Optical fiber fault detection and separation method and system
Through CNN model and Gramma matrix technology, the fiber fault location is quickly positioned, solving the problem of low fault detection efficiency of existing fibers, and achieving rapid and automatic fault location and processing.
Patent Information
- Application Number
- CN202411102528.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-12
- Publication Date
- 2024-12-27
- Estimated Expiration
- 2044-08-12
AI Technical Summary
The existing fiber fault detection methods are inefficient and rely on manual segmentation testing, and cannot quickly locate the fault location, resulting in difficulty in repairing in important application networks.
The CNN model is used to combine Gramma matrix and two-dimensional image conversion technology to calculate deviations by detecting the light intensity sequence and reference intensity of the detection point, and train the convolutional neural network model to achieve rapid positioning of the fiber fault location.
It greatly reduces labor investment, reduces labor expenditure, improves fault handling efficiency, and achieves rapid positioning of fiber fault locations.
Smart Images
Figure CN118921112B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of optical fiber fault detection, and in particular to an optical fiber fault detection separation method and system. Background Art
[0002] Fiber optic is short for optical fiber, which is a fiber made of glass or plastic that can be used as a light transmission tool. Fiber optic networks are very useful in life, and once a failure occurs, it will cause great trouble. How to troubleshoot common fiber optic network failures and troubleshooting methods are becoming increasingly important.
[0003] In the communication systems of various businesses, optical fiber has gradually replaced electrical cables due to its low cost, long transmission distance of optical signals and low loss. Therefore, when an optical cable line fails, it must be repaired every second, especially in important application network systems.
[0004] In order to ensure the long-distance and low-loss application characteristics of optical transmission signals, an optical cable line must meet certain physical environmental conditions. In the prior art, optical fiber fault detection mainly adopts manual segmentation testing, which is inefficient. Summary of the invention
[0005] The embodiment of the present application provides a fiber optic fault detection and separation method and system, which are used to combine the CNN model to realize the judgment of the fault location, greatly reduce the manual input, and realize the rapid positioning of the fault location.
[0006] The present application embodiment provides a method for detecting and separating optical fiber faults, including:
[0007] Loading light source data with preset parameters at the end of the sample optical fiber;
[0008] At different positions of the sample optical fiber line, multiple detection points are configured, and the light intensity at the rear end and output end of each detection point is detected to obtain a light intensity sequence at each detection point position, wherein the position of any detection point corresponds to a corresponding transmission distance interval, and the sample optical fiber line covers the transmission distance interval to be detected;
[0009] The Gram matrix is calculated based on the intensity at the sample optical fiber tip as the reference intensity;
[0010] According to the calculated Gram matrix, the light intensity sequence is converted into a two-dimensional image, and a corresponding normal or fault description mark is added to any detection point to construct a training sample, wherein any pixel value in the two-dimensional image is used to describe the deviation between the intensity at the corresponding detection point and the reference intensity;
[0011] Extracting image features of the two-dimensional image in the training sample to train a preset CNN model based on the extracted image features, wherein the CNN model comprises at least a convolutional layer, a pooling layer, a fully connected layer and a Softmax classification layer, wherein the fully connected layer is used to perform feature combination and fault prediction using multiple layers of fully connected neurons, and the Softmax classification layer is used to convert the output of the fully connected layer into a probability distribution;
[0012] Loading light source data of preset parameters at the head end of the optical fiber to be detected, and obtaining the light intensity at the output end of the optical fiber to be detected, and determining the positional relationship of the fitting detection points of the optical fiber to be detected and estimating the light intensity sequence to be detected based on the fitting positional relationship;
[0013] According to the estimated sequence of light intensities to be detected, the Gram matrix is calculated and converted into a two-dimensional image;
[0014] Based on the converted two-dimensional image and the trained CNN model, a fault detection result of the optical fiber to be detected is determined.
