CN113869240A - Transmission optical cable real-time warning method and device based on deep learning algorithm - Google Patents
Transmission optical cable real-time warning method and device based on deep learning algorithm Download PDFInfo
- Publication number
- CN113869240A CN113869240A CN202111160835.XA CN202111160835A CN113869240A CN 113869240 A CN113869240 A CN 113869240A CN 202111160835 A CN202111160835 A CN 202111160835A CN 113869240 A CN113869240 A CN 113869240A
- Authority
- CN
- China
- Prior art keywords
- time
- optical cable
- data
- space
- feature extraction
- 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
- 230000003287 optical effect Effects 0.000 title claims abstract description 207
- 230000005540 biological transmission Effects 0.000 title claims abstract description 39
- 238000000034 method Methods 0.000 title claims abstract description 24
- 238000013135 deep learning Methods 0.000 title claims abstract description 19
- 238000000605 extraction Methods 0.000 claims abstract description 82
- 238000011176 pooling Methods 0.000 claims description 36
- 238000013527 convolutional neural network Methods 0.000 claims description 17
- 230000015654 memory Effects 0.000 claims description 12
- 239000013307 optical fiber Substances 0.000 claims description 10
- 239000011159 matrix material Substances 0.000 claims description 9
- 230000010287 polarization Effects 0.000 claims description 8
- 230000007787 long-term memory Effects 0.000 abstract description 5
- 230000000246 remedial effect Effects 0.000 abstract description 5
- 230000006403 short-term memory Effects 0.000 abstract description 5
- 230000006870 function Effects 0.000 abstract description 4
- 238000012544 monitoring process Methods 0.000 description 6
- 238000013528 artificial neural network Methods 0.000 description 5
- 239000000835 fiber Substances 0.000 description 5
- 238000010276 construction Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000013598 vector Substances 0.000 description 2
- 238000009412 basement excavation Methods 0.000 description 1
- 238000005553 drilling Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01H—MEASUREMENT OF MECHANICAL VIBRATIONS OR ULTRASONIC, SONIC OR INFRASONIC WAVES
- G01H9/00—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves by using radiation-sensitive means, e.g. optical means
- G01H9/004—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves by using radiation-sensitive means, e.g. optical means using fibre optic sensors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Optical Communication System (AREA)
Abstract
The invention discloses a transmission optical cable real-time warning method and device based on a deep learning algorithm. The method comprises the steps of constructing an optical cable time-space data set of a transmission optical cable; performing space-time characteristic extraction processing on the cable space-time data set; predicting vibration data by using a long-term and short-term memory network to obtain a vibration data prediction value; and comparing the obtained vibration data predicted value with a preset vibration data threshold value, and sending an alarm signal when the vibration data predicted value is greater than the vibration data threshold value. The invention improves the accuracy of vibration data prediction, can more reliably monitor the condition of the transmission optical cable, and more accurately realize the real-time warning function of the transmission optical cable, so that the working personnel can timely take precautionary or remedial measures to protect the transmission optical cable.
Description
Technical Field
The invention relates to the technical field of optical cable monitoring, in particular to a transmission optical cable real-time warning method and device based on a deep learning algorithm.
Background
The traditional optical cable monitoring technology is generally used for collecting optical basic parameters (such as wavelength, frequency, intensity, polarization state, phase and the like) of optical signals transmitted in optical fibers so as to monitor physical quantities such as vibration along the optical cable and establish the relationship between the optical parameters and the vibration of an optical cable line, thereby realizing the monitoring of the safety state of the optical cable and avoiding external force damage and theft. However, the vibration of the optical cable has many factors, including pipe jacking construction, road excavation, underground drilling, external random factors and the like, and also includes the influence of the optical signal optical basic parameter noise, so that the reliability of the optical cable monitoring in the prior art is low.
Disclosure of Invention
The technical problem to be solved by the embodiments of the present invention is to provide a method and a device for real-time warning of a transmission optical cable based on a deep learning algorithm, so as to solve the problem that the existing optical cable monitoring technology is easily affected by noise and has low reliability.
