Distributed optical fiber sensing event identification method and device based on multi-parameter signal fusion
The multi-parameter signal is preprocessed and feature extraction through one-dimensional convolutional neural network, which solves the problem of vibration and temperature signals confusion under low signal-to-noise ratio, and realizes effective fusion of multi-parameter signals and high-accurate event recognition.
Patent Information
- Application Number
- CN202510528815.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-25
- Publication Date
- 2025-08-08
AI Technical Summary
The existing distributed fiber sensing technology is difficult to accurately identify multi-parameter events under low signal-to-noise ratios, especially the confusion of vibration and temperature signals leads to a decrease in recognition accuracy, and the sampling rate difference of multi-parameter signals leads to mismatch in signal lengths, making it difficult to effectively fusion.
A one-dimensional convolutional neural network is used to preprocess and feature extraction of multi-parameter signals, including the fusion of dynamic vibration, static temperature and low-frequency static vibration data, and feature fusion and recognition are performed through a convolutional neural network.
It improves the accuracy of event recognition under low signal-to-noise ratio, effectively combines the characteristics of signals of different parameters, reduces the impact of noise, and improves the recognition accuracy.
Smart Images
Figure CN120449087A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of optical fiber sensing, and in particular relates to a distributed optical fiber sensing event recognition method and device based on multi-parameter signal fusion. Background Art
[0002] Distributed fiber optic sensors (DOFS) offer real-time, distributed, and long-distance monitoring capabilities, making them suitable for a variety of scenarios, including pipelines, cables, and tunnels. In practical applications, distributed fiber optic sensing often requires accurate event recognition despite low signal-to-noise ratios. When the signal-to-noise ratio of the sensing signal decreases, sensor signals such as vibration, temperature, and strain can be easily confused with other interfering signals with similar characteristics, leading to misidentification by intelligent signal processing algorithms and reduced event recognition accuracy. In actual sensing applications, abnormal events often cause changes in multiple parameters. For example, cable icing can cause simultaneous changes in temperature and strain, while pipeline leaks can cause simultaneous changes in vibration and temperature.
[0003] Current research on multi-parameter distributed fiber-optic sensing primarily focuses on building multi-mechanism distributed fiber-optic sensing systems (see the literature "Single-end hybrid Rayleigh Brillouin and Raman distributed fiber-optic sensing system"). However, there is a lack of subsequent data processing methods for the multi-parameter data collected by the system. In studies combining distributed fiber-optic sensing with neural networks, only a single parameter is typically used for event recognition (see the literature "One-dimensional CNN-based intelligent recognition of vibrations in pipeline monitoring with DAS"). This lack of multi-parameter complementarity makes it difficult to achieve high recognition accuracy in low signal-to-noise ratio conditions.
[0004] However, there are still some technical issues in the fusion processing of multi-parameter sensor signals. The first is the time scale issue. The sampling rate of the vibration signal of the φ-OTDR system is usually greater than or equal to 1kHz, while the sampling rate of the temperature signal of the ROTDR system is generally less than 1Hz. The huge difference in sampling rate leads to a significant difference in signal length, making it difficult to use the general direct fusion method for multi-parameter signals. The second is the problem of vibration signal utilization. Considering the huge difference in the sampling rates of vibration and temperature signals, the time length of the vibration signal is generally much smaller than that of the temperature signal. For example, when a vibration signal of time length t1 is matched with a temperature signal of time length t2, the vibration signal information within the time period (t2-t1) will not be used. Summary of the Invention
[0005] In view of the above, the purpose of the present invention is to provide a distributed fiber optic sensing event recognition method and device based on multi-parameter signal fusion, which can effectively fuse multi-parameter dynamic and static sensing signals through a one-dimensional convolutional neural network, realize the complementarity of dynamic and static parameters, fully reflect event changes, and improve the event recognition accuracy under low signal-to-noise ratio, which is suitable for accurate event recognition under low signal-to-noise ratio conditions.
