SVM-based optical cable knocking positioning method and system

Abstract

The invention relates to the technical field of distributed optical fiber sensing, in particular to an SVM-based optical cable knocking positioning method and system, and the method comprises the following steps: knocking a manhole cover or the ground above a target optical cable through a hammer; collecting and caching vibration signal data of the optical cable; extracting the vibration signal data to obtain knocking waveform feature information; and inputting the extracted knocking waveform feature information into the SVM model to obtain a judgment result. According to the method, six types of features including an average value, a standard deviation, a variance, a maximum value, a median and a variable coefficient in a to-be-identified signal are extracted and then sent to an SVM model for identification. Non-manual knocking interference is eliminated, optical cable positioning is efficiently and rapidly achieved, and the false alarm rate of the system is reduced. Moreover, the vibration threshold value does not need to be set manually, the scene adaptability is enhanced, and the intelligence of the optical cable routing detector system is further improved. Therefore, the problem that an existing cable knocking positioning method is low in recognition precision is solved.

Description

Technical Field

[0001] The present invention relates to the field of distributed optical fiber sensing technology, and in particular to an optical cable knocking positioning method and system based on SVM. Background Art

[0002] With the construction of my country's optical communication network, more and more optical cables are laid. At the same time, the manufacturers and models of optical cables are becoming more and more consistent. The identification and positioning of target optical cables in communication operation and maintenance is becoming more and more difficult.

[0003] The commonly used positioning algorithm of the optical cable routing detector is to actively knock on the ground and manhole covers to cause changes in the optical cable vibration signal, and then compare the amplitude threshold set in advance to locate and identify the target optical cable. The urban environment is noisy, and a large number of optical cable wells exist on the road. The movement of vehicles will cause the manhole covers to vibrate, which is easy to cause false alarms and inaccurate identification. Using fixed threshold judgment, different sensitivities need to be set in different scenarios, and the adaptability of the scenarios is low. In addition, too many parameter settings have high requirements for users, which is not conducive to the promotion and use of optical cable routing detectors. Summary of the invention

[0004] The purpose of the present invention is to provide a method and system for locating optical cable knocks based on SVM, aiming to solve the problem of low recognition accuracy of existing cable knock locating methods.

[0005] To achieve the above object, in a first aspect, the present invention provides an optical cable knocking positioning method based on SVM, comprising the following steps:

[0006] Use a hammer to hit the manhole cover or the ground above the target optical cable;

[0007] Collect and cache the vibration signal data of the optical cable;

[0008] Extract the vibration signal data to obtain the characteristic information of the knocking waveform;

[0009] The extracted tapping waveform feature information is input into the SVM model to obtain the judgment result.

[0010] Among them, in “using a hammer to hit the manhole cover or the ground above the target optical cable”, the frequency of the hammer hitting is 4-6 times within 10 seconds.

[0011] Among them, in "collecting and caching vibration signal data of optical cable", the sampling rate of the collection is 32 Hz, the data cache time is 10 seconds, and the array frame length is 320.

[0012] Among them, in "extracting vibration signal data to obtain knocking waveform feature information", the extraction includes six types of features: mean, standard deviation, variance, maximum value, median and coefficient of variation.

[0013] In the second aspect, a SVM-based optical cable knock positioning system adopts the SVM-based optical cable knock positioning method described in the first aspect, including an optical path unit, a signal acquisition unit, a signal processing unit, a positioning unit and a display unit, wherein the optical path unit, the signal acquisition unit, the signal processing unit and the display unit are connected in sequence, and the positioning unit is connected to the signal processing unit and to the display unit.

[0014] The present invention discloses an optical cable knocking positioning method based on SVM, comprising the following steps: using a hammer to knock the manhole cover or the ground above the target optical cable; collecting and caching the vibration signal data of the optical cable; extracting the vibration signal data to obtain the knocking waveform feature information; inputting the extracted knocking waveform feature information into the SVM model to obtain a determination result. The present invention extracts six types of features, namely the mean value, standard deviation, variance, maximum value, median, and coefficient of variation, from the signal to be identified, and then inputs them into the SVM model for identification. Non-artificial knocking interference is eliminated, and the optical cable positioning is realized efficiently and quickly, reducing the false alarm rate of the system. In addition, there is no need to manually set the vibration threshold, which enhances the scene adaptability and further improves the intelligence of the optical cable routing detector system. This solves the problem of low recognition accuracy of the existing cable knocking positioning method. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 The present invention provides a flow chart of a method for locating optical cable knocking based on SVM.