In order to solve the above technical problems, an embodiment of the present invention provides a transmission optical cable real-time alarm method based on a deep learning algorithm, including:
step S1, constructing an optical cable space-time data set of a transmission optical cable, wherein the optical cable space-time data in the optical cable space-time data set comprises data of a plurality of optical signal optical basic parameters at N positions on an optical cable optical fiber at 1-T moment, wherein T is more than 1, and N is more than 1;
step S2, performing space-time feature extraction processing on the optical cable space-time data at each moment in the optical cable space-time data set to obtain T N-dimensional high-level semantic feature data at 1-T moments;
step S3, the obtained N-dimensional high-level semantic feature data at the time of 1-T is used as input, and a vibration data prediction value at the time of T +1 is obtained by using a long-short term memory network;
and step S4, comparing the obtained vibration data predicted value with a preset vibration data threshold value, and sending an alarm signal when the vibration data predicted value is larger than the vibration data threshold value.
Further, the plurality of optical signal optical basic parameters include part or all of wavelength, frequency, intensity, polarization state, and phase.
Further, the step S2 further includes: and performing space-time feature extraction processing on the optical cable space-time data at each moment in the optical cable space-time data set by using a convolutional neural network to obtain T N-dimensional high-level semantic feature data at 1-T moment.
Further, the step S2 further includes:
the convolutional neural network performs three different convolution operations on the optical cable time-space data at each moment to obtain three feature matrixes with different sizes;
and forming the N-dimensional high-level semantic feature data by tiling and splicing the obtained three feature matrixes with different sizes.
Further, the three different convolution operations include:
a first convolution operation comprising:
performing convolution on the optical cable space-time data at each moment by using a convolution kernel of 3 multiplied by 3, and performing pooling by using a pooling layer of 2 multiplied by 2 to complete a first characteristic extraction process;
performing convolution on the features extracted in the first feature extraction process by adopting a convolution kernel of 3 multiplied by 3 to finish the second feature extraction process;
and (5) performing convolution on the features extracted in the second feature extraction process by adopting a convolution kernel of 3 multiplied by 3 to finish the third feature extraction process so as to obtain a feature matrix with the first size.
Further, the three different convolution operations further include:
a second convolution operation comprising:
performing convolution on the optical cable space-time data at each moment by using a convolution kernel of 3 multiplied by 3, and performing pooling by using a pooling layer of 3 multiplied by 3 to complete a first characteristic extraction process;
performing convolution on the features extracted in the first feature extraction process by adopting a convolution kernel of 3 multiplied by 3, and performing pooling by adopting a pooling layer of 3 multiplied by 3 to finish the second feature extraction process;
and (5) performing convolution on the features extracted in the second feature extraction process by adopting a 2 multiplied by 2 convolution kernel to finish the third feature extraction process so as to obtain a feature matrix with a second size.
Further, the three different convolution operations further include:
a third convolution operation comprising:
performing convolution on the optical cable space-time data at each moment by using a convolution kernel of 3 multiplied by 3, and performing pooling by using a pooling layer of 2 multiplied by 2 to complete a first characteristic extraction process;
performing convolution on the features extracted in the first feature extraction process by adopting a convolution kernel of 3 multiplied by 3, and performing pooling by adopting a pooling layer of 2 multiplied by 2 to finish the second feature extraction process;
performing convolution on the features extracted in the second feature extraction process by adopting a convolution kernel of 3 multiplied by 3, and performing pooling by adopting a pooling layer of 2 multiplied by 2 to finish the third feature extraction process;
and (5) performing convolution on the features extracted in the third feature extraction process by adopting a 2 multiplied by 2 convolution kernel to finish the fourth feature extraction process so as to obtain a feature matrix with a third size.
