Method, system, device, medium and product for measuring structural parameters of optical fiber

Abstract

The application provides a method, system, device, medium and product for measuring structural parameters of an optical fiber. The method comprises the following steps: emitting laser to a to-be-detected optical fiber, acquiring spectral information of scattered light after the laser passes through the to-be-detected optical fiber, inputting the spectral information into a measurement model, acquiring a structural measurement model of the optical fiber, extracting low-frequency features of the spectral information to obtain a feature vector of the spectral information, and inferring structural parameters of the optical fiber according to the feature vector of the spectral information. The structural parameters of the optical fiber can be more effectively measured in the production and manufacturing process of the optical fiber.

Description

Technical Field

[0001] This application relates to the field of optical fiber geometry measurement technology, and in particular to a method, system, device, medium and product for measuring the structural parameters of optical fibers. Background Technology

[0002] Hollow-core anti-resonant fiber (HC-ARF) is a microstructured optical fiber with an air core. Its optical signals are mainly transmitted in the air holes. It has ultra-low latency, high power carrying capacity and wideband resonance characteristics, and is widely used in high-power laser transmission, nonlinear optics, quantum communication and biosensing.

[0003] The capillary cladding structure parameters within a hollow antiresonant fiber directly affect its light-guiding performance and mechanical strength. Therefore, it is necessary to measure the structural parameters of the capillary, such as its wall thickness and diameter, to ensure the effectiveness of the resulting hollow antiresonant fiber. However, existing methods for measuring fiber structural parameters cannot detect them without damaging the fiber, have poor real-time performance, and low accuracy, making it difficult to meet the requirements for real-time measurement of structural parameters in fibers such as hollow antiresonant fibers.

[0004] Therefore, how to measure the structural parameters of optical fibers in a timely and non-destructive manner during the production and manufacturing process, and how to non-destructively test the structural parameters of optical fibers during the inspection of finished optical fiber products, are technical problems that need to be solved in this field. Summary of the Invention

[0005] This application provides a method, system, device, medium, and product for measuring the structural parameters of optical fibers. By constructing and training a measurement model, the structural parameters of the capillaries inside the optical fiber are measured, thereby improving the real-time performance and accuracy of the measurement results without damaging the physical structure of the optical fiber. This addresses the shortcomings of existing technologies and enables more timely and non-destructive measurement of the structural parameters of optical fibers during the manufacturing process. It also allows for non-destructive testing of the optical fiber's parameters during the inspection of finished optical fiber products.

[0006] The first aspect of this application provides a method for measuring the structural parameters of an optical fiber, comprising: emitting a laser into the optical fiber to be tested; acquiring spectral information of the scattered light after the laser passes through the optical fiber to be tested; inputting the spectral information into a measurement model, causing the measurement model to extract low-frequency features of the spectral information to obtain a feature vector of the spectral information, and inferring the structural parameters of the optical fiber to be tested based on the feature vector of the spectral information; wherein the measurement model is pre-trained based on the spectral data of the designed structure of the optical fiber to be tested, and the frequency range of the low-frequency features is determined during the pre-training of the measurement model.

[0007] A second aspect of this application provides a system for measuring the structural parameters of an optical fiber, comprising: a laser emitting device for emitting laser light into an optical fiber to be tested; a laser receiving device for receiving scattered light from the laser light after it passes through the optical fiber to be tested; and a processing device connected to the laser emitting device and the laser receiving device, for controlling the parameters of the laser light emitted by the laser emitting device into the optical fiber to be tested, and controlling the position of the laser receiving device to obtain spectral information of the scattered light from the laser light after it passes through the optical fiber to be tested; the processing device is further configured to perform the optical fiber structural parameter measurement method as described in the first aspect to infer the structural parameters of the optical fiber to be tested.

[0008] A third aspect of this application provides a method for measuring the structural parameters of an optical fiber, used to train a measurement model as described in the first or second aspect. The method includes: acquiring training data of the optical fiber to be tested, the training data including spectral information of the scattered light after a laser emitted into the optical fiber passes through the optical fiber, and structural parameters of the optical fiber to be tested, wherein the training data is obtained by training based on the designed structure of the optical fiber to be tested and / or by actual measurement of the optical fiber to be tested; sequentially using the spectral information of the optical fiber to be tested in the training data as input data and the structural parameters of the optical fiber to be tested as output data to train the initial model structure of the measurement model until the training error is less than a preset value, thereby obtaining the measurement model.

[0009] A fourth aspect of this application provides an electronic device, comprising: a processor and a memory communicatively connected to the processor; the memory storing computer-executable instructions; and the processor executing the computer-executable instructions stored in the memory to implement the method as described in the first or third aspect.

[0010] The fifth aspect of this application provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the method described in the first or third aspect.

[0011] The sixth aspect of this application provides a computer program product, including a computer program that, when executed by a processor, implements the method described in the first or third aspect.

[0012] In summary, the fiber optic structural parameter measurement method, system, equipment, medium, and product provided in this application emit laser light into the hollow fiber to be tested and acquire the spectral information of the scattered light after the laser passes through the fiber. This spectral information is then input into a measurement model to obtain the fiber's structural parameters output by the model. Since the measurement model extracts low-frequency features of the spectral information to obtain a feature vector, the structural parameters of the hollow fiber are inferred based on this feature vector. This application achieves non-destructive testing of hollow fibers by inferring fiber structural parameters based on spectral information without damaging the fiber structure.

[0013] Furthermore, this application effectively handles structures under complex nonlinear mappings, and compared with other neural network methods, it considers the fusion of local and global features, and the feature selection is more comprehensive, thereby improving the accuracy of the measurement results of the structural parameters of optical fibers. As a result, it can more effectively measure the structural parameters of optical fibers during the production and manufacturing process.

[0014] Meanwhile, the measurement model provided in this application filters out high-frequency features and extracts only low-frequency features as the basic data for inferring fiber structure parameters. This ensures that the measurement model and method of this application are insensitive to noise during the measurement process, have good noise resistance and robustness, and also incorporate physical properties to enhance the stability of the measurement model under different application scenarios. This ensures the reliability and stability of the prediction results of fiber structure parameters, and ultimately improves the robustness and automation level of the measurement results without damaging the physical structure of the fiber.

[0015] In particular, when this application is applied to real-time online calculations integrated into the hollow fiber manufacturing process, it can not only measure the structural parameters of the optical fiber in a timely and non-destructive manner during the production and manufacturing process of hollow optical fiber to meet the needs of continuous production, but also perform non-destructive testing of the structural parameters of the optical fiber without damaging its physical structure during the inspection of finished hollow optical fiber products, thereby improving the accuracy and real-time performance of the measurement results. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A schematic diagram of an optical fiber provided for this application;

[0018] Figure 2 A schematic diagram of the cross-section of an optical fiber provided in this application;

[0019] Figure 3 A schematic diagram of an embodiment of the optical fiber structural parameter measurement system provided in this application;

[0020] Figure 4 A schematic diagram showing the position of the optical fiber structural parameter measurement system provided in this application and the optical fiber itself.

