people or machines dig, the zero-crossing rate and energy characteristic change of the vibration signal are large, and the instantaneous signal intensity is large; when the vehicle runs, the zero crossing rate and the energy characteristics of the vibration signal are relatively stable relative to the signal change during the excavation event; when there is a weak vibration signal such as noise, the energy characteristic changes little, but the zero-crossing rate changes greatly. The different change states of the characteristics shown by the vibration signals caused by various events can be used as classification bases of machine learning training models.
The third step: because a large amount of manpower and material resources are consumed for artificially marking event type labels of various vibration signals, and effective training data is difficult to obtain, a supervised learning algorithm SVM is finally selected as a classification model, and the model has a good classification effect on small sample data. However, the SVM is a two-classification model, and in order to achieve multi-classification of various events, two-stage SVM mode classification is achieved in the whole classification process. The specific process is as follows: firstly, a zero-crossing rate signal diagram and an energy signal diagram are obtained by using the obtained vibration signals, data blocks with the size of 3 multiplied by 3 are taken as characteristic samples from the left, the right and the left, the characteristic sample labels of the damage events including the mining events are 1, the characteristic sample labels of the normal behaviors including the vehicle walking and noise events are 0, and the SVM algorithm is used for training a model to identify the damage events (the excavator mining events and the manual mining events) and the normal events (the vehicle walking events and the noise events). And in the second stage, on the basis of identifying the damage event data, identifying the excavator excavation event and the manual excavation event by using an SVM (support vector machine) model for training, wherein the kernel function used by the model is a Radial Basis Function (RBF), the kernel parameter is 100, the penalty factor is 10, and the characteristic sample label of the excavator excavation event is 1 and the characteristic sample label of the manual excavation is 0. On the data of the normal event, the data label of the vehicle walking is 1, the noise data label is 0, the SVM classification model is used for training, and finally the type of the specific event is preliminarily obtained after the intrusion signal is input through model training.
The fourth step: on the basis that the type of the specific event is obtained preliminarily in the previous step, namely on the classification result of the two-stage SVM classification model obtained through training, similarity matching comparison of the actually measured vibration waterfall diagram is carried out, so that the false alarm rate is further reduced, and a more accurate alarm signal is obtained. The method comprises the steps of digging by an excavator, manually digging, driving by a vehicle and corresponding images with different forms on a spectrogram of a noise event, establishing an image library for a certain number of various event images, wherein the images in each class of library comprise the similar forms of possible event waterfall graphs, and the forms of each class of event graphs have the characteristics. And (3) obtaining a frequency spectrum waterfall graph corresponding to the optical fiber vibration signals with different periods and fixed lengths from the vibration signals obtained through actual measurement through Labview software, wherein the horizontal axis of the waterfall graph represents the length of the optical fiber, the vertical axis represents time, and the similarity calculation is carried out by utilizing the waterfall graph of the actual measurement signals and an image library.
The fifth step: in order to calculate the similarity degree between the waterfall graph of the intrusion signal and the image library, all images are binarized, the foreground pixel value is 0, the background pixel value is 255, and then the images are subjected to median filtering. Zooming the test picture and the waterfall picture in the image library, and uniformly zooming the pictures to 9 x 8 in order to retain the structure and remove details at the same time to obtain pictures with 72 pixels; then, converting the zoomed picture into a 256-level gray scale image; subtracting adjacent pixels (subtracting the left pixel value from the right pixel value), and generating 8 different differences among 9 pixels in each row, so as to obtain 64 difference values; if the difference result is a positive number or zero, recording as 1, otherwise, recording as 0, and obtaining the fingerprint of the image; and finally, calculating the Hamming distance between the fingerprints of the two pictures, wherein the larger the Hamming distance is, the more inconsistent the pictures are, otherwise, the smaller the Hamming distance is, the more similar the pictures are, when the distance is 0, the description is completely the same, when the distance is more than 10, the two pictures are completely different, and finally, when the Hamming distance is less than 5, the two pictures are considered to be similar.
