Glass fiber breakage identification method and device, electronic equipment and storage medium

Abstract

The invention relates to the technical field of computer vision, and discloses a broken glass fiber identification method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining a glass fiber production image of a target production region, carrying out the liquid drop identification of the glass fiber production image through a preset liquid drop identification model, and obtaining a liquid drop identification result, the liquid drop identification model is obtained by training a first image sample with a liquid drop existence label, and if the liquid drop identification result is that liquid drops exist, it is judged that glass fibers in the target production area are broken; according to the method, whether the glass fiber is broken or not can be effectively detected, the accuracy of a broken fiber identification result can be improved, and the method is simple and efficient.

Description

Technical Field

[0001] The present application relates to the field of computer vision technology, and in particular to a method, device, electronic device, and storage medium for identifying broken glass fibers. Background Art

[0002] In the glass fiber manufacturing process, the process of spinning molten glass through a spinning plate to form glass fibers is a critical step. The stability and continuity of this production process are crucial to the quality of the final product and the control of production costs. In actual production scenarios, glass fiber production often faces many challenges, among which the problem of broken fibers is particularly prominent.

[0003] When glass melt breaks during the spinning process, if it is not detected immediately and manual intervention measures are not taken in time to restore the broken fiber to normal state, a series of adverse consequences will occur. From the perspective of product quality, broken fibers will cause problems such as uneven structure and substandard strength in the subsequently produced glass fibers, seriously affecting the overall quality of the product and making it difficult to meet the strict requirements of downstream applications for glass fiber performance. From the perspective of production costs, if broken fibers are not handled in a timely manner, it will result in waste of raw materials, because glass melt that fails to spin normally cannot be converted into qualified products; at the same time, the production equipment continues to operate after the break, energy consumption does not decrease, and effective products cannot be produced, which undoubtedly increases energy waste and further increases production costs.

[0004] At present, the production monitoring method adopted by glass fiber production enterprises mainly relies on manual supervision. In actual operation, one worker is often responsible for monitoring multiple workstations at the same time. This model of one person on duty at multiple workstations has obvious disadvantages. On the one hand, the labor intensity of workers is extremely high, and they are easily fatigued by concentrating their attention for a long time to observe the production conditions of multiple workstations, which in turn affects their work efficiency and judgment accuracy. On the other hand, due to the limited energy of humans, it is difficult to achieve real-time and comprehensive observation of each workstation when monitoring multiple workstations at the same time. As a result, broken wires are often not discovered in time, making it difficult to effectively solve the problem of broken wires in the production process, seriously restricting the quality and efficiency of glass fiber production, and also increasing the operating costs of enterprises. Obviously, there is an urgent need for a new method for identifying broken glass fibers to solve at least one of the above problems.

[0005] It should be noted that the above content only provides background technical information related to this application and does not necessarily constitute prior art. Summary of the Invention

[0006] In view of the above-mentioned shortcomings of the prior art, the present application provides a broken glass filament identification method, device, electronic device and storage medium to improve the accuracy and efficiency of broken glass filament identification.

[0007] Other features and advantages of the present application will become apparent from the following detailed description, or may be learned in part by practice of the present application.

[0008] According to one aspect of an embodiment of the present application, a method for identifying broken glass filaments is provided, comprising: acquiring a glass filament production image of a target production area; performing droplet recognition on the glass filament production image using a preset droplet recognition model to obtain a droplet recognition result, wherein the droplet recognition model is trained based on a first image sample with a droplet presence label; if the droplet recognition result is that droplets are present, it is determined that the glass filaments in the target production area are broken.

[0009] In one embodiment of the present application, based on the aforementioned scheme, after droplet recognition is performed on the glass filament production image using a preset droplet recognition model and a droplet recognition result is obtained, the method further includes: if the droplet recognition result is that droplets exist, calculating the number of droplets and the rectangular area corresponding to the largest droplet; if the number of droplets is greater than a preset number threshold, or the rectangular area corresponding to the largest droplet is greater than a preset area threshold, performing perturbation recognition on the glass filament production image using a preset perturbation recognition model to obtain a perturbation recognition result, wherein the perturbation recognition model is trained based on a second image sample with a disturbance occurrence label.

[0010] In one embodiment of the present application, based on the aforementioned scheme, disturbance recognition is performed on the glass filament production image through a preset disturbance recognition model. After obtaining the disturbance recognition result, the method further includes: if the disturbance recognition result is that a disturbance occurs, the number of broken glass filament frames is increased by one, wherein the initial value of the number of broken glass filament frames is a first preset value; if the number of broken glass filament frames is greater than the preset broken glass filament frame threshold, it is determined that the glass filament in the target production area is broken.