[0006] To achieve the above-mentioned object of the invention, an embodiment provides a distributed optical fiber sensing event recognition method based on multi-parameter signal fusion, comprising the following steps:
[0007] Acquiring multi-parameter signal data collected by a multi-mechanism distributed optical fiber sensing system, wherein the multi-parameter signal data includes dynamic signal data and static signal data;
[0008] After preprocessing the multi-parameter signal data, static temperature data is extracted from low-sampling-rate static signal data over a long period of time, dynamic vibration data is extracted from high-sampling-rate dynamic signal data over a short period of time, and signal envelopes are extracted from high-sampling-rate dynamic signal data over a long period of time and low-frequency static vibration data is extracted from the envelopes.
[0009] Convolutional neural networks are used to fuse static temperature data, dynamic vibration data, and low-frequency static vibration data for distributed fiber optic sensing event recognition.
[0010] Preferably, the long time range is greater than 1 minute, and the short time range is less than 1 second.
[0011] Preferably, the convolutional neural network includes a vibration feature extraction module, a temperature feature extraction module, a feature fusion module and an event recognition module.
[0012] The vibration feature extraction module is used to extract features from dynamic vibration data and low-frequency static vibration data respectively, and then obtain vibration features through feature fusion and feature extraction;
[0013] The temperature feature extraction module is used to extract features from static temperature data to obtain temperature features;
[0014] The feature fusion module is used to fuse the vibration feature and the temperature feature to obtain a fusion feature;
[0015] The event recognition module is used to perform distributed optical fiber sensing event recognition based on fusion features to obtain recognition results.
[0016] Preferably, the feature extraction is implemented using at least one convolution unit, and each convolution unit includes a one-dimensional convolution layer and a pooling layer.
[0017] Preferably, the feature fusion is implemented by using at least one fusion unit, and each fusion unit first expands the high-dimensional vector of the input feature into a one-dimensional vector, and then performs one-dimensional vector splicing to achieve fusion.
[0018] Preferably, the event recognition module uses a fully connected layer and a SoftMax function to perform distributed optical fiber sensing event recognition based on fusion features.
[0019] Preferably, the convolutional neural network undergoes parameter optimization before being applied, and the specific optimization process is as follows:
[0020] Collect multi-parameter signal data of known distributed optical fiber sensing events and construct a sample set;
[0021] The parameters of the convolutional neural network are optimized based on the sample set. The loss function used in the parameter optimization is the cross entropy loss between the identified output distributed fiber optic sensing events and the known distributed fiber optic sensing events.
[0022] To achieve the above-mentioned object of the invention, an embodiment of the present invention further provides a distributed optical fiber sensing event recognition device based on multi-parameter signal fusion, comprising:
[0023] A data acquisition unit, which is used to acquire multi-parameter signal data collected by the multi-mechanism distributed optical fiber sensing system, wherein the multi-parameter signal data includes dynamic signal data and static signal data;
[0024] a data processing unit configured to extract static temperature data over a long period of time from static signal data at a low sampling rate, extract dynamic vibration data over a short period of time from dynamic signal data at a high sampling rate, and extract a signal envelope over a long period of time from dynamic signal data at a high sampling rate and extract low-frequency static vibration data from the envelope after preprocessing the multi-parameter signal data;
[0025] The event recognition unit is used to use a convolutional neural network to fuse static temperature data, dynamic vibration data, and low-frequency static vibration data to perform distributed optical fiber sensing event recognition.
[0026] To achieve the above-mentioned purpose of the invention, an embodiment further provides a computing device, including a memory and one or more processors, wherein the memory stores executable code, and when the one or more processors execute the executable code, they are used to implement the above-mentioned distributed optical fiber sensing event identification method based on multi-parameter signal fusion.
[0027] To achieve the above-mentioned purpose of the invention, an embodiment further provides a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the above-mentioned distributed optical fiber sensing event identification method based on multi-parameter signal fusion.
[0028] The present invention provides a distributed fiber optic sensing event recognition method and device based on multi-parameter signal fusion. Compared with the traditional method of event recognition relying on only a single parameter, the present invention effectively improves the accuracy of event recognition under low signal-to-noise ratio conditions. Its technical effects are reflected in:
[0029] 1. Extract features from dynamic vibration signals from both static and dynamic perspectives, which improves the utilization rate of vibration signals during multi-parameter signal fusion and facilitates the fusion of multi-parameter signals;
[0030] 2. Use a multi-input convolutional neural network to fuse multi-parameter signals with large differences in sampling rates, effectively combining the characteristics of different parameter signals and improving recognition accuracy;
[0031] 3. The multi-parameter signals of different distributed fiber optic sensors are effectively fused to reduce the impact of noise on the recognition results. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0033] Figure 1 is a flow chart of a distributed optical fiber sensing event recognition method based on multi-parameter signal fusion provided in an embodiment;
[0034] Figure 2 is a data extraction flow chart provided by the embodiment;
[0035] Figure 3 Schematic diagram of the structure of a convolutional neural network provided in the embodiment;
[0036] Figure 4It is a structural diagram of a distributed optical fiber sensing event recognition device based on multi-parameter signal fusion provided in an embodiment.
