CN104794839B - One kind is based on POTDR fiber optic intrusion recognizer - Google Patents

Abstract

The invention belongs to technical field of optical fiber sensing, disclose one kind and be based on POTDR fiber optic intrusion recognizer, including：Read the initial data in the 2N cycle that optical fibre vibration sensor collects from data acquisition end and carry out denoising；Array X after denoising is divided two groups, and asks for difference DELTA X；Difference array Δ X is divided into multiple equal length window arrays, forms array Y；Calculate the fourth central of the window array in array Y away from and compare with given threshold D, demarcate the superthreshold Value Data Y in array Y_aj, and by subscript aj typing coordinate array A；According to minima a in array A_minLocating alarming point, if no superthreshold array, repeat the above steps carry out lower whorl identification；Wherein, the arrangement mode of initial data is consistent with described optical fibre vibration sensor sampled point arrangement mode, and the subscript of window array as answers the coordinate of sampled point；The coordinate of sampled point characterizes sampled point position light path in a fiber.The present invention is lifted into reliability and the positioning precision invading recognizer by window array.

Description

POTDR-based optical fiber intrusion recognition algorithm

Technical Field

The invention relates to the technical field of optical fiber sensing, in particular to an optical fiber intrusion identification algorithm based on POTDR.

Background

The distributed optical fiber vibration sensing system has the advantages of electromagnetic interference resistance, good electrical insulation, corrosion resistance, high sensitivity, passive sensing units and the like, and is widely used for monitoring and protecting important areas and places such as national boundaries, military bases, oil and gas pipelines, power transmission cables and the like.

The system hardware comprises a computer, a data acquisition card, an optical detector, a pulse light source, a sensing optical fiber link and an optical circulator and an on-line analyzer, wherein the data acquisition card acquires data and transmits the data to the computer for further algorithm processing. The signal light of the standard POTDR sensing system comes from the backward rayleigh scattering of the forward propagating pulsed light. The biggest characteristics of such signal light are: the backward Rayleigh scattering coefficient of the common single mode fiber is very small and is only about 10^-7Per meter, the signal light generated is weak, on the order of a few to tens of nanowatts. Meanwhile, since the frequency of the signal light is very high, ranging from tens to hundreds of megahertz, the noise equivalent power of the photodetector is also tens of millionsNano-watt. Ultimately resulting in a system with a low signal-to-noise ratio. In addition, because rayleigh scattering is a random process, fluctuation exists, and signal noise is caused.

Due to the characteristics of the signals, in practical engineering application, due to the complexity of an engineering environment, more external interference, and the incompleteness of a light source, a light detector and the like of a system, sudden change of the signals is easily caused, false alarm of the system can be caused, the reliability is low, and meanwhile, an identification algorithm cannot accurately position an alarm point, so that the positioning precision is low.

Disclosure of Invention

The invention provides a POTDR-based optical fiber intrusion identification algorithm, which solves the problems of low intrusion identification reliability and low positioning precision in the prior art, and greatly improves the reliability and positioning precision of intrusion identification.

In order to solve the above technical problems, the present invention provides an intrusion identification algorithm based on a POTDR optical fiber, comprising:

reading 2N periods of original data acquired by the optical fiber vibration sensor from a data acquisition end and denoising;

dividing the denoised array X into two groups, and solving a difference value delta X;

dividing the difference array delta X into a plurality of equal-length window arrays to form an array Y;

calculating the fourth-order center distance of the window array in the array Y, comparing the four-order center distance with a set threshold value D, and calibrating the over-threshold value array Y in the array Y_ajAnd recording the subscript aj into a coordinate array A;

according to the minimum value a in the array A_minPositioning an alarm point, and if no over-threshold array exists, repeating the steps to perform next round identification;

the arrangement mode of the original data is consistent with the arrangement mode of sampling points of the optical fiber vibration sensor, and subscripts of the window array are coordinates of the sampling points; and the coordinates of the sampling points represent the optical path lengths of the positions of the sampling points in the optical fiber.

Further, the data length of each period of the original data is L, and a matrix array X is formed:

wherein, each pulse period collects a frame of data, 2N frames in total.

