Monitoring method and system based on multi-source optical cable data fusion

Abstract

The invention relates to the field of optical fiber sensing, in particular to a monitoring method and system based on multi-source optical cable data fusion. The method comprises the following steps: acquiring a plurality of monitoring data of the optical cable, the plurality of monitoring data being respectively acquired according to a unified acquisition time period through different monitoring devices; aligning the plurality of types of monitoring data in space and time, and acquiring a plurality of types of aligned monitoring data; preprocessing the plurality of aligned monitoring data to obtain an input matrix; key features and standard parameters are preset and input into the deep learning prediction model; inputting the input matrix into the deep learning prediction model for fusion to obtain fusion data, and analyzing the fusion data according to the key features and the standard parameters to obtain a monitoring result; according to the method, alignment and fusion of various optical cable data can be realized, and the precision and accuracy of optical cable monitoring are enhanced.

Description

Technical Field

[0001] The present invention relates to the field of optical fiber sensing, and more specifically, to a monitoring method and system based on multi-source optical cable data fusion. Background Art

[0002] In the field of fiber optic sensing today, a variety of different monitoring technologies are usually used to achieve multi-dimensional monitoring of optical cables, such as using optical time domain reflectometry (OTDR) technology to obtain the location of optical cable breakpoints and attenuation anomalies; using distributed vibration sensing (DVS) technology to monitor changes in the external environment or disturbances along the optical cable, such as earthquakes and construction intrusions; using distributed temperature sensing (DTS) technology to monitor cable overheating or ambient temperature changes. These technologies are widely used in optical cable monitoring, but they usually operate independently and lack an effective coordination mechanism, resulting in the limitation of information fragmentation in practical applications. Moreover, the data formats, sampling frequencies, and physical meanings collected by different technologies are different, and they cannot be effectively integrated; the time and space of the data sources are inconsistent, making it difficult to form comprehensive analysis and decision support; the independent use of the above technologies has a weak response capability to rapidly changing events, affecting the real-time monitoring effect. Therefore, it is necessary to make certain improvements to the display modes of different monitoring technologies. Summary of the invention

[0003] The present invention aims to overcome at least one defect (shortcoming) of the above-mentioned prior art and provide a monitoring method and system based on multi-source optical cable data fusion to achieve alignment and fusion of multiple optical cable data and enhance the precision and accuracy of optical cable monitoring.

[0004] According to a first aspect of the present application, a monitoring method based on multi-source optical cable data fusion is provided, the method comprising:

[0005] Acquire several types of monitoring data of the optical cable, wherein the several types of monitoring data are collected by different monitoring devices according to a unified collection time period;

[0006] Aligning the several types of monitoring data in space and time to obtain the aligned several types of monitoring data;

[0007] Preprocessing the aligned monitoring data to obtain input matrices respectively;

[0008] Presetting key features and standard parameters and entering them into the deep learning prediction model;

[0009] The input matrix is ​​input into the deep learning prediction model for fusion to obtain fused data, and the fused data is analyzed according to the key features and the standard parameters to obtain monitoring results.

[0010] It is understandable that by collecting multiple monitoring data, aligning the collected multiple monitoring data in space and time, and then pre-processing and inputting them into the deep learning prediction model for fusion; by monitoring the optical cable through fusion data, it is possible to achieve multi-dimensional fusion of multiple monitoring data, improve the accuracy of monitoring work, avoid subjective judgments and errors in manual analysis, and improve the overall analysis ability of monitoring work, thereby improving the efficiency of monitoring.

[0011] Optionally, the several types of monitoring data specifically include two or more of OTDR monitoring data, DVS monitoring data, temperature monitoring data, water level or gas monitoring data; wherein the OTDR monitoring data is acquired through an optical time domain reflectometer, the DVS monitoring data is acquired through a distributed optical fiber vibration sensing device, the temperature monitoring data is acquired through a distributed optical fiber temperature sensing device, and the water level or gas monitoring data is acquired through a water level sensing device or a gas sensing device.

[0012] It is understandable that because different monitoring data are collected using different monitoring equipment, and different monitoring equipment is based on different monitoring technologies, the collected data will also be different in format and focus. In general monitoring work, using only one monitoring equipment to monitor the optical cable cannot meet the needs of accurate monitoring of the optical cable, so multiple monitoring data are needed as monitoring materials to comprehensively analyze the overall operation of the optical cable.

[0013] Optionally, the unified acquisition time period is obtained in the following manner:

[0014] Acquire two or more of the original acquisition period of the OTDR monitoring data, the original acquisition period of the DVS monitoring data, the original acquisition period of the temperature monitoring data, and the original acquisition period of the water level or gas monitoring data as the initial acquisition period;

[0015] The greatest common divisor of all the initial acquisition periods is taken as the unified acquisition time period.

[0016] It is understandable that because different monitoring devices collect monitoring data at different frequencies, the collection cycles are different. If multiple types of monitoring data need to be monitored horizontally, the collection cycles of different types of monitoring data need to be aligned. The value-added or average operation using mathematical methods makes the collection cycle of each monitoring device a unified collection cycle, which can ensure that the monitoring devices can send monitoring sources at the same time, realize the alignment of the collection cycles, and facilitate the horizontal comparison of different monitoring data.

[0017] Optionally, the aligning the several types of monitoring data spatially and temporally to obtain the aligned several types of monitoring data specifically includes:

[0018] Each type of monitoring data includes the location information and the collection time of each collection point on the optical cable;

[0019] Obtain the location information l of each collection point of each monitoring data, and map the location information l to the standard distance L according to the following formula: common :

[0020]

[0021] Among them, the ΔL common is the standard distance resolution, l is the set of location information of each collection point of each monitoring data, and round(.) is a rounding function;

[0022] Obtain the collection time t of each collection point for each type of monitoring data, and map the collection time t to the standard time T according to the following formula:

[0023]

[0024] Among them, t is the collection time of each collection point of each monitoring data, T common To unify the collection time period.

[0025] It is understandable that for each collection point in different monitoring devices, the returned location distance and time will exist as floating-point numbers, which adds a lot of calculation work to mapping on the map and subsequent data processing. The above-mentioned function using rounding converts the location information and the numerical value of the collection time into positive integers, provides accurate monitoring data for the three-dimensional coordinates mapped on the map, reduces unnecessary rounding errors in subsequent calculations, and improves data processing efficiency.