[0011] In one embodiment of the present application, based on the aforementioned scheme, droplet recognition is performed on the glass fiber production image through a preset droplet recognition model. After the droplet recognition result is obtained, the method further includes: if the droplet recognition result is that a droplet exists, obtaining the first coordinate of the rectangular box corresponding to the droplet, and obtaining the second coordinate of the target production area; if the droplet is determined to be within the target production area based on the first coordinate and the second coordinate, the number of frames in which the droplet exists is increased by one; if the number of frames in which the droplet exists is greater than the preset number of frames in which the droplet exists threshold, the number of droplets is increased by one, wherein the initial value of the number of frames in which the droplet exists is a second preset value, and the initial value of the number of droplets is a third preset value; if the droplet is determined to be outside the target production area based on the first coordinate and the second coordinate, the droplet recognition result is updated to the absence of a droplet.

[0012] In one embodiment of the present application, based on the aforementioned scheme, the droplet recognition model is trained based on a first image sample with a droplet presence label, including: obtaining a droplet presence label and a sample production image; annotating the sample production image according to the droplet presence label to generate a first image sample with a droplet presence label; training a preset droplet recognition model based on the first image sample with the droplet presence label to obtain the droplet recognition model.

[0013] In one embodiment of the present application, based on the aforementioned scheme, the disturbance recognition model is trained based on a second image sample with a disturbance occurrence label, including: obtaining a disturbance occurrence label and a sample production image; annotating the sample production image according to the disturbance occurrence label to generate a second image sample with the disturbance occurrence label; and training a preset disturbance recognition model based on the second image sample with the disturbance occurrence label to obtain the disturbance recognition model.

[0014] In one embodiment of the present application, based on the aforementioned solution, after obtaining the glass fiber production image of the target production area, the method further includes: performing image preprocessing on the glass fiber production image, wherein the image preprocessing includes image slicing, image cleaning, image annotation, and image review.

[0015] According to one aspect of an embodiment of the present application, a broken glass filament identification device is provided, comprising: an image acquisition module for acquiring a glass filament production image of a target production area; an identification module for performing droplet identification on the glass filament production image using a preset droplet identification model to obtain a droplet identification result, wherein the droplet identification model is trained based on a first image sample with a droplet presence label; and a determination module for determining that the glass filaments in the target production area are broken if the droplet identification result indicates the presence of droplets.

[0016] According to one aspect of an embodiment of the present application, an electronic device is provided, comprising: one or more processors; a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the electronic device implements the broken glass fiber identification method as described in any one of the above embodiments.

[0017] The present application further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor of a computer, the computer is caused to execute the broken glass filament identification method as described in any one of the above embodiments.

[0018] The beneficial effects of the present application are as follows: the present application obtains a glass fiber production image of a target production area, and performs droplet recognition on the glass fiber production image through a preset droplet recognition model to obtain a droplet recognition result, wherein the droplet recognition model is trained based on a first image sample with a droplet presence label. If the droplet recognition result is that droplets exist, it is determined that the glass fiber in the target production area is broken. By timely and accurate detection and identification of broken glass fibers, the quality and efficiency of glass fiber production can be improved, thereby reducing production costs.

[0019] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The accompanying drawings are incorporated into and constitute a part of the specification, illustrating embodiments consistent with the present application and, together with the specification, serving to explain the principles of the present application. It is obvious that the drawings described below are merely some embodiments of the present application, and a person of ordinary skill in the art can derive other drawings based on these drawings without inventive effort. In the drawings: Figure 1 is a schematic diagram of an exemplary system architecture shown in an exemplary embodiment of the present application; Figure 2 1 is a flow chart of a method for identifying broken glass fibers according to an exemplary embodiment of the present application; Figure 3 1 is a schematic diagram of the overall flow of a method for identifying broken glass fibers, shown in an exemplary embodiment of the present application; Figure 4 1 is a detailed flow chart of a method for identifying broken glass fibers according to an exemplary embodiment of the present application; Figure 5 1 is a schematic diagram of a normal production of a broken glass filament identification method according to an exemplary embodiment of the present application; Figure 6 1 is a schematic diagram of a broken wire anomaly in a method for identifying broken glass wires according to an exemplary embodiment of the present application; Figure 7 is a block diagram of a broken glass fiber identification device shown in an exemplary embodiment of the present application; Figure 8 A schematic diagram of the structure of a computer system suitable for implementing an electronic device according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0021] The following will describe the embodiments of the present application with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand the other advantages and effects of the present application from the contents disclosed in this specification. The present application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be understood that the preferred embodiments are only for the purpose of illustrating the present application and are not intended to limit the scope of protection of the present application.

[0022] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present application. Therefore, the illustrations only show components related to the present application and are not drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component can be changed at will, and the component layout type may also be more complicated.