[0015] Optionally, based on the intensity of the sample optical fiber tip as a reference intensity, calculating the Gram matrix includes:
[0016] Pre-configure the positions of each detection point and the theoretical strength of the output end under normal conditions, and calculate the theoretical strength deviation:
[0017]
[0018] Where L represents the intensity at the end of the sample fiber, Indicates the detection point i The theoretical strength of
[0019] According to the angle position obtained by detection and the light intensity at the output end, the detection intensity deviation is calculated:
[0020]
[0021] in, Indicates the detection point i The detection strength;
[0022] The calculated theoretical intensity deviation and the detected intensity deviation are enhanced by using a nonlinear function, wherein the nonlinear function is configured such that the enhanced value increases nonlinearly as the intensity deviation increases.
[0023] Optionally, converting the light intensity sequence into a two-dimensional image according to the calculated Gram matrix comprises:
[0024] According to the intensity deviation of each detection point after enhancement, the Gram matrix is constructed using the NumPy library;
[0025] The elements of the Gram matrix are converted into pixel values of a two-dimensional image using a mapping function.
[0026] Optionally, based on the intensity deviation of each detection point after enhancement, the Gram matrix is constructed using the NumPy library including:
[0027] According to the detection intensity deviation of each detection point after enhancement, the similarity between the theoretical intensity deviation of each detection point is calculated to satisfy:
[0028]
[0029] The similarity calculated as the element at the corresponding position of the Gram matrix.
[0030] Optionally, converting the light intensity sequence into a two-dimensional image according to the calculated Gram matrix comprises:
[0031] Using the imshow function, the matrix values are mapped to pixel color values based on a given color spectrum to describe different angle differences through corresponding color values.
[0032] Optionally, determining the positional relationship of the fitting detection points of the optical fiber to be detected includes:
[0033] According to the transmission distance interval covered by the detection points in the sample, the detection points are fitted in the corresponding position interval, and the detection points at the location of any possible fault area are marked;
[0034] Estimating the intensity sequence of the light to be detected based on the fitted position relationship includes:
[0035] Based on the light intensities at the input and output ends of the optical fiber to be detected, bidirectional fitting is performed on the detection points corresponding to any possible fault area to estimate two light intensity sequences to be detected for any possible fault area.
[0036] Optionally, according to the estimated sequence of light intensities to be detected, calculating the Gram matrix includes:
[0037] For any possible fault area, two light intensity sequences to be detected are estimated, and the Gram matrices are calculated and converted into two-dimensional images respectively.
[0038] Optionally, determining the fault detection result of the optical fiber to be detected based on the converted two-dimensional image and the trained CNN model includes:
[0039] When the detection result output by the CNN model matches the identified possible fault area, the corresponding area is determined to be the fault area according to the matched sequence of light intensity to be detected and the fitting direction.
[0040] The embodiment of the present application further provides an optical fiber fault detection and separation system, comprising a processor and a memory, wherein the memory stores a computer program, and when the computer program is executed by the processor, the steps of the optical fiber fault detection and separation method as described above are implemented.
[0041] The method implemented in this application combines the CNN model to realize the judgment of the fault location, greatly reducing the input of manpower, reducing labor expenses, improving the efficiency of fault handling, and realizing the rapid positioning of the fault location of the optical fiber.
[0042] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS
[0043] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:
[0044] Figure 1 The figure is a schematic diagram of the basic process of the optical fiber fault detection and separation method of this embodiment. DETAILED DESCRIPTION
[0045] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.
[0046] The present application embodiment proposes a method for separating optical fiber fault detection, such as Figure 1 As shown, including:
[0047] In step S101, light source data of preset parameters are loaded at the end of the sample optical fiber. For example, in some examples, light source data of multiple intensities can be loaded separately to obtain more sample data.
[0048] In step S102, multiple detection points are configured at different positions of the sample optical fiber line, and the light intensity at the rear end and output end of each detection point is detected to obtain a light intensity sequence at each detection point position, wherein the position of any detection point corresponds to a corresponding transmission distance interval, and the sample optical fiber line covers the transmission distance interval to be detected. In some specific examples, the distance can be simulated by an optical attenuator, so that multiple detection points are configured under the same line in the test environment, and the light intensity at the rear end and output end of each detection point is detected. The embodiment of the present application detects the light intensity at the rear end of the detection point through the test environment, so as to adapt it to the measured light intensity in actual application, so as to match the possible fault location.
