Machine vision-based fiber bragg grating processing method and system

Abstract

The invention relates to the technical field of laser micro-nano processing, and discloses a method and a system for processing a fiber bragg grating based on machine vision, which realize automatic identification of a focal plane of an optical fiber core by combining advanced machine vision technology and neural network model technology. According to the invention, the laser focus can be automatically focused to the focal plane of the fiber core, so that the automation and intelligence of the processing process of the fiber bragg grating are realized, the method has the characteristics of high precision and high efficiency, also shows a wide application prospect, and provides powerful support for the development of the fields of fiber communication, sensing and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of laser micro-nano processing, and in particular to a processing method and system of a fiber Bragg grating based on machine vision. Background Art

[0002] Fiber Bragg Grating (FBG) has become an indispensable component in many high-tech fields such as communication network systems, sensor technology, medical applications, and fiber lasers due to its excellent precision, compact size, excellent electromagnetic tolerance, chemical inertness, excellent heat resistance, and powerful multiplexing capabilities. FBG is formed by implementing permanent periodic refractive index (RI) modulation inside the fiber core and exhibits unique narrowband Bragg reflection characteristics. This reflection characteristic is directly determined by the spacing of the periodic RI modulation, which can be precisely controlled during the manufacturing process.

[0003] Traditionally, FBGs are mainly manufactured by ultraviolet exposure using phase masks and interference lithography. These technologies have achieved remarkable success in the commercial field due to their simplicity and efficiency. However, with the continuous expansion of the application scope of FBGs and the growing demand, FBGs manufactured by traditional technologies have gradually exposed some limitations. Specifically, traditional FBGs cannot withstand high temperature environments exceeding 450°C and are prone to thermal attenuation in such environments. In addition, the choice of optical fiber core materials is limited by photosensitive materials, and additional hydrogen loading is usually required to enhance the photosensitivity of the optical fiber. Furthermore, different masks are required to make FBGs with different characteristics (such as period length and spectral characteristics), which undoubtedly increases the complexity and cost of manufacturing.

[0004] In view of this, researchers continue to explore new technologies in order to produce new FBGs that can overcome the above limitations. In 2001, Oi et al. pioneered the method of using near-infrared femtosecond (fs) lasers to make FBGs. With its extremely high peak intensity and ultra-short pulse width, femtosecond lasers can induce nonlinear absorption effects within the focal volume when used with high numerical aperture (NA) objectives, thereby achieving RI modification of almost all types of optical materials (including non-photosensitive materials such as pure silica, sapphire, etc.).

[0005] In addition, femtosecond laser direct writing (DLW) technology has also developed to the stage where FBG can be produced without a phase mask, and can achieve precise and flexible control of the grating period, shape, and RI modulation position. Using fs DLW technology, high-quality FBGs with high reflectivity, low insertion loss, stable operation in harsh environments such as high voltage and ionizing radiation, and high temperature resistance up to 1900°C have been successfully produced.

[0006] fs DLW technology can be subdivided into point-by-point (PbP), line-by-line (LbL) and facet-by-face technology. In PbP technology, the femtosecond laser is precisely focused to the center of the fiber core to permanently change the refractive index in the focal volume; LbL technology uses a focused femtosecond laser beam to write a series of parallel lines in the fiber core; and facet-by-face technology forms the grating plane by irradiating the entire core. Compared with the other two methods, PbP technology is favored due to its many advantages, including relatively low femtosecond pulse energy (10-100nJ) during the production process, higher production efficiency can be achieved by only point production, and more convenient control of the RI modulation period by changing the translation speed of the fiber or simply adjusting the femtosecond pulse repetition rate.

[0007] In PbP technology, a sub-micron femtosecond laser spot is focused to the center of the fiber core (specifically, the focal plane of the fiber core) with high precision to produce a periodic RI modulation spot array with a finite diameter. However, the current processing scheme still relies on the human eye to identify the center of the fiber core, which depends entirely on the operator's experience level. At the same time, this method has problems such as long processing time and low repetition rate because it cannot achieve automated processing.

[0008] Therefore, how to accurately and automatically align the laser focus with the center position of the optical fiber core has become a key technical difficulty and challenge in processing high-quality fiber Bragg gratings using PbP fsDLW technology, especially in achieving fast and high-repetition rate processing.

[0009] The above information is presented as background information only to assist with understanding the present disclosure and no determination or admission is made as to whether any of the above may be used as prior art with respect to the present disclosure. Summary of the invention

[0010] The present invention provides a method and system for processing a fiber Bragg grating based on machine vision to solve the problems existing in the prior art.

[0011] To achieve the above object, the present invention provides the following technical solutions:

[0012] In a first aspect, the present invention provides a method for processing a fiber Bragg grating based on machine vision, the method comprising:

[0013] At the starting point, scanning the optical fiber core and collecting an image at the starting point;

[0014] Inputting the image at the starting point into the trained neural network model, locating the focal plane of the optical fiber core at the starting point, and obtaining the optical fiber tilt angle;

[0015] Determining a termination point according to the optical fiber tilt angle and a given processing length;

[0016] At the termination point, scanning the optical fiber core and collecting an image at the termination point;

[0017] Inputting the image at the termination point into a trained neural network model to locate the focal plane of the optical fiber core at the termination point;

[0018] The laser focus is controlled to focus on the focal plane of the optical fiber core at the starting point, and the optical fiber core is processed until the focal plane of the optical fiber core at the ending point is processed to obtain a fiber Bragg grating.