[0026] Optionally, the method further comprises:

[0027] The standard distance L according to the several monitoring data common , the geographic coordinates (x, y, z) of the collection point in the map are converted and mapped using the following formula:

[0028] (x,y,z)=CIS_map(L common )

[0029] Wherein x is the longitude of the geographic coordinate, y is the latitude of the geographic coordinate, z is the height of the optical cable from the ground, and GIS_map(.) is a coordinate conversion function;

[0030] Placing the aligned monitoring data in the same coordinate system;

[0031] According to the geographic coordinates, the several types of monitoring data in the same coordinate system are displayed correspondingly on the map.

[0032] It is understandable that different types of monitoring data that have been aligned in time and space can be displayed on the same coordinate system. By using the geographic coordinate conversion function, the monitoring data can be displayed directly on the map, thereby enhancing the intuitiveness of monitoring and making it convenient for operation and maintenance personnel to conduct operation and maintenance monitoring directly on the map at the corresponding collection points one by one, thereby improving the efficiency of monitoring.

[0033] Optionally, the key features specifically include:

[0034] The key feature of the OTDR monitoring data is the slope of change of the signal reflection intensity;

[0035] The key feature of the DVS monitoring data is the vibration spectrum energy distribution;

[0036] The key feature of the temperature monitoring data is the temperature value;

[0037] The key feature of the water level or gas monitoring data is the water level height value or the gas concentration value;

[0038] And / or, the standard parameters include a slope threshold corresponding to the slope of the change of the signal reflection intensity, an energy threshold corresponding to the vibration spectrum energy distribution, a standard temperature range corresponding to the temperature value, and a standard water level height range or gas concentration range corresponding to the water level height value or gas concentration value.

[0039] It is understandable that the monitoring data collected by different monitoring devices include a large amount of basic and important information. Locating the focus of different types of monitoring data helps to more quickly identify whether the optical cable environment is normal. These focuses are the selected key features, which provide a clear direction for the processing of the deep learning prediction model and improve the efficiency of the deep learning prediction model processing. At the same time, setting standard parameters can provide the deep learning prediction model with a benchmark for the normal environment of the optical cable, accelerate the deep learning prediction model's learning of the normal environment of the optical cable, and accurately identify the abnormal environment.

[0040] Optionally, the inputting the input matrix into the deep learning prediction model for fusion to obtain fusion data, and analyzing the fusion data according to the key features and the standard parameters to obtain the monitoring results specifically includes:

[0041] Inputting the input matrix into the deep learning prediction model, extracting key features in the input matrix using the deep learning prediction model, and generating a feature matrix;

[0042] Weighting each key feature in the feature matrix to obtain a feature matrix with weights;

[0043] Performing weighted fusion on the feature matrix with weights to obtain fused data;

[0044] The deep learning prediction model monitors the key features in the fused data according to the standard parameters and records a number of abnormal points, wherein the abnormal points correspond one to one with the weights of the corresponding key features;

[0045] Acquire the correlation information of the plurality of abnormal points according to the fused data;

[0046] All the abnormal points are analyzed according to the weights corresponding to the abnormal points and the associated information to obtain the overall monitoring result of the optical cable.

[0047] It can be understood that the input matrix is ​​input into the deep learning prediction model, and the key features are obtained from the input matrix and the weighting steps are performed, and finally the different types of monitoring data are fused to obtain fused data to obtain the monitoring results; this method can extract the correlation between the different types of monitoring data in the optical cable using the deep learning prediction model, more accurately capture the abnormal points of the optical cable and the correlation between various abnormal points, reduce complicated manual comparisons and errors, and improve the efficiency and intelligence of monitoring work.

[0048] Optionally, the extracting key features in the input matrix using a deep learning prediction model specifically includes:

[0049] Acquire the signal reflection intensity of each acquisition point according to the OTDR monitoring data in the input matrix, and extract the change slope of the signal reflection intensity of the two acquisition points as the key feature corresponding to the OTDR monitoring data;

[0050] Acquire the vibration spectrum amplitude at the corresponding acquisition time according to the DVS monitoring data in the input matrix, and extract the vibration spectrum energy distribution at the corresponding acquisition time as the key feature corresponding to the DVS monitoring data;

[0051] Acquire the temperature value of each collection point according to the temperature monitoring data in the input matrix, and use the temperature value of each collection point as the key feature corresponding to the temperature monitoring data;

[0052] The water level height value or gas concentration value of each collection point is obtained according to the water level or gas monitoring data of the input matrix, and the water level height value or gas concentration value of each collection point is used as the key feature corresponding to the water level or gas monitoring data.

[0053] It is understandable that the deep learning prediction model can obtain different key features of different monitoring data based on the input matrix and obtain the data that needs to be monitored. This helps the deep learning prediction model to monitor different monitoring points of different monitoring data more quickly and improve monitoring accuracy and efficiency.

[0054] Optionally, the abnormal point is obtained in the following manner:

[0055] If the signal reflection intensity change slope of two acquisition points in the OTDR monitoring data is greater than a preset slope threshold, it is determined that an abnormal reflection intensity point appears in the optical cable between the two acquisition points; and / or,

[0056] If the vibration spectrum energy distribution in the DVS monitoring data is greater than a preset energy threshold, it is determined that the optical cable has a vibration abnormality point; and / or,

[0057] If the temperature value in the temperature monitoring data is not within the preset standard temperature range, the collection point is determined to be a temperature abnormality point; and / or,

[0058] If the water level height value in the water level monitoring data is not within the preset standard water level height range, the collection point is determined to be a water level abnormal point; or, if the gas concentration value in the gas monitoring data is not within the preset gas concentration range, the collection point is determined to be a gas abnormal point.

[0059] It is understandable that by using pre-set standard parameters as auxiliary parameters of the convolutional neural network and monitoring the key features, a benchmark indicator can be provided for the convolutional neural network. If the key features do not meet the standard parameters, it is necessary to focus on monitoring other monitoring data of the collection point to obtain the abnormal phenomena of the collection point, so as to locate the abnormal point of the optical cable.