[0037] Figure 5 is the confusion matrix of the recognition results of single-class input and three-class input combination, where (a) is the confusion matrix of the recognition result of single vibration data input, (b) is the confusion matrix of the recognition result of single temperature data input, (c) is the confusion matrix of the recognition result of vibration and temperature data combination, and (d) is the confusion matrix of the recognition result of dynamic and static vibration and temperature data combination;
[0038] Figure 6 The following is a comparison of the recognition results of single-class input and three-class input combination, where (a) is the accuracy comparison chart, (b) is the precision comparison chart, (c) is the recall comparison chart, and (d) is the F-value comparison chart;
[0039] Figure 7 It is the confusion matrix of the three types of data combined with the recognition results of different input models, where (a) is the confusion matrix of the recognition results of the input ANN model, (b) is the confusion matrix of the recognition results of the input RNN model, and (c) is the confusion matrix of the recognition results of the input 1d-CNN model;
[0040] Figure 8 The following is a comparison chart of the recognition results of three types of data combined with different input models, where (a) is the accuracy comparison chart, (b) is the precision comparison chart, (c) is the recall comparison chart, and (d) is the F-value comparison chart. DETAILED DESCRIPTION
[0041] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and do not limit the scope of protection of the present invention.
[0042] The inventive concept of the present invention is: In response to the problem that the existing distributed fiber optic sensing technology has insufficient accuracy in identifying low signal-to-noise ratio events, the present invention proposes a solution for multi-parameter dynamic and static signal fusion, which effectively fuses and uses multi-parameter dynamic and static sensing signals through a convolutional neural network to improve the accuracy of event recognition under low signal-to-noise ratio conditions, and is suitable for accurate event recognition under low signal-to-noise ratio conditions.
[0043] like Figure 1 As shown, the distributed optical fiber sensing event recognition method based on multi-parameter signal fusion provided by the embodiment includes the following steps:
[0044] S1, acquiring multi-parameter signal data collected by a multi-mechanism distributed optical fiber sensing system, wherein the multi-parameter signal data includes dynamic signal data and static signal data.
[0045] In an embodiment, the obtained multi-parameter signal data includes vibration, temperature or strain sensing signals, wherein the vibration signal belongs to dynamic signal data, and the temperature or strain signal belongs to static signal data.
[0046] S2, after preprocessing the multi-parameter signal data, extracts static temperature data in a long time range for the static signal data with a low sampling rate, extracts dynamic vibration data in a short time range for the dynamic signal data with a high sampling rate, and also extracts the signal envelope in a long time range for the dynamic signal data with a high sampling rate and extracts low-frequency static vibration data from the envelope.
[0047] In this embodiment, the collected multi-parameter signal data is preprocessed. Specifically, high-pass filtering is performed on the dynamic signal data and the static signal data. The filter cutoff frequency is 10 Hz to filter out DC and low-frequency noise. Valid temperature signal data is also intercepted from the filtered static signal data.
[0048] In the embodiment, after data preprocessing, data extraction is also performed, such as Figure 2 As shown, including:
[0049] Process low-sampling-rate static signal data over a long period of time to reduce high-frequency noise interference and obtain more accurate data trends. Segment the processed data to obtain static temperature data, with each data item having a size of 1×60 sampling points.
[0050] The high-sampling-rate dynamic signal data is processed within a short timeframe to reduce low-frequency noise interference. The processed data is segmented to obtain dynamic vibration data, with each data item having a size of 1×1000 sampling points.
[0051] The signal envelope of the high-sampling-rate dynamic signal data is also extracted over a long period of time. The signal envelope data size is 1×60,000 sampling points. Downsampling is performed from the low-frequency static envelope to extract low-frequency static vibration data with unchanged main frequency but shorter data length. The size of each data after downsampling is 1×1000 sampling points.