[0019] Furthermore, in the method for processing a fiber Bragg grating based on machine vision, before the step of scanning the fiber core at the starting point and collecting the image at the starting point, the method further includes:

[0020] Choose any point in the fiber core as the starting point.

[0021] Furthermore, in the method for processing a fiber Bragg grating based on machine vision, the step of scanning the fiber core at the starting point and collecting the image at the starting point is specifically:

[0022] At the starting point, the optical fiber core is scanned within a range of ±30 μm, and an image at the starting point is collected.

[0023] Furthermore, in the machine vision-based fiber Bragg grating processing method, the step of determining the termination point according to the fiber inclination angle and the given processing length is specifically:

[0024] A given processing length is moved in a direction of tilting the optical fiber tilt angle to determine a termination point.

[0025] Furthermore, in the machine vision-based fiber Bragg grating processing method, the step of scanning the optical fiber core at the termination point and collecting the image at the termination point is specifically:

[0026] At the termination point, the optical fiber core is scanned within a range of ±30 μm, and an image at the termination point is collected.

[0027] Furthermore, in the method for processing a fiber Bragg grating based on machine vision, the method further comprises:

[0028] Build a neural network model;

[0029] Dividing the collected out-of-focus images and in-focus images into a training set and a validation set; the out-of-focus images include images that are focused too deeply and images that are focused too shallowly;

[0030] Inputting the training set into the neural network model to train the neural network model;

[0031] The verification set is input into the neural network model to verify the neural network model and obtain the trained neural network model.

[0032] Furthermore, in the method for processing a fiber Bragg grating based on machine vision, before the step of inputting a training set into the neural network model to train the neural network model, the method further includes:

[0033] Classify and label each image;

[0034] Perform background subtraction on each image;

[0035] The region of interest containing the two boundary lines of the fiber core was cropped from the background-subtracted image;

[0036] Grayscale values ​​of the boundary lines in the cropped region of interest are calculated to extract the calculation results as features of the boundary lines.

[0037] Further, in the machine vision-based fiber Bragg grating processing method, the neural network model includes an input layer, four dense layers and an output layer;

[0038] A nonlinear activation function is used between each two adjacent dense layers;

[0039] The first dense layer contains 32 neurons;

[0040] The second dense layer contains 64 neurons;

[0041] The third dense layer contains 64 neurons;

[0042] The fourth dense layer contains 3 neurons and uses the SoftMax function output.

[0043] Furthermore, in the method for processing a fiber Bragg grating based on machine vision, the step of inputting a training set into the neural network model to train the neural network model includes:

[0044] Assigning random values ​​to weights of the neural network model;

[0045] Input an image from the training set into the neural network model and perform forward propagation to generate and output a predicted value;

[0046] Calculate the loss between the predicted value and the corresponding true value, the loss function is the mean square error;

[0047] Using the calculated loss to update the weights of the neural network model;

[0048] Continue to input the next image in the training set into the neural network model for forward propagation, loss calculation and weight update until the preset termination condition is met.

[0049] In a second aspect, the present invention provides a fiber Bragg grating processing system based on machine vision, the system comprising a laser, a beam splitter, an oil mirror, a red light source, a CCD camera, a 3D electric stage and a computer; wherein,

[0050] The 3D electric stage is used to carry the optical fiber and drive the optical fiber to move;

[0051] The laser is located above the 3D electric stage and is used to emit a laser beam;

[0052] The oil mirror is located between the 3D electric stage and the laser, and is used to focus the laser beam;

[0053] The red light source is located below the 3D electric stage and is used to emit red illumination light to illuminate the optical fiber;

[0054] The beam splitter is located between the oil mirror and the laser, and is used to reflect the red illumination light to the CCD camera;

[0055] The CCD camera is located on the light-reflecting surface of the beam splitter and is used to collect the image of the optical fiber;

[0056] The computer is used to:

[0057] Controlling the movement of the 3D electric stage, and controlling the CCD camera to scan the core of the optical fiber at the starting point, and collecting the image at the starting point;

[0058] Inputting the image at the starting point into the trained neural network model, locating the focal plane of the core of the optical fiber at the starting point, and obtaining the optical fiber tilt angle;

[0059] Determining a termination point according to the optical fiber tilt angle and a given processing length;

[0060] Controlling the movement of the 3D electric stage, and controlling the CCD camera to scan the core of the optical fiber at the end point, and collecting an image at the end point;

[0061] Inputting the image at the termination point into a trained neural network model to locate the focal plane of the core of the optical fiber at the termination point;

[0062] The 3D electric stage is controlled to move so as to focus the laser focus of the laser to the focal plane of the core of the optical fiber at the starting point, and start processing the core of the optical fiber until the focal plane of the core of the optical fiber at the ending point is reached to obtain a fiber Bragg grating.

[0063] Compared with the prior art, the present invention has the following beneficial effects:

[0064] The present invention provides a method and system for processing a fiber Bragg grating based on machine vision, which realizes automatic recognition of the focal plane of the optical fiber core by combining advanced machine vision technology and neural network model technology, so as to automatically focus the laser focus on the focal plane of the optical fiber core, thereby realizing the automation and intelligence of the fiber Bragg grating processing process. It not only has the characteristics of high precision and high efficiency, but also shows broad application prospects, providing strong support for the development of optical fiber communication, sensing and other fields.