The embodiment of the invention also provides a transmission optical cable real-time warning device based on the deep learning algorithm, which is characterized by comprising the following steps: the optical cable space-time data set comprises data of a plurality of optical signal optical basic parameters at N positions on an optical cable optical fiber at 1-T moment, wherein T is more than 1, and N is more than 1; the space-time feature extraction unit is used for performing space-time feature extraction processing on the optical cable space-time data at each moment in the optical cable space-time data set to obtain T N-dimensional high-level semantic feature data at 1-T moments; the prediction unit is used for taking the obtained N-dimensional high-level semantic feature data at the 1-T moment as input and obtaining a vibration data prediction value at the T +1 moment by utilizing a long-short term memory network; and the comparison alarm unit is used for comparing the obtained vibration data predicted value with a preset vibration data threshold value and sending an alarm signal when the vibration data predicted value is larger than the vibration data threshold value.
Furthermore, the space-time feature extraction unit is used for extracting the space-time features of the optical cable space-time data at each moment in the optical cable space-time data set by using a convolutional neural network to obtain T N-dimensional high-level semantic feature data at 1-T moment.
Further, the plurality of optical signal optical basic parameters include part or all of wavelength, frequency, intensity, polarization state, and phase.
The embodiment of the invention has the following beneficial effects: the embodiment of the invention constructs the optical cable time-space data set of the transmission optical cable by acquiring the data of the optical signal optical basic parameters within a section of the optical cable at a plurality of moments in a time period, then performs time-space characteristic extraction processing on the optical cable time-space characteristic set, and performs vibration data prediction at the next moment by using a long-short term memory network, thereby avoiding the defect that single-point data is easily influenced, improving the accuracy of vibration data prediction, more reliably monitoring the condition of the transmission optical cable, more accurately realizing the real-time alarm function of the transmission optical cable, and improving the effectiveness of real-time alarm, so that a worker can timely take precautionary or remedial measures to protect the transmission optical cable; the embodiment of the invention integrates the convolutional neural network and the long-short term memory network to establish the relation between the optical basic parameters of the optical signals and the vibration of the optical cable, and realizes the nonlinear expression of the space-time data of the optical cable through a plurality of convolutional kernels of the convolutional neural network, so that a deep learning algorithm can express more complicated characteristics, and the noise influence existing in the optical basic parameters of the optical signals can be greatly reduced.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a flowchart of a transmission optical cable real-time warning method based on a deep learning algorithm according to an embodiment of the present invention.
Fig. 2 is a schematic diagram of a processing procedure of performing spatio-temporal feature extraction processing by using a convolutional neural network according to an embodiment of the present invention.
FIG. 3 is a diagram illustrating vibration data prediction using a long term and short term memory network according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments refers to the accompanying drawings, which are included to illustrate specific embodiments in which the invention may be practiced.
The embodiment of the invention provides a transmission optical cable real-time alarm method and device based on a deep learning algorithm.
As shown in fig. 1, the method for real-time warning of a transmission optical cable based on a deep learning algorithm in the embodiment of the present invention includes the following steps:
and step S1, constructing an optical cable time-space data set of the transmission optical cable.
In order to establish a relationship between an optical signal optical basic parameter and optical cable vibration, the embodiment of the invention collects time series data of a plurality of optical signal optical basic parameters in a section of optical cable, namely data of a plurality of optical signal optical basic parameters in a section of optical cable at a plurality of moments in a time period, and constructs an optical cable time-space data set of a transmission optical cable for establishing the relationship with the optical cable vibration, so that the defect that single-point data is easily influenced can be avoided, and the prediction accuracy is improved. The optical basic parameters of the optical signals may include, but are not limited to, wavelength, frequency, intensity, polarization state, and phase.
In specific implementation, each position on the optical cable fiber can be regarded as a sensor, and each position can simultaneously collect optical signal optical basic parameter data, the embodiment of the invention constructs the optical cable time-space data set by collecting data of a plurality of optical signal optical basic parameters at N positions on the optical cable fiber at 1-T moment, that is to say, the optical cable time-space data in the optical cable time-space data set comprises data of a plurality of optical signal optical basic parameters at N positions on the optical cable fiber at 1-T moment, wherein T & gt 1 and N & gt 1.
