Optical fiber co-cable detection method and system based on deep learning

Abstract

The invention belongs to the technical field of optical fiber communication, and provides an optical fiber co-cable detection method and system based on deep learning, and the method comprises the steps: obtaining the signal characteristics of a to-be-detected optical fiber; constructing a same-cable optical fiber data set and a different-cable optical fiber data set based on the acquired signal features; constructing a deep learning optical fiber co-cable detection model based on the attention mechanism based on the constructed data set; and detecting a to-be-detected optical fiber according to the constructed optical fiber co-cable detection model to complete optical fiber co-cable detection. According to the invention, the deep learning model is combined with optical fiber co-cable detection, the signal feature information is learned, and the accuracy and efficiency of optical fiber co-cable identification detection are effectively improved.

Description

Technical Field

[0001] The present invention belongs to the field of optical fiber communication technology, and specifically relates to an optical fiber co-cable detection method and system based on deep learning. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] Optical fiber communication plays an important role in contemporary information transmission and communication infrastructure, and is widely used in Internet backbone networks, data centers, enterprise networks, urban networks, etc. As the scale of networks expands and bandwidth requirements increase, the structure of optical fiber networks becomes more and more complex. Many optical fiber cables contain multiple optical fibers, so it is very common to lay multiple optical fibers in the same cable or pipe.

[0004] Fiber co-cable detection is a very important task in fiber optic network maintenance and troubleshooting. With the popularization of fiber optic communication technology, optical fiber is widely used in data transmission, communication and various critical infrastructure. In order to ensure the efficient and safe operation of optical fiber, fiber co-cable detection has become a necessary step; it is usually used to identify, locate and analyze different events and faults on the optical fiber line, especially when multiple optical fibers are laid together in the same optical cable. When both the primary and backup services are transmitted on the same optical fiber, if the optical fiber fails or is damaged, whether due to physical damage, environmental factors or technical problems, it may cause the interruption of both the primary and backup services, increasing the time and complexity of service recovery.

[0005] Traditional fiber fault detection mainly relies on optical time domain reflectometer (OTDR) equipment, which generates the attenuation spectrum of the optical fiber by sending light pulses and measuring the returned reflected signals. However, OTDR signal analysis often relies on the experience of professional technicians, especially in the fiber co-cable environment, where signals from different optical fibers overlap and reflection events are complex, and manual analysis and diagnosis efficiency is low. In addition, the detection accuracy of traditional algorithms may not be ideal when faced with high noise, complex reflection signals and other situations. Summary of the invention

[0006] To solve the above problems, the present invention proposes a method and system for optical fiber co-cable detection based on deep learning, which combines the deep learning model with optical fiber co-cable detection, learns signal feature information, and effectively improves the accuracy and efficiency of optical fiber co-cable identification and detection.

[0007] According to some embodiments, a first solution of the present invention provides a method for detecting optical fiber co-cables based on deep learning, which adopts the following technical solution: A method for detecting optical fiber co-cables based on deep learning, comprising: Acquire signal characteristics of the optical fiber to be detected; Based on the acquired signal characteristics, constructing a same-cable optical fiber data set and a different-cable optical fiber data set; Based on the constructed dataset, a deep learning optical fiber co-cable detection model based on attention mechanism is constructed; The optical fiber to be detected is detected according to the constructed optical fiber co-cable detection model to complete the optical fiber co-cable detection.

[0008] As a further technical limitation, the constructed deep learning optical fiber co-cable detection model based on the attention mechanism adopts a parallel attention mechanism module; based on the feature aggregation in the parallel attention mechanism module, the key feature information of the signal features is extracted, based on the convolutional neural network, the local feature information of the signal features is obtained, the parallel attention mechanism module is combined with the convolutional neural network to obtain fusion features, and the deep learning optical fiber co-cable detection model based on the attention mechanism is trained according to the obtained fusion features.

[0009] Furthermore, in the process of training the deep learning optical fiber co-cable detection model based on the attention mechanism, the coarse-grained features are extracted from the obtained fusion features by adding a long short-term memory network to identify the key information in the fusion features, thereby completing the construction of the deep learning optical fiber co-cable detection model based on the attention mechanism.

[0010] As a further technical limitation, during the optical fiber co-cable detection process, the optical fiber co-cable detection model is used to determine whether the optical fiber to be detected belongs to the same cable optical fiber signal or the different cable optical fiber signal, and the optical fiber co-cable detection result is evaluated through evaluation indicators including precision, recall rate, F1-score and accuracy.