[0052] S3 uses convolutional neural networks to fuse static temperature data, dynamic vibration data, and low-frequency static vibration data for distributed fiber optic sensing event recognition.
[0053] In the embodiment, Figure 3As shown in the figure, the convolutional neural network used includes a vibration feature extraction module, a temperature feature extraction module, a feature fusion module and an event recognition module, wherein the vibration feature extraction module is used to extract features from dynamic vibration data and low-frequency static vibration data respectively, and then obtain vibration features through feature fusion and feature extraction; the temperature feature extraction module is used to extract features from static temperature data to obtain temperature features; the feature fusion module is used to fuse vibration features and temperature features to obtain fusion features; the event recognition module is used to perform distributed optical fiber sensing event recognition based on fusion features to obtain recognition results.
[0054] Specifically, feature extraction is implemented using at least one convolution unit, each of which includes a one-dimensional convolution layer and a maximum pooling layer. For example, for low-frequency static vibration data and dynamic vibration data, features are extracted using convolution unit 1 and convolution unit 2, respectively. The fusion feature of the two extracted vibration-like features is also extracted using convolution unit 3. For static temperature data, features are extracted using convolution units 4 and 5. Specifically, the convolution kernel sizes in the convolution layers during feature extraction are different. The convolution kernel size of the convolution layers in convolution units 1, 2, and 4 is 5*5, while the convolution kernel size of the convolution layers in convolution units 3 and 5 is 3*3.
[0055] Specifically, feature fusion is implemented using at least one fusion unit. Each fusion unit first expands the high-dimensional vectors of two input features into one-dimensional vectors, and then splices the one-dimensional vectors and outputs them to fuse data features of different parameters.
[0056] Specifically, the event recognition module uses a fully connected layer and a SoftMax function to identify distributed fiber optic sensing events based on fused features. The fully connected layer scales down the features extracted by the convolutional layer layer by layer, using them as output for judgment. The SoftMax function normalizes the output of the fully connected layer into a probability value, thereby determining the type of event.
[0057] The above convolutional neural network undergoes parameter optimization before being applied. The specific optimization process is as follows:
[0058] Multi-parameter signal data for known distributed fiber optic sensing events is collected and a sample set is constructed. The collected events include four categories: slow leaks, fast leaks, vibration interference, and temperature interference. Slow leaks are defined as leaks occurring with an interval greater than 1 second between leaks, fast leaks are defined as leaks occurring with an interval less than 1 second between leaks, vibration interference includes knocking and digging around the fiber, and temperature interference includes abnormal temperature changes caused by noise and temperature fluctuations caused by ambient temperature changes. The multi-mechanism distributed fiber optic sensing system used in this embodiment has a fiber length of 4000m, with one spatial point sampled every 4m. The time sampling rate for the vibration signal is 1000Hz, and the time sampling rate for the temperature signal is 1Hz.
[0059] The convolutional neural network parameters were optimized based on the sample set. The loss function used for parameter optimization was the cross-entropy loss between the output distributed fiber optic sensing events and known distributed fiber optic sensing events. 20 training rounds were used with a learning rate of 0.001. The accuracy, precision, recall, f-score, and loss curve of the training results were recorded. The convolutional neural network and its parameters that performed best on the test set were selected as the application recognition model.
[0060] In specific applications, the model is deployed on a computer connected to a distributed fiber optic sensor that combines multiple mechanisms to identify data with abnormal changes and determine the type of event.
[0061] like Figure 5 and Figure 6 As shown in the figure, the recognition effect of the combined vibration and temperature data is better than that of the single vibration data and single temperature data. The recognition effect of the combined dynamic and static vibration and temperature data is better than that of the combined vibration and temperature data in terms of accuracy, F value and other indicators. This shows that taking the three types of data as input can achieve better results than inputting the three types of data separately.
[0062] like Figure 7 and Figure 8 As shown in the figure, the three types of data are combined and input into artificial neural network (ANN), recurrent neural network (RNN) and one-dimensional convolutional neural network (1d-CNN) respectively. The recognition effect of RNN model is better than that of ANN model. The effect of 1d-CNN model is better than that of ANN model and RNN model in multiple indicators, indicating that the one-dimensional convolutional neural network event recognition effect designed by the present invention is better than that of general neural network model.