And a sixth step: firstly, obtaining an actually measured signal initial event type according to an svm model trained in the third step, then carrying out similarity calculation on an actually measured signal waterfall diagram and a corresponding event image library according to a fifth step similarity calculation method, if the number of the image libraries is more than two thirds, considering that an intrusion signal has high similarity with the event library, and the classification result has higher accuracy, wherein the intrusion event is the type of model prediction; and when the result is less than two-thirds of the number of the image libraries, the similarity with other image libraries is checked next, if the image library type with the largest number of similar pictures is inconsistent with the event type predicted by the classification model, the event library is subtracted from the similar picture number calculated by the event library corresponding to the svm model prediction event, when the value is greater than 1/4 image library numbers, the final intrusion signal type at the moment is the event type corresponding to the image library with the largest number of similar pictures, and otherwise, the final intrusion signal type is the event type predicted by the svm model. And when the finally identified event types are excavator excavation and manual excavation, carrying out intrusion signal alarm.
The invention adopts the technical scheme of comparing and detecting the vibration energy and the zero-crossing information with the image information of the vibration waterfall diagram, thereby greatly reducing the false alarm rate of the vibration type.
Claims (5)
1. The utility model provides a method for discerning optic fibre vibration detection system intrusion signal, obtains real-time signal from the monitoring signal of monitoring optic fibre perimeter monitoring system, handles the back to this real-time signal, judges this real-time signal classification, its characterized in that: the classification models of different vibration events are identified through training and learning of computer software according to vibration signals of different vibration events and labels thereof obtained in advance, the initial event type of the real-time monitoring signals is obtained according to the classification models, and meanwhile, the final intrusion signal event type is determined by combining the graph matching result of the vibration waterfall graph characteristics converted by the real-time signals.
2. The method according to claim 1, wherein the method comprises inputting vibration signals of fixed duration and sampling frequency of different vibration events obtained in advance, extracting zero-crossing rate and energy characteristics of the vibration signals, generating corresponding zero-crossing rate signal diagram and energy signal diagram, labeling the vibration signal characteristic samples according to the destructive event and normal event, respectively, training classification models by using a supervised learning algorithm Support Vector Machine (SVM) to identify the destructive event or normal event of the real-time signals, labeling labels of four different vibration signal characteristic samples of excavator excavation event, artificial excavation event, vehicle traveling event and noise event obtained in advance on the basis of classification data of two types of event models, repeating the training models, and classifying and identifying the type of the intrusion event, and according to the result of the classification model identification, a corresponding vibration signal waterfall graph obtained by combining real-time signals is subjected to similarity matching with an image library established in advance, and the matching process is as follows: and binarizing all images, carrying out median filtering to obtain the fingerprint of each image, calculating the Hamming distance between the image fingerprint in an image library and the waterfall fingerprint of the real-time monitoring signal obtained from the monitoring signal of the monitored optical fiber perimeter monitoring system, determining the similarity between the two images according to the calculated Hamming distance value, obtaining the similarity between the image of the real-time monitoring signal and the image library according to the result, and finally determining the final intrusion signal event type through a judgment rule.
3. The method according to claim 2, wherein when the supervised learning algorithm svm classification model is used for training, the kernel function used by svm selects a Radial Basis Function (RBF), the kernel parameter is 100, and the penalty factor is 10.
4. The method for identifying the intrusion signal of the optical fiber vibration detection system according to claim 1,2 or 3, wherein: according to the optical fiber vibration signals with fixed optical fiber length and sampling frequency, obtaining corresponding waterfall image images, and then carrying out similarity calculation on the images, wherein the calculation process is as follows: carrying out binarization on a waterfall graph of a real-time signal and a waterfall graph in an event library, wherein a foreground pixel value is 0, a background pixel value is 255, then carrying out median filtering, and then uniformly scaling pictures to 9 x 8 to obtain pictures with 72 pixels; converting the zoomed picture into a 256-level gray scale image; subtracting adjacent pixels, and generating 8 different differences among 9 pixels in each row to obtain 64 difference values; and when the difference result is positive number or zero, recording as 1, otherwise, recording as 0, thus obtaining the fingerprints of the images, calculating to obtain the Hamming distance according to the fingerprints of the two images, and finally judging that the two images are similar when the Hamming distance is less than 5.