[0017] Figure 2 It is a schematic diagram of an optical cable knocking positioning system based on SVM provided by the present invention.

[0018] Figure 3 The vibration signal is obtained by hitting the manhole cover above the target optical cable with a hammer.

[0019] Figure 4 It is the vibration signal caused by other interference factors of the manhole cover above the target optical cable.

[0020] 1-optical path unit, 2-signal acquisition unit, 3-signal processing unit, 4-positioning unit, 5-display unit. DETAILED DESCRIPTION

[0021] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0022] See also Figure 1 In a first aspect, the present invention provides a method for locating an optical cable knocking based on SVM, comprising the following steps:

[0023] S1 uses a hammer to hit the manhole cover or the ground above the target optical cable;

[0024] The frequency of the hammer striking is 4-6 times within 10 seconds.

[0025] Specifically, the tester arrives at the location to be tested and uses a hammer to hit the manhole cover or the ground above the target optical cable. The tester maintains a certain force and hits at a fixed frequency.

[0026] S2 collects the vibration signal data of the optical cable and caches it;

[0027] The sampling rate of the acquisition is 32 Hz, the data buffer time is 10 seconds, and the array frame length is 320.

[0028] Specifically, the signal data to be determined of a specified length is cached and average difference preprocessed. First, the original signal data is accumulated and averaged, then the data of each two adjacent points after the accumulation and averaging are differentially calculated, and finally the absolute value is taken to convert the phase change of the reflected light in the optical cable due to vibration into the change of the reflected light intensity amplitude, and the data is cached.

[0029] S3 extracts the vibration signal data to obtain the knocking waveform feature information;

[0030] The extraction includes six types of features: mean value, standard deviation, variance, maximum value, median and coefficient of variation.

[0031] Specifically, the tapping waveform features extracted include mean value (Mean), standard deviation (Std), variance (Var), maximum value (Max), median (Median), and coefficient of variation (Coefficient of Variaition).

[0032] The calculation formula of the coefficient of variation is:

[0033]

[0034] In formula (1), c vrepresents the coefficient of variation; σ represents the mean value of the preprocessed signal; μ represents the standard deviation of the preprocessed signal.

[0035] A total of six types of features are extracted, namely, mean, standard deviation, variance, maximum value, median, and coefficient of variation. Among them, the mean and median are used to estimate the central trend of the overall data. The mean is greatly affected by extreme values, and the median is insensitive to extreme values. Variance and standard deviation are used to estimate the degree of dispersion of the overall data, indicating the degree of fluctuation of the overall data. The variation system also reflects the degree of dispersion of the data, but is also affected by the mean value. The interference sources that cause false alarms during positioning are mainly car driving, optical cable shaking, etc. The above six types of features can well distinguish between knocking signals and interference signals. At the same time, these features are simple to calculate and take a short time, which is conducive to the real-time positioning of the system.

[0036] S4 inputs the extracted tapping waveform feature information into the SVM model to obtain a determination result.

[0037] Specifically, the construction method of the SVM model includes: using an optical cable routing detector to collect data on artificial knocking on manhole covers or ground signals, and other noise interference signal data (such as cars running over manhole covers, hanging optical cables shaking, etc.), and dividing the collected data into two categories: artificial knocking and non-artificial knocking; preprocessing and feature extraction, constructing training data sets and test data sets; configuring SVM model parameters, inputting the training set into the SVM model for training, and using the test set to verify the effect.

[0038] First, collect the training data set of the SVM model. The data types are divided into knocking signals and other interference signals, among which the interference signals are various types of non-knocking signals in the real optical cable line. Mark the interference signal as 0 and the knocking signal as 1. Perform average difference preprocessing on the marked type of signal, and then extract six types of eigenvalues ​​to form the training set of the SVM model. 70% of the data is used to train the model, and the remaining 30% of the data is used as the test set. Configure the relevant parameters to complete the training of the SVM model. Finally, the waveform features to be identified are input into the SVM model, and the output is the probability of the knocking waveform signal. If the probability is greater than 0.6, it is judged as a knocking signal. Finally, determine the location of the corresponding sampling point, and then convert it into the knocking position of the target optical cable.