[0021] Figure 5 A flowchart illustrating an embodiment of the optical fiber structural parameter measurement method provided in this application;

[0022] Figure 6 A schematic diagram of the spectrum of the scattered light provided in this application;

[0023] Figure 7 A schematic diagram of the spectrum processing flow of the processing device provided in this application;

[0024] Figure 8 A schematic diagram of the structure of an embodiment of the measurement model provided in this application;

[0025] Figure 9 A frequency domain schematic diagram of the signal for the spectrum provided in this application;

[0026] Figure 10 The correspondence between the low-frequency retention terms and the loss of the spectrum provided in this application;

[0027] Figure 11 A schematic diagram comparing the MSE of the measurement model provided in this application with the MSE of calculation results from other methods;

[0028] Figure 12 A schematic diagram comparing the RMSE of the calculation results of the measurement model provided in this application with the RMSE of calculation results from other methods;

[0029] Figure 13 A schematic diagram comparing the MAE of the measurement model provided in this application with the MEA of the calculation results from other methods;

[0030] Figure 14 R-squared of the calculation results of the measurement model provided in this application 2 R-squared results from calculations using other methods 2 A comparison diagram;

[0031] Figure 15 A flowchart illustrating an embodiment of the training method for the measurement model provided in this application;

[0032] Figure 16 A schematic diagram of an embodiment of the optical fiber structural parameter measuring device provided in this application;

[0033] Figure 17 A schematic diagram of the structure of an embodiment of the electronic device provided in this application. Detailed Implementation

[0034] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0035] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented, for example, in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0036] Optical fiber is an optical transmission medium consisting of a core, cladding, and outer sheath. Based on the principle of total internal reflection, it can transmit optical signals over long distances with extremely low loss. With its significant advantages such as large capacity, high bandwidth, and resistance to electromagnetic interference, it has become a key technology carrier for achieving efficient information transmission and precise control in many fields such as modern communications, medicine, and industry.

[0037] Hollow-core anti-resonant fiber (HC-ARF) is a microstructured optical fiber with an air core. Its optical signals are mainly transmitted in the air holes. It has ultra-low latency, high power carrying capacity and wideband resonance characteristics, and is widely used in high-power laser transmission, nonlinear optics, quantum communication and biosensing.

[0038] In actual production, the capillary cladding structural parameters (such as wall thickness, diameter, and circumference) of hollow antiresonant optical fibers directly affect their light guiding performance and mechanical strength. For example, in high-power laser transmission scenarios, even slight deviations in capillary wall thickness can lead to optical signal loss or fiber melting damage; in the field of biosensing, the accuracy of the capillary diameter determines the sensitivity and resolution of the sensor.

[0039] For example, Figure 1 A schematic diagram of an optical fiber provided in this application is shown below. Figure 1 The optical fiber shown extends horizontally, as... Figure 1 The optical fiber 10 shown is a hollow-core optical fiber, specifically a hollow-core anti-resonant optical fiber. The optical fiber 10 includes multiple sets of capillaries and / or capillary sleeve groups, wherein the multiple sets of capillaries are arranged at equal angular intervals, and / or the multiple sets of capillary sleeve groups are arranged at equal angular intervals.

[0040] In one embodiment, the optical fiber 10 may also be a hollow-core optical fiber preform, which includes a prefabricated structure, also referred to as a capillary. Specifically, the hollow-core optical fiber preform may include multiple sets of capillary tubes and / or capillary sleeves.

[0041] Figure 2 A schematic diagram of the cross-section of an optical fiber provided in this application shows... Figure 1 The interface of fiber 10 in the A direction.

[0042] Specifically, such as Figure 2 The optical fiber 10 shown exhibits a circular cross-sectional structure, with four groups of capillaries at different positions arranged at equal angles inside the fiber, for example... Figure 2 The optical fiber is composed of four groups of capillaries spaced 90 degrees apart. This equiangular arrangement gives the fiber a high degree of structural symmetry, which helps light propagate uniformly within the fiber and reduces light scattering and loss caused by structural asymmetry. F1-F4, as labeled in the diagram, represent a group of capillaries, designated as the first capillary group Z1. Each capillary has a perfectly circular cross-section. Capillary F1 is set alone, while capillaries F2, F3, and F4 are nested sequentially to form a capillary group. The other groups of capillaries are designated as the second group Z2, the third group Z3, and the fourth group Z4, respectively. The circular capillaries are relatively easy to precisely control during manufacturing, and their regular geometry allows for better anti-resonance effects. The relatively simple boundary conditions of the circular shape result in a more regular interaction between light and the capillary walls during light transmission, which is beneficial for optimizing the fiber's transmission performance. Meanwhile, the different radii of the nested capillaries allow for precise control of optical fiber transmission characteristics, such as transmission bandwidth and loss, by setting the radius of each ring of capillaries and the spacing between rings.

[0043] exist Figure 2 In the example, the cross-section of fiber 10, the capillary tube, and the capillary tube assembly are all circular. In other examples, the cross-section of fiber 10, the capillary tube, and the capillary tube assembly can also be elliptical. Different structural parameters need to be inferred for different fiber cross-sections. For example, if the cross-section is circular, the inferred structural parameters are the circumference or diameter of the capillary tube; if the cross-section is elliptical, the inferred structural parameters are the major axis, minor axis, and circumference of the capillary tube.

[0044] More specifically, such as Figure 2 The core principle of fiber optic signal transmission shown is based on the anti-resonance effect. When light propagates in the fiber, if the wavelength of the light matches the resonant wavelength of the capillary, the light will undergo strong reflection and scattering at the capillary wall, leading to increased optical loss. However, when the wavelength of the light is in the anti-resonance region, the light can pass through the capillary wall more smoothly, achieving low-loss transmission in the hollow core. The circular and equally angled capillary structure helps to form stable anti-resonance conditions, enabling light within a specific wavelength range to transmit efficiently in the fiber.

[0045] Furthermore, the light-guiding performance of optical fibers during signal transmission depends heavily on the geometric accuracy of their cladding structure. However, due to the influence of various factors such as pressure, temperature, rotation, and traction speed, the wall thickness and diameter of the capillaries in optical fibers often exhibit slight deviations. Therefore, during the manufacturing process of optical fibers, it is necessary to measure structural parameters such as the wall thickness and diameter of the capillaries to ensure the effectiveness of the resulting optical fibers.