[0060] According to a second aspect of the present application, a monitoring system based on multi-source optical cable data fusion is provided, the system comprising:

[0061] A collection module, used to obtain several types of monitoring data of the optical cable, wherein the several types of monitoring data are collected through different monitoring devices according to a unified collection time period;

[0062] An alignment module, used to align the several types of monitoring data in space and time to obtain the aligned several types of monitoring data;

[0063] A preprocessing module, used for preprocessing the aligned monitoring data to obtain input matrices;

[0064] A pre-entry module, used to pre-set key features and standard parameters and enter them into the deep learning prediction model;

[0065] The monitoring module is used to input the input matrix into the deep learning prediction model for fusion to obtain fusion data, and analyze the fusion data according to the key features and the standard parameters to obtain monitoring results.

[0066] Based on any of the above aspects, the embodiment of the present application provides a monitoring method and system based on multi-source optical cable data fusion, which aligns and preprocesses different types of monitoring data collected by different monitoring devices, and then inputs them into a deep learning prediction model for fusion to obtain fused data, and analyzes and monitors the fused data to identify abnormal conditions of the optical cable operating environment. This method can fuse different types of monitoring data to achieve horizontal comparison, and use a deep learning prediction model to extract key features and correlation points, which can more accurately capture abnormal conditions of the optical cable operating environment, improve the efficiency of monitoring work, and enhance the precision and accuracy of optical cable monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0068] Figure 1 A flow chart of a monitoring method based on multi-source optical cable data fusion provided in this embodiment.

[0069] Figure 2 A flow chart of a method for adjusting the collection period of different monitoring devices provided in this embodiment.

[0070] Figure 3 A flow chart of a method for obtaining monitoring results provided in this embodiment.

[0071] Figure 4 A diagram of a situation of extracting key features provided in this embodiment.

[0072] Figure 5 This is a diagram of a situation of determining an abnormal point provided in this embodiment.

[0073] Figure 6 A flow chart of a method for mapping monitoring data to a map provided in this embodiment.

[0074] Figure 7 This is a module diagram of a monitoring system based on multi-source optical cable data fusion provided in this embodiment.

[0075] Figure 8 This is a device structure diagram of an electronic device provided in this embodiment. DETAILED DESCRIPTION

[0076] The drawings of this application are only used for illustrative purposes and should not be construed as limiting the present application. In order to better illustrate the following embodiments, some parts of the drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product; it is understandable to those skilled in the art that some well-known structures and their descriptions in the drawings may be omitted.

[0077] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0078] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0079] In the field of optical cable monitoring today, different types of monitoring data are collected using monitoring equipment based on different technologies according to different monitoring needs, so as to monitor different aspects of optical cables. However, the collection frequency, sampling period and physical meaning of different types of monitoring data are different. In the scenario where the overall operating environment of the optical cable needs to be analyzed, the monitoring data obtained using a single technology cannot achieve the effect of comprehensive monitoring; the simple integration and comparison of multiple monitoring data has a weak response capability to rapidly changing events in the optical cable, which affects the effect of real-time monitoring and is less intuitive; manual comparison and manual monitoring are complicated and prone to errors, and it is difficult to identify correlation effects and form comprehensive analysis and decision support. Therefore, it is necessary to improve the method of comprehensive monitoring.

[0080] This embodiment provides a technical solution that can solve the above-mentioned problem. The specific implementation methods of this application are described in detail below in conjunction with the accompanying drawings.

[0081] In this embodiment, if Figure 1, provides a flow chart of a monitoring method based on multi-source optical cable data fusion, which can be specifically divided into the following steps:

[0082] S110, obtaining several types of monitoring data of the optical cable, wherein the several types of monitoring data are collected by different monitoring devices respectively according to a unified collection time period;

[0083] In this embodiment, it is necessary to first collect several types of real-time monitoring data to provide an original data source for the method of this application, which can ensure the subsequent multi-dimensional monitoring of the environmental status of the operation of the optical cable, and obtain the operating details and basic information of the optical cable environment in all directions, so as to ensure the accurate capture of abnormal points of the optical cable.

[0084] Specifically, the several types of real-time monitoring data specifically include two or more of OTDR monitoring data, DVS monitoring data, temperature monitoring data, water level or gas monitoring data; wherein the OTDR monitoring data is acquired through an optical time domain reflectometer, the DVS monitoring data is acquired through a distributed optical fiber vibration sensing device, the temperature monitoring data is acquired through a distributed optical fiber temperature sensing device, and the water level or gas monitoring data is acquired through a water level sensing device or a gas sensing device.

[0085] It is understandable that the OTDR (Optical Time Domain Reflectometer) is a precise optoelectronic integrated instrument made by using the backscattering produced by Rayleigh scattering and Fresnel reflection when light is transmitted in the optical fiber. It is widely used in the maintenance and construction of optical cable lines, and can measure the optical fiber length, optical fiber transmission attenuation, joint attenuation and fault location. The DVS (Distributed Optical Fiber Vibration Sensing System) distributed optical fiber vibration sensor is an optical instrument that uses optical fiber as a sensor for vibration perception. The system uses a single optical fiber to simultaneously realize vibration monitoring and signal transmission. The DTS (Distributed Optical Fiber Temperature Sensing) distributed optical fiber temperature measurement system realizes temperature monitoring based on the principle of optical time domain reflectometry (OTDR) and the sensitivity of Raman scattering effect to temperature.

[0086] It is understandable that the OTDR monitoring data of the optical cable is collected by a monitoring device based on the optical time domain reflectometry (OTDR) technology, which is used for optical cable breakpoint detection and attenuation anomaly location by analyzing the signal of the light scattering echo;

[0087] The DVS monitoring data of the optical cable is collected by a monitoring sensor device based on the distributed optical fiber vibration sensing DVS technology. The sensing device monitors the external environment changes of the optical cable or the disturbances along the optical cable, such as earthquakes, construction intrusions, etc., by extracting vibration signals;

[0088] The temperature monitoring data of the optical cable is collected by a monitoring sensor device based on distributed optical fiber temperature sensing DTS technology, which uses Raman scattering in the optical fiber to achieve distributed measurement of temperature along the optical cable, and is used to monitor cable overheating or ambient temperature changes;

[0089] The water level and / or gas monitoring data of the optical cable are collected by a liquid sensing device or a gas concentration sensing device, and the water level and gas data are encoded into a specific vibration signal by an optical sensing communication terminal and transmitted on the optical cable in the form of a vibration frequency, so as to monitor whether the liquid water level or air concentration of the underwater optical cable or the exposed optical cable is normal;

[0090] In this embodiment, several types of the above-mentioned monitoring data can be selected for collection and use, and there is no limitation in this application. However, the more types of monitoring data are selected as the original data source, the more aspects of information on the optical cable can be used to monitor, which can improve the monitoring accuracy.