[0065] The present invention has other features and advantages, which will be apparent from or will be described in detail in the accompanying drawings and the following detailed description incorporated herein, which together serve to explain the specific principles of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0067] Figure 1It is a flow chart of a method for processing a fiber Bragg grating based on machine vision provided in the first embodiment of the present invention;

[0068] Figure 2 are example images of different categories mentioned in the first embodiment of the present invention;

[0069] Figure 3 is a schematic diagram of the process of denoising an image mentioned in the first embodiment of the present invention;

[0070] Figure 4 is a schematic diagram of the structure of the neural network model mentioned in the first embodiment of the present invention;

[0071] Figure 5 In the figure, a is a graph of training and validation loss versus the number of training cycles, b is a statistical graph of fiber core identification time based on repeated measurements, and c is a comparison of fiber core identification accuracy with and without denoising.

[0072] Figure 6 a is a microscopic image of a periodic RI modulation point in a fabricated FBG with a period of 1.6 μm, and b is a reflection spectrum of a 2 mm FBG with a measured period of 1.6 μm;

[0073] Figure 7 is the measured signal-to-noise ratio plot of FBG reflectivity;

[0074] Figure 8 a is a diagram of measured reflection spectra of six FBGs with the same manufacturing parameters, b is a diagram of measured Bragg wavelength of each FBG, c is a diagram of measured FWHM of each reflection spectrum, and d is a diagram of measured signal-to-noise ratio of each reflection spectrum;

[0075] Fig. 9 It is a structural schematic diagram of a machine vision-based fiber Bragg grating processing system provided in Embodiment 2 of the present invention.

[0076] Reference numerals:

[0077] Laser 1, beam splitter 2, oil mirror 3, red light source 4, CCD camera 5, 3D electric stage 6, computer 7. DETAILED DESCRIPTION

[0078] In order to explain in detail the possible application scenarios, technical principles, specific schemes that can be implemented, and the purposes and effects that can be achieved, the following is a detailed description of the specific embodiments listed in conjunction with the accompanying drawings. The embodiments described herein are only used to more clearly illustrate the technical solutions of the present application, and are therefore only used as examples, and cannot be used to limit the scope of protection of the present application.

[0079] Reference to "embodiment" herein means that the specific features, structures or characteristics described in conjunction with the embodiment may be included in at least one embodiment of the present application. The term "embodiment" appearing in various places in the specification does not necessarily refer to the same embodiment, nor does it particularly limit its independence or association with other embodiments. In principle, in the present application, as long as there is no technical contradiction or conflict, the various technical features mentioned in the embodiments can be combined in any way to form a corresponding implementable technical solution.

[0080] Unless otherwise defined, the technical terms used in this document have the same meanings as those generally understood by those skilled in the art to which this application belongs; the use of relevant terms in this document is only for describing specific embodiments and is not intended to limit this application.

[0081] In the description of this application, the term "and / or" is an expression used to describe the logical relationship between objects, indicating that three relationships may exist, for example, A and / or B, which means: A exists, B exists, and A and B exist at the same time. In addition, the character " / " in this article generally indicates that the objects before and after are in an "or" logical relationship.

[0082] In the present application, terms such as “first” and “second” are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship of quantity, priority or sequence between these entities or operations.

[0083] Without further limitations, in this application, the words "include", "comprises", "has" or other similar expressions used in the sentences are intended to cover non-exclusive inclusion. These expressions do not exclude the presence of additional elements in the process, method or product including the elements, so that the process, method or product including a series of elements may include not only those limited elements, but also other elements not explicitly listed, or also include elements inherent to such process, method or product.

[0084] In this application, expressions such as "greater than", "less than", "exceed" and the like are understood to exclude the number itself; expressions such as "above", "below", "within" and the like are understood to include the number itself. In addition, in the description of the embodiments of this application, "multiple" means more than two (including two), and similar expressions related to "multiple" are also understood in this way, such as "multiple groups", "multiple times", etc., unless otherwise clearly and specifically limited.

[0085] In the description of the embodiments of the present application, space-related expressions used, such as "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "vertical", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc., indicate the orientation or position relationship based on the orientation or position relationship shown in the specific embodiments or drawings, and are only for the convenience of describing the specific embodiments of the present application or facilitating the reader's understanding, and do not indicate or imply that the referred device or component must have a specific position, a specific orientation, or be constructed or operated in a specific orientation, and therefore cannot be understood as a limitation on the embodiments of the present application.

[0086] Unless otherwise expressly specified or limited, in the description of the embodiments of the present application, the terms such as "install", "connect", "connect", "fix", "set", etc. used should be understood in a broad sense. For example, the "connection" can be a fixed connection, a detachable connection, or an integrated setting; it can be a mechanical connection, an electrical connection, or a communication connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium; it can be the internal connection of two elements or the interaction relationship between two elements. For technicians in the technical field to which the present application belongs, the specific meanings of the above terms in the embodiments of the present application can be understood according to the specific circumstances.

[0087] Embodiment 1

[0088] In view of the above-mentioned defects in the prior art, the applicant, based on many years of rich practical experience and professional knowledge in design and manufacturing in this field, and in conjunction with the application of theory, actively conducts research and innovation in the hope of creating a technology that can solve the defects in the prior art. After continuous research and design, and after repeated trial production and improvement, the present invention with real practical value was finally created.

[0089] Please refer to Figure 1 , is a flow chart of a method for processing a fiber Bragg grating based on machine vision provided in Embodiment 1 of the present invention, which is applicable to the scenario of processing the fiber core using a femtosecond laser. The method specifically comprises the following steps:

[0090] S1. At a starting point, scan the optical fiber core and collect an image at the starting point.