[0049] In step S103, the Gram matrix is calculated based on the intensity of the sample optical fiber head end as the reference intensity. In the specific example, the sample optical fiber head end is used as the loading end of the light source and as the calculation reference, so that the same light source parameters and the same calculation reference can be guaranteed in the subsequent judgment of the optical fiber to be tested.
[0050] In step S104, the light intensity sequence is converted into a two-dimensional image according to the calculated Gram matrix, and a corresponding normal or fault description mark is added to any detection point to construct a training sample, wherein any pixel value in the two-dimensional image is used to describe the deviation between the intensity at the corresponding detection point and the reference intensity. In some examples, the deviation between the intensity at the corresponding detection point and the reference intensity can be obtained by enhancement processing.
[0051] In step S105, the image features of the two-dimensional image in the training sample are extracted to train a preset CNN model based on the extracted image features, wherein the CNN model includes at least a convolutional layer, a pooling layer, a fully connected layer and a Softmax classification layer, wherein the fully connected layer is used to perform feature combination and fault prediction using multiple layers of fully connected neurons, and the Softmax classification layer is used to convert the output of the fully connected layer into a probability distribution. In some specific examples, the CNN model may include multiple layers of convolutional layers, such as three layers, each layer using different filters to capture features of different scales and complexities. The pooling layer is used to reduce the spatial dimension and retain important features, the fully connected layer is used to flatten the features extracted by the convolutional layer and the pooling layer, and perform feature combination and classification prediction using multiple layers of fully connected neurons, and the Softmax classification layer is used to convert the output of the fully connected layer into a probability distribution.
[0052] In step S106, light source data with preset parameters is loaded at the head end of the optical fiber to be detected, and the light intensity at the output end of the optical fiber to be detected is obtained, and the positional relationship of the fitting detection points of the optical fiber to be detected is determined and the light intensity sequence to be detected is estimated based on the fitting positional relationship. In some examples, light source data with preset parameters can be loaded according to the actual situation of the optical fiber to be detected, and the preset parameters can be matched and obtained from the light source data of multiple intensities of the sample.
[0053] In step S107, a Gram matrix is calculated based on the estimated sequence of light intensities to be detected, and converted into a two-dimensional image.
[0054] In step S108, a fault detection result of the optical fiber to be detected is determined based on the converted two-dimensional image and the trained CNN model.
[0055] The method implemented in this application combines the CNN model to realize the judgment of the fault location, greatly reducing the input of manpower, reducing labor expenses, improving the efficiency of fault handling, and realizing the rapid positioning of the fault location of the optical fiber.
[0056] In some embodiments, based on the intensity of the sample optical fiber tip as the reference intensity, calculating the Gram matrix includes:
[0057] Pre-configure the positions of each detection point and the theoretical strength of the output end under normal conditions, and calculate the theoretical strength deviation:
[0058]
[0059] Where L represents the intensity at the end of the sample fiber, Indicates the detection point i The theoretical strength.
[0060] According to the angle position obtained by detection and the light intensity at the output end, the detection intensity deviation is calculated:
[0061]
[0062] in, Indicates the detection point i detection strength.
[0063] The calculated theoretical intensity deviation and the detected intensity deviation are enhanced using a nonlinear function, wherein the nonlinear function is configured so that the enhanced value increases nonlinearly as the intensity deviation increases. In some embodiments, the specific nonlinear function can be set according to actual needs, for example, the nonlinear function can be determined based on the intensity deviation configuration power term and the basic term.
[0064] In some embodiments, converting the light intensity sequence into a two-dimensional image according to the calculated Gram matrix comprises:
[0065] According to the intensity deviation of each detection point after enhancement, a Gram matrix is constructed using the NumPy library, and the elements of the Gram matrix are converted into pixel values of a two-dimensional image using a mapping function.
[0066] In some embodiments, constructing a Gram matrix using the NumPy library according to the intensity deviation of each detection point after enhancement includes:
[0067] According to the detection intensity deviation of each detection point after enhancement, the similarity between the theoretical intensity deviation of each detection point is calculated to satisfy:
[0068]
[0069] The similarity calculated As the element of the corresponding position of the Gram matrix. In the embodiment of the present application, the detection intensity deviation and the theoretical intensity deviation are amplified by using a nonlinear function for enhancement, and further, the distribution of the deviations can be further determined by calculating the similarity between the deviations.