[0011] As a further technical limitation, in the process of obtaining the signal characteristics of the optical fiber to be detected, the light intensity signal reflected from various positions on the optical fiber by the short pulse laser sent by the optical fiber in different optical cables is measured by an optical time domain reflectometer, so as to obtain the signal characteristics of the optical fiber to be detected including the optical fiber path and optical fiber ID.

[0012] As a further technical limitation, after obtaining the signal characteristics of the optical fiber to be detected, the obtained signal characteristics are preprocessed, that is, the amplitude of the signal characteristics is scaled to a uniform range and then the signal characteristic length is normalized, and then the noise signal in the normalized signal characteristics is removed.

[0013] According to some embodiments, a second solution of the present invention provides a fiber co-cable detection system based on deep learning, which adopts the following technical solution: An optical fiber co-cable detection system based on deep learning, comprising: An acquisition module, which is configured to acquire signal characteristics of the optical fiber to be detected; A first construction module, configured to construct a same-cable optical fiber dataset and a different-cable optical fiber dataset based on the acquired signal characteristics; A second building module is configured to build a deep learning optical fiber co-cable detection model based on an attention mechanism based on the constructed data set; The detection module is configured to detect the optical fiber to be detected according to the constructed optical fiber co-cable detection model to complete the optical fiber co-cable detection.

[0014] According to some embodiments, a third solution of the present invention provides a computer-readable storage medium, which adopts the following technical solution: A computer-readable storage medium stores a program, which, when executed by a processor, implements the steps in the optical fiber co-cable detection method based on deep learning as described in the first solution of the present invention.

[0015] According to some embodiments, a fourth solution of the present invention provides an electronic device, which adopts the following technical solution: An electronic device comprises a memory, a processor and a program stored in the memory and running on the processor, wherein when the processor executes the program, the steps in the optical fiber co-cable detection method based on deep learning as described in the first scheme of the present invention are implemented.

[0016] According to some embodiments, a fifth solution of the present invention provides a computer program product, which adopts the following technical solution: A computer program product includes software codes, wherein the program in the software codes executes the steps in the optical fiber co-cable detection method based on deep learning as described in the first embodiment of the present invention.

[0017] Compared with the prior art, the present invention has the following beneficial effects: The present invention uses statistical features such as mean, standard deviation, and maximum value to directly extract the core information of the signal, reduce the high-dimensional redundancy of the original OTDR signal, and facilitate the discovery of the commonalities and differences of the signal from statistical laws. Compared with directly inputting the OTDR signal curve, the input data after feature extraction has a lower dimension, reducing the risk of model overfitting.

[0018] The present invention adopts a CNN-LSTM network model based on an attention mechanism. On the basis of a deeply separable CNN module, a parallel attention mechanism branch is added to extract significant features, and the original input is segmented according to two different scales. The length of each attention mechanism module input is twice the length of the corresponding CNN module input, so that the attention module can more comprehensively grasp the context information and accurately obtain significant features. There will be overlaps in adjacent input subsequences. The output features of the CNN module are multiplied element by element with the significant features output by the corresponding attention mechanism module, thereby completing the identification of important features. At the end of the model, an LSTM unit is added to extract coarse-grained features from the significant fine-grained features extracted by the CNN based on the attention mechanism, and while finely processing the features of each dimension, memory loss and gradient diffusion problems caused by too long a step length are prevented. The model ensures that the similarity or difference features between optical fibers can be fully learned, and through comparative analysis of two optical fibers, key information can be more efficiently identified from feature fusion, rather than processing each optical fiber signal separately. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings in the specification that constitute a part of this embodiment are used to provide a further understanding of this embodiment. The schematic embodiments of this embodiment and their descriptions are used to explain this embodiment and do not constitute improper limitations on this embodiment.

[0020] Figure 1 This is a flow chart of the optical fiber co-cable detection method based on deep learning in Embodiment 1 of the present invention; Figure 2 This is an architecture diagram of a CNN structure based on an attention mechanism in Embodiment 1 of the present invention; Figure 3 This is an architecture diagram of the CNN-LSTM structure based on the attention mechanism in Embodiment 1 of the present invention; Figure 4 This is a structural block diagram of an optical fiber co-cable detection system based on deep learning in Example 2 of the present invention. DETAILED DESCRIPTION

[0021] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0022] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0023] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0024] In the present invention, terms such as "upper", "lower", "left", "right", "front", "back", "vertical", "horizontal", "side", "bottom", etc. indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings. They are relational words determined only for the convenience of describing the structural relationships of the various parts or elements of the present invention, and do not specifically refer to any part or element in the present invention and should not be understood as limitations on the present invention.

