Positioning detection system for fabric defects

Abstract

The invention relates to a fabric defect spot positioning detection system, method and equipment and a medium. The system comprises an image acquisition module, an image processing module, a defect detection module and a detection result generation module, the image acquisition module is used for acquiring an original detection image of a to-be-detected fabric; the image processing module is used for segmenting the original detection image to obtain a to-be-detected image, and the to-be-detected image comprises a core to-be-detected image; the defect detection module is used for inputting the core to-be-detected image into a core defect point detection model to obtain core defect point detection information; the detection result generation module is used for generating a real-time early-warning defect detection result in response to the condition that the core defect point detection information meets the defect detection early-warning condition; and generating a fabric defect initial judgment grade detection result based on the core defect point detection information. By adopting the system, the fabric defect spots can be automatically detected, the actual distribution condition of the fabric defect spots can be accurately reflected, and the fabric quality is ensured to be in a controllable state.

Description

Technical Field

[0001] The present application relates to the field of intelligent manufacturing technology, and in particular to a system for locating and detecting fabric defect points. Background Art

[0002] The continuous development of technologies such as the Internet of Things and artificial intelligence has driven digital transformation in the field of intelligent manufacturing. During digital production, quality inspections of raw materials and semi-finished products are necessary to ensure product quality throughout the entire digital production process. In the modern textile industry, using advanced inspection technologies instead of the human eye to automatically detect fabric defects can improve inspection efficiency, reduce labor intensity, and further enhance fabric quality.

[0003] However, currently, manual inspection is often used to spot-check fabrics and define their quality grades. This defect detection method is labor-intensive, difficult to implement in real time during the digital production process, and suffers from poor stability and accuracy of test results, as well as low reproducibility, making it incompatible with digital production. Summary of the Invention

[0004] Based on this, it is necessary to provide a fabric defect point positioning detection system, method, computer equipment and storage medium that can automatically detect fabric defect points in real time during the digital generation process to address the above technical problems.

[0005] In a first aspect, the present application provides a fabric defect location detection system, comprising: an image acquisition module, an image processing module, a defect detection module, and a detection result generation module;

[0006] The image acquisition module is used to acquire the original detection image of the fabric to be detected, and the original detection image is used to represent the color and texture characteristics of the fabric to be detected in the fabric defect point detection area;

[0007] The image processing module is used to segment the original detection image to obtain the image to be detected, which includes the core image to be detected. The core image to be detected is used to obtain the initial judgment information of the fabric defect;

[0008] The defect detection module is used to input the core image to be inspected into the core defect point detection model to obtain core defect point detection information. The core defect point detection information is used to characterize the characteristic information of the defect point that plays a core and leading role in the inspection and evaluation of the grade of the fabric to be inspected;

[0009] The detection result generation module is used to generate a real-time warning defect detection result in response to the core defect point detection information meeting the defect detection warning conditions; and generate a fabric defect preliminary grade detection result based on the core defect point detection information.

[0010] In one embodiment, the fabric defect location detection system further includes: a parameter setting module;

[0011] The parameter setting module includes a component for acquiring the meta-information of the fabric to be detected, a component for matching the image segmentation coefficient, and a component for matching the feeding speed of the fabric to be detected;

[0012] The fabric meta-information acquisition component is used to acquire the fabric meta-information to be tested, including the fabric width to be tested, the parameters of the core defect point detection model applicable to the fabric to be tested, and the preset detection accuracy;

[0013] The image segmentation coefficient matching component is used to generate image segmentation coefficients based on the metadata of the fabric to be inspected. The image segmentation coefficients are used to represent the segmentation ratio relationship between the input image of the core defect point detection model and the original inspection image.

[0014] The fabric feed speed matching component to be detected is used to generate fabric feed speed control information to be detected based on the fabric meta-information to be detected and the image segmentation coefficient. The fabric feed speed control information to be detected is used to control the feed speed of the fabric to be detected in the fabric defect point detection area.

[0015] In one embodiment, the core defect point detection model is a first improved YOLOV8 model, which is obtained by replacing the neck network structure of the original YOLOV8 model with a bidirectional feature pyramid network structure with at least four layers.

[0016] In one embodiment, a global attention mechanism module is added between the spine network and the neck network of the first improved YOLOV8 model, and a replacement attention mechanism module is added between the neck network and the head network of the first improved YOLOV8 model.