[0023] In the following description, a large number of details are discussed to provide a more thorough explanation of the embodiments of the present application. However, it is obvious to those skilled in the art that the embodiments of the present application can be implemented without these specific details. In other embodiments, well-known structures and devices are shown in the form of block diagrams rather than in detail to avoid making the embodiments of the present application difficult to understand.

[0024] First, it's important to note that the K-means clustering algorithm is an iterative cluster analysis algorithm. It begins by partitioning the data into K groups, randomly selecting K objects as initial cluster centers, and then calculating the distance between each object and each seed cluster center, assigning each object to the cluster center closest to it. The cluster centers and the objects assigned to them represent a cluster. With each assignment, the cluster center is recalculated based on the existing objects in the cluster. This process repeats until a termination criterion is met. This termination criterion can include no (or a minimum number of) objects being reassigned to different clusters, no (or a minimum number of) cluster centers changing, or a local minimum in the sum of squared errors.

[0025] The Efficient Intersection over Union Loss (EIOU) loss, also known as the EIoU loss, is a loss function used in object detection tasks, designed to improve the accuracy and efficiency of the model in object detection. The EIOU loss is an improved version of the CIoU loss. By decomposing the width and height losses into independent terms, it simplifies computation and improves optimization efficiency, while maintaining precise control over the shape of the predicted box.

[0026] Figure 1 It is a schematic diagram of an exemplary system architecture shown in an exemplary embodiment of the present application.

[0027] Reference Figure 1 As shown, the system architecture may include an image acquisition device 101 and a computer device 102. Computer device 102 may be at least one of a desktop graphics processing unit (GPU) computer, a GPU computing cluster, a neural network computer, and the like. Image acquisition device 101 is used to capture images of glass filament production in a target production area. In this embodiment, image acquisition device 101 acquires the data and provides it to computer device 102 for processing. Relevant technicians can use computer device 102 to perform droplet recognition on the glass filament production images using a preset droplet recognition model to obtain droplet recognition results. The droplet recognition model is trained based on a first image sample labeled with a droplet presence. If the droplet recognition result indicates the presence of a droplet, it is determined that the glass filaments in the target production area are broken. It should be noted that the image acquisition device 101 and computer device 102 provided in this embodiment are merely examples and should not limit the functionality or scope of use of the embodiments of this application.

[0028] It should be noted that the broken glass filament identification method provided in the embodiment of the present application is generally executed by the computer device 102 , and accordingly, the broken glass filament identification device is generally disposed in the computer device 102 .

[0029] Figure 2 The following is a flow chart of a method for identifying broken glass filaments according to an exemplary embodiment of the present invention. The method for identifying broken glass filaments can be executed by a computing and processing device. The computing and processing device can be Figure 1 The computer device 102 shown in FIG. Figure 2 As shown, the broken glass fiber identification method includes at least steps S210 to S230, which are described in detail as follows: In step S210, a glass fiber production image of a target production area is acquired.

[0030] In one embodiment of the present application, an image acquisition device is used to acquire images of glass fiber production in a target production area. The image acquisition device can be a camera or a video camera, including but not limited to high-resolution industrial cameras, line array cameras, industrial cameras, and other industrial cameras; float glass streak meters, float glass streak meters, three-dimensional information, and color and brightness image acquisition devices, among other specialized image acquisition devices.

[0031] In one embodiment of the present application, after obtaining the glass fiber production image of the target production area, the process further includes the following steps: performing image preprocessing on the glass fiber production image, where the image preprocessing includes image slicing, image cleaning, image annotation, and image review.

[0032] In this embodiment, preprocessing glass fiber production images is a critical step in ensuring the accuracy of subsequent image analysis (such as defect detection and anomaly identification). Image slicing involves extracting a region of interest (ROI) from the original image and removing irrelevant background or noise. Image slicing can be performed manually or automatically. Manual slicing uses image processing software to manually select the glass fiber area. Automatic slicing utilizes edge detection algorithms, threshold segmentation, deep learning models, and other techniques. Image cleaning removes noise, enhances contrast, and corrects for uneven lighting. Image annotation identifies key features of the glass fiber (such as defect location, edges, and size) to provide labels for subsequent model training. Image review ensures that the preprocessed image meets quality requirements and contains no omissions or inaccuracies.

[0033] In step S220, droplet recognition is performed on the glass filament production image using a preset droplet recognition model to obtain a droplet recognition result.

[0034] The droplet recognition model is trained based on a first image sample with a droplet presence label.

[0035] In one embodiment of the present application, the process of training a droplet recognition model based on a first image sample with a droplet presence label includes the following steps: obtaining a droplet presence label and a sample production image; annotating the sample production image according to the droplet presence label to generate a first image sample with a droplet presence label; training a preset droplet recognition model based on the first image sample with the droplet presence label to obtain a droplet recognition model.