[0063] like Figure 4 As shown, the embodiment also provides a distributed optical fiber sensing event recognition device based on multi-parameter signal fusion, including: a data acquisition unit 41, a data processing unit 42, and an event recognition unit 43, wherein the data acquisition unit is used to obtain multi-parameter signal data collected by a multi-mechanism distributed optical fiber sensing system; the data processing unit 42 is used to pre-process the multi-parameter signal data, extract static temperature data from low-sampling-rate static signal data within a long time range, extract dynamic vibration data from high-sampling-rate dynamic signal data within a short time range, and extract signal envelope from high-sampling-rate dynamic signal data within a long time range and extract low-frequency static vibration data from the envelope; the event recognition unit 43 is used to use a convolutional neural network to fuse static temperature data, dynamic vibration data, and low-frequency static vibration data to perform distributed optical fiber sensing event recognition.
[0064] It should be noted that the distributed optical fiber sensing event identification device based on multi-parameter signal fusion provided in the above embodiment should be illustrated by the division of the above-mentioned functional modules when performing distributed optical fiber sensing event identification. The above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the terminal or server is divided into different functional modules to complete all or part of the functions described above. In addition, the distributed optical fiber sensing event identification device based on multi-parameter signal fusion provided in the above embodiment and the distributed optical fiber sensing event identification method embodiment based on multi-parameter signal fusion belong to the same concept. The specific implementation process is detailed in the distributed optical fiber sensing event identification method embodiment based on multi-parameter signal fusion, which will not be repeated here.
[0065] Based on the same inventive concept, an embodiment further provides a computing device including a memory and one or more processors. The memory stores executable code. When the one or more processors execute the executable code, the computing device is used to implement the above-mentioned distributed optical fiber sensing event recognition method based on multi-parameter signal fusion, which specifically includes the following steps:
[0066] S1, acquiring multi-parameter signal data collected by a multi-mechanism distributed optical fiber sensing system, wherein the multi-parameter signal data includes dynamic signal data and static signal data;
[0067] S2, after preprocessing the multi-parameter signal data, extracts static temperature data over a long period of time for the static signal data with a low sampling rate, extracts dynamic vibration data over a short period of time for the dynamic signal data with a high sampling rate, and further extracts signal envelopes over a long period of time for the dynamic signal data with a high sampling rate and extracts low-frequency static vibration data from the envelopes;
[0068] S3 uses convolutional neural networks to fuse static temperature data, dynamic vibration data, and low-frequency static vibration data for distributed fiber optic sensing event recognition.
[0069] The computing device provided in the embodiment, in addition to the processor and memory, also includes hardware required for other services such as internal bus, network interface, memory, etc. at the hardware level. The memory is a non-volatile memory, and the processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to implement the distributed fiber optic sensing event recognition method based on multi-parameter signal fusion described in S1-S3 above. Of course, in addition to software implementation, the present invention does not exclude other implementation methods, such as logic devices or a combination of software and hardware, etc., that is, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.
[0070] Based on the same inventive concept, an embodiment further provides a computer-readable storage medium having a program stored thereon. When the program is executed by a processor, the distributed optical fiber sensing event recognition method based on multi-parameter signal fusion is implemented, specifically comprising the following steps:
[0071] S1, acquiring multi-parameter signal data collected by a multi-mechanism distributed optical fiber sensing system, wherein the multi-parameter signal data includes dynamic signal data and static signal data;
[0072] S2, after preprocessing the multi-parameter signal data, extracts static temperature data over a long period of time for the static signal data with a low sampling rate, extracts dynamic vibration data over a short period of time for the dynamic signal data with a high sampling rate, and further extracts signal envelopes over a long period of time for the dynamic signal data with a high sampling rate and extracts low-frequency static vibration data from the envelopes;
[0073] S3 uses convolutional neural networks to fuse static temperature data, dynamic vibration data, and low-frequency static vibration data for distributed fiber optic sensing event recognition.
[0074] In the embodiment, computer-readable media includes permanent and non-permanent, removable and non-removable media and can implement information storage by any method or technology. The information can be computer-readable instructions, data structures, program modules or other data.