[0091] Specifically, Figure 2 As shown, the method for obtaining the unified acquisition time period is specifically divided into the following steps:

[0092] S111, obtaining two or more of the original collection period of the OTDR monitoring data, the original collection period of the DVS monitoring data, the original collection period of the temperature monitoring data, and the original collection period of the water level or gas monitoring data as the initial collection period;

[0093] In this embodiment, monitoring data is collected by monitoring equipment according to a collection cycle. However, due to different attributes of the collected data, such as OTDR monitoring data measures long-term periodic changes through light scattering echoes, while DVS monitoring data requires high-frequency and short-time time interval collection to respond to small external vibrations, it is necessary to unify the collection cycles of different monitoring data to facilitate horizontal comparison of different monitoring data.

[0094] S112: Taking the greatest common divisor of all the initial collection periods as the unified collection time period.

[0095] Preferably, in this embodiment, four types of monitoring data are selected, including OTDR monitoring data, DVS monitoring data, temperature monitoring data, and water level or gas monitoring data; then the unified acquisition time period T common It can be obtained through the following formula:

[0096] T common =gcd(T OTDR ,T vib ,T temp ,T sensor )

[0097] Among them, gcd(.) is the greatest common divisor function, T OTDR is the original acquisition period of OTDR monitoring data, T vib is the original acquisition period of DVS monitoring data, T temp is the original acquisition period of temperature monitoring data, T sensor It is the original collection period of water level or gas monitoring data;

[0098] In this embodiment, the collection periods of different monitoring data are unified using the greatest common divisor processing method, which can ensure that the monitoring data using a relatively short collection period still has a monitoring effect, and make the monitoring data using a relatively long collection period compatible with a relatively short collection period, while ensuring the monitoring effect, adding multiple collection times to achieve the unification of the collection period.

[0099] Specifically, for OTDR monitoring data, its sampling frequency is low, and it is usually used to monitor the loss or strain change of optical cables. The sampling period is long, roughly in the range of several seconds to tens of seconds. In this embodiment, a value-added mathematical method is used to perform linear interpolation or spline interpolation on the OTDR monitoring data. For example, assuming that the set T common is 0.01s; the OTDR monitoring data sampling time is: [0,10,20] seconds, the loss value is: [0.2,0.3,0.4], the time series after interpolation is: [0,0.01,0.02,...,10] seconds, and the loss value is calculated by linear interpolation to generate smooth trend data;

[0100] For DVS monitoring data, the sampling frequency is high, usually 2kHz or higher, to capture fast dynamic changes, so the high-frequency characteristics of the original signal should be maintained as much as possible. For example:

[0101] When T common =T vib When , no additional processing is required, and direct alignment is sufficient;

[0102] When T common <T vib (The unified acquisition cycle is lower than the original acquisition cycle), and the DVS monitoring data is downsampled. For example: if the DVS monitoring data is 1kHz and the unified acquisition cycle is 0.01 seconds (100Hz), it is necessary to take one point every 10 points as the adjusted DVS monitoring data.

[0103] When T common >T vib When (the unified acquisition cycle is higher than the original acquisition cycle), it is necessary to use the interpolation method to improve the data resolution and increase the number of acquisitions;

[0104] For temperature monitoring data, because the temperature changes are usually slow and the sampling period is low (such as 1 Hz or lower), the average method is usually selected to unify the acquisition period; illustratively, the low-frequency temperature data is divided into intervals according to the unified acquisition time period. For example, T is set common If the temperature is 0.01 seconds, the temperature value at 1 second is evenly distributed to all time points within [0,1] seconds; each time point is assigned the same temperature value;

[0105] For water level and / or gas monitoring data, they usually change slowly and have a low sampling frequency, so the unified collection time period is similar to that of temperature monitoring data and will not be repeated here.

[0106] S120, aligning the plurality of monitoring data in space and time to obtain the aligned plurality of monitoring data;

[0107] It is understandable that the position distance of each collection point also needs to be aligned to facilitate mapping of multiple different types of monitoring data onto the same spatial reference system.

[0108] In this embodiment, for OTDR monitoring data, the location information of each acquisition point from the monitoring light source is l OTDR By measuring the time difference T1 between the emission signal of the monitoring light source and the received reflected signal, we can obtain:

[0109]

[0110] Where c is the speed of light and n is the refractive index in the optical fiber;

[0111] For DVS monitoring data, the location information of each collection point from the DVS monitoring sensor device is l vib The principle of acquiring data is similar to that of OTDR monitoring, so I will not repeat it here:

[0112]

[0113] Wherein, T2 is the time difference between the transmission signal of the DVS monitoring sensor device and the received reflected signal;

[0114] For temperature monitoring data, temperature sensing uses the intensity ratio of Stokes light and anti-Stokes light in Raman scattering to achieve temperature demodulation; it can also be obtained through the same method as OTDR monitoring data, and the temperature sampling position information l is determined based on the time difference T3 between signal sending and signal returning.temp , the calculation formula is:

[0115]

[0116] For water level / gas monitoring data, the water level and gas monitoring data are encoded into specific vibration signals through the optical sensor communication terminal and transmitted on the optical cable in the form of vibration frequency. Since the modulated signal is transmitted in the form of vibration frequency, the water level / gas sampling location information can also be determined by the same acquisition method as the OTDR monitoring data based on the time difference T4 between signal sending and signal return. sensor , the calculation formula is:

[0117]

[0118] Specifically, the spatially and temporally aligning the plurality of monitoring data to obtain the aligned plurality of monitoring data specifically includes:

[0119] Each type of monitoring data includes the location information and the collection time of each collection point on the optical cable;

[0120] Obtain the location distance l of each data point of each real-time monitoring data, and map the location distance l to the standard distance L according to the following formula: common :

[0121]

[0122] Among them, the ΔL common is the standard distance resolution, l∈(l OTDR ,l vib ,l temp ,l sensor ) is the set of location information of each collection point of each monitoring data, and round(.) is a rounding function;

[0123] Preferably, in this embodiment, four types of monitoring data are selected, including OTDR monitoring data, DVS monitoring data, temperature monitoring data, and water level or gas monitoring data; then l∈(l OTDR ,l vib ,l temp ,l sensor );

[0124] Obtain the collection time t of each collection point for each type of monitoring data, and map the collection time t to the standard time T according to the following formula:

[0125]

[0126] Among them, t is the collection time of each collection point of each monitoring data, Tcommon To unify the collection time period.