[0046] In one existing technology, the method for measuring the structural parameters of optical fibers is a destructive testing method, specifically scanning electron microscopy. This method requires cutting the optical fiber sample and placing it in a vacuum environment for observation. Although it can provide high-resolution structural images, it has drawbacks such as complex sample preparation, inability to monitor online, and damage to the integrity of the optical fiber, making it difficult to meet the needs of continuous production.

[0047] In another existing technology, the structural parameters of optical fibers are measured using a non-contact physical detection method, specifically whispering-gallery mode (WGM) spectroscopy: the WGM in the capillary wall is excited by lateral illumination, the fluorescence spectrum is recorded, and the diameter is inverted based on the spectral peaks. This method relies on precise control of the incident angle and polarization state and is easily affected by environmental noise, resulting in large fluctuations in the measurement results.

[0048] It can be seen that existing methods for measuring the structural parameters of optical fibers suffer from problems such as the inability to perform measurements without damaging the fiber, poor real-time performance, and low accuracy, making it difficult to meet the requirements for measuring the structural parameters of optical fibers. Therefore, how to measure the structural parameters of optical fibers in a timely and non-destructive manner during the manufacturing process, and how to perform non-destructive testing of the structural parameters of optical fibers during the inspection of finished optical fibers, are technical problems that need to be solved in this field.

[0049] Based on this, the technical solution provided in this application constructs and trains a measurement model to measure the structural parameters of the capillary inside the optical fiber, thereby improving the real-time performance and accuracy of the measurement results without damaging the physical structure of the optical fiber. This solves the above-mentioned defects of the prior art and enables more effective, timely and non-destructive measurement of the structural parameters of the optical fiber during the manufacturing process, as well as non-destructive testing of the optical fiber's parameters during the inspection of finished optical fiber products.

[0050] The technical solutions of this application will be described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0051] Figure 3 A schematic diagram of an embodiment of the optical fiber structural parameter measurement system provided in this application is shown below. Figure 3 The fiber optic structural parameter measurement system 20 shown includes:

[0052] The laser emitting device 201 is used to emit laser light to the optical fiber 10 (or simply optical fiber 10) to be tested.

[0053] The laser receiver 202 is used to receive the spectral information of the scattered light after the laser passes through the optical fiber 10 to be tested.

[0054] Specifically, when a laser is incident from the side of the optical fiber 10 under test, whispering gallery modes (WGMs) are excited in a set of capillary walls within the capillary tubes and / or capillary sleeves of the optical fiber 10. After the laser propagates for at least one revolution within this set of capillary walls and interferes, it is emitted as scattered light. The laser receiving device 202 can generate spectral information based on the scattered light, where the horizontal axis of the spectral information represents the frequency of the laser, and the vertical axis represents the intensity of the scattered light.

[0055] The processing device 200 is connected to the laser transmitting device 201 and the laser receiving device 202 respectively. It is used to control the parameters of the laser emitted by the laser transmitting device 201 to the optical fiber 10 to be tested. The laser parameters include the wavelength, frequency, and emission time of the laser. It also controls the position and other receiving parameters of the laser receiving device 202 so that the laser receiving device 202 receives the scattered light after the laser passes through the optical fiber 10 to be tested and determines the spectral information of the scattered light.

[0056] In one embodiment, the processing device 200 may be set independently of the laser emitting device 201 and the laser receiving device 202, or in another embodiment, the processing device 200 may also be set in the laser receiving device 202.

[0057] In one embodiment, the processing device 200 can control the position and other receiving parameters of the laser receiving device 202, so that the laser receiving device 202 can receive the scattered light at a more suitable position and improve the quality of the received scattered light.

[0058] Figure 4 The schematic diagram of the optical fiber structural parameter measurement system and the optical fiber provided in this application shows the relative positional relationship between the laser emitting device 201, the laser receiving device 202 and the optical fiber 10 in the optical fiber structural parameter measurement system.

[0059] like Figure 4 As shown, for fiber 10, according to Figure 2 The plane containing the cross-section shown in the figure establishes an xy coordinate system. The optical fiber 10 extends perpendicularly to the z direction of the xy plane, thus forming an xyz coordinate system.

[0060] A laser emitting device 201 is disposed on the first side of the optical fiber 10, and is used to emit laser light in a direction parallel to the cross-section of the optical fiber 10, toward the center of the cross-section of the optical fiber 10. Specifically, the first side is... Figure 4 If the laser emitter 201-1 is positioned on the left side of the diagram, it emits a laser along the x-direction towards the optical fiber 10 to be tested.

[0061] Meanwhile, the first capillary sleeve Z1 in the optical fiber 10 to be tested faces the laser emitting device 201. The laser emitted by the laser emitting device 201 into the optical fiber 10 to be tested enters the capillaries F1, F2, F3 and F4 in the first capillary sleeve Z1 and is emitted after traveling along the capillary wall for at least one cycle. At this time, the laser scattered by other groups of capillaries in the optical fiber 10 to be tested is equivalent to an interference signal and is not included in the subsequent process of determining the structural parameters of the optical fiber.

[0062] A laser receiver 202 is disposed on the second side of the optical fiber 10 and is used to receive laser light in a direction parallel to the cross-section of the optical fiber and perpendicular to the incident direction of the laser light, toward the center of the cross-section of the optical fiber 10. Figure 4 In the example shown, the second side is specifically in Figure 4 The laser receiver 202 is positioned on the upper side of the irradiation plane and is set perpendicular to the direction of light irradiation. Specifically, the laser receiver 202 can be a time-domain detector that continuously detects electromagnetic wave signals. The scattered light formed by the laser in the optical fiber 10 can be received by the time-domain detector. The laser receiver 202 can calculate the spectral information of the scattered light based on the received scattered light.

[0063] The laser receiver 202 is positioned in the positive y-axis direction. The laser receiver 202 can receive the scattered light generated by the laser emitted by the laser emitter 201-1 after scattering by capillaries F1, F2, F3, and F4 in the first capillary sleeve Z1. After determining the spectral information of the scattered light, the receiver 202 sends it to the processing device 200. The processing device 200 then determines the circumference of the capillaries F1, F2, F3, and F4 included in the first capillary sleeve Z1 that scatters the laser light in the optical fiber, based on the received spectral information of the scattered light.