[0017] In one embodiment, the image to be detected includes a detail image to be detected, and the pixel accuracy of the detail image to be detected is higher than that of the core image to be detected;

[0018] The defect detection module is further used to input the detail image to be detected into the detail defect point detection model to obtain detail defect point detection information;

[0019] The detection result generation module is also used to combine the core defect point detection information and the detail defect point detection information to generate a comprehensive fabric grade detection result; and, in response to the comprehensive fabric grade detection result meeting the defect area detection conditions, combine the core defect point detection information and the detail defect point detection information to generate a fabric defect area detection result, and the fabric defect area detection result is used to characterize the area where defects are concentrated in the fabric.

[0020] In one embodiment, the detail defect point detection model is a second improved YOLOV8 model. The second improved YOLOV8 model is obtained by deleting the convolution module and C2F module with the largest downsampling multiple in the spine network from the original YOLOV8 model, and replacing its neck network structure with a bidirectional feature pyramid network structure with three structural layers.

[0021] In one embodiment, the image processing module is further configured to pre-process the original detection image, including:

[0022] Adjust the resolution of the original detection image to generate a size-matched image;

[0023] Performing filtering on the size matching image to generate a filtered image;

[0024] The filtered image is subjected to denoising to generate an image to be segmented, and the original detection image is updated based on the image to be segmented.

[0025] In a second aspect, the present application also provides a method for locating and detecting fabric defect points, comprising:

[0026] Acquire an original inspection image of the fabric to be inspected, where the original inspection image is used to characterize the color and texture characteristics of the fabric to be inspected in the fabric defect point inspection area;

[0027] Segmenting the original inspection image to obtain an image to be inspected, wherein the image to be inspected includes a core image to be inspected, and the core image to be inspected is used to obtain preliminary judgment information of fabric defects;

[0028] Input the core image to be inspected into the core defect point detection model to obtain core defect point detection information. The core defect point detection information is used to characterize the characteristic information of the defect points that play a core leading role in the grade detection and evaluation of the fabric to be inspected.

[0029] In response to the core defect point detection information meeting the defect detection warning conditions, generating a real-time warning defect detection result;

[0030] Generate preliminary fabric defect grade detection results based on core defect point detection information.

[0031] In a third aspect, the present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0032] Acquire an original inspection image of the fabric to be inspected, where the original inspection image is used to characterize the color and texture characteristics of the fabric to be inspected in the fabric defect point inspection area;

[0033] Segmenting the original inspection image to obtain an image to be inspected, wherein the image to be inspected includes a core image to be inspected, and the core image to be inspected is used to obtain preliminary judgment information of fabric defects;

[0034] Input the core image to be inspected into the core defect point detection model to obtain core defect point detection information. The core defect point detection information is used to characterize the characteristic information of the defect points that play a core leading role in the grade detection and evaluation of the fabric to be inspected.

[0035] In response to the core defect point detection information meeting the defect detection warning conditions, generating a real-time warning defect detection result;

[0036] Generate preliminary fabric defect grade detection results based on core defect point detection information.

[0037] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the following steps:

[0038] Acquire an original inspection image of the fabric to be inspected, where the original inspection image is used to characterize the color and texture characteristics of the fabric to be inspected in the fabric defect point inspection area;

[0039] Segmenting the original inspection image to obtain an image to be inspected, wherein the image to be inspected includes a core image to be inspected, and the core image to be inspected is used to obtain preliminary judgment information of fabric defects;

[0040] Input the core image to be inspected into the core defect point detection model to obtain core defect point detection information. The core defect point detection information is used to characterize the characteristic information of the defect points that play a core leading role in the grade detection and evaluation of the fabric to be inspected.

[0041] In response to the core defect point detection information meeting the defect detection warning conditions, generating a real-time warning defect detection result;

[0042] Generate preliminary fabric defect grade detection results based on core defect point detection information.

[0043] The above-mentioned fabric defect positioning detection system, method, computer equipment and storage medium can realize automated detection of fabric defect points through the coordinated cooperation of image acquisition, image processing and defect detection modules. Compared with manual sampling, it can process a large number of fabrics to be inspected more quickly, greatly improving the overall efficiency of fabric defect point detection, and better meeting the requirements of fast and efficient digital production.