[0127] Preferably, in this embodiment, four types of monitoring data are selected, including OTDR monitoring data, DVS monitoring data, temperature monitoring data, and water level or gas monitoring data; then t∈(t OTDR ,t vib ,t temp ,t sensor ), where t OTDR is the acquisition time of each acquisition point of OTDR monitoring data, t vib is the collection time of each collection point of DVS monitoring data, t temp is the collection time of each collection point of temperature monitoring data, t sensor The collection time of each collection point for water level or gas monitoring data;

[0128] In this embodiment, the location information and collection time obtained by different types of monitoring data may contain floating-point numbers, but the data with floating-point numbers brings great computing pressure in the conversion of geographic coordinates and the processing of deep learning prediction models, and in a large amount of computing work, rounding errors will be invisibly magnified in the accumulated calculations, so it is necessary to convert the floating-point numbers of the original data source into integers, which reduces the computing pressure and reduces the errors to a certain extent. In this embodiment, the floating-point numbers are rounded using the rounding method, which is simple to operate, requires less calculation, and improves computing efficiency; and the standard distance resolution is quoted in terms of distance, in order to ensure that the distance resolution of different monitoring data is unified and facilitate subsequent cross-fusion.

[0129] S130, preprocessing the aligned monitoring data to obtain input matrices;

[0130] In this embodiment, the different types of monitoring data contain a large amount of basic information. It is necessary to select key monitoring data from the large amount of basic information to improve the monitoring efficiency and save computing resources.

[0131] In this embodiment, several types of monitoring data are preprocessed, including but not limited to performing conventional operations such as noise processing and normalization on the several types of monitoring data, which will not be described in detail here.

[0132] S140, presetting key features and standard parameters and entering them into the deep learning prediction model;

[0133] In this embodiment, key features and standard parameters are set and entered into the deep learning prediction model, which can provide the model with an accurate monitoring direction and monitoring focus, and enable the model to accurately learn the parameters of the optical cable under normal environment, thereby more accurately identifying abnormal points in the optical cable environment.

[0134] Specifically, the key features include:

[0135] The key feature of the OTDR monitoring data is the slope of change of the signal reflection intensity;

[0136] The key feature of the DVS monitoring data is the vibration spectrum energy distribution;

[0137] The key feature of the temperature monitoring data is the temperature value;

[0138] The key feature of the water level or gas monitoring data is the water level height value or the gas concentration value;

[0139] And / or, the standard parameters include a slope threshold corresponding to the slope of the change of the signal reflection intensity, an energy threshold corresponding to the vibration spectrum energy distribution, a standard temperature range corresponding to the temperature value, and a standard water level height range or gas concentration range corresponding to the water level height value or gas concentration value.

[0140] In this embodiment, the slope of change of signal reflection intensity, vibration spectrum energy distribution, temperature value, water level height value or gas concentration value are set as key features, and the key features are used as the key monitoring objects to provide a monitoring direction for the deep learning prediction model; in this embodiment, it is necessary to set standard parameters for the convolutional neural network CNN in advance; it is understandable that the standard parameters are relevant normal parameters recorded during the normal operation of the historical optical cable, including relevant thresholds and data ranges, which can guide the convolutional neural network CNN to learn the standard parameters for the normal operation of the optical cable and improve the sensitivity to abnormal data.

[0141] Preferably, in this embodiment, the data related to the key features also need to be dimensionalized, specifically including:

[0142] Normalize the reflection intensity r in the OTDR monitoring data to the [0,1] interval:

[0143]

[0144] Among them, r is the reflection intensity of every two collection points, r min is the minimum reflection intensity of the pre-recorded historical data, r max It is the maximum value of reflection intensity of pre-recorded historical data;

[0145] The vibration spectrum amplitude F(f) in the DVS monitoring data is standardized to a preset fixed range:

[0146]

[0147] Among them, F(f) minis the minimum amplitude of the pre-recorded vibration spectrum, F(f) max is the maximum value of the pre-recorded vibration spectrum amplitude; illustratively, [F(f) min ,F(f) max ]=[0,1];

[0148] The temperature value c of the temperature monitoring data is standardized to a preset fixed range:

[0149]

[0150] Among them, c min is the minimum pre-recorded temperature value, c max is the maximum value of the pre-recorded temperature value; illustratively, [c min ,c max ] = [-20,80];

[0151] The water level height value h in the water level monitoring data is standardized to a preset fixed range:

[0152]

[0153] Among them, h min is the minimum pre-recorded water level height value, h max The maximum value of the pre-recorded water level height;

[0154] Normalize the gas concentration value p in the gas monitoring data to a preset fixed range:

[0155]

[0156] Among them, p min is the minimum pre-recorded gas concentration value, p max It is the maximum value of the pre-recorded gas concentration.

[0157] S150, inputting the input matrix into the deep learning prediction model for fusion to obtain fused data, and analyzing the fused data according to the key features and the standard parameters to obtain monitoring results.

[0158] In this embodiment, the deep learning prediction model is used to integrate different types of monitoring data, directly fuse various types of monitoring data at the data layer to form an original data matrix, and the deep learning prediction model is used to extract key features. The convolutional neural network is used to extract related information from key features and aggregate similar information. After data and feature processing, it is determined whether the optical cable status is abnormal according to the set standard parameters.

[0159] Preferably, in this embodiment, the deep learning prediction model can be a convolutional neural network CNN.