[0091] It should be noted that at the starting point of the optical fiber core processing, a machine vision system (such as a CCD camera) is used to scan the optical fiber core, and images are continuously collected during the scanning process. This image will be used for subsequent analysis and processing to determine the focal plane of the optical fiber core at the starting point.

[0092] In one implementation of this embodiment, step S1 can be specifically broken down into the following contents:

[0093] At the starting point, the optical fiber core is scanned within a range of ±30 μm, and an image at the starting point is collected.

[0094] It should be noted that the selection of a scanning range of ±30μm centered on the starting point is intended to fully take into account possible slight deviations or irregularities in the optical fiber core, thereby ensuring that image information at the starting point and its surrounding areas can be fully and accurately captured.

[0095] In one implementation of this embodiment, before the entire processing flow is officially carried out in step S1, a crucial preparatory step is added, namely, starting point selection:

[0096] Choose any point in the fiber core as the starting point.

[0097] It should be noted that in this step, the operator needs to select any point on the fiber core and set it as the starting point of the entire processing flow. This selection process needs to fully consider the physical properties of the fiber, processing requirements, and the connectivity of subsequent steps to ensure the smooth progress of the entire processing flow. The setting of this preparatory step not only lays a solid foundation for the image acquisition, focal plane positioning, and laser processing in the subsequent steps, but also further enhances the flexibility and operability of the entire processing process.

[0098] After the selection of the starting point is completed, the entire processing flow formally enters step S1, i.e. the starting point scanning and image acquisition stage, and then proceeds step by step according to the established process until the processing of the fiber Bragg grating is finally completed. The introduction of this implementation method not only optimizes the processing flow, but also significantly improves the processing efficiency and accuracy, providing more reliable technical support for the manufacture of fiber Bragg gratings.

[0099] S2. Input the image at the starting point into the trained neural network model, locate the focal plane of the optical fiber core at the starting point, and obtain the optical fiber tilt angle.

[0100] It should be noted that this step is to input the image collected at the starting point into a pre-trained neural network model. This model can recognize and analyze the image to determine the focal plane of the optical fiber core at the starting point, that is, the precise position where the laser should be focused at the starting point. At the same time, the model can further recognize and analyze the image to determine whether the optical fiber is tilted, and then determine the tilt angle of the optical fiber, which is crucial to ensure that the laser is correctly focused.

[0101] S3. Determine a termination point according to the optical fiber tilt angle and a given processing length.

[0102] It should be noted that the end point of the processing is determined using the fiber tilt angle provided by the neural network model and the given processing length. This process takes into account the tilt of the fiber to ensure that the path of the laser processing is accurate.

[0103] In one implementation of this embodiment, step S3 can be specifically refined into the following contents:

[0104] A given processing length is moved in a direction of tilting the optical fiber tilt angle to determine a termination point.

[0105] It should be noted that the termination point is the end point of the processing path. The process of determining the termination point is not simply to extend the processing length in a straight line, but to fully consider the possible inclination of the optical fiber.

[0106] Specifically, the given processing length is moved along the direction of the optical fiber inclination angle, that is, the tangent direction of the actual direction of the optical fiber. This movement process not only ensures that the processing path is consistent with the actual shape of the optical fiber, but also avoids processing errors that may be caused by straight line extension.

[0107] Through this refined calculation and movement process, the location of the end point can be accurately determined, providing a precise target for subsequent steps such as fiber core scanning, image acquisition, and laser processing. The introduction of this refined implementation method not only improves the precision and accuracy of the entire processing flow, but also further enhances the processing quality and consistency of the fiber Bragg grating.

[0108] S4. Scanning the optical fiber core at the termination point and acquiring an image at the termination point.

[0109] It should be noted that at the determined end point, the machine vision system is used again to scan the optical fiber core and collect an image at this position, which will be used for subsequent focal plane positioning of the optical fiber core at the end point.

[0110] In one implementation of this embodiment, step S4 can be specifically refined into the following contents:

[0111] At the termination point, the optical fiber core is scanned within a range of ±30 μm, and an image at the termination point is collected.

[0112] It should be noted that when the end point of the processing path is accurately determined, the fiber core is scanned carefully within a small range of ±30μm centered on the end point. This scanning step is similar to the scanning process and considerations of the starting point, but the purpose is different. At the starting point, the scanning is mainly to capture the focal plane of the fiber core at the starting point, while at the end point, the scanning is to capture the focal plane of the fiber core at the end point.

[0113] S5. Input the image at the termination point into the trained neural network model to locate the focal plane of the optical fiber core at the termination point.

[0114] It should be noted that this step is to input the image collected at the termination point into the same neural network model. This model will recognize and analyze the image and determine the focal plane position of the optical fiber core at the termination point, that is, the precise position where the laser should be focused at the termination point.

[0115] S6, controlling the laser focus to focus on the focal plane of the optical fiber core at the starting point, and starting to process the optical fiber core until the focal plane of the optical fiber core at the ending point is reached to obtain a fiber Bragg grating.

[0116] It should be noted that according to the focal plane coordinates of the fiber core at the starting and ending points provided by the neural network model, the focus of the femtosecond laser can be precisely controlled to focus on the focal plane of the fiber core and process along the length of the fiber until it reaches the end point. This process will create a periodic refractive index change and form a fiber Bragg grating.