[0064] It should be noted that, as Figure 4 and Figure 5 In the illustrated system, taking the first capillary sleeve Z1 of the optical fiber 10 facing the laser emitting device 201 as an example, the optical fiber structural parameter measurement system can be used to measure the circumference of multiple capillaries in the first capillary sleeve Z1 of the optical fiber 10. The optical fiber structural parameter measurement system can adjust the positional relationship between the optical fiber 10 and the laser emitting device 201 so that any group of capillary sleeves in the optical fiber 10 faces the laser emitting device 201, thereby achieving the measurement of the structural parameters of different groups of capillaries in the optical fiber 10. For example, in one embodiment, in the optical fiber structural parameter measurement system, the processing device 200 can be used to keep the position of the optical fiber 10 constant and control the laser emitting device 201 to adjust to different positions.

[0065] Figure 5 A flowchart illustrating an embodiment of the optical fiber structural parameter measurement method provided in this application is shown below. Figure 5 The method shown can be specifically applied to, for example... Figure 3 The structural parameter measurement device 20 for the optical fiber is shown, and is specifically executed by the processing device 200. Specifically, as... Figure 5 The methods for measuring the structural parameters of the optical fiber shown include:

[0066] S101: Emits a laser beam into the optical fiber to be tested.

[0067] In one embodiment, in response to receiving a measurement command for the optical fiber 10 to be tested, the processing device 200 controls the laser emitting device 201 to emit a laser beam toward the optical fiber 10 to be tested. (See reference...) Figure 4 As shown in the structure, the laser emitting device 201 can emit a short-pulse femtosecond laser beam perpendicularly to the side of the optical fiber 10.

[0068] In one embodiment, the laser emitted by the laser emitting device 201 to the optical fiber 10 to be tested can be an equally spaced laser light wave.

[0069] S102: Obtain the spectral information of the scattered light after the laser passes through the optical fiber to be tested.

[0070] In one embodiment, the processing device 200 responds to the laser emitting device 201 emitting a laser towards the optical fiber 10 to be tested, controls the laser receiving device 202 to receive the scattered light after the laser passes through the optical fiber to be tested, and determines the spectrum of the scattered light based on the scattered light.

[0071] For example, Figure 6 A schematic diagram of the scattered light spectrum provided in this application, as shown below. Figure 6 The image shows the spectrum of scattered light formed after scattering by capillaries F1, F2, F3, and F4 in the first capillary assembly Z1, acquired by the laser receiving device 202. It can be seen that within a certain frequency range of the laser light wave, capillaries with different structures produce different spectra of scattered light for the same laser. The spectrum should ideally include the superposition information of all laser light waves during refraction. By selecting 1000 discrete points within the aforementioned laser light wave frequency range for irradiation, the scattered light is scattered as shown in the image. Figure 6 The spectrum shown has the laser frequency on the horizontal axis and the intensity of the scattered light on the vertical axis. Therefore, the processing device 200 can acquire the spectrum of the scattered light transmitted by the laser receiving device 202, and determine the structural parameters of the optical fiber 10 based on the spectrum of the scattered light. The structural parameters specifically include the circumference of each capillary in the corresponding set of capillaries.

[0072] S103: Input the spectral information of the scattered light obtained in S102 into the measurement model to obtain the structural parameters of the fiber 10 to be tested output by the measurement model.

[0073] Figure 7 The schematic diagram of the spectral processing flow of the processing apparatus provided in this application is shown below. Figure 7 As shown, after acquiring the spectrum of the scattered light obtained by the laser receiving device 202, the processing device 200 can input the spectrum into the measurement model so that the measurement model can process the spectrum.

[0074] In one embodiment, the measurement model may specifically be a Fourier neural operator network. The measurement model specifically determines the correlation between spectral information and the fiber capillary through a deep learning convolutional model with Fourier transform, and predicts the structural parameters of the capillary. The processing device 200 can then obtain the structural parameters of the fiber 10 output by the neural operator network.

[0075] Specifically, Figure 8 A schematic diagram of the structure of an embodiment of the measurement model provided in this application is shown below. Figure 8 As shown, the measurement model provided in this application embodiment includes:

[0076] The input processing module is used to map spectral information to a target space, where the target space can be a high-dimensional feature space.

[0077] At least one convolutional module is connected to the input processing module, wherein the convolutional module is used to extract low-frequency features of spectral information in the target space through frequency domain convolution, and multiple convolutional modules sequentially extract low-frequency features of the spectrum to form a feature vector of spectral information.

[0078] The output inference module, connected to at least one convolutional module, is used to infer the structural parameters of the optical fiber to be detected corresponding to the feature vector of the spectral information.

[0079] In one embodiment, the input processing module provided in this embodiment includes a linear projection layer (or Linear layer) for mapping the original signal of the spectrum to a high-dimensional feature space, thereby providing sufficient expressive capacity for operator operations in the subsequent convolution module.

[0080] For example, taking a one-dimensional spectral array [1, 1000] of size 1000 obtained by a measurement model as an example, a linear projection layer can be used to map the one-dimensional spectral array [1, 1000] to a high-dimensional feature space to obtain a high-dimensional spectral array [1, width, 1000]. Here, width is the size of the mapped high-dimensional feature space, which can also be called the number of channels. It can be a preset value or can be configured in advance. In this application embodiment, the specific size of the number of channels is not limited.

[0081] In one embodiment, the at least one convolutional module provided in this embodiment consists of N consecutive convolutional modules, each of which adopts a dual-path feature extraction structure, combining global dependency and global feature capture.

[0082] Specifically, such as Figure 8 As shown, the convolution module provided in this embodiment specifically includes:

[0083] Local convolutional units are used to extract local features of the spectrum through local convolution.

[0084] The frequency domain convolution unit is used to extract low-frequency features of the spectrum through frequency domain convolution, thereby obtaining the feature vector of spectral information.

[0085] The normalization unit (or BatchNorm1d unit) connects the local convolutional unit and the frequency domain convolutional unit respectively. It is used to fuse the local features of the spectral information output by the local convolutional unit and the low-frequency features of the spectral information output by the frequency domain convolutional unit through residual connections to obtain the feature vector of the spectral information. The fused feature vector is then subjected to one-dimensional batch normalization to stabilize the numerical distribution.

[0086] The activation function unit (or ReLU unit) is connected to the normalization unit and is used to introduce nonlinear features into the feature vector of spectral information.

[0087] More specifically, such as Figure 8 As shown, the frequency domain convolution unit in the convolution module provided in this embodiment specifically includes:

[0088] The Fourier transform subunit (or FFT subunit) is used to perform Fourier transform on spectral information. Specifically, by performing Fourier transform on spectral information, the spectral information can be simplified, which is equivalent to subtracting from the spectral information and removing content that is irrelevant to the features.

[0089] In one embodiment, since the frequency domain of the spectral information is used to indicate the correspondence between the frequency and intensity of light, after performing a Fourier transform on the spectral information, the distribution with τ (time / optical path) as an independent variable can be obtained based on the spectrum S(ν) of the frequency ν of the spectral information.