[0070] In some embodiments, converting the light intensity sequence into a two-dimensional image according to the calculated Gram matrix comprises:
[0071] Using an imshow function, such as the imshow function of the Matplotlib library, the values of the matrix are mapped to color values of pixels based on a given color spectrum to describe different angle differences through corresponding color values. In some embodiments, the given color spectrum can be a heat map color spectrum to represent different similarities, such as warm colors represent low similarity and cold colors represent high similarity.
[0072] In some embodiments, determining the positional relationship of the fitting detection points of the optical fiber to be detected includes:
[0073] According to the transmission distance interval covered by the detection points in the sample, detection points are fitted in the corresponding position interval, and the detection points at the location of any possible fault area are marked. In some embodiments, the coverage distance of the sample optical fiber is longer than the laying distance of the optical fiber to be detected, so that the fitting detection points can be set in the corresponding position interval of the optical fiber to be detected according to the sequence of the detection points selected in the sample optical fiber, and the corresponding position interval can be set as a possible fault area according to meteorological information, such as local heavy rain or strong wind, and the possible fault area can be marked.
[0074] Estimating the intensity sequence of the light to be detected based on the fitted position relationship includes:
[0075] Based on the light intensity at the input and output ends of the optical fiber to be detected, a bidirectional fitting is performed on the detection points corresponding to any possible fault area to estimate two light intensity sequences to be detected for any possible fault area. That is, in some examples, the bidirectional fitting is based on the light source data with preset parameters loaded at the head end of the optical fiber to be detected, and the light intensity at the output end of the optical fiber to be detected, which are respectively used as references for bidirectional intensity fitting, so that two light intensity sequences to be detected can be estimated for any possible fault area.
[0076] In some embodiments, calculating the Gram matrix according to the estimated sequence of light intensities to be detected includes:
[0077] For any possible fault area, two light intensity sequences to be detected are estimated, and the Gram matrices are calculated and converted into two-dimensional images respectively.
[0078] In some embodiments, based on the converted two-dimensional image and the trained CNN model, determining the fault detection result of the optical fiber to be detected includes: when the detection result output by the CNN model matches the identified possible fault area, determining the corresponding area as the fault area according to the matched light intensity sequence to be detected and the fitting direction. In a specific example, the detection result output by the CNN model includes the probability of possible faults at each detection point, and the corresponding area is determined to be the fault area by selecting the fitting detection point with the highest fault probability and matching it with the possible fault area identified in the light intensity sequence to be detected.
[0079] The embodiment of the present application further provides an optical fiber fault detection and separation system, comprising a processor and a memory, wherein the memory stores a computer program, and when the computer program is executed by the processor, the steps of the optical fiber fault detection and separation method as described above are implemented.
[0080] In addition, although exemplary embodiments have been described herein, the scope includes any and all embodiments based on the present disclosure with equivalent elements, modifications, omissions, combinations (e.g., various embodiments intersecting schemes), adaptations or changes. It is not limited to the examples described in this specification or during the implementation of this application, and its examples will be interpreted as non-exclusive.
[0081] The above description is intended to be illustrative rather than restrictive. For example, the above examples (or one or more of them) can be used in combination with each other. For example, those of ordinary skill in the art may use other embodiments when reading the above description.
[0082] The above embodiments are merely exemplary embodiments of the present disclosure. Those skilled in the art may make various modifications or equivalent substitutions to the present invention within the essence and protection scope of the present disclosure, and such modifications or equivalent substitutions shall also be deemed to fall within the protection scope of the present invention.