[0117] In one implementation of this embodiment, the method further includes the process of establishing, training and verifying a neural network model, namely:

[0118] Build a neural network model;

[0119] Dividing the collected out-of-focus images and in-focus images into a training set and a validation set; the out-of-focus images include images that are focused too deeply and images that are focused too shallowly;

[0120] Inputting the training set into the neural network model to train the neural network model;

[0121] The verification set is input into the neural network model to verify the neural network model and obtain the trained neural network model.

[0122] It should be noted that, specifically, first, according to the image recognition requirements in the fiber Bragg grating processing process, a suitable neural network model is designed and constructed. The model should have strong image processing capabilities, be able to accurately distinguish out-of-focus images from focused images, and locate the focal plane of the fiber core.

[0123] Secondly, a large number of out-of-focus images and in-focus images are collected, which should cover the fiber core morphology under different focusing states. Among them, out-of-focus images include images with too deep focus (i.e., the focus is below the fiber core) and too shallow focus (i.e., the focus is above the fiber core) to ensure that the neural network model can fully learn the changes in the focusing state.

[0124] These images are divided into training sets and validation sets according to a certain ratio. The training set is used for the training process of the neural network, while the validation set is used to evaluate the performance of the model and ensure the generalization ability of the model.

[0125] Furthermore, the training set is input into the neural network model, and the model parameters are continuously adjusted through multiple iterations and back-propagation algorithms, so that the model can accurately identify and locate the focal plane of the optical fiber core. During the training process, indicators such as the model's loss function and accuracy can be monitored to evaluate the model's training effect.

[0126] Finally, the validation set is input into the trained neural network model, and the performance of the model is evaluated by comparing the model prediction results with the actual results.

[0127] After verification, if the model performance meets the requirements, it is regarded as a trained neural network model and applied to the processing of fiber Bragg gratings. If the model performance is not good, it is necessary to readjust the model structure, increase training data or optimize the training strategy to improve the recognition ability of the model.

[0128] This detailed implementation method not only reveals the important role of neural network models in the fiber Bragg grating processing process, but also provides a complete process of model establishment, training and verification, providing strong technical support for practical applications.

[0129] In one implementation of this embodiment, before the step of inputting the training set into the neural network model to train the neural network model, the method further includes a process of preprocessing the image, namely:

[0130] Classify and label each image;

[0131] Perform background subtraction on each image;

[0132] The region of interest containing the two boundary lines of the fiber core was cropped from the background-subtracted image;

[0133] Grayscale values ​​of the boundary lines in the cropped region of interest are calculated to extract the calculation results as features of the boundary lines.

[0134] It should be noted that by performing data preprocessing on the image, the accuracy and success rate of the evaluation can be improved, which is crucial for the deep learning process. Specifically, first, the interference information (such as dust or background noise) is minimized to ensure that the neural network model can pick up key features. At the same time, the key features that are useful for the neural network model to extract discriminative information should be retained. Therefore, the image is preprocessed in two steps, including labeling and denoising.

[0135] During the processing, the fiber core is located on a 3D motorized stage, which moves in the z direction according to the feedback of the neural network model, which tells the 3D motorized stage whether it has reached the focus position, that is, whether the laser focus is focused on the focal plane of the fiber core. The neural network model only needs to provide a reference for the movement direction of the 3D motorized stage. In other words, the neural network model needs to identify whether the position of the laser focus is too deep (out of focus, too deep focus, negative) or too shallow (out of focus, too shallow focus, positive) relative to the focus point position, so that the 3D motorized stage can move in the opposite direction to achieve the focus state. Therefore, the evaluation results are divided into three categories: out of focus (negative), in focus, and out of focus (positive). In order to capture a large number of images for deep learning, continuous videos (resolution of 1,024×1,024 and frame rate of 23) are captured while the 3D motorized stage moves in the z direction to scan the fiber core. In order for the neural network model to know the classification of each image, each image in the video needs to be labeled (can be manually labeled). Figure 2 Examples of images of different categories are shown. Due to diffraction at the fiber core boundary, two white lines appear when the fiber core is out of focus (negative), e.g. Figure 2 As shown in a in Figure 1, when the fiber core is out of focus (positive), two black lines will appear, such as Figure 2 When the fiber core is focused, the diffraction disappears, as shown in c. Figure 2 These features provide useful information to identify the focus position.

[0136] Each image is then subjected to background subtraction to minimize noise information that may interfere with the deep learning process, e.g. Figure 3 As shown. In this way, the significant features that need to be identified can be highlighted. In addition, in order to further reduce the impact of areas without boundary lines and speed up the deep learning process, it is necessary to crop the region of interest containing the two boundary lines of the optical fiber core for deep learning processing, and then calculate the average grayscale value of the boundary line to extract the calculation result as the feature of the boundary line, which is input into the model for training.

[0137] In one implementation of this embodiment, Figure 4 As shown in the figure, the neural network model adopts a multi-layer perceptron (MLP) model, including an input layer, four dense layers and an output layer; each dense layer consists of multiple neurons and is responsible for calculating hidden information. A nonlinear activation function (such as the rectified linear unit (ReLU) activation function) is used between each two adjacent dense layers to make the information nonlinear, attempting to capture more nonlinear information during the information propagation process;

[0138] like Figure 4 As shown, in this embodiment, the first dense layer contains 32 neurons, and then the ReLU activation function is used; then, the result of the first dense layer passes through the second dense layer and the third dense layer with different numbers of neurons, and the ReLU activation function is also used;

[0139] The second dense layer contains 64 neurons;

[0140] The third dense layer contains 64 neurons;

[0141] The fourth dense layer contains 3 neurons and uses the SoftMax function output.