[0090] The frequency domain selection subunit, connected to the Fourier transform subunit, is used to select low-frequency features of spectral information in the frequency domain and remove high-frequency features of spectral information.

[0091] The inverse Fourier transform subunit (or IFFT subunit) is connected to the frequency domain selection subunit and is used to perform an inverse Fourier transform on the selected low-frequency features to obtain the feature vector of the spectral information, thereby maintaining the consistency of the physical quantity properties of the input and output information of the frequency domain convolution unit.

[0092] Figure 9 A frequency domain schematic diagram of the signal for the spectrum provided in this application, as shown below. Figure 9 As shown, the frequency domain signal after Fourier transform of the spectral information shows that reading the corresponding peaks after Fourier transform of the spectral information can obtain more accurate features. Moreover, from the frequency domain distribution of the spectral information, it can be seen that the data features are concentrated in the low-frequency region. Therefore, the frequency domain selection unit can be used to select the low-frequency part of the spectral information and remove the high-frequency part of the spectral information to obtain the low-frequency features of the spectrum. Subsequent calculations using the low-frequency features can effectively reduce the amount of computation, improve the training speed, and eliminate high-frequency noise interference.

[0093] Figure 10 The correspondence between the low-frequency retention terms and the loss of the spectral signal provided in this application is extracted. Figure 9 For low-frequency information points in the mid-frequency range of [16~64], the model performs well when the value of the low-frequency retention term k is in the range of [37, 50]. The low-frequency retention term can be preset or configured through experiments. For example, after experiments on different optical fibers, it is preferable that the value of the low-frequency retention term k is in the range of [36, 48], which can ensure a more effective overall prediction result of the measurement model.

[0094] In one embodiment, if the proportion of high-frequency feature data removed by the frequency domain selection subunit in the feature data of the spectral information is greater than 50%, then the proportion of low-frequency feature data retained in the feature data of the spectral information is less than 50%, thereby more effectively removing high-frequency jitter interference in the spectral information, and improving the speed and efficiency of processing the spectral information while ensuring the feature accuracy of the spectral information.

[0095] In one embodiment, the frequency domain selection subunit retains the lower frequency information before the determined low-frequency retention item k, and sets the higher frequency information after the low-frequency retention item k to a preset value, wherein the preset value can be 0.

[0096] In one embodiment, the frequency domain selection subunit processes the low-frequency portion of the spectrum obtained after removing the high frequencies using a linear convolution kernel to obtain the low-frequency features of the spectrum. These low-frequency features contain the main trends and global structural information of the spectrum. Through linear convolution kernel processing, the frequency domain selection subunit can further extract key modes and related information from these low-frequency features, remove redundant low-frequency noise or unimportant low-frequency components, and increase the nonlinear expressive power of the network. When the low-frequency features processed by the linear convolution kernel are converted back to the time domain using an inverse Fourier transform, the task-related time-domain signal can be reconstructed more accurately, which helps to more accurately determine the structural parameters of the optical fiber in subsequent steps.

[0097] In one embodiment, local convolutional units are used to extract local features of the spectrum, which are then used as residuals in the convolutional kernel. This can prevent gradient vanishing and network degradation problems, make training more stable, and provide local features on the basis of low-frequency features, so that the model can obtain more sufficient feature information.

[0098] Understandably, in Figure 8 In the measurement model shown, each of the multiple convolutional modules, such as convolutional module 1, convolutional module 2, ..., convolutional module N, can be implemented in the same way as convolutional module 1. Multiple convolutional modules extract features sequentially, and can repeatedly refine frequency domain features in different dimensions, thereby learning complex nonlinear mapping operators.

[0099] In one specific implementation, for example... Figure 2 The optical fiber shown includes four sets of capillaries. The number of multiple convolution modules can be set to 4, or it can be set in advance.

[0100] For example, taking the high-dimensional spectral array [1, width, 1000] provided by the input processing module in the measurement model as an example, after processing by multiple convolution modules, the dimension of the feature vector of the output spectrum is also [1, width, 1000].

[0101] In one embodiment, the output prediction module provided in this application includes:

[0102] Pooling layers are used to compress the feature vectors obtained by multiple convolutional modules in a high-dimensional feature space into feature vectors of a preset dimension, while retaining the most significant feature components and enhancing the model's robustness to changes in input length.

[0103] In one embodiment, the pooling layer may specifically be an adaptive average pooling layer (or referred to as an AdaptiveAvgPool1d layer).

[0104] For example, if the feature vector of the spectrum output by multiple convolutional modules also has dimensions [1, width, 1000], after processing by the pooling layer, the feature vector with the preset dimension is [1, width].

[0105] The regression module (or Regression) is used to map the extracted feature vectors of a preset dimension to the output space, forming and outputting the predicted results of the structural parameters of the fiber 10 to be detected.

[0106] In one embodiment, the structural parameters include at least one of parameters such as the perimeter and diameter of the optical fiber 10 to be tested. When the optical fiber 10 to be tested is an elliptical hollow-core optical fiber, the structural parameters may also include at least one of parameters such as the major axis, minor axis, and roundness.

[0107] In one embodiment, the regression module may specifically be a multilayer perceptron (MLP), which, by introducing hidden layers and nonlinear activation functions, can learn the complex nonlinear relationship between input and output, and is suitable for regression and classification tasks.

[0108] For example, taking a feature vector with a preset dimension of [1, width] as an example, after processing by the regression module, the predicted dimension of the fiber's structural parameters is obtained as [B, out_dim], where out_dim is the output dimension. Specifically, this is combined with... Figure 2 The structure of the optical fiber shown can have its output dimension out_dim set to 4, that is, 4 dimensions, which are used to indicate the structural parameters of capillaries F1, F2, F3 and F4 in the first capillary sleeve Z1.

[0109] In one embodiment, the structural parameter is specifically the perimeter of the capillary. The structural parameters of the optical fiber that the measurement model can output include: the perimeter of capillary F1, the perimeter of capillary F2, the perimeter of capillary F3, and the perimeter of capillary F4.

[0110] In a specific implementation, the circumference of the capillary calculated through the embodiments of this application is close to the actual circumference of the capillary, with an error of about 0.5%.

[0111] More specifically, Table 1 shows an error metric between the perimeter of the capillary calculated in the embodiments of this application and the actual perimeter of the capillary.

[0112]

[0113] Where MSE (Mean Squared Error) is the average of the squared errors, RMSE (Root Mean Squared Error) is the square root of MSE, MAE (Mean Absolute Error) is the average of the absolute values ​​of the errors, and R... 2 (Coefficient of Determination) is the proportion of data variance explained by the model.