Table 1 below provides an example of a cable time-space data set in which the locations on the cable fiber are represented by sensors, i.e., sensors 1-N represent locations 1-N on the cable fiber, and data for five optical signal optical basis parameters, including wavelength data, frequency data, intensity data, polarization state data, and phase data, are collected at each location. The data of the five optical signal optical basic parameters at the positions 1-N of the optical fiber of the optical cable are respectively collected at the moments 1-T, and the optical cable time-space data set shown in the table 1 can be constructed, so that the optical cable time-space data of the optical cable time-space data set in the table 1 comprise the data of the five optical signal optical basic parameters at the positions N on the optical fiber of the optical cable at the moments T.
TABLE 1
And step S2, performing space-time feature extraction processing on the cable space-time data set.
The embodiment of the invention carries out space-time feature extraction processing on the optical cable space-time data at each moment in the optical cable space-time data set to obtain multi-dimensional high-level semantic feature data (namely multi-dimensional feature vectors) at each moment, namely T N-dimensional high-level semantic feature data at 1-T moments.
As an implementation manner, in the embodiment of the present invention, a Convolutional Neural Network (CNN) is used to perform spatio-temporal feature extraction processing on the optical cable spatio-temporal data at each time in the optical cable spatio-temporal data set, so as to obtain T N-dimensional high-level semantic feature data at 1-T times.
The spatio-temporal feature extraction processing procedure specifically may include; performing three different convolution operations on the optical cable time-space data at each moment by using a convolution neural network to obtain three characteristic matrixes with different sizes; and forming N-dimensional high-level semantic feature data by tiling and splicing the obtained three feature matrixes with different sizes. Of course, the specific way of performing the spatio-temporal feature extraction processing by using the convolutional neural network is not limited to this, and other feasible ways may also be adopted to perform the spatio-temporal feature extraction processing on the cable spatio-temporal data set to obtain the high-level semantic feature data.
FIG. 2 shows an embodiment of the present invention in which a convolutional neural network is used to perform three different convolution operations on cable spatio-temporal data at each time (e.g., T time) to obtain N-dimensional high-level semantic feature data. As shown in fig. 2, the first convolution operation performed by the convolutional neural network on the cable spatio-temporal data at time T includes: performing convolution (Conv) on the optical cable space-time data at the time T by using a convolution kernel of 3 multiplied by 3, and performing pooling (MP) by using a pooling layer of 2 multiplied by 2 to complete a first characteristic extraction process; performing convolution on the features extracted in the first feature extraction process by adopting a convolution kernel of 3 multiplied by 3 to finish the second feature extraction process; and (5) performing convolution on the features extracted in the second feature extraction process by adopting a convolution kernel of 3 multiplied by 3 to finish the third feature extraction process, so as to obtain a feature matrix S1 with a first size (M1 multiplied by N1).
The second convolution operation of the convolution neural network on the optical cable space-time data at the T moment comprises the following steps: performing convolution on the optical cable space-time data at each moment by using a convolution kernel of 3 multiplied by 3, and performing pooling by using a pooling layer of 3 multiplied by 3 to complete a first characteristic extraction process; performing convolution on the features extracted in the first feature extraction process by adopting a convolution kernel of 3 multiplied by 3, and performing pooling by adopting a pooling layer of 3 multiplied by 3 to finish the second feature extraction process; and (5) performing convolution on the features extracted in the second feature extraction process by adopting a 2 multiplied by 2 convolution kernel to finish the third feature extraction process, so as to obtain a feature matrix S2 with a second size (M2 multiplied by N2).
The third convolution operation of the convolution neural network on the optical cable space-time data at the T moment comprises the following steps: performing convolution on the optical cable space-time data at each moment by using a convolution kernel of 3 multiplied by 3, and performing pooling by using a pooling layer of 2 multiplied by 2 to complete a first characteristic extraction process; performing convolution on the features extracted in the first feature extraction process by adopting a convolution kernel of 3 multiplied by 3, and performing pooling by adopting a pooling layer of 2 multiplied by 2 to finish the second feature extraction process; performing convolution on the features extracted in the second feature extraction process by adopting a convolution kernel of 3 multiplied by 3, and performing pooling by adopting a pooling layer of 2 multiplied by 2 to finish the third feature extraction process; and (5) performing convolution on the features extracted in the third feature extraction process by adopting a 2 multiplied by 2 convolution kernel to finish the fourth feature extraction process, so as to obtain a feature matrix S3 with a third size (M3 multiplied by N3).