[0142] It should be noted that the purpose of this example is to enable the MLP model to learn deep features through different dense layers, where different numbers of neurons capture different aspects of deep features, and the ReLU activation function can make the features nonlinear. The last dense layer consists of 3 neurons corresponding to the three classes, followed by a SoftMax function (mapping the results to the probability of each class).

[0143] ……(1);

[0144] ……(2);

[0145] Here, k is equal to the number of categories and x is the result before ReLU function activation.

[0146] In this example, the MLP model is trained using a training set of 8,000 images. An early stopping strategy is used to select the best model based on the loss of the validation set (2,000 images). The goal of training is to minimize the loss through weight updates.

[0147] In one implementation of this embodiment, the step of inputting the training set into the neural network model to train the neural network model can be further refined to include the following contents:

[0148] Assigning random values ​​to the weights of the neural network model; the selection of these initial values ​​has an important impact on the training effect of the model;

[0149] Input an image from the training set into the neural network model and perform forward propagation to generate and output a predicted value; this step simulates the prediction process of the model in actual application;

[0150] Calculating the loss between the predicted value and the corresponding true value;

[0151] Using the calculated loss to update the weights of the neural network model; this step aims to reduce the difference between the predicted value and the true value by adjusting the weights, thereby improving the accuracy of the model;

[0152] Continue to input the next image in the training set into the neural network model for forward propagation, loss calculation and weight update until the preset termination condition is met. The termination condition may include reaching a specified number of iterations, the loss value is less than a certain threshold, or the performance of the model no longer improves on the validation set.

[0153] It should be noted that the loss function used in the training process is the mean square error (MSE), and the learning rate (λ) is set to 0.0001. Formula (3) defines the calculation of MSE, and formula (4) shows the training optimization process. Figure 5 Shows the loss monitoring during training.

[0154] ……(3);

[0155] ……(4);

[0156] in, is the true value of the sample, is the prediction of the sample, For the Layer parameters, τ is the learning rate, For the The updated parameters of the layer, L is the loss. The training is optimized by monitoring the loss at each step, such as Figure 5 As shown in a in . In the step where the validation loss reaches the minimum value, the optimal model is selected.

[0157] Accuracy and efficiency are key parameters for evaluating the quality of the model. Accuracy is determined by using 1408 test images (out-of-focus (positive): 627; in-focus: 84; out-of-focus (negative): 717) that were not used in model training to identify the fiber core. Table 1 shows the classification confusion matrix.

[0158] Table 1:

[0159]

[0160] Accuracy is a measure of the performance of the focus recognition model and is defined as:

[0161] ……(5);

[0162] Among them, TP is true positive, TN is true negative, FP is false positive, and FN is false negative. According to Table 1, the accuracy of the fiber core identification model is 96.09%, which is very high and confirms the success of the model.

[0163] In addition, the fiber core identification efficiency is measured by the output time of each image, e.g. Figure 5 The average prediction time is about 0.002 seconds (using a Windows laptop equipped with an Intel Core i7-10875H @ 2.30 GHz CPU, 32GB RAM and RGX3070 GPU), which is equivalent to about 500 FPS. This fast response time makes the model suitable for real-time fiber core identification. In order to further understand the method, this example conducted experiments to observe the effect of denoising (such as Figure 5 (c in Figure 1). Under the same conditions, the recognition accuracy of the denoised image is much higher than that of the image without denoising, which confirms the necessity of denoising.

[0164] By using the processing method provided in this embodiment, the inventors successfully produced a series of FBGs in commercial Corning SMF-28 optical fibers. Figure 6 Figure a shows a local microscopic image of a 2 mm long FBG, with a period of 1.6 μm for the RI modulation point. The corresponding fabrication parameters are: femtosecond laser power of 70 μw, femtosecond laser pulse width of 100 fs, repetition rate of 1 kHz, fabrication speed of 8 μm / s, fabrication frequency of 5 Hz, and fabrication duty cycle of 50%. A tunable fiber laser source (ANDO AQ4321D, spectral region of 1520-1620) is directly introduced into the FBG to measure the reflection spectrum of the FBG. Figure 6 Figure (b) shows the measured reflectance spectrum of the FBG, with a full width at half maximum (FWHM) of about 0.4 nm, a signal-to-noise ratio of about 30 dB, and a Bragg wavelength of 1537.65 nm, following the equation mλFBG=2neffΛ, where m is the grating order, neff is the effective refractive index of the fiber, and Λ is the period of the point. Therefore, the new AI-driven PbP fs DLW system demonstrates its ability to manufacture high-quality FBGs.

[0165] In addition, in order to use this method to produce higher quality FBG, this embodiment optimizes the production parameters. The two main parameters that need to be optimized here are the length of the FBG and the power of the femtosecond laser. In this embodiment, the period of the RI modulation point is still set to 1.6μm, so the production speed is 8μm / s, the production frequency is 5Hz, and the production duty cycle is 50%. This embodiment produces a series of FBGs with lengths of 1, 2, 3 and 5mm, and each length of FBG is produced with different femtosecond laser powers: from 65 to 90μw. Figure 7 The measured signal-to-noise ratio of the reflectivity of these FBGs is plotted. Two obvious conclusions can be drawn within the experimental parameter range: first, the longer the FBG, the greater the reflectivity; second, when the femtosecond laser power is set to 70μW, FBGs of various lengths with the highest reflectivity can be produced.