[0039] See also Figures 2 to 4 In the second aspect, a SVM-based optical cable knocking positioning system adopts the SVM-based optical cable knocking positioning method described in the first aspect, including an optical path unit 1, a signal acquisition unit 2, a signal processing unit 3, a positioning unit 4 and a display unit 5, wherein the optical path unit 1, the signal acquisition unit 2, the signal processing unit 3 and the display unit 5 are connected in sequence, and the positioning unit 4 is connected to the signal processing unit 3 and to the display unit 5.

[0040] In this embodiment, in the optical path unit 1, the narrow line width laser emits a high coherence narrow band laser, which is modulated into a pulse form by the modulator and enters the optical cable to be tested, and the generated backward Rayleigh scattered light is returned to the signal acquisition unit 2 through the circulator; the signal acquisition unit 2 converts the received backward Rayleigh scattered light signal into an electrical signal and transmits it to the signal processing unit 3; the signal processing unit 3 pre-processes the original signal it receives, that is, converts the optical cable vibration into a corresponding amplitude change, and sends the waveform to the display unit 5. At the same time, after receiving the positioning instruction, the signal processing unit 3 caches the signal data of the specified length and sends it to the positioning unit 4 for positioning identification; after receiving the cached data to be determined, the positioning unit 4 extracts the relevant features and sends them to the SVM model to determine whether it is a valid knocking signal, and then calculates the position of the acquisition point corresponding to the valid knocking signal on the optical cable, and sends the position information to the display unit 5 to complete the knocking positioning. Among them, the display unit 5 displays the vibration waveform signal on the monitoring optical cable line and the knocking position information determined by the positioning unit 4 in real time.

[0041] Figure 3 and Figure 4 The waveform diagrams of the artificial knocking waveform signal and the noise interference signal in the embodiment are respectively within a length of 10 seconds. Figure 3 The vibration signal is obtained by hitting the manhole cover above the target optical cable with a hammer. Figure 4 This is the vibration signal caused by other interference factors of the manhole cover above the target optical cable. Figure 3 and Figure 4 It can be seen that Figure 4 The vibration signal curve in the image is chaotic, which is caused by the irregular vibration of the outside world. Figure 3 The waveform in the figure is caused by regular knocking on the manhole cover. It can be clearly seen that this is a vibration signal with a specific pattern.

[0042] Beneficial effects:

[0043] 1. The present invention extracts six types of features, namely, mean value, standard deviation, variance, maximum value, median and coefficient of variation, from the signal to be identified, and then feeds them into the SVM model for identification. This eliminates non-artificial knocking interference, achieves optical cable positioning efficiently and quickly, and reduces the false alarm rate of the system.

[0044] 2. There is no need to manually set the vibration threshold, which enhances the adaptability to the scene and further improves the intelligence of the optical cable routing detector system.

[0045] What is disclosed above is only a preferred embodiment of the optical cable tapping positioning method and system based on SVM of the present invention. Of course, it cannot be used to limit the scope of rights of the present invention. Ordinary technicians in this field can understand that all or part of the processes of the above embodiments and equivalent changes made according to the claims of the present invention still fall within the scope of the invention.

Claims

1. A method for locating optical cable knocking based on SVM, characterized in that: The following steps are involved: Use a hammer to hit the manhole cover or the ground above the target optical cable; Collect and cache the vibration signal data of the optical cable; Extract the vibration signal data to obtain the characteristic information of the knocking waveform; The extracted tapping waveform feature information is input into the SVM model to obtain the judgment result.

2. The optical cable knocking positioning method based on SVM as claimed in claim 1, characterized in that: In "using a hammer to hit the manhole cover or the ground above the target optical cable", the frequency of the hammer hitting is 4-6 times within 10 seconds.

3. The optical cable knocking positioning method based on SVM as claimed in claim 1, characterized in that: In "collecting and caching vibration signal data of optical cable", the sampling rate of the collection is 32 Hz, the data cache time is 10 seconds, and the array frame length is 320.

4. The optical cable knocking positioning method based on SVM as claimed in claim 1, characterized in that: In "extracting vibration signal data to obtain knocking waveform feature information", the extraction includes six types of features: mean, standard deviation, variance, maximum value, median and coefficient of variation.

5. An optical cable knocking positioning system based on SVM, adopting the optical cable knocking positioning method based on SVM according to any one of claims 1 to 4, characterized in that: The optical path unit comprises a signal collection unit, a signal processing unit, a positioning unit and a display unit. The optical path unit, the signal collection unit, the signal processing unit and the display unit are connected in sequence. The positioning unit is connected to the signal processing unit and is also connected to the display unit.

Images