Further, denoising the array X in an average denoising mode;

setting the front N frame data and the back N frame data of the array X as independent arrays respectively for independent average denoising to obtain two-frame array X_NAnd array X_2N(ii) a The array X is processed_NAnd array X_2NThe difference value is found to be the array Δ X:

[ΔX(1),ΔX(2),ΔX(3),...,ΔX(L-1),ΔX(L)]。

further, the array Y comprises a plurality of continuous equal-length window arrays, and the window arrays are selected according to the following rules;

Y₁＝[ΔX(1),ΔX(2),...,ΔX(n)]

Y₂＝[ΔX(2),ΔX(3),...,ΔX(n+1)]

Y_L-n+1＝[ΔX(L-n+1),ΔX(L-n+2),...,ΔX(L)]

wherein n is a natural number greater than 0 and less than L.

Further, the fourth-order center distance is:

Z＝[z₁,z₂,...,z_L-n+1]

wherein,k＝1,2,…,L-n+1。

further, according to the minimum value a in the array A_minPositioning an alarm point; by the formula:

wherein τ is the time interval between each sampling point, and ν is the propagation speed of light in the optical fiber.

Further, the algorithm further comprises: controlling ringing width J and ringing frequency I;

the ringing times I represent the actual length of the array A, namely the number of the super-threshold window arrays; recording at most one super-threshold array in each round of identification operation process;

entering next round of identification after the super-threshold array is not found in the round of identification or a super-threshold array appears and a super-threshold array subscript is recorded;

the ringing width J represents the number of turns for carrying out intrusion identification;

and according to the minimum value a in the array A, only under the condition that the number I of the actually recorded super-threshold arrays is greater than or equal to a set value I and the number J of the rounds of intrusion identification is equal to a set value J_minPositioning an alarm point;

wherein, the set value of the ringing width J is more than or equal to the set value of the ringing times I.

One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

1. the POTDR-based optical fiber intrusion identification algorithm provided by the embodiment of the invention acquires two frames of data by grouping and averaging the data in multiple periods, thereby effectively avoiding the influence of disturbance on the data acquisition of the sensor, minimizing the influence of accidental fluctuation of the data, reducing the system error of the sensor, such as the influence of thermal noise of devices and the like, and greatly improving the reliability of the detected data.

2. The POTDR-based optical fiber intrusion identification algorithm provided by the embodiment of the invention represents the noise distribution state in the window area range by taking the difference of the data after grouping average denoising and grouping the window arrays on the basis, thereby avoiding the accidental defect that the difference of two frames of data is taken as an alarm identification algorithm in the prior art, and improving the reliability of alarm judgment through the multi-data state in the window area.

3. The POTDR-based optical fiber intrusion identification algorithm provided by the embodiment of the invention perfects the setting rule of the window array on the basis of the window array, sets a plurality of equal-length window arrays, advances along the propagation direction of the pulse optical signal, simultaneously partially covers the window range of the adjacent window array, finds out the interval where an intrusion point is located by utilizing the characteristic that noise distribution obeys normal distribution, definitely finds out the window array where the intrusion point is located by calculating the four-step center distance, and greatly improves the positioning accuracy of alarm information;

4. the POTDR-based optical fiber intrusion identification algorithm provided by the embodiment of the invention accurately obtains all window arrays meeting alarm conditions by comparing the four-order center distance with the set threshold value, and uses the minimum subscript a_minFinding out an intrusion point for reference, namely determining a place where disturbance occurs earliest on the basis of the characteristics of data subscripts corresponding to sampling point coordinates of a sensing system in the prior art, and greatly improving the positioning accuracy; in conclusion, through multi-period data acquisition and average denoising, data disturbance and accidental fluctuation are reduced, the reliability of data acquisition is improved, and through the setting of a window array; through window array setting, calculation of the four-step center distance and minimum subscript positioning, accuracy of intrusion information identification and judgment and accuracy of alarm information positioning are improved.

5. The control process can further reduce the risk of misinformation through the ringing width J and the ringing times I, and simultaneously further improves the positioning precision and greatly improves the efficiency of the intrusion recognition algorithm.