5. The method of claim, wherein the method further comprises the steps of: on the basis of the judgment rule that the similarity calculation result of an image library corresponding to a real-time signal waterfall graph and a signal classification model result is obtained, if the number of similar images is more than two thirds of the number of the image libraries, the intrusion signal is considered to have high similarity with the event library, the classification result has high accuracy, the final intrusion event type is the event type predicted by the classification model, when the similarity result is less than two thirds of the number of the image libraries, the similarity with other libraries is checked next, if the image library type with the largest number of similar images is not consistent with the event type predicted by the classification model, the number of similar images calculated by the event library corresponding to the classification model prediction event is subtracted from the number of similar images calculated by the event library corresponding to the classification model prediction event to obtain an absolute value, and when the value is more than 1/4, the final signal type is the event type corresponding to the image library with the largest number of similar images, otherwise, the event type predicted by the classification model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010011195.5A CN111222461A (en) | 2020-01-06 | 2020-01-06 | Method for identifying invasion signal of optical fiber vibration detection system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010011195.5A CN111222461A (en) | 2020-01-06 | 2020-01-06 | Method for identifying invasion signal of optical fiber vibration detection system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111222461A true CN111222461A (en) | 2020-06-02 |
Family
ID=70828143
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010011195.5A Pending CN111222461A (en) | 2020-01-06 | 2020-01-06 | Method for identifying invasion signal of optical fiber vibration detection system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111222461A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112836591A (en) * | 2021-01-14 | 2021-05-25 | 清华大学深圳国际研究生院 | Method for extracting optical fiber early warning signal characteristics of oil and gas long-distance pipeline |
| CN112836574A (en) * | 2020-12-30 | 2021-05-25 | 江苏光为传感设备有限公司 | Optical fiber vibration event identification method based on spectrum similarity |
| CN112883802A (en) * | 2021-01-20 | 2021-06-01 | 刘信 | Method for identifying destructive event of pipeline optical fiber vibration safety early warning system |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0104228A1 (en) * | 1982-03-31 | 1984-04-04 | Gould Inc. | Fiber optic acoustic transducer intrusion detection system |
| US20020094531A1 (en) * | 1999-06-14 | 2002-07-18 | Frederic Zenhausern | Apparatus and method for monitoring molecular species within a medium |
| CN102045120A (en) * | 2010-10-29 | 2011-05-04 | 成都九洲电子信息系统有限责任公司 | Vibration signal identification method for optical fiber perimeter system |
| CN102663032A (en) * | 2012-03-26 | 2012-09-12 | 电子科技大学 | Fiber grating fence invasion event mode recognition method |
| CN102830719A (en) * | 2012-09-23 | 2012-12-19 | 兰州奥普信息技术有限公司 | Optical fiber type anti-oil spillage control system and control method |
| 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 |
| CN105931402A (en) * | 2016-06-27 | 2016-09-07 | 上海波汇科技股份有限公司 | Optical fiber perimeter intrusion monitoring method based on image recognition |
| CN108709633A (en) * | 2018-08-29 | 2018-10-26 | 中国科学院上海光学精密机械研究所 | Distributed optical fiber vibration sensing intelligent and safe monitoring method based on deep learning |
| US20190074896A1 (en) * | 2017-06-12 | 2019-03-07 | Network Integrity Systems, Inc. | Monitoring a Fiber Optic Cable for Intrusion Using a Weighted Algorithm |
| CN110570613A (en) * | 2019-09-18 | 2019-12-13 | 广州亓行智能科技有限公司 | Fence vibration intrusion positioning and mode identification method based on distributed optical fiber system |
-
2020
- 2020-01-06 CN CN202010011195.5A patent/CN111222461A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0104228A1 (en) * | 1982-03-31 | 1984-04-04 | Gould Inc. | Fiber optic acoustic transducer intrusion detection system |
| US20020094531A1 (en) * | 1999-06-14 | 2002-07-18 | Frederic Zenhausern | Apparatus and method for monitoring molecular species within a medium |
| CN102045120A (en) * | 2010-10-29 | 2011-05-04 | 成都九洲电子信息系统有限责任公司 | Vibration signal identification method for optical fiber perimeter system |
| CN102663032A (en) * | 2012-03-26 | 2012-09-12 | 电子科技大学 | Fiber grating fence invasion event mode recognition method |