[0160] Specifically, Figure 3 As shown, the input matrix is ​​input into the deep learning prediction model for fusion to obtain fusion data, and the fusion data is analyzed according to the standard parameters to obtain the monitoring results, which specifically include:

[0161] S151, inputting the input matrix into the deep learning prediction model, using the deep learning prediction model to extract key features in the input matrix, and generating a feature matrix;

[0162] In this embodiment, it is necessary to extract key features from the input matrix as the focus of monitoring based on pre-set key features, so as to provide a monitoring direction for the deep learning prediction model;

[0163] Specifically, Figure 4 As shown, the use of the deep learning prediction model to extract key features in the input matrix specifically includes the following situations:

[0164] Acquire the signal reflection intensity of each acquisition point according to the OTDR monitoring data in the input matrix, and extract the change slope of the signal reflection intensity of the two acquisition points as the key feature corresponding to the OTDR monitoring data;

[0165] In this embodiment, the calculation formula of the signal reflection intensity change slope ΔR is:

[0166]

[0167] Among them, r i represents the signal reflection intensity of the i-th acquisition point, r i+1 represents the signal reflection intensity of the i+1th acquisition point, l i is the location information of the i-th collection point, l i+1 is the location information of the i+1th collection point;

[0168] And / or, obtaining the vibration spectrum amplitude F(f) at the corresponding acquisition time according to the DVS monitoring data in the input matrix, and extracting the vibration spectrum energy distribution at the corresponding acquisition time as the key feature corresponding to the DVS monitoring data;

[0169] In this embodiment, the calculation formula of the vibration spectrum energy distribution E(f) is:

[0170]

[0171] Among them, f1 and f2 are the minimum and maximum values ​​of the vibration spectrum amplitude at the acquisition moment;

[0172] and / or, obtaining a temperature value of each collection point according to the temperature monitoring data in the input matrix, and using the temperature value of each collection point as a key feature corresponding to the temperature monitoring data;

[0173] And / or, the water level height value or gas concentration value of each collection point is obtained according to the water level or gas monitoring data of the input matrix, and the water level height value or gas concentration value of each collection point is used as the key feature corresponding to the water level or gas monitoring data.

[0174] Preferably, if there is a type of monitoring data with a higher dimension than other types of monitoring data, it needs to be reduced in dimension; a dimensionality reduction algorithm such as principal component analysis (PCA) or linear discriminant analysis (LDA) is used to simplify the feature space;

[0175] S152, weighting each key feature in the feature matrix to obtain a feature matrix with weights;

[0176] In this embodiment, it is necessary to set the weight wi of each key feature, and the weight can be set according to importance or obtained through training and learning;

[0177] S153, performing weighted fusion on the feature matrix with weights to obtain fused data;

[0178] In this embodiment, the deep learning prediction model is processed using a convolutional neural network (CNN). The convolution process may include: a first layer of convolution may be used to extract a feature pattern of a single data source of an input matrix, such as a frequency characteristic of a vibration signal; a second layer of convolution may be used to extract a cross-source feature association of the input matrix, such as vibration enhancement caused by temperature increase;

[0179] Then use the ReLU activation function to increase the nonlinear expression ability of the model:

[0180] f(x)=max(0,x)

[0181] In the pooling layer, use maximum pooling or average pooling to reduce the feature dimension and retain the main information;

[0182] Finally, the fused data is obtained;

[0183] S154, the deep learning prediction model monitors the key features in the fused data according to the standard parameters, and records a number of abnormal points, wherein the abnormal points correspond one to one with the weights of the corresponding key features;

[0184] In this embodiment, the key features of each monitoring data are analyzed according to the standard parameters to obtain abnormal points. It can be understood that the abnormal points obtained in this step exist independently and have no connection with other abnormal points;

[0185] Specifically, Figure 5As shown, the acquisition method of the abnormal point can be divided into the following cases:

[0186] If the signal reflection intensity change slope of two acquisition points in the OTDR monitoring data is greater than a preset slope threshold, it is determined that an abnormal reflection intensity point appears in the optical cable between the two acquisition points; and / or,

[0187] If the vibration spectrum energy distribution in the DVS monitoring data is greater than a preset energy threshold, it is determined that the optical cable has a vibration abnormality point; and / or,

[0188] If the temperature value in the temperature monitoring data is not within the preset standard temperature range, the collection point is determined to be a temperature abnormality point; and / or,

[0189] If the water level height value in the water level monitoring data is not within the preset standard water level height range, the collection point is determined to be a water level abnormality point; or, if the gas concentration value in the gas monitoring data is not within the preset gas concentration range, the collection point is determined to be a gas abnormality point.

[0190] In this embodiment, according to the above-mentioned steps, several abnormal points can be obtained, and the output result of the deep learning prediction model can reflect the situation of the abnormal points; preferably, the output result can be in a binary classification form such as normal / abnormal; it can be in a multi-classification form, including different abnormal levels; it can also be in a probability form, indicating the degree of abnormality.

[0191] S155, obtaining association information of the plurality of abnormal points according to the fused data;

[0192] In this embodiment, the correlation information obtains the optical cable information of multiple abnormal points according to the correlation in the fused data through the deep learning prediction model, which can provide the inferred cause of the abnormality of the abnormal point and provide the operation and maintenance personnel with an inference idea combining different monitoring data and directions, so that the root cause of the abnormality can be found more quickly.

[0193] S156: Analyze all the abnormal points according to the weights corresponding to the abnormal points and the associated information to obtain an overall monitoring result of the optical cable.

[0194] In this embodiment, a deep learning prediction model is used to analyze the output results and related information of the abnormal points, obtain the environmental status of the overall operation of the optical cable, and realize all-round operation and maintenance of the optical cable operating environment, which helps to enhance the precision and accuracy of optical cable monitoring.

[0195] Preferably, in this embodiment, the processing of historical monitoring data can also be realized, so as to realize the prediction of future trends, which specifically includes:

[0196] Collecting several types of historical monitoring data at multiple collection moments, and marking the several types of historical monitoring data according to corresponding monitoring results to form a sample set;

[0197] Inputting the sample set into a deep learning prediction model, training and testing the deep learning prediction model, and obtaining a trained deep learning prediction model;

[0198] The trained deep learning prediction model is used to predict the several types of monitoring data to generate prediction results for optical cable operation.