[0036] In some embodiments, the preset droplet recognition model is a modified YOLOv5 model, wherein the modified YOLOv5 model is obtained by introducing a K-means clustering algorithm, a SE attention mechanism, and an EIOU loss function into the YOLOv5 model. The K-means clustering algorithm is introduced during the anchor frame setting phase of the YOLOv5 model. The YOLOv5 model requires a set of anchor frames to be pre-set before training. The size and proportion of these anchor frames have a significant impact on the detection effect. The K-means clustering algorithm is used to automatically generate a set of more appropriate anchor frames based on the size distribution of target droplets in the dataset to replace the default anchor frames of the YOLOv5 model. The SE attention mechanism is introduced in the backbone network (Backbone) or neck network (Neck) of the YOLOv5 model. The SE attention mechanism can be flexibly embedded in various layers of the convolutional neural network to enhance the network's ability to focus on important features. Preferably, this embodiment adds the SE attention mechanism after each convolutional layer (Conv) in the backbone network of the YOLOv5 model. The EIOU loss function is introduced in the loss function calculation part of the YOLOv5 model. The loss function is used to measure the difference between the model prediction result and the true label. The EIOU loss function is used to replace the original loss function of YOLOv5 to optimize the model training process.

[0037] In this embodiment, prior box optimization based on K-means clustering aims to generate anchor boxes that better fit glass fiber droplet targets through cluster analysis, thereby improving small object detection accuracy. The specific implementation steps include data preparation, K-means clustering, and replacing the YOLOv5 default anchor boxes. Data preparation involves collecting images of droplets from glass fiber production and annotating their bounding boxes (x, y, w, h). This ensures that the dataset includes droplet samples of varying scales and angles, including abnormal morphologies such as elongated and dispersed droplets when the fiber breaks. K-means clustering uses the K-means algorithm to cluster the annotated boxes based on their aspect ratio (w / h) and area (w×h) to determine the optimal number of anchor boxes (e.g., k=9). The anchor box size is determined by minimizing the IoU distance between the cluster center and all bounding boxes. Replacing the YOLOv5 default anchor boxes involves modifying the YOLOv5 configuration file (e.g., yolov5s.yaml) to replace the original anchor boxes generated by clustering, adapting them to the scale distribution of the droplet targets.

[0038] In this example, the SE attention mechanism is introduced to enhance the model's focus on key droplet features (such as edges and texture) and suppress background interference. Its specific implementation steps include module integration and model configuration modification. Module integration involves inserting the SE module after the C3 module (or Bottleneck layer) of YOLOv5 to weight the feature map channels. In the SE module structure, channel compression (Squeeze) is used for global average pooling to generate channel descriptors (1×1×C); channel excitation (Excitation) is used to learn channel weights through fully connected layers and normalize them using the Sigmoid activation function; and channel scaling (Scale) is used to multiply the weights with the original feature map to emphasize important channels. Model configuration modifications include adding the SE module to the key layers of the backbone and neck networks in yolov5s.yaml and adjusting the network depth (depth_multiple) and width (width_multiple) to balance computational overhead and performance.

[0039] In this embodiment, the EIOU loss function is introduced to optimize bounding box regression, improve droplet positioning accuracy, and reduce false detection of broken wires. The EIOU loss is obtained by calculating the IoU loss, center point distance loss, and aspect ratio loss. In the YOLOv5 training process, the original CIOU (Complete IoU) loss function is replaced with the EIOU loss function to complete the introduction of the EIOU loss function. It can be understood that the overlapping area loss measures the ratio of the intersection over union (IoU) between the predicted box and the true box, the center point distance loss measures the Euclidean distance between the center point of the predicted box and the true box, and the aspect ratio loss measures the difference in width and height between the predicted box and the true box respectively.

[0040] It can be understood that this embodiment is only a preferred embodiment. In actual application, the preset droplet recognition model can be obtained by other means, and the steps, parameters and methods of obtaining the preset droplet recognition model in this embodiment can also be modified. This application does not limit this, nor should it bring any limitations to the functions and scope of use of the embodiments of this application.

[0041] In step S230 , if the droplet identification result indicates that a droplet exists, it is determined that the glass filaments in the target production area are broken.

[0042] In one embodiment of the present application, if the droplet identification result is that no droplets exist, it is determined that the glass filaments in the target production area are not broken.

[0043] In this embodiment, the droplets can be white, meaning the droplet recognition model identifies white droplets. This is because glass filaments are typically produced by melting raw glass at high temperatures and then drawing them into a filament. Insufficient melting temperatures, excessive drawing speeds, or uneven cooling can lead to stress concentration within the glass filaments, which can cause them to break. Unsolidified glass droplets remain at the ends of the broken glass filaments, appearing white after cooling (due to light scattering caused by oxidation or surface roughness).