| CN102830719A (en) * | 2012-09-23 | 2012-12-19 | 兰州奥普信息技术有限公司 | Optical fiber type anti-oil spillage control system and control method |
| 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 |
| CN105931402A (en) * | 2016-06-27 | 2016-09-07 | 上海波汇科技股份有限公司 | Optical fiber perimeter intrusion monitoring method based on image recognition |
| US20190074896A1 (en) * | 2017-06-12 | 2019-03-07 | Network Integrity Systems, Inc. | Monitoring a Fiber Optic Cable for Intrusion Using a Weighted Algorithm |
| CN108709633A (en) * | 2018-08-29 | 2018-10-26 | 中国科学院上海光学精密机械研究所 | Distributed optical fiber vibration sensing intelligent and safe monitoring method based on deep learning |
| CN110570613A (en) * | 2019-09-18 | 2019-12-13 | 广州亓行智能科技有限公司 | Fence vibration intrusion positioning and mode identification method based on distributed optical fiber system |
Non-Patent Citations (1)
| Title |
|---|
| 邹柏贤等: "基于ELM算法的光纤振动信号识别研究", 《计算机工程与应用》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112836574A (en) * | 2020-12-30 | 2021-05-25 | 江苏光为传感设备有限公司 | Optical fiber vibration event identification method based on spectrum similarity |
| CN112836574B (en) * | 2020-12-30 | 2023-08-29 | 江苏光为传感设备有限公司 | Fiber vibration event identification method based on spectrum similarity |
| CN112836591A (en) * | 2021-01-14 | 2021-05-25 | 清华大学深圳国际研究生院 | Method for extracting optical fiber early warning signal characteristics of oil and gas long-distance pipeline |
| CN112836591B (en) * | 2021-01-14 | 2024-02-27 | 清华大学深圳国际研究生院 | Method for extracting optical fiber early warning signal characteristics of oil gas long-distance pipeline |
| CN112883802A (en) * | 2021-01-20 | 2021-06-01 | 刘信 | Method for identifying destructive event of pipeline optical fiber vibration safety early warning system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhou et al. | Safety helmet detection based on YOLOv5 | |
| Tejedor et al. | Real field deployment of a smart fiber-optic surveillance system for pipeline integrity threat detection: Architectural issues and blind field test results | |
| CN105931402B (en) | Optical fiber perimeter intrusion detection method based on image identification | |
| CN105809679B (en) | Mountain railway side slope rockfall detection method based on visual analysis | |
| CN111222461A (en) | Method for identifying invasion signal of optical fiber vibration detection system | |
| CN110570613A (en) | Fence vibration intrusion positioning and mode identification method based on distributed optical fiber system | |
| CN104766099B (en) | Distribution type fiber-optic circumference vibration signal processing and recognition methods based on image | |
| CN111649817B (en) | Distributed optical fiber vibration sensor system and mode identification method thereof | |
| CN104063719A (en) | Method and device for pedestrian detection based on depth convolutional network | |
| CN102163290A (en) | Method for modeling abnormal events in multi-visual angle video monitoring based on temporal-spatial correlation information | |
| CN110519637A (en) | The method for monitoring abnormality combined based on audio frequency and video monitoring | |
| CN102928435A (en) | Aircraft skin damage identification method and device based on image and ultrasound information fusion | |
| Zhou et al. | Deep learning-based crack segmentation for civil infrastructure: Data types, architectures, and benchmarked performance | |
| CN112883802A (en) | Method for identifying destructive event of pipeline optical fiber vibration safety early warning system | |
| CN115620212B (en) | Behavior identification method and system based on monitoring video | |
| CN111505632A (en) | Ultra-wideband radar action attitude identification method based on power spectrum and Doppler characteristics | |
| Elharrouss et al. | Video summarization based on motion detection for surveillance systems | |
| CN115762023A (en) | Deep learning-based positioning type intrusion behavior recognition method | |
| US20230152150A1 (en) | Road surface conditions detection by distributed optic fiber system | |
| CN113780094A (en) | Perimeter intrusion identification method based on Gaussian mixture model and hidden Markov model | |
| Diao et al. | Image sequence measures for automatic target tracking | |
| Karunakaran | Deep learning based object detection using mask RCNN | |
| Liu et al. | Video monitoring of Landslide based on background subtraction with Gaussian mixture model algorithm | |
| CN113534276B (en) | Railway intrusion behavior detection method based on fast R-CNN | |
| CN114925722A (en) | Perimeter security intrusion signal detection method based on generalized S transformation and transfer learning |
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: 20200602 |