[0025] In the present invention, terms such as "fixed connection", "connected", "connection", etc. should be understood in a broad sense, indicating that it can be fixedly connected, integrally connected or detachably connected; it can be directly connected or indirectly connected through an intermediate medium. For relevant scientific research or technical personnel in this field, the specific meanings of the above terms in the present invention can be determined according to specific circumstances, and they cannot be understood as limitations on the present invention.

[0026] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0027] Embodiment 1 Embodiment 1 of the present invention introduces a method for detecting optical fiber co-cables based on deep learning.

[0028] like Figure 1 A method for detecting optical fiber co-cables based on deep learning is shown, comprising: Acquire signal characteristics of the optical fiber to be detected; Based on the acquired signal characteristics, constructing a same-cable optical fiber data set and a different-cable optical fiber data set; Based on the constructed dataset, a deep learning optical fiber co-cable detection model based on attention mechanism is constructed; The optical fiber to be detected is detected according to the constructed optical fiber co-cable detection model to complete the optical fiber co-cable detection.

[0029] This embodiment analyzes the signal data of multiple optical fibers from the same optical cable and different optical cables, and constructs a model that can identify optical fiber event information and determine whether these optical fibers belong to the same optical cable. The model needs to identify the commonalities and differences of optical fiber signals in the optical cable. At the same time, a deep learning model is used to automatically extract meaningful features from optical fiber signals, avoiding the limitations of traditional manually designed features and improving the judgment accuracy and robustness of the model; it provides an efficient and practical solution for the identification and event detection of optical fibers in the same cable, which can be applied to practical fields such as optical fiber communications and optical cable maintenance.

[0030] In view of the shortcomings of the prior art, this embodiment uses the CNN-LSTM model based on the attention mechanism as the feature extraction network of the OTDR signal, and realizes the accurate classification of whether the same cable is used by extracting the signal features. To achieve the above purpose, the detailed scheme of the optical fiber same cable detection method adopted in this embodiment is as follows: (1) Data collection An optical time domain reflectometer (OTDR) is used to connect to the optical fibers in different optical cables, send short pulse lasers to the optical fibers, measure the light intensity signal reflected from various points on the optical fibers, and record the cable path and optical fiber ID.

[0031] (2) Data preprocessing OTDR signal data needs to be preprocessed, including normalizing the signal data, scaling the signal amplitude to a uniform range, and normalizing the length; then using a denoising algorithm to remove noise from the signal so that the deep learning model can focus more on useful signal features.

[0032] (3) Constructing data sets of optical fibers in the same cable and optical fibers in different cables For each processed signal data segment, the mean, standard deviation, maximum value, minimum value, number of reflection points, location of reflection points and reflection intensity are calculated to form a feature vector. The feature vectors of the two signals are then concatenated and labeled. The label for the same cable fiber is 0, and the label for different cables is 1.

[0033] (4) Constructing a CNN structure module based on the attention mechanism This embodiment adopts Figure 2 The CNN structure module based on the attention mechanism shown in the figure; specifically: The task of the CNN module is to extract more and more abstract features from the input signal layer by layer. The CNN module has three layers, each of which contains a depthwise separable convolution layer (depthwise convolution + pointwise convolution). The number of input channels of the input layer is set to 1, the size of the convolution kernel is set to 3×3×1, 3×3×8, 3×3×16, the step size is set to 1, and the network is of the same padding type to ensure that the size of the feature map remains unchanged. By stacking multiple layers of depthwise separable convolution, CNN can gradually learn from low-level features (such as morphological changes of signals) to more advanced abstract features (reflection events, fault locations). Depthwise separable convolution significantly reduces the number of parameters and computations by separating spatial filtering and channel mixing. Depthwise convolution performs 3×3 spatial convolution on each input channel independently to extract local features (such as local reflection peaks and signal attenuation areas). Pointwise convolution fuses cross-channel information through 1×1 convolution kernels and adjusts the number of output channels. This design retains the ability of convolution operations to capture local patterns while improving model efficiency through lightweight structures.