Then, feature fusion is carried out on the obtained three feature matrixes S1, S2 and S3 with different sizes in a mode of tiling (converting the size of the feature matrixes into N multiplied by 1) and then splicing to form N-dimensional high-level semantic feature data.
Through the spatio-temporal feature extraction processing process, T N-dimensional high-level semantic feature data at 1-T moment can be obtained. The specific process of extracting the spatiotemporal features by using the convolutional neural network is not limited to this, and other convolutional operation processes may be adopted to extract the spatiotemporal features of the cable spatiotemporal data set to obtain high-level semantic feature data.
The embodiment of the invention adopts a plurality of convolution kernels to carry out convolution operation, can increase the nonlinear expression of space-time data, enables deep learning to express more complicated characteristics and effectively reduces the influence of the self noise of optical basic parameters of optical signals.
In addition, the specific manner of performing the space-time feature extraction processing on the cable space-time data set is not limited to the above, and other feasible deep learning algorithms may also be adopted to perform the space-time feature extraction processing on the cable space-time data set to obtain the high-level semantic feature data.
And step S3, predicting the vibration data by using a Long-Short Term Memory network (LSTM) to obtain a vibration data prediction value.
The N-dimensional high-level semantic feature data at the time point 1-T obtained in step S2 is input to the long-short term memory network as an input, and the vibration data prediction value at the next time point, i.e., the time point T +1, can be obtained using the long-short term memory network.
In one embodiment of the present invention, the vibration data may be represented by an amplitude value of the vibration signal of the optical fiber, and the larger the vibration data value at a certain position, the larger the amplitude of the optical fiber at the position, and vice versa. But not limited thereto.
This process is illustrated in FIG. 3, where x1,x2,…xTHigh-level semantic feature data respectively representing optical basic parameters of the optical signal at 1-T moments (the high-level semantic feature data at each moment has N dimensions, namely xiComprises an N-dimensional vector; t times in total, so i is 1, …, T), which is input as input to LSTM and trained (h in the figure)c 1,…hc TRepresenting the predicted value of vibration data at time 1-T)), based on the trained LSTM processing, the predicted value of vibration data at time T + 1, i.e., output y'T+1。
And step S4, comparing the obtained vibration data predicted value with a preset vibration data threshold value, and sending an alarm signal when the vibration data predicted value is greater than the vibration data threshold value.
Wherein the vibration data threshold is preset to indicate that the optical cable may be damaged if the vibration data exceeds the vibration data threshold. Therefore, when the predicted value of the vibration data is larger than the vibration data threshold value, an alarm signal is sent out to prompt staff that the optical cable is possibly damaged, so that measures can be taken in time to prevent damage or remedial measures can be taken.
As can be seen from the above description of the embodiments of the present invention, the embodiments of the present invention utilize a long and short term memory network (LSTM) to predict a situation at a certain future time by using data of optical basic parameters of optical signals at multiple times, and can establish a relatively comprehensive correlation between the optical basic parameters of optical signals and the vibration of the optical cable. Because the high-level semantic feature data of a certain optical signal optical basic parameter is possibly influenced by external noise, the vibration value at a single moment is larger, but the vibration value is not particularly reliable as the basis of alarming, the embodiment of the invention does not adopt the high-level semantic feature data of the optical signal optical basic parameter at the single moment to judge the vibration condition of the optical cable, but adopts the high-level semantic feature data of the optical signal optical basic parameter at a plurality of moments (1-T) to predict the vibration data at a future moment (T +1), realizes the prediction of the vibration value from the angle of a time domain, improves the accuracy of the prediction of the vibration value, can more accurately realize the real-time alarming function of the optical cable, and improves the effectiveness of real-time alarming. Meanwhile, the embodiment of the invention compares the vibration data predicted value obtained by the long-term and short-term memory network with the preset vibration data threshold value to judge whether the alarm is needed or not, thereby enhancing the operability of real-time alarm.