Claims
1. A method for detecting and separating optical fiber faults, characterized in that: include: Loading light source data with preset parameters at the end of the sample optical fiber; At different positions of the sample optical fiber line, multiple detection points are configured, and the light intensity at the rear end and output end of each detection point is detected to obtain a light intensity sequence at each detection point position, wherein the position of any detection point corresponds to a corresponding transmission distance interval, and the sample optical fiber line covers the transmission distance interval to be detected; The Gram matrix is calculated based on the intensity of the sample optical fiber tip as the reference intensity; According to the calculated Gram matrix, the light intensity sequence is converted into a two-dimensional image, and a corresponding normal or fault description mark is added to any detection point to construct a training sample, wherein any pixel value in the two-dimensional image is used to describe the deviation between the intensity at the corresponding detection point and the reference intensity; Extracting image features of the two-dimensional image in the training sample to train a preset CNN model based on the extracted image features, wherein the CNN model comprises at least a convolutional layer, a pooling layer, a fully connected layer and a Softmax classification layer, wherein the fully connected layer is used to perform feature combination and fault prediction using multiple layers of fully connected neurons, and the Softmax classification layer is used to convert the output of the fully connected layer into a probability distribution; Loading light source data of preset parameters at the head end of the optical fiber to be detected, and obtaining the light intensity at the output end of the optical fiber to be detected, and determining the positional relationship of the fitting detection points of the optical fiber to be detected and estimating the light intensity sequence to be detected based on the fitting positional relationship; According to the estimated sequence of light intensities to be detected, the Gram matrix is calculated and converted into a two-dimensional image; Based on the converted two-dimensional image and the trained CNN model, a fault detection result of the optical fiber to be detected is determined.
2. The optical fiber fault detection and separation method according to claim 1, characterized in that: Based on the intensity of the sample fiber tip as the reference intensity, the calculation of the Gram matrix includes: Pre-configure the positions of each detection point and the theoretical strength of the output end under normal conditions, and calculate the theoretical strength deviation: Where L represents the intensity at the end of the sample fiber, Indicates the detection point i The theoretical strength of According to the angle position obtained by detection and the light intensity at the output end, the detection intensity deviation is calculated: in, Indicates the detection point i The detection intensity; The calculated theoretical intensity deviation and the detected intensity deviation are enhanced by using a nonlinear function, wherein the nonlinear function is configured such that the enhanced value increases nonlinearly as the intensity deviation increases.
3. The optical fiber fault detection and separation method according to claim 2, characterized in that: According to the calculated Gram matrix, converting the light intensity sequence into a two-dimensional image comprises: According to the intensity deviation of each detection point after enhancement, the Gram matrix is constructed using the NumPy library; The elements of the Gram matrix are converted into pixel values of a two-dimensional image using a mapping function.
4. The optical fiber fault detection and separation method according to claim 3, characterized in that: According to the intensity deviation of each detection point after enhancement, the Gram matrix is constructed using the NumPy library, including: According to the detection intensity deviation of each detection point after enhancement, the similarity between the theoretical intensity deviation of each detection point is calculated to satisfy: The similarity calculated as the element at the corresponding position of the Gram matrix.
5. The optical fiber fault detection and separation method according to claim 4, characterized in that: According to the calculated Gram matrix, converting the light intensity sequence into a two-dimensional image comprises: Using the imshow function, the matrix values are mapped to pixel color values based on a given color spectrum to describe different angle differences through corresponding color values.
6. The optical fiber fault detection and separation method according to claim 4, characterized in that: Determining the positional relationship of the fitting detection points of the optical fiber to be detected includes: According to the transmission distance interval covered by the detection points in the sample, the detection points are fitted in the corresponding position interval, and the detection points at the location of any possible fault area are marked; Estimating the intensity sequence of the light to be detected based on the fitted position relationship includes: Based on the light intensities at the input and output ends of the optical fiber to be detected, bidirectional fitting is performed on the detection points corresponding to any possible fault area to estimate two light intensity sequences to be detected for any possible fault area.
7. The optical fiber fault detection and separation method according to claim 6, characterized in that: According to the estimated sequence of light intensities to be detected, the calculation of the Gram matrix includes: For any possible fault area, two light intensity sequences to be detected are estimated, and the Gram matrices are calculated and converted into two-dimensional images respectively.
8. The optical fiber fault detection and separation method according to claim 6, characterized in that: Determining the fault detection result of the optical fiber to be detected based on the converted two-dimensional image and the trained CNN model includes: When the detection result output by the CNN model matches the identified possible fault area, the corresponding area is determined to be the fault area according to the matched sequence of light intensity to be detected and the fitting direction.
9. An optical fiber fault detection and separation system, characterized in that: The method comprises a processor and a memory, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the steps of the optical fiber fault detection and separation method according to any one of claims 1 to 8 are implemented.