[0034] In the parallel attention mechanism module, key information is first extracted from subsequences of multiple scales through feature aggregation operations. The stacking of each layer of convolution and pooling helps capture cross-scale features. Each part of the network consists of two consecutive layers of depth-separable convolution (3×3 depth convolution + 1×1 point-by-point convolution) and a pooling layer, called a convolutional block. There are 4 such convolutional blocks in total, each of which contains two consecutive convolutional layers with a convolution kernel size of 3×3, a stride of 1, and zero padding on the edges. After each convolution, a batch normalization layer is connected and nonlinearity is introduced using the ReLU activation function. At the end of each convolutional block is a maximum pooling layer with a pooling window size of 2×2 and a stride of 2, which is used to downsample the feature map. In the last convolutional block, a 1×1 convolution layer is used to mine the linear relationship between features.

[0035] The scale recovery part uses upsampling to enlarge the size of the feature map according to a certain ratio. The interpolation method used is linear interpolation (mode='linear'), which is suitable for feature reconstruction of one-dimensional signals. These key features are restored to the same size as the CNN module output, thereby ensuring that the availability of the features matches the CNN module output. Finally, the Sigmoid function is applied to the upsampled feature map to limit the feature value of each channel to the range of [0,1]. These extracted significant features are fused with the features extracted by CNN for subsequent classification or event detection tasks.

[0036] The output features of the CNN module are multiplied element by element with the significant features output by the corresponding attention mechanism module. The attention mechanism assigns a weight to each position in the feature map extracted by CNN, and this weight indicates the importance of the feature. The higher the weight, the greater the contribution of the feature to the final judgment result, and the network will pay more attention to these features; conversely, the lower the weight, the smaller the impact of the feature on the final result, and the network will pay less attention to these features. Therefore, through this weighted method, the attention mechanism helps the model automatically identify and strengthen the features of key areas in the signal, and improves the sensitivity to fiber events (reflection, breakage, bending, etc.).

[0037] This embodiment combines the attention mechanism with the deep separable CNN module to construct a lightweight and efficient dual-branch network, and further enhances the feature extraction capability by introducing a parallel attention mechanism branch. Specifically, the original input signal is divided into multiple subsequences according to different scales, and the input of the attention module covers the input of the entire CNN, so that the contextual information of the signal can be more comprehensively grasped. This design ensures that the attention mechanism can effectively focus on those significant features that have an important impact on the task (fiber event detection). The feature fusion of the two significantly improves the accuracy and robustness of fiber event detection, while reducing the demand for computing resources.

[0038] (5) Constructing a CNN-LSTM architecture based on the attention mechanism like Figure 3 As shown in the figure, the fusion result is used as the input of the LSTM unit, and the number of hidden units in each layer is 64. The deep separable CNN provides spatial and local features, the attention mechanism provides significant information, and the fused feature representation is richer. These multi-level features provide high-quality input for LSTM. The number of neurons in the last fully connected layer is kept as 1, and the Sigmoid activation function is used for classification output.

[0039] (6) Training the CNN-LSTM network based on the attention mechanism First, train the deep separable CNN and attention mechanism parts separately, and then train the entire model jointly. First, train the deep separable CNN module separately so that it can learn to extract effective local features from the input signal. The input is the joint features of the two signals, and the batch size is 128. Use Binary Cross-Entropy Loss as the loss function to measure the gap between the probability distribution predicted by the model and the true label. Update the parameters of the CNN module through the gradient descent algorithm (Adam optimizer), and the initial learning rate is set to 1e-3. During training, freeze the attention mechanism module and only optimize the CNN module until the loss function converges. The Binary Cross-Entropy loss function is as follows: Loss=- (1) in, is the true label (0 or 1), is the model prediction probability, ranging from 0 to 1.

[0040] The attention mechanism module is trained separately to learn to generate effective attention weights and focus on the key areas in the signal. The input is also the joint features of the two signals, and the batch size is 128. Binary Cross-EntropyLoss is used as the loss function, and the initial learning rate is set to 1e-3. During the training process, the CNN module is frozen and only the attention mechanism module is optimized until the loss function converges.