[0044] In one embodiment of the present application, droplet recognition is performed on a glass filament production image using a preset droplet recognition model, and the process after obtaining the droplet recognition result further includes the following steps: if the droplet recognition result is that droplets exist, the number of droplets and the rectangular area corresponding to the largest droplet are calculated; if the number of droplets is greater than a preset number threshold, or the rectangular area corresponding to the largest droplet is greater than a preset area threshold, disturbance recognition is performed on the glass filament production image using a preset disturbance recognition model to obtain a disturbance recognition result, wherein the disturbance recognition model is trained based on a second image sample with a disturbance occurrence label.

[0045] In this embodiment, the area of ​​the rectangle corresponding to the largest droplet can be understood as the area of ​​the circumscribed rectangle corresponding to the largest droplet. The circumscribed rectangle refers to the smallest rectangle that can completely enclose a droplet, with its sides parallel to the coordinate axes. For a droplet, the area of ​​the circumscribed rectangle is the minimum area of ​​the droplet's enclosed region in the image. The preset number threshold and the preset area threshold can be set as needed and are not limited in this application.

[0046] In one embodiment of the present application, the process of training a disturbance recognition model based on a second image sample with a disturbance occurrence label includes the following steps: obtaining a disturbance occurrence label and a sample production image; annotating the sample production image according to the disturbance occurrence label to generate a second image sample with a disturbance occurrence label; and training a preset disturbance recognition model based on the second image sample with the disturbance occurrence label to obtain a disturbance recognition model.

[0047] In one embodiment, the disturbance recognition model can also determine whether a disturbance has occurred by determining whether the image is shaking. Image disturbances are usually manifested as inconsistencies in inter-frame motion, that is, pixel displacements or structural changes between adjacent frames or key frames exceed normal thresholds. Detection methods can be divided into the following three categories: The first is a method based on global motion estimation. Its principle is to determine whether there is unexpected translation, rotation, or scaling by calculating the global motion vectors (such as affine transformation, homography matrix) of adjacent frames. The second is a method based on optical flow. Its principle is to analyze the consistency of the overall motion by calculating the motion vector field of pixel points. The third is a method based on frequency domain analysis. Its principle is that disturbances will cause changes in the spectral energy distribution of the image sequence (such as an increase in high-frequency components).

[0048] In one embodiment of the present application, disturbance recognition is performed on a glass filament production image through a preset disturbance recognition model, and the process after obtaining the disturbance recognition result also includes the following steps: if the disturbance recognition result is that a disturbance occurs, the number of broken glass filament frames is increased by one, wherein the initial value of the number of broken glass filament frames is a first preset value; if the number of broken glass filament frames is greater than the preset broken glass filament frame threshold, it is determined that the glass filament in the target production area is broken.

[0049] In this embodiment, the first preset value can be set to 0, that is, the initial value of the number of broken glass fiber frames is 0. It is understandable that the first preset value can also be set to other values, and the preset broken glass fiber frame number threshold can be set according to actual needs or experience, and this application does not limit this.

[0050] In one embodiment, if the disturbance identification result is that no disturbance has occurred, it is determined that the glass filaments in the target production area are not broken.

[0051] In one embodiment of the present application, droplet recognition is performed on a glass fiber production image through a preset droplet recognition model, and the process after obtaining the droplet recognition result also includes the following steps: if the droplet recognition result is that a droplet exists, the first coordinate of the rectangular box corresponding to the droplet is obtained, and the second coordinate of the target production area is obtained; if the droplet is determined to be within the target production area based on the first coordinate and the second coordinate, the number of frames in which the droplet exists is increased by one; if the number of frames in which the droplet exists is greater than the preset number of frames in which the droplet exists threshold, the number of droplets is increased by one, wherein the initial value of the number of frames in which the droplet exists is a second preset value, and the initial value of the number of droplets is a third preset value; if the droplet is determined to be outside the target production area based on the first coordinate and the second coordinate, the droplet recognition result is updated to the absence of a droplet.

[0052] In this embodiment, if the number of frames in which droplets are present is less than or equal to the preset threshold number of frames, the number of droplets remains unchanged. The second and third preset values ​​can be set to 0, i.e., the initial values ​​of the number of frames in which droplets are present and the initial values ​​of the number of droplets are 0. It is understood that the second and third preset values ​​can also be set to other values, and this application is not limited thereto.

[0053] In one embodiment of the present application, when it is determined that the glass filaments in the target production area are broken, an alarm message and a control instruction are generated; based on the control instruction, the target device is controlled to stop running and the target user is prompted through an alarm message. The target device can be a glass filament drawing machine, a glass fiber drawing machine, etc.