Before the vibration data prediction is carried out by using the long and short term memory network, the embodiment of the invention adopts the deep learning algorithm to carry out space-time feature extraction on the optical cable space-time data set, can effectively extract the features of the optical signal optical basic parameters and reduce the noise influence of the optical signal optical basic parameters, particularly adopts the convolutional neural network to carry out space-time feature extraction processing on the optical cable space-time data, and realizes the nonlinear expression of the optical cable space-time data by using a plurality of convolutional kernels, so that the deep learning algorithm can express more complex features, and can greatly reduce the noise influence of the optical signal optical basic parameters.
Correspondingly, the embodiment of the invention also provides a transmission optical cable real-time warning device based on the deep learning algorithm, and the method can be realized. The real-time warning device for the transmission optical cable comprises a construction unit, a space-time feature extraction unit, a prediction unit and a comparison warning unit. The construction unit is used for constructing an optical cable space-time data set of a transmission optical cable, and the optical cable space-time data in the optical cable space-time data set comprises data of a plurality of optical signal optical basic parameters at N positions on an optical cable optical fiber at 1-T moment, wherein T is more than 1, and N is more than 1. The plurality of optical signal optical basis parameters include some or all of wavelength, frequency, intensity, polarization, and phase. The space-time feature extraction unit is used for performing space-time feature extraction processing on the optical cable space-time data at each moment in the optical cable space-time data set to obtain T N-dimensional high-level semantic feature data at 1-T moments. And the prediction unit is used for taking the obtained N-dimensional high-level semantic feature data at the 1-T moment as input and obtaining the vibration data prediction value at the T +1 moment by using the long-short term memory network. The comparison alarm unit is used for comparing the obtained vibration data predicted value with a preset vibration data threshold value and sending an alarm signal when the vibration data predicted value is larger than the vibration data threshold value so as to prompt staff that the optical cable is possibly damaged, so that measures can be taken in time to prevent damage or remedial measures can be taken.
As an implementation mode, the space-time feature extraction unit is used for performing space-time feature extraction processing on the optical cable space-time data at each moment in the optical cable space-time data set by using a convolutional neural network to obtain T N-dimensional high-level semantic feature data at 1-T moment.
Compared with the prior art, the invention has the beneficial effects that: the optical cable time-space data set of the transmission optical cable is constructed by collecting data of optical signal optical basic parameters in a section of the optical cable at a plurality of moments in a time period, then the optical cable time-space feature set is subjected to time-space feature extraction processing, and a long-short term memory network is utilized to predict vibration data at the next moment, so that the accuracy of vibration data prediction is improved, the condition of the transmission optical cable can be more reliably monitored, the real-time alarm function of the transmission optical cable can be more accurately realized, and the effectiveness of real-time alarm is improved, so that a worker can timely take precautionary or remedial measures to protect the transmission optical cable; the invention integrates the convolution neural network and the long-short term memory network to establish the relation between the optical basic parameters of the optical signals and the vibration of the optical cable, and realizes the nonlinear expression of the space-time data of the optical cable through a plurality of convolution kernels of the convolution neural network, so that the deep learning algorithm can express more complicated characteristics, and the noise influence existing in the optical basic parameters of the optical signals can be greatly reduced.
The above disclosure is only for the purpose of illustrating the preferred embodiments of the present invention, and it is therefore to be understood that the invention is not limited by the scope of the appended claims.
Claims (10)
1. A transmission optical cable real-time warning method based on a deep learning algorithm is characterized by comprising the following steps:
step S1, constructing an optical cable space-time data set of a transmission optical cable, wherein the optical cable space-time data in the optical cable space-time data set comprises data of a plurality of optical signal optical basic parameters at N positions on an optical cable optical fiber at 1-T moment, wherein T is more than 1, and N is more than 1;
step S2, performing space-time feature extraction processing on the optical cable space-time data at each moment in the optical cable space-time data set to obtain T N-dimensional high-level semantic feature data at 1-T moments;
step S3, the obtained N-dimensional high-level semantic feature data at the time of 1-T is used as input, and a vibration data prediction value at the time of T +1 is obtained by using a long-short term memory network;
and step S4, comparing the obtained vibration data predicted value with a preset vibration data threshold value, and sending an alarm signal when the vibration data predicted value is larger than the vibration data threshold value.