[0166] Repeatability is one of the most important characteristics of a mature and reliable processing method. Therefore, based on the optimized processing parameters, this embodiment tests the repeatability of the processing method by processing six 2 mm FBGs. The processing parameters are the same: the period of the RI modulation point is 1.6 μm; the fs laser power is 70 μw; the fs laser pulse width is 100 fs; the repetition rate is 1 kHz; the processing speed is 8 μm / s; the processing frequency is 5 Hz; and the processing duty cycle is 50%. The measured reflectance spectra of these FBGs are as follows: Figure 8 As shown in a, they are almost overlapped, ignoring the fluctuation caused by the connection between the FBG and the fiber laser source. Figure 8 b in the figure is the Bragg wavelength of the six FBGs, which is 1537.6 nm. The measurement error of ±0.05 nm is negligible. This embodiment uses Gaussian distribution to fit these reflection spectra and calculates their full width at half maximum (FWHM). Figure 8 As shown in c, the FWHM is 0.4±0.03nm. Figure 8 Figure d shows the signal-to-noise ratio of the reflectance spectrum, which is as high as 30 ± 2 dB. Therefore, this processing method has demonstrated its reproducibility in high-quality FBG fabrication, which will become one of the most effective and simplest techniques for large-scale FBG fabrication.

[0167] This invention demonstrates the broad application prospects of machine vision in the field of femtosecond laser FBG processing. The advancement of this low-cost and high-speed technology will further promote the commercialization and application of femtosecond FBG processing.

[0168] Although the terms such as machine vision, focal plane, and optical fiber are used more frequently in this application, the possibility of using other terms is not excluded. These terms are used only to more conveniently describe and explain the essence of the present invention; interpreting them as any additional restrictions is contrary to the spirit of the present invention.

[0169] The embodiment of the present invention provides a method for processing a fiber Bragg grating based on machine vision. By combining advanced machine vision technology and neural network model technology, the method realizes automatic recognition of the focal plane of the optical fiber core, so as to automatically focus the laser focus on the focal plane of the optical fiber core, thereby realizing the automation and intelligence of the fiber Bragg grating processing process. It not only has the characteristics of high precision and high efficiency, but also shows broad application prospects, providing strong support for the development of optical fiber communication, sensing and other fields.

[0170] Embodiment 2

[0171] Please refer to Fig. 9 Embodiment 2 of the present invention provides a fiber Bragg grating processing system based on machine vision, the system comprising a laser 1, a beam splitter 2, an oil mirror 3, a red light source 4, a CCD camera 5, a 3D electric stage 6 and a computer 7; wherein,

[0172] The 3D electric stage 6 is used to carry the optical fiber and drive the optical fiber to move;

[0173] The laser 1 is located above the 3D electric stage 6 and is used to emit a laser beam;

[0174] The oil mirror 3 is located between the 3D electric stage 6 and the laser 1, and is used to focus the laser beam;

[0175] The red light source 4 is located below the 3D electric stage 6 and is used to emit red illumination light to illuminate the optical fiber;

[0176] The beam splitter 2 is located between the oil mirror 3 and the laser 1, and is used to reflect the red illumination light to the CCD camera 5;

[0177] The CCD camera 5 is located on the light-reflecting surface of the beam splitter 2 and is used to collect the image of the optical fiber;

[0178] The computer 7 is used for:

[0179] Control the 3D electric stage 6 to move, and control the CCD camera 5 to scan the core of the optical fiber at the starting point, and collect the image at the starting point;

[0180] Inputting the image at the starting point into the trained neural network model, locating the focal plane of the core of the optical fiber at the starting point, and obtaining the optical fiber tilt angle;

[0181] Determining a termination point according to the optical fiber tilt angle and a given processing length;

[0182] Control the 3D electric stage 6 to move, and control the CCD camera 5 to scan the core of the optical fiber at the end point, and collect the image at the end point;

[0183] Inputting the image at the termination point into a trained neural network model to locate the focal plane of the core of the optical fiber at the termination point;

[0184] The 3D electric stage 6 is controlled to move so as to focus the laser focus of the laser 1 to the focal plane of the core of the optical fiber at the starting point position, and start processing the core of the optical fiber until the focal plane of the core of the optical fiber at the ending point position is reached to obtain a fiber Bragg grating.

[0185] It should be noted that the laser 1 can be an 800nm ​​femtosecond laser with a pulse width of 100fs and an adjustable repetition rate. The collimated fs laser beam is focused into the fiber core through a high NA (1.45) oil mirror 3 (magnification of 100 times). The optical fiber can be installed in a V-groove bracket, which is fixed on a 3D electric stage 6 with nanometer resolution. A 650nm red light source 4 (LED light source) is placed under the optical fiber for illumination, and the illumination light is reflected into the CCD camera 5 through a beam splitter 2. During the FBG manufacturing process, the fiber core image collected by the CCD camera 5 is transmitted to the computer 7 in real time. The computer 7 realizes the automatic recognition of the focal plane of the fiber core by combining advanced machine vision technology and neural network model technology, so as to automatically focus the laser focus to the focal plane of the fiber core, thereby realizing the automation and intelligence of the fiber Bragg grating processing process, which not only has the characteristics of high precision and high efficiency, but also shows broad application prospects, and provides strong support for the development of optical fiber communication, sensing and other fields.