[0044] Furthermore, the above-mentioned fabric defect point positioning detection system, method, computer equipment and storage medium obtain core defect point detection information through the core defect point detection model. Based on the established technical process and model algorithm, the interference of human factors is reduced, so that the accuracy, stability and reproducibility of the detection results can be significantly improved, and the actual distribution of fabric defect points can be more accurately reflected.

[0045] Furthermore, the fabric defect location detection system, method, computer device, and storage medium described above, through a detection result generation module, can promptly generate real-time warning defect detection results, enabling the immediate identification of fabric defects during the production process, allowing for the rapid and accurate implementation of appropriate intervention measures. Furthermore, the system can also generate preliminary fabric defect grade detection results, helping to accurately assess fabric quality and rationally manage fabric quality grades, thereby ensuring that product quality remains under control throughout the digital production process. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0047] Figure 1 A schematic diagram of an application environment in one embodiment of the present application;

[0048] Figure 2 This is a structural diagram of a fabric defect location detection system in one embodiment of the present application;

[0049] Figure 3 This is a structural diagram of another fabric defect location detection system in one embodiment of the present application;

[0050] Figure 4 This is a schematic structural diagram of a core defect point detection model in one embodiment of the present application;

[0051] Figure 5 This is a schematic structural diagram of another core defect point detection model in one embodiment of the present application;

[0052] Figure 6 This is a schematic structural diagram of a detail defect point detection model in one embodiment of the present application;

[0053] Figure 7 This is a flowchart of an image preprocessing method in one embodiment of the present application;

[0054] Figure 8 The figure is a flow chart of a method for locating and detecting fabric defect points in one embodiment of the present application. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0056] The fabric defect point positioning detection system provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. The computing device 101 communicates with the visual acquisition device 102 and the fabric transmission device 103 through a communication channel. The computing device 101 may be, but is not limited to, various edge computing devices, various distributed computing devices, personal computers, laptops, smart phones, tablet computers, Internet of Things devices, and portable wearable devices. The computing device 101 can be implemented using an independent server or a server cluster consisting of multiple servers. The visual acquisition device 102 may be, but is not limited to, an industrial camera, a high-definition camera, an intelligent visual sensor, and a visual acquisition device. The fabric transmission device 103 may be, but is not limited to, a conveyor belt system, a roller conveyor device, a suspended transmission device, and a pneumatic transmission system.

[0057] For example, the computing device 101 may obtain fabric defect distribution information based on the visual acquisition device 102 , and may generate a fabric grade evaluation result based on the fabric defect distribution information, thereby controlling the working state of the fabric transmission device 103 .

[0058] In an exemplary embodiment, Figure 2 As shown, a fabric defect location detection system 200 is provided, which is applied to Figure 1 The computing device 101 in FIG. 1 is used as an example to illustrate the embodiment, which includes an image acquisition module 201 , an image processing module 202 , a defect detection module 203 and a detection result generation module 204 .

[0059] Specifically, the computing device 101 can obtain the original detection image of the fabric to be detected through the communication network via the visual acquisition device based on the image acquisition module 201. The original detection image is used to represent the color and texture characteristics of the fabric to be detected in the fabric defect point detection area.

[0060] For example, the expression of the resolution of the original detection image can be:

[0061] [NK-(2N-2)U]×(K-2V)

[0062]

[0063] Where [NK-(2N-2)U] is the radial resolution of the fabric movement, (K-2V) is the resolution in the direction of the fabric movement, N is the segmentation coefficient of the original detection image, K is the input resolution coefficient of the fabric defect detection model, D is the segmentation compensation coefficient of the original detection image along the radial direction of the fabric movement, V is the segmentation compensation coefficient of the original detection image along the direction of the fabric movement, L is the width of the fabric, To round up, is the rounding down term, D is the pixel precision coefficient, and S is the rounding adjustment threshold.

[0064] Specifically, the computing device 101 may segment the original detection image based on the image processing module 202 to obtain the image to be detected, where the image to be detected includes a core image to be detected, and the core image to be detected is used to obtain preliminary judgment information of fabric defects.