[0054] Figure 3 FIG. 1 is a schematic diagram of the overall flow of a method for identifying broken glass fibers, shown in an exemplary embodiment of the present application. Figure 4 FIG. 1 is a detailed flow chart of a method for identifying broken glass fibers according to an exemplary embodiment of the present invention. Figure 3 and Figure 4 In an exemplary embodiment, image data of white droplets in a composite glass fiber production area is acquired as a sample production image through an image acquisition device and preprocessed, including one or more of slicing, cleaning, labeling, and reviewing. After preprocessing, a training sample is formed. Based on the training sample, a droplet recognition model is obtained through model training and optimization. When using the droplet recognition model, the preprocessed glass fiber production image of the target production area is input into the droplet recognition model for inference. The inference result is analyzed, and the presence of white droplets is determined based on whether white droplets exist in a preset judgment area, i.e., the target production area. When white droplets exist, the number of white droplets and the rectangular area corresponding to the largest white droplet are calculated to determine whether a disturbance recognition model is triggered. The disturbance recognition model is used to determine whether disturbances occur in the composite glass fiber. When disturbances occur, the composite glass fiber is judged to be broken by comparing the number of broken fiber frames of the composite glass fiber with a preset broken fiber frame number threshold. The present application can effectively detect whether glass fiber breakage occurs and promptly handle the breakage. The broken fiber recognition result is accurate, and the method is simple and efficient.

[0055] In this embodiment, the droplet recognition model infers the rectangular box corresponding to the white droplet. The rectangular box is represented by the coordinates of the upper left corner and lower right corner of the white droplet: ,in, is the coordinate of the upper left corner of the white droplet on the x-axis of the image, is the coordinate of the upper left corner of the white droplet on the y-axis of the image, is the coordinate of the lower right corner of the white droplet on the x-axis of the image, is the coordinate of the lower right corner of the white droplet on the y-axis of the image. The preset judgment area is represented by a rectangular box, which is represented by the coordinates of the upper left corner and lower right corner of the preset judgment area: ,in, It is the coordinate of the upper left corner of the preset judgment area on the x-axis of the image. The coordinate of the upper left corner of the preset judgment area on the y-axis of the image, It is the coordinate of the lower right corner of the preset judgment area on the x-axis of the image. The coordinate of the lower right corner of the preset judgment area on the y-axis of the image. The white droplet is within the judgment area and the following conditions must be met: , , , If the white droplets in the current image frame are determined to be within the preset judgment area, the number of frames in which white droplets exist increases by 1. When the number of frames in which white droplets exist is greater than the preset threshold value of frames in which white droplets exist, it is determined that white droplets exist and the number of white droplets N increases by 1. The initial value of N is 0, which is represented as follows: Formula (1) Among them, check is the sign of the presence of white droplets. When it is 1, white droplets exist, N=N+1; when it is 0, white droplets do not exist, N=N+0; m is the number of frames in which white droplets exist, and n is the preset frame number threshold.

[0056] Calculate the rectangular area corresponding to the largest white droplet and combine it with the number of white droplets N to determine whether to trigger the disturbance recognition model. This can be expressed by the following formula: Formula (2) Formula (3) Where S is the rectangular area corresponding to the largest white droplet, and N is the number of white droplets; and are the coordinates of the upper left corner of the white droplet i on the x-axis and y-axis of the image, and are the coordinates of the lower right corner of the white droplet i on the x-axis and y-axis of the image; test is the flag that triggers the perturbation recognition model. When it is 1, further analysis is performed through the perturbation recognition model; when it is 0, no further analysis is performed through the perturbation recognition model; S0 is the preset area threshold, and N0 is the preset number threshold.

[0057] The triggered disturbance recognition model is used to determine whether the composite glass fibers are disturbed, and further to determine whether the glass fibers are broken. This can be expressed by the following formula: Formula (4) Among them, check' is the sign of the composite material glass yarn disturbance. When check' is 1, the composite material glass yarn is disturbed, and the composite material glass yarn broken frame number increases by 1; when check' is 0, the composite material glass yarn is not disturbed.

[0058] Formula (5) Among them, test' is the sign of composite material glass fiber broken. When test' is 0, composite material glass fiber is not broken and no alarm is given. Figure 5 , Figure 5 This is a normal production diagram of a glass fiber broken wire identification method shown in an exemplary embodiment of the present application; when test' is 1, the composite material glass fiber is broken, an alarm is issued, and reference is made to Figure 6 , Figure 6 1 is a schematic diagram of a broken wire anomaly in a glass wire broken wire identification method according to an exemplary embodiment of the present application; j is the number of broken wire frames of the composite material glass wire, and k is the preset broken wire frame number threshold.