[0199] In this embodiment, the above method can be used to predict future trends. For example, the time series of temperature monitoring data obtained by the deep learning prediction model can predict the temperature rise trend and judge the fire risk in advance. The abnormal detection algorithm of the deep learning prediction model can be used to identify abnormal points and discover emergencies. For example, a sudden increase in the intensity of the vibration signal may indicate external force intrusion. Then, based on the monitoring results and prediction results, automatic risk warning information is established to push environmental abnormality data to operation and maintenance personnel in real time.

[0200] Preferably, the deep learning prediction model can also be trained and updated based on historical monitoring data to continuously optimize the prediction accuracy; the model can be retrained using real-time feedback from operation and maintenance personnel to improve detection sensitivity.

[0201] Specifically, Figure 6 As shown in the figure, mapping different types of monitoring data into the map can be broken down into the following steps:

[0202] S210, based on the standard distance L of the several types of monitoring data common , the geographic coordinates (x, y, z) of the collection point in the map are converted and mapped using the following formula:

[0203] (x,y,z)=GIS_map(L common )

[0204] Wherein x is the longitude of the geographic coordinate, y is the latitude of the geographic coordinate, z is the height of the optical cable from the ground, and GIS_map(.) is a coordinate conversion function;

[0205] S220, placing the aligned monitoring data in the same coordinate system;

[0206] S230. Display the several types of monitoring data in the same coordinate system on the map accordingly according to the geographic coordinates.

[0207] In this embodiment, by displaying the aligned monitoring data on the same map, real-time comprehensive monitoring of different types of monitoring data can be achieved; preferably, the geographic coordinates of different collection points are obtained through the geographic conversion formula provided by the GIS system, and different types of detection data are displayed in the same coordinate system, which can provide real-time monitoring, early warning and historical trend analysis for operation and maintenance personnel, thereby improving monitoring efficiency.

[0208] Based on the same inventive concept, the embodiment of the present application also provides a monitoring system based on multi-source optical cable data fusion, such as Figure 3 As shown, the monitoring system based on multi-source optical cable data fusion may include:

[0209] Acquisition module 311, alignment module 312, pre-processing module 313, pre-entry module 314, monitoring module 315. Among them:

[0210] The acquisition module 311 is used to obtain several types of monitoring data of the optical cable, wherein the several types of monitoring data are collected by different monitoring sensor devices according to a unified collection time period;

[0211] In this embodiment, the acquisition module 311 can be used to perform Figure 1 As shown in step S110 , for a detailed description of the acquisition module 311 , reference may be made to the description of step S110 .

[0212] The alignment module 312 is used to align the several types of monitoring data in space and time to obtain the aligned several types of monitoring data;

[0213] In this embodiment, the alignment module 312 may be used to perform Figure 1 As shown in step S120 , for a detailed description of the alignment module 312 , reference may be made to the description of step S120 .

[0214] The preprocessing module 313 is used to preprocess the aligned monitoring data to obtain input matrices.

[0215] In this embodiment, the preprocessing module 313 can be used to perform Figure 1 As shown in step S130, for the detailed description of the preprocessing module 313, reference may be made to the description of step S130.

[0216] The pre-entry module 314 is used to pre-set key features and standard parameters and enter them into the deep learning prediction model;

[0217] In this embodiment, the pre-entry module 314 can be used to perform Figure 1As shown in step S140 , for a detailed description of the pre-entry module 314 , reference may be made to the description of step S140 .

[0218] The monitoring module 315 is used to input the input matrix into the deep learning prediction model for fusion to obtain fusion data, and analyze the fusion data according to the key features and the standard parameters to obtain monitoring results;

[0219] In this embodiment, the monitoring module 315 can be used to perform Figure 1 As shown in step S150, for the detailed description of the monitoring module 315, reference may be made to the description of the step S150.

[0220] Based on the same inventive concept, this embodiment also provides an electronic device, Figure 8 The structural diagram of the electronic device of this embodiment is shown, including a memory 411 and a processor 412, the memory 411 stores computer-readable instructions, and the processor 412 executes the computer-readable instructions to implement the monitoring method based on multi-source optical cable data fusion of this embodiment.

[0221] Preferably, the electronic device further includes a bus 415 and an input / output interface 414 , and the processor 412 , the input / output interface 414 and the memory 411 are connected via the bus 415 .

[0222] The memory 411 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk storage. The communication connection between the system network element and at least one other network element is realized through at least one input / output interface 414 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used. The bus 415 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus 415 may be divided into an address bus, a data bus, a control bus, etc. (not fully drawn in the figure).

[0223] The processor 412 may be an integrated circuit chip with signal processing capability. In the specific implementation process, each step in the above method embodiment may be completed by the hardware integrated logic circuit or software instructions in the processor 412. The above processor 412 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, which may implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of the present invention. The general-purpose processor may be a microprocessor or the processor 412 may also be any conventional processor 412, etc. The steps of the method disclosed in conjunction with the embodiments of the present invention may be directly embodied as being executed by a hardware decoding processor, or may be executed by a combination of hardware and software modules in a decoding processor. The software module may be located in a storage medium mature in the art, such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 411, and the processor 412 reads the information in the memory 411 and completes the steps of the method of the above embodiment in combination with its hardware.

[0224] An embodiment of the present invention further provides a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are called and executed by the processor 412, the computer-executable instructions prompt the processor 412 to implement the above-mentioned monitoring method based on multi-source optical cable data fusion. The specific implementation can be found in the embodiment, which will not be repeated here.

[0225] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0226] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the technical solution of the present invention, and are not intended to limit the specific implementation methods of the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the claims of the present invention shall be included in the protection scope of the claims of the present invention.

Claims

1. A monitoring method based on multi-source optical cable data fusion, characterized in that: The method comprises: Acquire several types of monitoring data of the optical cable, wherein the several types of monitoring data are collected by different monitoring devices according to a unified collection time period; Aligning the several types of monitoring data in space and time to obtain the aligned several types of monitoring data; Preprocessing the aligned monitoring data to obtain input matrices respectively; Presetting key features and standard parameters and entering them into the deep learning prediction model; The input matrix is ​​input into the deep learning prediction model for fusion to obtain fused data, and the fused data is analyzed according to the key features and the standard parameters to obtain monitoring results.