[0186] Finally, it should be noted that although the above embodiments have been described in the specification and drawings of this application, this does not limit the scope of patent protection of this application. All technical solutions generated by replacing or modifying equivalent structures or equivalent processes based on the essential concept of this application using the contents recorded in the specification and drawings of this application, as well as directly or indirectly implementing the technical solutions of the above embodiments in other related technical fields, are included in the scope of patent protection of this application.

Claims

1. A method for processing a fiber Bragg grating based on machine vision, characterized in that: The method comprises: At the starting point, scanning the optical fiber core and collecting an image at the starting point; Inputting the image at the starting point into the trained neural network model, locating the focal plane of the optical fiber core at the starting point, and obtaining the optical fiber tilt angle; Determining a termination point according to the optical fiber tilt angle and a given processing length; At the termination point, scanning the optical fiber core and collecting an image at the termination point; Inputting the image at the termination point into a trained neural network model to locate the focal plane of the optical fiber core at the termination point; The laser focus is controlled to focus on the focal plane of the optical fiber core at the starting point, and the optical fiber core is processed until the focal plane of the optical fiber core at the ending point is processed to obtain a fiber Bragg grating.

2. The method for processing a fiber Bragg grating based on machine vision according to claim 1, characterized in that: Before the step of scanning the optical fiber core at the starting point and collecting the image at the starting point, the method further includes: Choose any point in the fiber core as the starting point.

3. The method for processing a fiber Bragg grating based on machine vision according to claim 1, characterized in that: The step of scanning the optical fiber core at the starting point and collecting the image at the starting point is specifically: At the starting point, the optical fiber core is scanned within a range of ±30 μm, and an image at the starting point is collected.

4. The method for processing a fiber Bragg grating based on machine vision according to claim 1, characterized in that: The step of determining the termination point according to the optical fiber tilt angle and the given processing length is specifically as follows: A given processing length is moved in a direction of tilting the optical fiber tilt angle to determine a termination point.

5. The method for processing a fiber Bragg grating based on machine vision according to claim 1, characterized in that: The step of scanning the optical fiber core at the termination point and collecting the image at the termination point is specifically: At the termination point, the optical fiber core is scanned within a range of ±30 μm, and an image at the termination point is collected.

6. The method for processing a fiber Bragg grating based on machine vision according to claim 1, characterized in that: The method further comprises: Build a neural network model; Dividing the collected out-of-focus images and in-focus images into a training set and a validation set; the out-of-focus images include images that are focused too deeply and images that are focused too shallowly; Inputting the training set into the neural network model to train the neural network model; The verification set is input into the neural network model to verify the neural network model and obtain the trained neural network model.

7. The method for processing a fiber Bragg grating based on machine vision according to claim 6, characterized in that: Before the step of inputting the training set into the neural network model to train the neural network model, the method further includes: Classify and label each image; Perform background subtraction on each image; The region of interest containing the two boundary lines of the fiber core was cropped from the background-subtracted image; Grayscale values ​​of the boundary lines in the cropped region of interest are calculated to extract the calculation results as features of the boundary lines.

8. The method for processing a fiber Bragg grating based on machine vision according to claim 6, characterized in that: The neural network model includes an input layer, four dense layers and an output layer; A nonlinear activation function is used between each two adjacent dense layers; The first dense layer contains 32 neurons; The second dense layer contains 64 neurons; The third dense layer contains 64 neurons; The fourth dense layer contains 3 neurons and uses the SoftMax function output.

9. The method for processing a fiber Bragg grating based on machine vision according to claim 6, characterized in that: The step of inputting the training set into the neural network model to train the neural network model includes: Assigning random values ​​to weights of the neural network model; Input an image from the training set into the neural network model and perform forward propagation to generate and output a predicted value; Calculate the loss between the predicted value and the corresponding true value, the loss function is the mean square error; Using the calculated loss to update the weights of the neural network model; Continue to input the next image in the training set into the neural network model for forward propagation, loss calculation and weight update until the preset termination condition is met.

10. A fiber Bragg grating processing system based on machine vision, characterized in that: The system includes a laser, a beam splitter, an oil mirror, a red light source, a CCD camera, a 3D electric stage and a computer; wherein, The 3D electric stage is used to carry the optical fiber and drive the optical fiber to move; The laser is located above the 3D electric stage and is used to emit a laser beam; The oil mirror is located between the 3D electric stage and the laser, and is used to focus the laser beam; The red light source is located below the 3D electric stage and is used to emit red illumination light to illuminate the optical fiber; The beam splitter is located between the oil mirror and the laser, and is used to reflect the red illumination light to the CCD camera; The CCD camera is located on the light-reflecting surface of the beam splitter and is used to collect the image of the optical fiber; The computer is used to: Controlling the movement of the 3D electric stage, and controlling the CCD camera to scan the core of the optical fiber at the starting point, and collecting the image at the starting point; Inputting the image at the starting point into the trained neural network model, locating the focal plane of the core of the optical fiber at the starting point, and obtaining the optical fiber tilt angle; Determining a termination point according to the optical fiber tilt angle and a given processing length; Controlling the movement of the 3D electric stage, and controlling the CCD camera to scan the core of the optical fiber at the end point, and collecting an image at the end point; Inputting the image at the termination point into a trained neural network model to locate the focal plane of the core of the optical fiber at the termination point; The 3D electric stage is controlled to move so as to focus the laser focus of the laser to the focal plane of the core of the optical fiber at the starting point, and start processing the core of the optical fiber until the focal plane of the core of the optical fiber at the ending point is reached to obtain a fiber Bragg grating.

Images