[0065] Schematically, the computing device 101 can copy part of the pixels of the adjacent images at the segmented edges to the corresponding positions of the other edge during the process of segmenting the original detection image to obtain the image to be detected, so as to supplement the edge information and make the segmented edge more complete and continuous. The pixel size copied at the edge segmented along the radial direction of the fabric movement can be U, and the pixel size copied at the edge segmented along the direction of the fabric movement can be V.

[0066] Exemplarily, the resolution of the core image to be detected obtained by the computing device 101 through the image processing module 202 may be 640×640, and the core image to be detected may be stored in the computing device 101 in RGB format.

[0067] Specifically, the computing device 101 can input the core image to be inspected into the core defect point detection model mounted on the computing device 101 based on the defect detection module 203 to obtain core defect point detection information. The core defect point detection information is used to characterize the characteristic information of the defect points that play a core and dominant role in the detection and evaluation of the grade of the fabric to be inspected.

[0068] Schematically, the core defect point detection model can be a YOLO series model. The core defect point detection model can use the following loss function expression to obtain the defect point pre-selection box information:

[0069]

[0070] Where IoU is the intersection-over-union loss term, (x, y) is the center coordinate of the prediction box, w, h are the width and height of the prediction box, (x gt ,y gt ) is the center coordinate of the real frame, w gt and h gt is the width and height of the real box, and c is the diagonal length of the minimum closure area that can contain both the predicted box and the real box.

[0071] Furthermore, the core defect point detection model can use the following loss function expression to obtain defect point classification information:

[0072]

[0073] Where, L KD is the temperature distillation loss based on KL divergence, T is the knowledge distillation temperature, is the output probability distribution of the k-th classification in the teacher model, is the output probability distribution of the kth classification in the student model.

[0074] Schematically, the core defect point detection information represents the characteristic information of the defect points that play a core and dominant role in the inspection and evaluation of the grade of the fabric to be inspected. Its data format can be a structured data form containing relevant parameters such as the location, type, severity, etc. of the defect points, such as a specific data table or feature vector.

[0075] Specifically, the computing device 101 can generate a real-time warning defect detection result based on the detection result generation module 204 in response to the core defect point detection information meeting the defect detection warning condition, and generate a fabric defect preliminary grade detection result based on the core defect point detection information.

[0076] Illustratively, computing device 101 can pre-configure defect detection and warning conditions based on industry standards and enterprise production requirements. These conditions may include, but are not limited to, defect area exceeding a certain threshold or identification of a specific defect type. These conditions can be stored in a configuration file of computing device 101 or in a database hosted on computing device 101 in the form of a rule list, logical expression, or the like.

[0077] Furthermore, the computing device 101 can perform a comprehensive evaluation of each of the core defect point detection information according to the grading rules, and can determine the fabric defect preliminary judgment grade through the fabric defect preliminary judgment algorithm. The feasible expression of the fabric defect preliminary judgment algorithm is as follows:

[0078]

[0079] Where, α is the initial grade test result of fabric defects, τ A , τ B and τ C is the possible initial judgment level of fabric defects, S is the area of ​​the pre-selected box of the defect point, ρ is the weight coefficient corresponding to the type of defect point, π A , π B and π C It is the initial grade threshold for fabric defects.

[0080] In the above-mentioned fabric defect location detection system, with the coordinated cooperation of image acquisition, image processing, and defect detection modules, automated detection of fabric defects can be achieved, and a large number of fabrics to be tested can be quickly processed, greatly improving the overall efficiency of fabric defect detection. In addition, the above-mentioned fabric defect location detection system can obtain information through the core defect detection model and, based on established technical processes and model algorithms, significantly improve the accuracy, stability, and reproducibility of the detection results, more accurately reflecting the actual distribution of fabric defects. In addition, the above-mentioned fabric defect location detection system utilizes the detection result generation module to not only generate real-time early warning defect detection results in a timely manner, discovering defective fabrics at the first time during production so that intervention measures can be taken quickly and accurately, but also generate preliminary fabric defect grade detection results, which helps to understand the quality status of the fabric and ensure that product quality is controllable during the digital production process.

[0081] In one embodiment, please refer to Figure 3 The fabric defect point positioning detection system 3000 also includes: a parameter setting module 3010.