Drawings

Fig. 1 is a schematic flowchart of a POTDR-based optical fiber intrusion identification algorithm according to an embodiment of the present invention;

fig. 2 is a process curve diagram of a POTDR-based optical fiber intrusion identification algorithm according to an embodiment of the present invention.

Detailed Description

The embodiment provides an optical fiber intrusion identification algorithm based on POTDR, and solves the problems that in the prior art, the signal to noise ratio is too low, the intrusion identification false alarm rate is high, the alarm reliability is low, and the positioning accuracy is not high. The technical effects of improving the intrusion recognition efficiency and improving the reliability and the positioning precision of the health care function are achieved.

In order to solve the above technical problem, the general idea of the technical solution provided in the embodiments of the present application is as follows:

a POTDR-based optical fiber intrusion identification algorithm comprises the following steps:

reading 2N periods of original data acquired by the optical fiber vibration sensor from a data acquisition end and denoising;

dividing the denoised array X into two groups, and solving a difference value delta X;

dividing the difference array delta X into a plurality of equal-length window arrays to form an array Y;

calculating the fourth-order center distance of the window array in the array Y, comparing the four-order center distance with a set threshold value D, and calibrating the over-threshold value array Y in the array Y_ajAnd recording the subscript aj into a coordinate array A;

according to the minimum value a in the array A_minPositioning an alarm point, and if no over-threshold array exists, repeating the steps to perform next round identification;

the arrangement mode of the original data is consistent with the arrangement mode of sampling points of the optical fiber vibration sensor, and subscripts of the window array are coordinates of the sampling points; and the coordinates of the sampling points represent the optical path of the positions of the sampling points in the optical fiber.

According to the method, the collected data in 2N sampling periods are subjected to average denoising, so that the influence of random disturbance, accidental data fluctuation and other factors on the sampled data is reduced to the minimum, and the reliability of the original data is improved; then obtaining a difference array delta X by grouping and differencing, wherein in the difference array delta X, along the data arrangement direction, the difference value array is divided into a plurality of continuous equal-length window arrays Y by a sectional interception mode, partial data of adjacent window arrays are the same, and according to the principle that noise obeys normal distribution, the noise state of each window array can be accurately reflected by calculating the four-order center distance of each window array, the influence caused by the data change in the window array is directly reflected, under the condition that no invasion signal occurs, the four-step center distance change of the sequentially arranged window arrays is not obvious or is below a limit value, when the intrusion signal appears, the four-order center distance of the window array is obviously changed relative to the four-order center distance of the previous window array, therefore, the existence of obvious intrusion signals can be judged, and meanwhile, the specific intrusion positioning information can be accurately calibrated. Compared with the prior art, the method for reflecting the intrusion information and positioning by the difference value of two frames of data is simple, and the algorithm provided by the embodiment has higher reliability and positioning accuracy.

In order to better understand the technical solutions, the technical solutions will be described in detail below with reference to the drawings and the specific embodiments of the present disclosure, and it should be understood that the specific features in the embodiments and examples of the present disclosure are detailed descriptions of the technical solutions of the present disclosure, but not limitations of the technical solutions of the present disclosure, and the technical features in the embodiments and examples of the present disclosure may be combined with each other without conflict.

Referring to fig. 1, an intrusion identification algorithm based on a POTDR optical fiber provided by an embodiment of the present invention includes:

reading 2N periods of original data acquired by the optical fiber vibration sensor from a data acquisition end and denoising; the data length of each period of the original data is L, and a matrix array X is formed:

wherein, every pulse cycle gathers a frame data, gathers 2N frames altogether, from top to bottom according to gathering the order and constitutes the matrix array.

In the normal collection process, various disturbances generally affect some data, on the premise of acknowledging the randomness of the disturbances, the influence of the disturbances on the whole data can be fully reduced through multiple sampling, and the reliability of the data can be improved through average denoising.

In the prior art, the difference operation is performed on two frames of data, and whether intrusion exists is reflected by the size relationship between the difference value and a preset threshold value, so that the randomness is high, and the false alarm risk is high.