2. The transmission optical cable real-time warning method of claim 1, wherein the optical basic parameters of the plurality of optical signals include some or all of wavelength, frequency, intensity, polarization state, and phase.
3. The transmission optical cable real-time warning method of claim 1, wherein the step S2 further comprises:
and performing space-time feature extraction processing on the optical cable space-time data at each moment in the optical cable space-time data set by using a convolutional neural network to obtain T N-dimensional high-level semantic feature data at 1-T moment.
4. The transmission cable real-time warning method of claim 3, wherein the step S2 further comprises:
the convolutional neural network performs three different convolution operations on the optical cable time-space data at each moment to obtain three feature matrixes with different sizes;
and forming the N-dimensional high-level semantic feature data by tiling and splicing the obtained three feature matrixes with different sizes.
5. The transmission cable real-time warning method of claim 4, wherein the three different convolution operations include:
a first convolution operation comprising:
performing convolution on the optical cable space-time data at each moment by using a convolution kernel of 3 multiplied by 3, and performing pooling by using a pooling layer of 2 multiplied by 2 to complete a first characteristic extraction process;
performing convolution on the features extracted in the first feature extraction process by adopting a convolution kernel of 3 multiplied by 3 to finish the second feature extraction process;
and (5) performing convolution on the features extracted in the second feature extraction process by adopting a convolution kernel of 3 multiplied by 3 to finish the third feature extraction process so as to obtain a feature matrix with the first size.
6. The transmission cable real-time warning method of claim 5, wherein the three different convolution operations further comprise:
a second convolution operation comprising:
performing convolution on the optical cable space-time data at each moment by using a convolution kernel of 3 multiplied by 3, and performing pooling by using a pooling layer of 3 multiplied by 3 to complete a first characteristic extraction process;
performing convolution on the features extracted in the first feature extraction process by adopting a convolution kernel of 3 multiplied by 3, and performing pooling by adopting a pooling layer of 3 multiplied by 3 to finish the second feature extraction process;
and (5) performing convolution on the features extracted in the second feature extraction process by adopting a 2 multiplied by 2 convolution kernel to finish the third feature extraction process so as to obtain a feature matrix with a second size.
7. The transmission cable real-time warning method of claim 6, wherein the three different convolution operations further comprise:
a third convolution operation comprising:
performing convolution on the optical cable space-time data at each moment by using a convolution kernel of 3 multiplied by 3, and performing pooling by using a pooling layer of 2 multiplied by 2 to complete a first characteristic extraction process;
performing convolution on the features extracted in the first feature extraction process by adopting a convolution kernel of 3 multiplied by 3, and performing pooling by adopting a pooling layer of 2 multiplied by 2 to finish the second feature extraction process;
performing convolution on the features extracted in the second feature extraction process by adopting a convolution kernel of 3 multiplied by 3, and performing pooling by adopting a pooling layer of 2 multiplied by 2 to finish the third feature extraction process;
and (5) performing convolution on the features extracted in the third feature extraction process by adopting a 2 multiplied by 2 convolution kernel to finish the fourth feature extraction process so as to obtain a feature matrix with a third size.
8. A transmission optical cable real-time warning device based on a deep learning algorithm is characterized by comprising:
the optical cable space-time data set comprises data of a plurality of optical signal optical basic parameters at N positions on an optical cable optical fiber at 1-T moment, wherein T is more than 1, and N is more than 1;
the space-time feature extraction unit is used for performing space-time feature extraction processing on the optical cable space-time data at each moment in the optical cable space-time data set to obtain T N-dimensional high-level semantic feature data at 1-T moments;
the prediction unit is used for taking the obtained N-dimensional high-level semantic feature data at the 1-T moment as input and obtaining a vibration data prediction value at the T +1 moment by utilizing a long-short term memory network;
and the comparison alarm unit is used for comparing the obtained vibration data predicted value with a preset vibration data threshold value and sending an alarm signal when the vibration data predicted value is larger than the vibration data threshold value.