[0041] The entire model is trained jointly so that the CNN module, the attention mechanism module, and the LSTM layer work together. The output of the CNN module is multiplied element-by-element with the output of the attention mechanism module, and the fused features are input into the LSTM layer to capture the temporal dependency. The output of the LSTM layer is mapped to the binary classification result through the fully connected layer. Binary Cross-Entropy Loss is used as the loss function, and the initial learning rate is set to a small value (1e-4) to avoid destroying the learned features. All parameters are updated by the gradient descent algorithm until the loss function converges. Through the staged training strategy, the CNN module and the attention mechanism module are first allowed to learn effective feature extraction and weight allocation respectively, and then the overall performance is optimized through joint training, and finally the binary classification task of the two signal segments is achieved (0 indicates different cables, 1 indicates the same cable).

[0042] (7) Testing The mean, standard deviation, maximum value, minimum value, number of reflection points, location of reflection points and reflection intensity of the collected original OTDR signal data are calculated and formed into feature vectors. The feature vectors of the two optical fiber data are then spliced ​​and input into the trained CNN-LSTM network based on the attention mechanism to determine whether each sample belongs to the same cable optical fiber signal or different cable optical fiber signals. After the test is completed, the evaluation is performed. The actual labels of the test samples are compared with the labels predicted by our test, and the performance of the model is evaluated using four commonly used evaluation indicators: Precision, Recall, F1-score and Accuracy. Precision and Recall focus on the accuracy of the model and its ability to capture positive class samples. Accuracy and F1-score are indicators to measure the overall performance of the model; the details are as follows: (2) (3) (4) (5) Here, the "same cable fiber" signal is considered as "P (Positive)", and the "different cable fiber" signal is considered as "N (Negative)". Therefore, TP represents the fiber pair correctly classified as "same cable fiber"; TN represents the fiber pair correctly classified as "different cable fiber"; FP represents the different cable fiber pair incorrectly classified as "same cable fiber"; FN represents the same cable fiber pair incorrectly classified as "different cable fiber".

[0043] This embodiment adopts a CNN-LSTM network model based on the attention mechanism. On the basis of the deeply separable CNN module, a parallel attention mechanism branch is added to extract significant features, and the original input is divided according to two different scales. The length of each attention mechanism module input is twice the length of the corresponding CNN module input, so that the attention module can more comprehensively grasp the context information and accurately obtain significant features. There will be overlaps in adjacent input subsequences. The output features of the CNN module are multiplied element by element with the significant features output by the corresponding attention mechanism module to complete the identification of important features. At the end of the model, an LSTM unit is added to extract coarse-grained features from the significant fine-grained features extracted by the CNN based on the attention mechanism, and the memory loss and gradient diffusion problems caused by too long a step length are prevented while finely processing the features of each dimension. The model ensures that the similarity or difference features between optical fibers can be fully learned, and through comparative analysis of the two optical fibers, key information can be more efficiently identified from feature fusion, rather than processing each optical fiber signal separately.

[0044] Embodiment 2 Embodiment 2 of the present invention introduces an optical fiber co-cable detection system based on deep learning.

[0045] like Figure 4 The optical fiber co-cable detection system based on deep learning is shown, comprising: An acquisition module, which is configured to acquire signal characteristics of the optical fiber to be detected; A first construction module, configured to construct a same-cable optical fiber dataset and a different-cable optical fiber dataset based on the acquired signal characteristics; A second building module is configured to build a deep learning optical fiber co-cable detection model based on an attention mechanism based on the constructed data set; The detection module is configured to detect the optical fiber to be detected according to the constructed optical fiber co-cable detection model to complete the optical fiber co-cable detection.

[0046] The detailed steps are the same as those of the optical fiber co-cable detection method based on deep learning provided in Example 1 and will not be repeated here.

[0047] Embodiment 3 Embodiment 3 of the present invention provides a computer-readable storage medium.

[0048] A computer-readable storage medium stores a program, which, when executed by a processor, implements the steps in the optical fiber co-cable detection method based on deep learning as described in Embodiment 1 of the present invention.

[0049] The detailed steps are the same as those of the optical fiber co-cable detection method based on deep learning provided in Example 1 and will not be repeated here.

[0050] Embodiment 4 A fourth embodiment of the present invention provides an electronic device.

[0051] An electronic device comprises a memory, a processor and a program stored in the memory and running on the processor, wherein when the processor executes the program, the steps in the optical fiber co-cable detection method based on deep learning as described in Embodiment 1 of the present invention are implemented.

[0052] The detailed steps are the same as those of the optical fiber co-cable detection method based on deep learning provided in Example 1 and will not be repeated here.