[0082] The parameter setting module 3010 includes a to-be-detected fabric meta-information acquisition component 3011, an image segmentation coefficient matching component 3012, and a to-be-detected fabric feed speed matching component 3013;

[0083] Specifically, the fabric meta-information acquisition component 3011 can be used to acquire the fabric meta-information to be detected, which may include the width of the fabric to be detected, parameters of the core defect point detection model applicable to the fabric to be detected, and preset detection accuracy.

[0084] Specifically, the image segmentation coefficient matching component 3012 can be used to generate image segmentation coefficients based on the metadata of the fabric to be detected. The image segmentation coefficients can be used to represent the segmentation ratio relationship between the input image of the core defect point detection model and the original detection image.

[0085] Specifically, the fabric feed speed matching component 3013 can be used to generate fabric feed speed control information based on the fabric meta-information and image segmentation coefficient. The fabric feed speed control information is used to control the feed speed of the fabric in the fabric defect detection area.

[0086] For example, a feasible expression for the feed speed control information of the fabric to be detected is as follows:

[0087]

[0088] Where μ is the feed speed control information of the fabric to be detected, K' is the input pixel parameter of the core defect point detection model, D' is the preset detection accuracy, N' is the original detection image segmentation coefficient, σ is the fabric feed redundancy coefficient, and T is the response time parameter of the core defect point detection model.

[0089] Illustratively, the fabric defect location detection system 3000 may include an image processing module 3020 , an image processing module 3030 , a defect detection module 3040 and a detection result generation module 3050 .

[0090] In the above-mentioned fabric defect point positioning detection system, the fabric feed speed matching component can be used to accurately adjust the feed speed of fabric defect point detection, thereby improving the stability and real-time performance of the fabric defect point positioning detection system.

[0091] In one embodiment, please refer to Figure 4 The core defect point detection model can be a first improved YOLOV8 model. The first improved YOLOV8 model is obtained by replacing the neck network structure of the original YOLOV8 model with a bidirectional feature pyramid network structure with at least four layers.

[0092] Schematically, the bidirectional feature pyramid network structure realizes cross-scale connection by constructing a bidirectional channel, fusing the features in the feature extraction network with the relative size features in the bottom-up path, retaining the shallow semantic information without losing too much deep semantic information. The bidirectional feature pyramid network structure uses a fast normalized fusion method for weighted fusion, which can scale the weights to the range of 0 to 1, thereby increasing the training speed. Cross-scale connection is achieved by adding skip connections and bidirectional paths, realizing weighted fusion and bidirectional cross-scale connection. The specific calculation process is as follows:

[0093]

[0094] Among them, O is the output feature, I i is the input feature, ω is the weight of each node, and ε is the learning rate.

[0095] Furthermore, the learning rate ε can be set to 0.0001 to ensure the stability of the training process and avoid fluctuations in the results.

[0096] In one embodiment, please refer to Figure 5 A global attention mechanism module can be added between the spine network and the neck network of the first improved YOLOV8 model, and a replacement attention mechanism module can be added between the neck network and the head network of the first improved YOLOV8 model.

[0097] Specifically, the global attention mechanism module can include a channel attention mechanism submodule and a spatial attention mechanism submodule. The channel attention mechanism submodule uses a three-dimensional arrangement to retain information in three dimensions, and then uses a two-layer MLP to amplify the cross-dimensional channel-spatial dependency. The feasible expression of the channel attention mechanism submodule is as follows:

[0098]

[0099] In the formula, F1 is the input feature map, F2 is the intermediate feature map, F3 is the output feature map, and M c (F1) is the channel feature map, M s (F2) is the spatial feature map, Represents element-wise multiplication.

[0100] For example, the permutation attention mechanism module can first group the channel dimension into multiple sub-features and then process them in parallel. For each sub-feature, the permutation attention mechanism module uses a channel permutation unit to describe the feature dependencies in the spatial and channel dimensions. It then aggregates all sub-features and uses the channel permutation operator to communicate information between different sub-features, achieving better performance without increasing computational cost.

[0101] In one embodiment, the image to be detected may include a detail image to be detected, and the pixel accuracy of the detail image to be detected is higher than that of the core image to be detected.

[0102] The defect detection module can also be used to input the detail image to be detected into the detail defect point detection model to obtain detail defect point detection information.