In this embodiment, the denoised array X is divided into two groups, and a difference Δ X is obtained; setting the front N frame data and the back N frame data of the matrix array X as independent arrays respectively for independent average denoising to obtain two frame arrays X_NAnd array X_2N(ii) a The array X is processed_NAnd array X_2NThe difference value is found to be the array Δ X:

[ΔX(1),ΔX(2),ΔX(3),...,ΔX(L-1),ΔX(L)]。

meanwhile, dividing the difference array delta X into a plurality of equal-length window arrays to form an array Y; along the data arrangement direction, the difference array delta X is divided into a plurality of window arrays Y with continuous equal length in a segmentation intercepting mode, and partial data of adjacent window arrays are the same, namely the window arrays are selected according to the following rules:

Y₁＝[ΔX(1),ΔX(2),...,ΔX(n)]

Y₂＝[ΔX(2),ΔX(3),...,ΔX(n+1)]

Y_L-n+1＝[ΔX(L-n+1),ΔX(L-n+2),...,ΔX(L)]

wherein n is a natural number greater than 0 and less than L; most data of a rear array and a front array in the adjacent window arrays are the same, and a new data is added to the rear array; and the subscript of the window array is exactly consistent with the coordinates of the sampling points. The method advances along the propagation direction of the pulse optical signal, simultaneously the window range of the adjacent window arrays is partially covered, the interval where the intrusion point is located is found out by utilizing the characteristic that the noise distribution obeys normal distribution, the window array where the intrusion point is located is definitely found out through the calculation of the four-step center distance, and the positioning precision of the alarm information is greatly improved. The window array reflects the noise distribution state in a range, and under the condition that the data parts of the continuous window arrays are the same, if larger noise state change occurs, the intrusion signal in the subsequent window array can be judged, so that the alarm is realized.

Calculating the four-order center distance of the window array in the array Y, comparing the four-order center distance with a set threshold value D, and calibrating the over-threshold value array Y in the array Y_ajAnd recording the subscript aj into a coordinate array A; the fourth-order center distance is as follows:

Z＝[z₁,z₂,...,z_L-n+1]

wherein,k＝1,2,…,L-n+1。

according to the minimum value a in the array A_minPositioning an alarm point, and if no over-threshold array exists, repeating the steps to perform next round identification; according to the minimum value a in the array A_minPositioning an alarm point; by the formula:

wherein τ is a time interval between each sampling point, and ν is a propagation speed of light in the optical fiber; the arrangement mode of the original data is consistent with the arrangement mode of the sampling points of the optical fiber vibration sensor, and the subscript of the window array is the coordinate of the sampling points; the coordinates of the sampling points represent the optical path of the sampling point positions in the optical fiber; therefore, the specific position information of the previous point can be directly obtained through the formula, and accurate positioning is realized.

Referring to fig. 1, in order to further enhance reliability and positioning accuracy of intrusion identification, a ring identification loop algorithm is set, that is, a ring width J and a ring frequency I are set to respectively represent the number of times of judging the super-threshold array and the number of the super-threshold array, so that flow management and control are performed, and automatic loop execution is realized.

The specific algorithm rule is as follows: the ringing times I represent the actual length of the array A, namely the number of the super-threshold window arrays; recording at most one super-threshold array in each round of identification operation process;

and entering the next round of identification after the super-threshold array is not found in the round of identification or a super-threshold array appears and the subscript of the super-threshold array is recorded.

The ringing width J represents the number of turns for carrying out intrusion identification;

and according to the minimum value a in the array A, only under the condition that the number I of the actually recorded super-threshold arrays is greater than or equal to a set value I and the number J of the rounds of intrusion identification is equal to a set value J_minAnd (5) positioning an alarm point.

Wherein, the set value of the ringing width J is more than or equal to the set value of the ringing times I.