[0059] This application provides a technology that can timely and accurately detect broken glass fibers in composite materials and implement automated intervention, which has important practical significance for improving the quality and efficiency of glass fiber production and reducing production costs.

[0060] Figure 7 This is a block diagram of a glass fiber broken wire identification device shown in an exemplary embodiment of the present application. The device can be applied to Figure 1 The implementation environment shown is specifically configured in the computer device 102. The apparatus may also be applicable to other exemplary implementation environments and specifically configured in other devices. This embodiment does not limit the implementation environment to which the apparatus is applicable.

[0061] like Figure 7 As shown, the exemplary broken glass filament identification device includes: an image acquisition module 710 , an identification module 720 and a determination module 730 .

[0062] Among them, the image acquisition module 710 is used to obtain the glass fiber production image of the target production area; the recognition module 720 is used to perform droplet recognition on the glass fiber production image through a preset droplet recognition model to obtain a droplet recognition result, wherein the droplet recognition model is trained based on a first image sample with a droplet presence label; the judgment module 730 is used to determine that the glass fiber in the target production area is broken if the droplet recognition result is that droplets exist.

[0063] It should be noted that the broken glass wire identification device provided in the above embodiment and the broken glass wire identification method provided in the above embodiment are based on the same concept. The specific manner in which each module and unit performs operations has been described in detail in the method embodiment and will not be repeated here. In actual applications, the broken glass wire identification device provided in the above embodiment can, as needed, allocate the above functions to different functional modules, that is, divide the internal structure of the device into different functional modules to complete all or part of the functions described above, and this is not limited here.

[0064] An embodiment of the present application further provides an electronic device, comprising: one or more processors; a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the electronic device implements the glass wire broken wire identification method provided in the above-mentioned embodiments.

[0065] Figure 8 The following is a schematic diagram showing the structure of a computer system suitable for implementing an electronic device according to an embodiment of the present application. Figure 8 The computer system 800 of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0066] like Figure 8 As shown, computer system 800 includes a central processing unit (CPU) 801, which can perform various appropriate actions and processes, such as executing the methods provided in the various embodiments described above, based on programs stored in read-only memory (ROM) 802 or programs loaded from storage 808 into random access memory (RAM) 803. RAM 803 also stores various programs and data required for system operation. CPU 801, ROM 802, and RAM 803 are interconnected via a bus 804. An input / output (I / O) interface 805 is also connected to bus 804.

[0067] The following components are connected to the I / O interface 805: an input section 806 including a keyboard, mouse, and the like; an output section 807 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and speakers; a storage section 808 including a hard disk; and a communication section 809 including a network interface card such as a LAN (Local Area Network) card or a modem. The communication section 809 performs communication processing via a network such as the Internet. A drive 810 is also connected to the I / O interface 805 as needed. Removable media 811, such as a magnetic disk, optical disk, magneto-optical disk, or semiconductor memory, is installed in the drive 810 as needed, so that computer programs read from the media can be installed in the storage section 808 as needed.

[0068] In particular, according to embodiments of the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program including a computer program for executing the methods illustrated in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via the communication section 809 and / or installed from removable media 811. When executed by the central processing unit (CPU) 801, the computer program performs the various functions defined in the system of the present application.

[0069] It should be noted that the computer-readable medium described in the embodiments of this application may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may, for example, be an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or component, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying a computer-readable computer program. This propagated data signal may take a variety of forms, including, but not limited to, an electromagnetic signal, an optical signal, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. A computer program embodied on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, or any suitable combination thereof.

[0070] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. Among them, each box in the flowchart or block diagram can represent a module, program segment, or part of the code, and the above-mentioned module, program segment, or part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0071] The units involved in the embodiments described in this application may be implemented by software or hardware, and the units described may also be set in a processor. In some cases, the names of these units do not constitute limitations on the units themselves.

[0072] Another aspect of the present application provides a computer-readable storage medium storing a computer program. When executed by a computer processor, the computer program causes the computer to perform the broken glass fiber identification method provided in the above embodiments. The computer-readable storage medium may be included in the electronic device described in the above embodiments, or may exist independently and not be incorporated into the electronic device.

[0073] It should be noted that, although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiment of the application, the features and functions of two or more modules or units described above can be concretized in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into multiple modules or units to be concretized.

[0074] Another aspect of the present application provides a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the broken glass filament identification method provided in each of the above embodiments.

[0075] Through the description of the above embodiments, it is easy for those skilled in the art to understand that the example embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solution according to the embodiments of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a touch terminal, or a network device, etc.) to execute the method according to the embodiments of the present application.