[0075] The specific implementation methods described above provide a detailed description of the technical solutions and beneficial effects of the present invention. It should be understood that the above is only the most preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, supplements and equivalent substitutions made within the scope of the principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. A distributed optical fiber sensing event recognition method based on multi-parameter signal fusion, characterized in that: The following steps are involved: Acquiring multi-parameter signal data collected by a multi-mechanism distributed optical fiber sensing system, wherein the multi-parameter signal data includes dynamic signal data and static signal data; After preprocessing the multi-parameter signal data, static temperature data is extracted from low-sampling-rate static signal data over a long period of time, dynamic vibration data is extracted from high-sampling-rate dynamic signal data over a short period of time, and signal envelopes are extracted from high-sampling-rate dynamic signal data over a long period of time and low-frequency static vibration data is extracted from the envelopes. Convolutional neural networks are used to fuse static temperature data, dynamic vibration data, and low-frequency static vibration data for distributed fiber optic sensing event recognition.
2. The distributed optical fiber sensing event recognition method based on multi-parameter signal fusion according to claim 1 is characterized in that: The long time range is greater than 1 minute, and the short time range is less than 1 second.
3. The distributed optical fiber sensing event recognition method based on multi-parameter signal fusion according to claim 1 is characterized in that: The convolutional neural network includes a vibration feature extraction module, a temperature feature extraction module, a feature fusion module and an event recognition module. The vibration feature extraction module is used to extract features from dynamic vibration data and low-frequency static vibration data respectively, and then obtain vibration features through feature fusion and feature extraction; The temperature feature extraction module is used to extract features from static temperature data to obtain temperature features; The feature fusion module is used to fuse the vibration feature and the temperature feature to obtain a fusion feature; The event recognition module is used to perform distributed optical fiber sensing event recognition based on fusion features to obtain recognition results.
4. The distributed optical fiber sensing event recognition method based on multi-parameter signal fusion according to claim 2 is characterized in that: The feature extraction is implemented by using at least one convolution unit, and each convolution unit includes a one-dimensional convolution layer and a pooling layer.
5. The distributed optical fiber sensing event recognition method based on multi-parameter signal fusion according to claim 2 is characterized in that: The feature fusion is implemented by using at least one fusion unit. Each fusion unit first expands the high-dimensional vector of the input feature into a one-dimensional vector and then performs one-dimensional vector splicing to achieve fusion.
6. The distributed optical fiber sensing event recognition method based on multi-parameter signal fusion according to claim 2, characterized in that: The event recognition module uses a fully connected layer and a SoftMax function to perform distributed optical fiber sensing event recognition based on fusion features.
7. The distributed optical fiber sensing event recognition method based on multi-parameter signal fusion according to claim 1 is characterized in that: Before being applied, the convolutional neural network undergoes parameter optimization. The specific optimization process is as follows: Collect multi-parameter signal data of known distributed optical fiber sensing events and construct a sample set; The parameters of the convolutional neural network are optimized based on the sample set. The loss function used in the parameter optimization is the cross entropy loss between the identified output distributed fiber optic sensing events and the known distributed fiber optic sensing events.
8. A distributed optical fiber sensing event recognition device based on multi-parameter signal fusion, characterized in that: include: A data acquisition unit, which is used to acquire multi-parameter signal data collected by the multi-mechanism distributed optical fiber sensing system, wherein the multi-parameter signal data includes dynamic signal data and static signal data; a data processing unit configured to extract static temperature data over a long period of time from static signal data at a low sampling rate, extract dynamic vibration data over a short period of time from dynamic signal data at a high sampling rate, and extract a signal envelope over a long period of time from dynamic signal data at a high sampling rate and extract low-frequency static vibration data from the envelope after preprocessing the multi-parameter signal data; The event recognition unit is used to use a convolutional neural network to fuse static temperature data, dynamic vibration data, and low-frequency static vibration data to perform distributed optical fiber sensing event recognition.
9. A computing device comprising a memory and one or more processors, wherein the memory stores executable code, characterized in that: When the one or more processors execute the executable code, they are used to implement the distributed optical fiber sensing event recognition method based on multi-parameter signal fusion according to any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that A program is stored thereon, and when the program is executed by a processor, the distributed optical fiber sensing event recognition method based on multi-parameter signal fusion according to any one of claims 1 to 7 is implemented.