As can be clearly seen from the above algorithm, the specific alarm positioning rule is as follows:

in the process of intrusion identification, firstly screening and positioning a super-threshold value array are completed, and positioning information is recorded in an array A, wherein the process comprises two results, namely a super-threshold value window array exists, so that a sampling alarm point can be positioned, the ringing frequency i is added with 1, or the super-threshold value window array is not found, no alarm positioning information exists, and the ringing frequency i is unchanged; this round of intrusion recognition ends and the count of the ring width j is incremented by 1. Repeating the above operation to perform the next round of identification until the ringing width J is increased to a set value J, namely performing J rounds of intrusion identification processes, completing all the intrusion identification algorithm flows, and then obtaining two results: the ringing times I are more than or equal to a set value I, namely an over-threshold window array exceeding the set requirement number is found, the number of the sampling alarm points meets the alarm requirement, the alarm operation can be executed, and the alarm positioning is executed through a minimum subscript positioning formula; or the ringing frequency I is less than the set value I, namely the number of the sampling alarm points is too small, so that the accidental risk is considered to be too high and not suitable for alarming, and the intrusion risk is eliminated; and clearing the real-time counting of the ringing width j and the ringing times i at the same time, and finishing the algorithm flow.

The intrusion identification process is expanded to a plurality of continuous identification periods through a ringing width J control algorithm and a ringing frequency I control algorithm, namely, multi-round continuous identification is carried out, on one hand, false alarm caused by instant change of data which can appear in single-round identification can be avoided, the risk can be reduced through multi-round continuous intrusion identification, and the reliability of intrusion identification is improved; on the other hand, the sampling range of the super-threshold window array is expanded through multi-round continuous intrusion identification, and the positioning precision is further enhanced by matching with a minimum positioning algorithm.

Referring to fig. 2, a data process curve of the algorithm, a graph a and a graph B are data curves of the first N frames and the last N frames respectively, and a graph C, which is used for alarm judgment in the prior art, is seen to be within a threshold range, and the reliability of disturbing many alarms is not high; and the graph D is an intrusion identification curve for calculating the four-order center distance through the window array, so that the alarm reliability is obviously higher. In field applications, the threshold D is set according to actual conditions.

One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

1. the POTDR-based optical fiber intrusion identification algorithm provided by the embodiment of the invention acquires two frames of data by grouping and averaging the data in multiple periods, thereby effectively avoiding the influence of disturbance on the data acquisition of the sensor, minimizing the influence of accidental fluctuation of the data, reducing the system error of the sensor, such as the influence of thermal noise of devices and the like, and greatly improving the reliability of the detected data.

2. The POTDR-based optical fiber intrusion identification algorithm provided by the embodiment of the invention represents the noise distribution state in the window area range by taking the difference of the data after grouping average denoising and grouping the window arrays on the basis, thereby avoiding the accidental defect that the difference of two frames of data is taken as an alarm identification algorithm in the prior art, and improving the reliability of alarm judgment through the multi-data state in the window area.

3. The POTDR-based optical fiber intrusion identification algorithm provided by the embodiment of the invention perfects the setting rule of the window array on the basis of the window array, sets a plurality of equal-length window arrays, advances along the propagation direction of the pulse optical signal, simultaneously partially covers the window range of the adjacent window array, finds out the interval where an intrusion point is located by utilizing the characteristic that noise distribution obeys normal distribution, definitely finds out the window array where the intrusion point is located by calculating the four-step center distance, and greatly improves the positioning accuracy of alarm information;

4. the POTDR-based optical fiber intrusion identification algorithm provided by the embodiment of the invention accurately obtains all window arrays meeting alarm conditions by comparing the four-order center distance with the set threshold value, and uses the minimum subscript a_minFinding out an intrusion point for reference, namely determining a place where disturbance occurs earliest on the basis of the characteristics of data subscripts corresponding to sampling point coordinates of a sensing system in the prior art, and greatly improving the positioning accuracy; in conclusion, through multi-period data acquisition and average denoising, data disturbance and accidental fluctuation are reduced, the reliability of data acquisition is improved, and through the setting of a window array; through window array setting, calculation of the four-step center distance and minimum subscript positioning, accuracy of intrusion information identification and judgment and accuracy of alarm information positioning are improved.