[0103] The test result generation module can also be used to combine the core defect point detection information and the detail defect point detection information to generate a comprehensive fabric grade test result. Furthermore, in response to the comprehensive fabric grade test result meeting the defect area detection conditions, the module can be used to combine the core defect point detection information and the detail defect point detection information to generate a fabric defect area detection result. The fabric defect area detection result is used to characterize the area in the fabric where defects are concentrated.

[0104] In one embodiment, please refer to Figure 6 The detail defect detection model is the second improved YOLOV8 model. The second improved YOLOV8 model is obtained by deleting the convolution module and C2F module with the largest downsampling multiple in the spine network from the original YOLOV8 model, and replacing its neck network structure with a three-layer bidirectional feature pyramid network structure.

[0105] In one embodiment, please refer to Figure 7 ,The image processing module is also used to pre-process the original detection image, including:

[0106] Step S701: Adjust the resolution of the original detection image to generate a size-matched image.

[0107] Step S702, filtering the size matching image to generate a filtered image;

[0108] Step S703 , performing noise reduction processing on the filtered image to generate an image to be segmented, and updating the original detection image based on the image to be segmented.

[0109] The above-mentioned fabric defect positioning and detection system can realize automatic detection of fabric defect points, quickly process a large number of fabrics to be tested, greatly improve the overall efficiency of fabric defect detection, meet the requirements of digital production for fast and efficient detection, and effectively reduce the time and labor costs consumed by traditional methods such as manual sampling.

[0110] Furthermore, the above-mentioned fabric defect positioning detection system can supplement edge information by copying part of the pixels of adjacent images at the segmentation edge to the corresponding position of the other edge during the process of segmenting the original detection image to obtain the image to be detected, which helps to improve the accuracy of subsequent defect detection based on the image to be detected and avoid detection errors caused by missing edge information.

[0111] Furthermore, the above-mentioned fabric defect point positioning detection system can use the fabric feed speed matching component to be detected in the parameter setting module to generate fabric feed speed control information to be detected based on the fabric metadata to be detected and the image segmentation coefficient, and can accurately adjust the feed speed of fabric defect point detection, thereby improving the stability and real-time performance of the fabric defect point positioning detection system and adapting to the needs of different fabrics and detection scenarios.

[0112] Furthermore, the above-mentioned fabric defect positioning and detection system can retain and amplify information in different dimensions by adding a global attention mechanism module and a replacement attention mechanism module between different network structures of the first improved YOLOV8 model, thereby enhancing the model's attention to and extraction capabilities of key features, improving performance without increasing computing costs, and enabling the model to focus more on features related to fabric defects, thereby improving detection accuracy and efficiency.

[0113] Based on the same inventive concept, the present application also provides a method for locating and detecting fabric defects, which is used to implement the aforementioned system for locating and detecting fabric defects. The solution provided by this method is similar to the solution described in the aforementioned method. Therefore, the specific limitations of the following embodiments of the method for locating and detecting one or more fabric defects can be found in the aforementioned limitations on the system for locating and detecting fabric defects, and will not be further elaborated here.

[0114] In an exemplary embodiment, Figure 8 As shown, a method for locating and detecting fabric defect points is provided, comprising:

[0115] Step S801: Acquire an original inspection image of the fabric to be inspected.

[0116] Step S802: Segment the original detection image to obtain an image to be detected.

[0117] Step S803: input the core image to be detected into the core defect point detection model to obtain core defect point detection information.

[0118] Step S804 : In response to the core defect point detection information meeting the defect detection warning condition, a real-time warning defect detection result is generated.

[0119] Step S805: Generate a fabric defect preliminary grade detection result based on the core defect point detection information.

[0120] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0121] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the method for locating and detecting fabric defect points as described above are implemented.

[0122] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0123] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to the partial description of the method embodiments. The device embodiments described above are merely illustrative, wherein the components described as separate parts may or may not be physically separated, and the parts displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the disclosed solution. A person of ordinary skill in the art can understand and implement it without expending creative work.

[0124] The above-described embodiments merely represent several implementation methods of the embodiments of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that a person skilled in the art may make various modifications and improvements without departing from the concept of the embodiments of the present application, and these modifications and improvements fall within the scope of protection of the embodiments of the present application.