2. The monitoring method based on multi-source optical cable data fusion according to claim 1 is characterized in that: The several types of monitoring data specifically include two or more of OTDR monitoring data, DVS monitoring data, temperature monitoring data, and water level or gas monitoring data; wherein the OTDR monitoring data is acquired through an optical time domain reflectometer, the DVS monitoring data is acquired through a distributed optical fiber vibration sensing device, the temperature monitoring data is acquired through a distributed optical fiber temperature sensing device, and the water level or gas monitoring data is acquired through a water level sensing device or a gas sensing device.

3. The monitoring method based on multi-source optical cable data fusion according to claim 2 is characterized in that: The unified acquisition time period is obtained as follows: Acquire two or more of the original acquisition period of the OTDR monitoring data, the original acquisition period of the DVS monitoring data, the original acquisition period of the temperature monitoring data, and the original acquisition period of the water level or gas monitoring data as the initial acquisition period; The greatest common divisor of all the initial acquisition periods is taken as the unified acquisition time period.

4. The monitoring method based on multi-source optical cable data fusion according to claim 2 is characterized in that: The spatially and temporally aligning the plurality of monitoring data to obtain the aligned plurality of monitoring data specifically includes: Each type of monitoring data includes the location information and the collection time of each collection point on the optical cable; Obtain the location information l of each collection point of each monitoring data, and map the location information l to the standard distance L according to the following formula: common : Among them, the ΔL common is the standard distance resolution, l is the set of location information of each collection point of each monitoring data, and round(.) is a rounding function; Obtain the collection time t of each collection point for each type of monitoring data, and map the collection time t to the standard time T according to the following formula: Among them, t is the collection time of each collection point of each monitoring data, T common To unify the collection time period.

5. A monitoring method for multi-source real-time optical cable data alignment and fusion according to claim 4, characterized in that: The method further comprises: The standard distance L according to the several monitoring data common , the geographic coordinates (x, y, z) of the collection point in the map are converted and mapped using the following formula: (x,y,z)=GIS_map(L common ) Wherein x is the longitude of the geographic coordinate, y is the latitude of the geographic coordinate, z is the height of the optical cable from the ground, and GIS_map(.) is a coordinate conversion function; Placing the aligned monitoring data in the same coordinate system; According to the geographic coordinates, the several types of monitoring data in the same coordinate system are displayed correspondingly on the map.

6. A monitoring method based on multi-source optical cable data fusion according to any one of claims 1 to 5, characterized in that: The key features specifically include: The key feature of the OTDR monitoring data is the slope of change of the signal reflection intensity; The key feature of the DVS monitoring data is the vibration spectrum energy distribution; The key feature of the temperature monitoring data is the temperature value; The key feature of the water level or gas monitoring data is the water level height value or the gas concentration value; And / or, the standard parameters include a slope threshold corresponding to the slope of the change of the signal reflection intensity, an energy threshold corresponding to the vibration spectrum energy distribution, a standard temperature range corresponding to the temperature value, and a standard water level height range or gas concentration range corresponding to the water level height value or gas concentration value.

7. A monitoring method based on multi-source optical cable data fusion according to any one of claims 1 to 5, characterized in that: The input matrix is ​​input into the deep learning prediction model for fusion to obtain fusion data, and the fusion data is analyzed according to the key features and the standard parameters to obtain the monitoring results, specifically including: Inputting the input matrix into the deep learning prediction model, extracting key features in the input matrix using the deep learning prediction model, and generating a feature matrix; Weighting each key feature in the feature matrix to obtain a feature matrix with weights; Performing weighted fusion on the feature matrix with weights to obtain fused data; The deep learning prediction model monitors the key features in the fused data according to the standard parameters and records a number of abnormal points, wherein the abnormal points correspond one to one with the weights of the corresponding key features; Acquire the correlation information of the plurality of abnormal points according to the fused data; All the abnormal points are analyzed according to the weights corresponding to the abnormal points and the associated information to obtain the overall monitoring result of the optical cable.

8. The monitoring method based on multi-source optical cable data fusion according to claim 7 is characterized in that: The extracting key features from the input matrix using a deep learning prediction model specifically includes: Acquire the signal reflection intensity of each acquisition point according to the OTDR monitoring data in the input matrix, and extract the signal reflection intensity change slope of the two acquisition points as the key feature corresponding to the OTDR monitoring data; and / or, Obtain the vibration spectrum amplitude at the corresponding acquisition time according to the DVS monitoring data in the input matrix, and extract the vibration spectrum energy distribution at the corresponding acquisition time as the key feature corresponding to the DVS monitoring data; and / or, Obtain the temperature value of each acquisition point according to the temperature monitoring data in the input matrix, and use the temperature value of each acquisition point as the key feature corresponding to the temperature monitoring data; and / or, The water level height value or gas concentration value of each collection point is obtained according to the water level or gas monitoring data of the input matrix, and the water level height value or gas concentration value of each collection point is used as the key feature corresponding to the water level or gas monitoring data.

9. The monitoring method based on multi-source optical cable data fusion according to claim 8, characterized in that: The method for obtaining the abnormal point is: If the signal reflection intensity change slope of two acquisition points in the OTDR monitoring data is greater than a preset slope threshold, it is determined that an abnormal reflection intensity point appears in the optical cable between the two acquisition points; and / or, If the vibration spectrum energy distribution in the DVS monitoring data is greater than a preset energy threshold, it is determined that the optical cable has a vibration abnormality point; and / or, If the temperature value in the temperature monitoring data is not within the preset standard temperature range, the collection point is determined to be a temperature abnormality point; and / or, If the water level height value in the water level monitoring data is not within the preset standard water level height range, the collection point is determined to be a water level abnormality point; or, if the gas concentration value in the gas monitoring data is not within the preset gas concentration range, the collection point is determined to be a gas abnormality point.

10. A monitoring system based on multi-source optical cable data fusion, the system comprising: A collection module, used to obtain several types of monitoring data of the optical cable, wherein the several types of monitoring data are collected through different monitoring devices according to a unified collection time period; An alignment module, used to align the several types of monitoring data in space and time to obtain the aligned several types of monitoring data; A preprocessing module, used for preprocessing the aligned monitoring data to obtain input matrices; A pre-entry module, used to pre-set key features and standard parameters and enter them into the deep learning prediction model; The monitoring module is used to input the input matrix into the deep learning prediction model for fusion to obtain fusion data, and analyze the fusion data according to the key features and the standard parameters to obtain monitoring results.

Images