[0053] Embodiment 5 Embodiment 5 of the present invention provides a computer program product.

[0054] A computer program product includes software code, wherein the program in the software code executes the steps in the optical fiber co-cable detection method based on deep learning as described in Embodiment 1 of the present invention.

[0055] The detailed steps are the same as those of the optical fiber co-cable detection method based on deep learning provided in Example 1 and will not be repeated here.

[0056] It should be understood by those skilled in the art that the embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. The solutions in the embodiments of the present invention may be implemented in various computer languages, for example, object-oriented programming language Java and interpreted scripting language JavaScript, etc.

[0057] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0058] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0059] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0060] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0061] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

[0062] The above description is only a preferred embodiment of the present embodiment and is not intended to limit the present embodiment. For those skilled in the art, the present embodiment may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present embodiment shall be included in the protection scope of the present embodiment.

Claims

1. A method for detecting optical fiber co-cable based on deep learning, characterized in that: include: Acquire signal characteristics of the optical fiber to be detected; Based on the acquired signal characteristics, constructing a same-cable optical fiber data set and a different-cable optical fiber data set; Based on the constructed dataset, a deep learning optical fiber co-cable detection model based on attention mechanism is constructed; The optical fiber to be detected is detected according to the constructed optical fiber co-cable detection model to complete the optical fiber co-cable detection.

2. The optical fiber co-cable detection method based on deep learning as described in claim 1, characterized in that: The constructed deep learning optical fiber co-cable detection model based on attention mechanism adopts a parallel attention mechanism module; based on the feature aggregation in the parallel attention mechanism module, the key feature information of the signal features is extracted, based on the convolutional neural network, the local feature information of the signal features is obtained, the parallel attention mechanism module is combined with the convolutional neural network to obtain fusion features, and the deep learning optical fiber co-cable detection model based on attention mechanism is trained according to the obtained fusion features.

3. The optical fiber co-cable detection method based on deep learning as described in claim 2, characterized in that: In the process of training the deep learning optical fiber co-cable detection model based on the attention mechanism, the coarse-grained features are extracted from the obtained fusion features by adding the long short-term memory network to identify the key information in the fusion features, thus completing the construction of the deep learning optical fiber co-cable detection model based on the attention mechanism.

4. The optical fiber co-cable detection method based on deep learning as described in claim 1, characterized in that: In the process of optical fiber co-cable detection, the optical fiber co-cable detection model is used to determine whether the optical fiber to be detected belongs to the same cable optical fiber signal or the different cable optical fiber signal, and the optical fiber co-cable detection result is evaluated through evaluation indicators including precision, recall rate, F1-score and accuracy.

5. The optical fiber co-cable detection method based on deep learning as described in claim 1, characterized in that: In the process of obtaining the signal characteristics of the optical fiber to be detected, the optical time domain reflectometer is used to measure the light intensity signal reflected from various positions on the optical fiber by the short pulse laser sent by the optical fiber in different optical cables, so as to obtain the signal characteristics of the optical fiber to be detected including the optical fiber path and optical fiber ID.

6. The optical fiber co-cable detection method based on deep learning as described in claim 1, characterized in that: After obtaining the signal characteristics of the optical fiber to be detected, the obtained signal characteristics are preprocessed, that is, the amplitude of the signal characteristics is scaled to a uniform range, the signal characteristic length is normalized, and then the noise signal in the normalized signal characteristics is removed.

7. A fiber co-cable detection system based on deep learning, characterized in that: include: An acquisition module, which is configured to acquire signal characteristics of the optical fiber to be detected; A first construction module, configured to construct a same-cable optical fiber dataset and a different-cable optical fiber dataset based on the acquired signal characteristics; A second building module is configured to build a deep learning optical fiber co-cable detection model based on an attention mechanism based on the constructed data set; The detection module is configured to detect the optical fiber to be detected according to the constructed optical fiber co-cable detection model to complete the optical fiber co-cable detection.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the optical fiber co-cable detection method based on deep learning as described in any one of claims 1 to 6 are implemented.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the program, the steps of the optical fiber co-cable detection method based on deep learning as described in any one of claims 1 to 6 are implemented.

10. A computer program product comprising software code, characterized in that The program in the software code executes the steps of the optical fiber co-cable detection method based on deep learning as described in any one of claims 1 to 6.

Images