Claims

1. A fabric defect location detection system, characterized in that: The system comprises: an image acquisition module, an image processing module, a defect detection module and a detection result generation module; The image acquisition module is used to acquire an original detection image of the fabric to be detected, and the original detection image is used to characterize the color and texture characteristics of the fabric to be detected in the fabric defect point detection area; The image processing module is used to segment the original detection image to obtain the image to be detected, and the image to be detected includes a core image to be detected, and the core image to be detected is used to obtain the initial judgment information of the fabric defect; The defect detection module is used to input the core image to be detected into the core defect point detection model to obtain core defect point detection information, and the core defect point detection information is used to characterize the characteristic information of the defect point that plays a core leading role in the grade detection and evaluation of the fabric to be detected; The detection result generation module is used to generate a real-time warning defect detection result in response to the core defect point detection information meeting the defect detection warning condition; and generate a fabric defect preliminary grade detection result based on the core defect point detection information.

2. The system according to claim 1, characterized in that The system further comprises: a parameter setting module; The parameter setting module includes a to-be-detected fabric meta-information acquisition component, an image segmentation coefficient matching component, and a to-be-detected fabric feed speed matching component; The fabric meta information acquisition component to be detected is used to acquire the fabric meta information to be detected, wherein the fabric meta information to be detected includes the width of the fabric to be detected, the parameters of the core defect point detection model applicable to the fabric to be detected, and the preset detection accuracy; The image segmentation coefficient matching component is used to generate an image segmentation coefficient based on the metadata of the fabric to be detected, and the image segmentation coefficient is used to represent the segmentation ratio relationship between the input end image of the core defect point detection model and the original detection image; The fabric feed speed matching component to be detected is used to generate fabric feed speed control information to be detected based on the fabric meta information to be detected and the image segmentation coefficient, and the fabric feed speed control information to be detected is used to control the feed speed of the fabric to be detected in the fabric defect point detection area.

3. The system according to claim 1, characterized in that The core defect point detection model is a first improved YOLOV8 model, which is obtained by replacing the neck network structure of the original YOLOV8 model with a bidirectional feature pyramid network structure with at least four structural layers.

4. The system according to claim 3, characterized in that A global attention mechanism module is added between the spine network and the neck network of the first improved YOLOV8 model, and a replacement attention mechanism module is added between the neck network and the head network of the first improved YOLOV8 model.

5. The system according to claim 1, characterized in that The image to be detected includes a detail image to be detected, and the pixel accuracy of the detail image to be detected is higher than that of the core image to be detected; The defect detection module is also used to input the detail image to be detected into the detail defect point detection model to obtain detail defect point detection information; The detection result generation module is also used to combine the core defect point detection information and the detail defect point detection information to generate a comprehensive fabric grade detection result; and, in response to the comprehensive fabric grade detection result meeting the defect area detection condition, combine the core defect point detection information and the detail defect point detection information to generate a fabric defect area detection result, wherein the fabric defect area detection result is used to characterize the area where defects are concentrated in the fabric.

6. The system according to claim 5, characterized in that The detail defect point detection model is the second improved YOLOV8 model, which is obtained by deleting the convolution module and C2F module with the largest downsampling multiple in the spine network from the original YOLOV8 model, and replacing its neck network structure with a bidirectional feature pyramid network structure with three structural layers.

7. The system according to any one of claims 1 to 6, characterized in that The image processing module is also used to pre-process the original detection image, including: Adjusting the resolution of the original detection image to generate a size-matched image; Performing filtering processing on the size matching image to generate a filtered image; The filtered image is subjected to denoising to generate an image to be segmented, and the original detection image is updated based on the image to be segmented.

8. A method for locating and detecting fabric defect points, characterized in that: The method comprises: Acquire an original detection image of the fabric to be detected, wherein the original detection image is used to characterize the color and texture characteristics of the fabric to be detected in the fabric defect point detection area; Segmenting the original detection image to obtain an image to be detected, wherein the image to be detected includes a core image to be detected, and the core image to be detected is used to obtain preliminary judgment information of fabric defects; Inputting the core image to be detected into the core defect point detection model to obtain core defect point detection information, wherein the core defect point detection information is used to characterize the characteristic information of the defect point that plays a core leading role in the grade detection and evaluation of the fabric to be detected; In response to the core defect point detection information meeting the defect detection warning condition, generating a real-time warning defect detection result; A fabric defect preliminary grade detection result is generated based on the core defect point detection information.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to claim 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to claim 8 are implemented.

Images