[0114] Figure 11 The diagram comparing the MSE of the measurement model provided in this application with the MSE of other calculation methods shows that the MSE of the measurement model in this application is smaller, and therefore the perimeter of the capillary obtained is closer to the actual perimeter of the capillary.

[0115] Other methods include:

[0116] The MLP model extracts nonlinear transformations and abstract representations of the spectral information of the optical fiber under test. Specifically, the MLP model can be used in fully connected layers to progressively transform the original input features into more abstract and discriminative high-level features. In its implementation, the MLP model consists of an input layer, hidden layers (multiple fully connected linear layers), and an output layer. After the input layer, z-score normalization is performed. The hidden layer structure also includes LayerNorm normalization, activation functions, and Dropout layers. The MLP model can then be used to output the structural parameters of the optical fiber under test through the output layer.

[0117] 1D-CNN models can be used to extract local features of the spectral information of an optical fiber under test. Specifically, 1D-CNN models can capture local dependencies and important patterns in sequential data through convolutional kernels. In a concrete implementation, a 1D-CNN model consists of an input layer, hidden layers, and an output layer. After the input layer, z-score normalization is performed. The hidden layer structure includes a one-dimensional convolutional layer, normalization, activation function, max-pooling layer, adaptive average pooling layer, and a fully connected regression head. The 1D-CNN model can then output the structural parameters of the optical fiber under test.

[0118] The Transformer model can be used to extract global sequence features of the spectral information of the optical fiber under test. Specifically, the Transformer model uses a multi-head attention mechanism to capture long-range dependencies between different wavelengths in the spectral sequence. In its implementation, the Transformer model includes a linear projection layer, a position encoding and dropout layer, and a Transformer encoder layer. The Transformer encoder layer consists of multiple layers (num_layers), internally containing a multi-head attention mechanism, a feedforward neural network, a final global average pooling layer, and an output layer. The Transformer model can then output the structural parameters of the optical fiber under test.

[0119] Figure 12 The diagram comparing the RMSE of the measurement model provided in this application with the RMSE of calculation results from other methods shows that the RMSE of the measurement model in this application is smaller, and therefore the perimeter of the capillary obtained is closer to the actual perimeter of the capillary.

[0120] Figure 13 A comparison diagram of the MAE of the measurement model provided in this application and the MEA of the calculation results from other methods shows that the MEA of the measurement model in this application is smaller, and therefore the perimeter of the capillary obtained is closer to the actual perimeter of the capillary.

[0121] Figure 14 R-squared of the calculation results of the measurement model provided in this application 2 R-squared results from calculations using other methods 2 The comparison diagram shows that the R of the measurement model in this application is... 2 It is smaller, so the circumference of the obtained capillary is closer to the actual circumference of the capillary.

[0122] In this embodiment, a circular cross-section of the optical fiber is used as an example. A measurement model is pre-trained based on the spectral data of the circular optical fiber, and the structural parameters of the optical fiber are inferred based on the measurement model. In other embodiments, the cross-section of the optical fiber can also be elliptical or other shapes. Only the spectral data of the elliptical optical fiber needs to be replaced to train the measurement model, thereby inferring the structural parameters of the optical fiber based on the measurement model corresponding to the design structure of the optical fiber.

[0123] In summary, the fiber optic structural parameter measurement method provided in this embodiment emits a laser into the fiber to be tested and obtains the spectral information of the scattered light after the laser passes through the fiber. The spectral information is then input into a measurement model to obtain the fiber optic structural parameters predicted by the measurement model. Since the measurement model extracts the low-frequency features of the spectral information to obtain the feature vector of the spectral information, and determines the fiber optic structural parameters based on the feature vector of the spectral information, this application achieves non-destructive testing of hollow fiber optics by predicting the fiber optic structural parameters based on spectral information without damaging the hollow fiber structure.

[0124] Furthermore, this application effectively handles structures under complex nonlinear mappings, and compared with other neural network methods, it considers the fusion of local and global features, resulting in more comprehensive feature selection. Through comparison of multiple error metrics, it can be seen that this embodiment significantly improves the accuracy of the measurement results of the structural parameters of optical fibers, thereby enabling more effective measurement of the structural parameters of optical fibers during the production and manufacturing process.

[0125] Meanwhile, the measurement model provided in this application filters out high-frequency features and extracts only low-frequency features as the basic data for inferring fiber structure parameters. This ensures that the measurement model and method of this application are insensitive to noise during the measurement process, have good noise resistance and robustness, and also incorporate physical properties to enhance the stability of the measurement model under different application scenarios. This ensures the reliability and stability of the prediction results of fiber structure parameters, and ultimately improves the robustness and automation level of the measurement results without damaging the physical structure of the fiber.

[0126] In particular, when this application is applied to real-time online calculations integrated into the hollow-core optical fiber manufacturing process, it can both measure the structural parameters of the optical fiber in a timely and non-destructive manner during the production and manufacturing process, meeting the needs of continuous production, and also perform non-destructive testing of the structural parameters of the optical fiber during the inspection of finished hollow-core optical fiber products without damaging the physical structure of the fiber, thus improving the accuracy and real-time performance of the measurement results. In one embodiment, such as Figure 3 The system shown includes a processing unit 200 that also feeds back the inferred structural parameters of the optical fiber to be tested to the target device or personnel, enabling the target device or personnel to determine the structural parameters more in real time and to perform subsequent operations more effectively based on the structural parameters.

[0127] In one embodiment, the target device may be an optical fiber drawing control module. The optical fiber drawing control module is used to draw the optical fiber to be tested. After receiving the structural parameters of the optical fiber to be tested provided by the processing device 200, the optical fiber drawing control module can adjust the drawing parameters when drawing the optical fiber to be tested according to the structural parameters. In this way, the structural parameters of the optical fiber can be measured in a timely and non-destructive manner during the drawing process, and the real-time adjustment of the drawing parameters can be realized, thereby improving the quality of the formed optical fiber.

[0128] In one embodiment, the target device may be an optical fiber testing module, which can be used to test the finished optical fiber. The optical fiber testing module can determine the structural parameters of the optical fiber to be tested based on the structural parameters of the optical fiber to be tested provided by the processing device 200, thereby non-destructively testing the structural parameters of the optical fiber during the testing process of the finished optical fiber.

[0129] Figure 15 A flowchart illustrating an embodiment of the method for measuring the structural parameters of an optical fiber provided in this application is shown below. Figure 15 The method shown can be used to train a measurement model as provided in any of the foregoing embodiments of this application. This training method can be executed by the processing device 200 or by other devices with relevant data processing capabilities, specifically, such as... Figure 5 The methods shown include:

[0130] S201: Acquire training data, which includes the spectral information of the scattered light after the laser emitted into the fiber under test passes through the fiber under test, as well as the structural parameters of the fiber under test.