5. The control process can further reduce the risk of misinformation through the ringing width J and the ringing times I, and simultaneously further improves the positioning precision and greatly improves the efficiency of the intrusion recognition algorithm.

Finally, it should be noted that the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting, and although the present invention has been described in detail with reference to examples, it should be understood by those skilled in the art that modifications or equivalent substitutions may be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, which should be covered by the claims of the present invention.

Claims (7)

1. A POTDR-based optical fiber intrusion identification algorithm is characterized by comprising the following steps:

reading 2N periods of original data acquired by the optical fiber vibration sensor from a data acquisition end and denoising;

dividing the denoised array X into two groups, and solving a difference value delta X;

dividing the difference array delta X into a plurality of equal-length window arrays to form an array Y;

calculating the four-order center distance of the window array in the array Y, comparing the four-order center distance with a set threshold value D, and calibrating the superordinate in the array YThreshold array Y_ajAnd recording the subscript aj into a coordinate array A;

according to the minimum value a in the array A_minPositioning an alarm point, and if no over-threshold array exists, repeating the steps to perform next round identification;

the arrangement mode of the original data is consistent with the arrangement mode of sampling points of the optical fiber vibration sensor, and subscripts of the window array are coordinates of the sampling points; and the coordinates of the sampling points represent the optical path lengths of the positions of the sampling points in the optical fiber.

2. The POTDR-based fiber intrusion identification algorithm of claim 1, wherein: the data length of each period of the original data is L, and a matrix array X is formed:

$\left[\begin{array}{cccc}{x}_{11}& x_{12}& ...& x_{1L}\\ x_{21}& x_{22}& ...& x_{2L}\\ ...& ...& ...& ...\\ x_{2N,1}& x_{2N,2}& ...& x_{2N,L}\end{array}\right]$

wherein, each pulse period collects a frame of data, 2N frames in total.

3. The POTDR-based fiber intrusion identification algorithm of claim 1, wherein: denoising the array X by adopting an average denoising method;

setting the front N frame data and the back N frame data of the array X as independent arrays respectively for independent average denoising to obtain two-frame array X_NAnd array X_2N(ii) a The array X is processed_NAnd array X_2NThe difference value is found to be the array Δ X:

[ΔX(1),ΔX(2),ΔX(3),…,ΔX(L-1),ΔX(L)]。

4. the POTDR-based fiber intrusion identification algorithm of claim 1, wherein: the array Y comprises a plurality of continuous equal-length window arrays, and the window arrays are selected according to the following rules;

Y₁＝[ΔX(1),ΔX(2),...,ΔX(n)]

Y₂＝[ΔX(2),ΔX(3),...,ΔX(n+1)]

...

Y_L-n+1＝[ΔX(L-n+1),ΔX(L-n+2),...,ΔX(L)]

wherein n is a natural number greater than 0 and less than L.

5. The POTDR-based fiber intrusion identification algorithm of claim 1, wherein: the four-step center distance is as follows:

Z＝[z₁,z₂,...,z_L-n+1]

wherein,

6. the POTDR-based optical fiber intrusion identification algorithm of any one of claims 1 to 5, wherein: according to the minimum value a in the array A_minPositioning an alarm point; by the formula:

$D_{alarm}=\frac{a_{min}\×\mathrm{\τ}\×v}{2}$

wherein tau is the time interval between each sampling point, and v is the propagation speed of light in the optical fiber.

7. The POTDR-based fiber intrusion identification algorithm of claim 1, further comprising: controlling ringing width J and ringing frequency I;

the ringing times I represent the actual length of the array A, namely the number of the super-threshold window arrays; recording at most one super-threshold array in each round of identification operation process;

entering next round of identification after the super-threshold array is not found in the round of identification or a super-threshold array appears and a super-threshold array subscript is recorded;

the ringing width J represents the number of turns for carrying out intrusion identification;

and according to the minimum value a in the array A, only under the condition that the number I of the actually recorded super-threshold arrays is greater than or equal to a set value I and the number J of the rounds of intrusion identification is equal to a set value J_minPositioning an alarm point;

wherein, the set value of the ringing width J is more than or equal to the set value of the ringing times I.