[0076] Those skilled in the art will readily conceive of other embodiments of the present application after considering the specification and practicing the embodiments disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present application that follow the general principles of this application and include common knowledge or customary techniques in the art that are not disclosed herein.

[0077] The above embodiments are merely illustrative of the principles and effects of this application and are not intended to limit this application. Anyone skilled in the art may modify or alter the above embodiments without departing from the spirit and scope of this application. Therefore, any equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concepts disclosed in this application shall be covered by the claims of this application.

Claims

1. A method for identifying broken glass fibers, characterized in that: include: Acquire glass fiber production images of the target production area; Performing droplet recognition on the glass filament production image using a preset droplet recognition model to obtain a droplet recognition result, wherein the droplet recognition model is trained based on a first image sample with a droplet presence label; If the droplet identification result indicates that a droplet exists, it is determined that the glass filaments in the target production area are broken.

2. The broken glass fiber identification method according to claim 1, characterized in that: After performing droplet recognition on the glass filament production image using a preset droplet recognition model to obtain a droplet recognition result, the method further includes: If the droplet recognition result is that a droplet exists, the number of droplets and the area of ​​the rectangle corresponding to the largest droplet are calculated; If the number of droplets is greater than a preset number threshold, or the rectangular area corresponding to the largest droplet is greater than a preset area threshold, disturbance recognition is performed on the glass fiber production image using a preset disturbance recognition model to obtain a disturbance recognition result, wherein the disturbance recognition model is trained based on a second image sample with a disturbance occurrence label.

3. The method for identifying broken glass fibers according to claim 2, wherein: After performing disturbance recognition on the glass yarn production image using a preset disturbance recognition model and obtaining a disturbance recognition result, the method further includes: If the disturbance identification result is that a disturbance occurs, the number of glass filament broken frames is increased by one, wherein the initial value of the number of glass filament broken frames is a first preset value; If the number of broken glass yarn frames is greater than a preset broken glass yarn frame number threshold, it is determined that the glass yarn in the target production area is broken.

4. The method for identifying broken glass fibers according to any one of claims 1 to 3, wherein: After performing droplet recognition on the glass filament production image using a preset droplet recognition model to obtain a droplet recognition result, the method further includes: If the droplet recognition result is that a droplet exists, obtaining a first coordinate of a rectangular box corresponding to the droplet and obtaining a second coordinate of a target production area; If the droplet is determined to be within the target production area based on the first coordinate and the second coordinate, the number of droplet existence frames is increased by one; if the number of droplet existence frames is greater than a preset existence frame threshold, the number of droplets is increased by one, wherein an initial value of the droplet existence frame number is a second preset value, and an initial value of the droplet number is a third preset value; If it is determined based on the first coordinate and the second coordinate that the droplet is outside the target production area, the droplet identification result is updated to indicate that no droplet exists.

5. The method for identifying broken glass fibers according to any one of claims 1 to 3, characterized in that: The droplet recognition model is trained based on a first image sample with a droplet presence label, including: Acquire droplet presence labels and sample production images; Annotating the sample production image according to the droplet presence label to generate a first image sample with the droplet presence label; A preset droplet recognition model is trained based on the first image sample with the droplet presence label to obtain the droplet recognition model.

6. The method for identifying broken glass fibers according to any one of claims 1 to 3, characterized in that: The disturbance recognition model is trained based on a second image sample with a disturbance occurrence label, and includes: Obtain perturbation occurrence labels and sample production images; Annotating the sample production image according to the disturbance occurrence label to generate a second image sample with the disturbance occurrence label; A preset disturbance recognition model is trained based on the second image sample with the disturbance occurrence label to obtain the disturbance recognition model.

7. The method for identifying broken glass fibers according to any one of claims 1 to 3, characterized in that: After acquiring the glass fiber production image of the target production area, the method further includes: The glass fiber production image is subjected to image preprocessing, wherein the image preprocessing includes image slicing, image cleaning, image annotation and image review.

8. A broken glass fiber identification device, characterized in that: include: An image acquisition module, used to acquire glass fiber production images of a target production area; an identification module, configured to perform droplet identification on the glass filament production image using a preset droplet identification model to obtain a droplet identification result, wherein the droplet identification model is trained based on a first image sample with a droplet presence label; The determination module is configured to determine whether the glass filaments in the target production area are broken if the droplet identification result indicates that a droplet exists.

9. An electronic device, characterized in that: The electronic device comprises: one or more processors; A storage device is used to store one or more programs, and when the one or more programs are executed by the one or more processors, the electronic device implements the broken glass fiber identification method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that A computer program is stored thereon, and when the computer program is executed by a processor of a computer, the computer is caused to execute the broken glass filament identification method according to any one of claims 1 to 7.

Images