[0131] In one embodiment, the training data is obtained based on the design structure of the optical fiber under test, or the training data is obtained through actual measurement of the optical fiber under test. Alternatively, the training data includes a portion of data obtained through training based on the design structure of the optical fiber under test, and another portion of data obtained through actual measurement of the optical fiber under test.

[0132] Specifically, in combination with, for example Figure 4 The fiber optic structural parameter measurement system shown has a processing unit 200 that can generate spectral information of the scattered light after the laser emitted by the laser emitting device 201 passes through the fiber optic cable, and the spectrum of the scattered light after the laser received by the laser receiving device 202 passes through the fiber optic cable. These data are then combined with known structural parameters of the fiber optic cable to form training data. For example, B training data can be represented as [B, 1000].

[0133] S202: The spectral information of the optical fiber to be tested in the training data is used as input data and the structural parameters of the optical fiber to be tested are used as output data to train the initial model structure of the measurement model until the training error is less than the preset value, and the training is completed to obtain the measurement model.

[0134] Specifically, in combination Figure 8 The measurement model structure shown is such that the processing device 200 sequentially uses the spectral information of the optical fiber to be tested in the training data as input data and the structural parameters of the optical fiber to be tested as output data to train the initial model structure of the measurement model until the training error is less than the preset value, thus completing the training and obtaining the measurement model.

[0135] In one embodiment, the emission angle of the laser emitted into the optical fiber to be tested in the training data is random. After generating the training data based on the design structure of the optical fiber to be tested, the processing device 200 can also normalize the training data to eliminate the influence of the angle of the laser emitted into the optical fiber on the structural parameters of the optical fiber.

[0136] Specifically, after acquiring training data and before training the model, during the pre-training stage, the processing device 200 can normalize each angle of the dataset separately to eliminate the influence of the spectral illumination angle θ on training and allow the model to focus on features other than the angle, thereby eliminating the offset differences caused by the angle and preserving the shape information.

[0137] In one embodiment, for the structure of the optical fiber provided in this application, the processing device 200 can specifically use the Z-score normalization method to achieve better normalization processing results.

[0138] Furthermore, since the measurement model trained from the normalized training data eliminates the influence of the incident angle θ, the processing device 200 does not need to know the specific angle θ of the incident light to perform calculations when determining structural parameters based on the measurement model. This eliminates the need for additional compensation for angle deviations during the training and prediction phases of the measurement model, thereby significantly improving the robustness of the model to the incident angle, reducing the complexity of the training data, and enhancing the stability of the model in different application scenarios, ensuring the reliability of the prediction results for the structural parameters of the optical fiber.

[0139] In one embodiment, the processing device 200 can also acquire training data formed by measuring the structural parameters of the optical fiber to be tested using the measurement model, and update the measurement model based on this training data to ensure the real-time performance and effectiveness of the measurement model, so that the measurement model can more accurately predict the structural parameters of the optical fiber.

[0140] In the foregoing embodiments of this application, the method for measuring the structural parameters of optical fibers provided in the embodiments of this application has been described. To implement the functions of the methods provided in the embodiments of this application, the processing device 200, as the executing entity, can be implemented through hardware structures and / or software modules, for example, in the form of hardware structures, software modules, or a combination of hardware structures and software modules. Whether a particular function is executed in the form of hardware structures, software modules, or a combination of hardware structures and software modules depends on the specific application and design constraints of the technical solution.

[0141] For example, Figure 16 A schematic diagram of an embodiment of the optical fiber structural parameter measuring device provided in this application is shown below. Figure 16 The device 1000 shown includes: a transmitting module 1001 for transmitting laser light to the optical fiber to be tested; an acquisition module 1002 for acquiring the spectral information of the scattered light after the laser passes through the optical fiber to be tested; and a measurement module 1003 for inputting the spectral information into a measurement model, enabling the measurement model to extract the low-frequency features of the spectral information, obtain the feature vector of the spectral information, and infer the structural parameters of the optical fiber to be tested based on the feature vector of the spectral information.

[0142] The optical fiber structural parameter measuring device provided in this embodiment can be used to perform the optical fiber structural parameter measuring method provided in the aforementioned embodiment. Its implementation method and principle are the same, and will not be described again.

[0143] It should be understood that the division of the various modules and units in the above device is merely a logical functional division. In actual implementation, they can be fully or partially integrated into a single physical entity, or they can be physically separated. Furthermore, these modules can be implemented entirely in software through processing element calls; they can be fully implemented in hardware; or some modules can be implemented by processing element calls to software, while others are implemented in hardware. For example, a module can be a separately established processing element, or it can be integrated into a chip in the above device. Alternatively, it can be stored as program code in the memory of the above device, and its functions can be called and executed by a processing element of the device. The implementation of other modules is similar. Moreover, these modules can be fully or partially integrated together, or they can be implemented independently. The processing element here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each of the above modules can be completed through the integrated logic circuits in the hardware of the processor element or through software instructions.

[0144] For example, these modules can be one or more integrated circuits configured to implement the above methods, such as one or more application-specific integrated circuits (ASICs), one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs). As another example, when a module is implemented using processing element scheduler code, the processing element can be a general-purpose processor, such as a central processing unit (CPU) or other processor capable of calling program code. Furthermore, these modules can be integrated together to implement a system-on-a-chip (SOC).

[0145] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0146] For example, Figure 17 A schematic diagram of the structure of an embodiment of the electronic device provided in this application is shown below. Figure 17 The electronic device 2000 shown can be used to execute the optical fiber structural parameter measurement method or the optical fiber structural parameter measurement model training method provided in any embodiment of this application.

[0147] In one embodiment, such as Figure 17The illustrated electronic device includes one or more processors 2001 and a memory 2002. The memory 2002 stores computer-executable instructions, and the processor 2001 can execute the computer-executable instructions stored in the memory 2002. When the computer-executable instructions are executed by the processor 2001, the processor 2001 implements the optical fiber structural parameter measurement method or the optical fiber structural parameter measurement model training method provided in any of the foregoing embodiments of this application.

[0148] In one embodiment, such as Figure 17 The electronic device shown also includes a communication interface 2003, through which the processor 2001 can communicate with other devices, such as sending and receiving data through the communication interface 2003.

[0149] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0150] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0151] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0152] This application also provides a chip for executing instructions, which is used to execute the optical fiber structural parameter measurement method or the optical fiber structural parameter measurement model training method provided in any of the foregoing embodiments of this application.