9. The real-time warning device for transmission cable according to claim 8, wherein the spatiotemporal feature extraction unit performs spatiotemporal feature extraction processing on the cable spatiotemporal data at each time in the cable spatiotemporal data set by using a convolutional neural network to obtain T N-dimensional high-level semantic feature data at 1-T times.
10. The transmission cable real-time warning apparatus of claim 8, wherein the plurality of optical signal optical basic parameters include some or all of wavelength, frequency, intensity, polarization state, and phase.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111160835.XA CN113869240A (en) | 2021-09-30 | 2021-09-30 | Transmission optical cable real-time warning method and device based on deep learning algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111160835.XA CN113869240A (en) | 2021-09-30 | 2021-09-30 | Transmission optical cable real-time warning method and device based on deep learning algorithm |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113869240A true CN113869240A (en) | 2021-12-31 |
Family
ID=79001184
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111160835.XA Pending CN113869240A (en) | 2021-09-30 | 2021-09-30 | Transmission optical cable real-time warning method and device based on deep learning algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113869240A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116760466A (en) * | 2023-08-23 | 2023-09-15 | 青岛诺克通信技术有限公司 | Optical cable positioning method and system |
-
2021
- 2021-09-30 CN CN202111160835.XA patent/CN113869240A/en active Pending
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116760466A (en) * | 2023-08-23 | 2023-09-15 | 青岛诺克通信技术有限公司 | Optical cable positioning method and system |
CN116760466B (en) * | 2023-08-23 | 2023-11-28 | 青岛诺克通信技术有限公司 | Optical cable positioning method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11562224B2 (en) | 1D-CNN-based distributed optical fiber sensing signal feature learning and classification method | |
CN111596604B (en) | Intelligent fault diagnosis and self-healing control system and method for engineering equipment based on digital twinning | |
US11346200B2 (en) | Method and system for guaranteeing safety of offshore oil well control equipment | |
CN111579243B (en) | Rolling bearing intelligent diagnosis system based on deep migration learning | |
CN112487356B (en) | Structural health monitoring data enhancement method | |
CN102355299B (en) | Analysis method of seabed photoelectrical compound cable fault type | |
CN112907901B (en) | Tunnel monitoring entity risk early warning assessment model | |
CN113670434B (en) | Method and device for identifying sound abnormality of substation equipment and computer equipment | |
FR3092191A1 (en) | CONSTRUCTION OF DIGITAL TWINS FOR OIL AND GAS RECOVERY USING A SET KALMAN FILTER | |
CN113869240A (en) | Transmission optical cable real-time warning method and device based on deep learning algorithm | |
CN112392458A (en) | Method and system for monitoring vibration of cutter head and drill rod of raise boring machine | |
CN117235443A (en) | Electric power operation safety monitoring method and system based on edge AI | |
CN108805427B (en) | Power distribution network running state risk early warning system based on big data | |
CN106774054A (en) | GIS device analysis system and method based on the identification of complicated unstructured data | |
CN116523722A (en) | Environment monitoring analysis system with machine learning capability | |
CN115499318B (en) | Tunnel monitoring data communication management and control method, system and terminal equipment | |
CN117423201A (en) | Intelligent fire-fighting state monitoring method and system for restaurant | |
CN114354744B (en) | Power cable external damage prevention monitoring method, system and device | |
CN116610996A (en) | Cable state detection method and device and computer equipment | |
CN110763955A (en) | Cable fault detection system and method based on high-voltage pulse signal injection | |
CN110632643A (en) | Detection alarm method for preventing third-party construction excavation | |
CN113486595B (en) | Well blowout intelligent early warning method, system, equipment and storage medium | |
CN115659243B (en) | Infrastructure risk monitoring method and monitoring system based on MEMS | |
CN117452128B (en) | Submarine cable health monitoring and early warning method and system | |
CN117744010B (en) | Small data driven real-time positioning method for pressure abnormality of coal mine support |
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 |