[0153] This application also provides a computer program product, including a computer program that, when executed, implements the optical fiber structural parameter measurement method or the optical fiber structural parameter measurement model training method provided in any of the foregoing embodiments of this application.

[0154] This application also provides a computer-readable storage medium storing computer-executable instructions. When executed, these computer-executable instructions can be used to implement the optical fiber structural parameter measurement method or the optical fiber structural parameter measurement model training method provided in any of the foregoing embodiments of this application.

[0155] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0156] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0157] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0158] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0159] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0160] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0161] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0162] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A method of measuring a structural parameter of an optical fiber, characterized by, include: Emit a laser beam toward the optical fiber to be tested; Obtain the spectral information of the scattered light after the laser passes through the optical fiber under test; The spectral information is input into the measurement model, which extracts the low-frequency features of the spectral information to obtain the feature vector of the spectral information. The structural parameters of the optical fiber to be tested are inferred based on the feature vector of the spectral information. The measurement model is pre-trained based on the spectral data of the design structure of the optical fiber to be tested, and the frequency range of the low-frequency features is determined during the pre-training of the measurement model.

2. The method according to claim 1, characterized in that, The measurement model includes: An input processing module is used to map the spectral information to a target space; At least one convolutional module is used to extract low-frequency features of the spectral information in the target space through frequency domain convolution to obtain a feature vector of the spectral information; An output inference module is used to infer the structural parameters of the optical fiber to be detected corresponding to the feature vector.

3. The method according to claim 2, characterized in that, The convolutional module includes: A frequency domain convolution unit is used to extract low-frequency features of the spectral information through frequency domain convolution to obtain a feature vector of the spectral information. The frequency domain convolutional unit specifically includes: The Fourier transform subunit is used to perform Fourier transform on the spectral information; A frequency domain selection subunit, connected to the Fourier transform subunit, is used to select a portion of low-frequency features in the frequency domain spectral information while removing a portion of high-frequency features from the spectral information, wherein the proportion of the removed high-frequency feature data in the feature data of the spectral information is greater than 50%. The inverse Fourier transform subunit, connected to the frequency domain selection subunit, is used to perform an inverse Fourier transform on the selected low-frequency features to obtain the feature vector of the spectral information.

4. The method according to claim 3, characterized in that, The convolution module also includes: Local convolutional units are used to extract local features of the spectral information through local convolution.

5. The method according to claim 4, characterized in that, The convolution module also includes: A normalization unit, connected to the frequency domain convolution unit and the local convolution unit respectively, is used to fuse the low-frequency features and local features of the spectral information to obtain the feature vector of the spectrum; And / or, activation function units, connected to the normalization units, are used to introduce nonlinear features into the feature vector of the spectral information.

6. The method according to claim 5, characterized in that, The output inference module includes: The pooling layer is used to compress the feature vector of the spectral information into a feature vector of a preset dimension. The regression module, connected to the pooling layer, is used to obtain the structural parameters of the optical fiber to be detected based on the feature vector of the preset dimension.

7. The method according to any one of claims 1-6, characterized in that, The optical fiber to be tested includes hollow optical fiber or hollow optical fiber preform; The optical fiber to be tested includes multiple sets of capillaries and / or multiple sets of capillary sleeves; The multiple sets of capillaries are arranged at equal angular intervals, and / or the multiple sets of capillary sleeves are arranged at equal angular intervals. The structural parameters of the optical fiber to be tested include the circumference of at least one set of capillaries in the capillary tube and / or capillary sleeve.

8. A system for measuring the structural parameters of an optical fiber, characterized in that, include: A laser emitting device used to emit laser light into the optical fiber to be tested; A laser receiving device is used to receive the scattered light of the laser after it passes through the optical fiber to be tested; The processing device is connected to the laser emitting device and the laser receiving device respectively, and is used to control the parameters of the laser emitted by the laser emitting device to the optical fiber to be tested, and to control the position of the laser receiving device, so as to obtain the spectral information of the scattered light after the laser passes through the optical fiber to be tested. The processing device is further configured to perform the structural parameter measurement method of the optical fiber as described in any one of claims 1-7, so as to infer the structural parameters of the optical fiber to be tested.

9. The system according to claim 8, characterized in that, The processing device is also used to feed back the structural parameters of the optical fiber to be tested to the target device. The target device includes an optical fiber drawing control module and an optical fiber detection module. The optical fiber drawing control module is used to adjust the optical fiber drawing parameters of the optical fiber to be tested according to the structural parameters. The optical fiber detection module is used to detect the structural parameters of the finished optical fiber.

10. The system according to claim 8 or 9, characterized in that, The laser emitting device is disposed on the first side of the optical fiber to be tested, and is used to emit laser light in a direction parallel to the cross-section of the optical fiber to be tested toward the center of the cross-section of the optical fiber to be tested. The laser receiving device is disposed on the second side of the optical fiber to be tested, and is used to receive the laser in a direction parallel to the cross-section of the optical fiber to be tested and perpendicular to the incident direction of the laser, toward the center of the cross-section of the optical fiber to be tested.

11. A method for measuring the structural parameters of an optical fiber, characterized in that, The method is used to train the measurement model in the optical fiber structural parameter measurement method as described in any one of claims 1-7, or to train the measurement model in the optical fiber structural parameter measurement system as described in any one of claims 8-10, the method comprising: The training data of the optical fiber to be tested is obtained. The training data includes the spectral information of the scattered light after the laser emitted into the optical fiber passes through the optical fiber, as well as the structural parameters of the optical fiber. The training data is obtained by simulation based on the design structure of the optical fiber and / or by actual measurement of the optical fiber. The initial model structure of the measurement model is trained by sequentially using the spectral information of the optical fiber to be tested in the training data as input data and the structural parameters of the optical fiber to be tested as output data until the training error is less than a preset value, thereby obtaining the measurement model.

12. The method according to claim 11, characterized in that, Also includes: Acquire training data formed by measuring the structural parameters of the optical fiber to be tested using the measurement model described above; The measurement model is updated based on the training data generated from the actual measurements.

13. The method according to claim 11 or 12, characterized in that, The emission angle of the laser emitted into the optical fiber to be detected in the training data is random; After generating training data based on the design structure of the fiber to be tested, the method further includes: The training data is normalized to eliminate the influence of the emission angle of the laser emitted into the optical fiber on the structural parameters of the optical fiber.

14. An electronic device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the method as described in any one of claims 1-7 or 11-13.

15. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1-7 or 11-13.

16. A computer program product, characterized in that, Includes a computer program, which, when executed by a processor, implements the method as described in any one of claims 1-7 or 11-13.

Images