A vision-based yarn defect detection system for sizing machines

Abstract

This invention discloses a vision-based yarn defect detection system for sizing machines, relating to the field of yarn defect detection technology. It solves the technical problem in existing technologies where the influence of conveying force cannot be eliminated during dynamic detection, thus reducing the accuracy of defect detection. Specifically, a static detection architecture performs static detection on the yarn, defining a static stage and a relatively static stage within the sizing machine's operating phase. Each stage is analyzed to infer the defect location in the yarn's static state and sent to a defect monitoring and management platform. After completing the static yarn detection, the yarn enters the dynamic stage when the sizing machine starts operating, and dynamic detection is performed. The influence of conveying force detection on yarn transport is analyzed, and after eliminating the transport influence, the yarn image is analyzed to obtain the defect location. Based on the location analysis, the defect type is defined and sent to the defect monitoring and management platform.

Description

Technical Field

[0001] This invention relates to the field of yarn defect detection technology, specifically a sizing machine yarn defect detection system based on visual analysis. Background Technology

[0002] Sizing machines are key equipment in the textile industry used to sizing warp yarns. Their core function is to improve the strength, abrasion resistance, and smoothness of yarns by coating the yarn surface with sizing agents, thereby reducing the breakage rate during weaving and improving production efficiency and fabric quality. Visual analytics is an interdisciplinary field that integrates technologies such as computer vision, data visualization, machine learning, and human-computer interaction. It aims to discover patterns or complete specific tasks through the intelligent processing and analysis of visual data such as images, videos, and visualizations.

[0003] However, in the existing technology, when the sizing machine is running, it is not possible to detect defects in both static and dynamic states of the yarn. In the static stage, it is not possible to detect defects based on the pre-operation inspection and the relatively static stage. In addition, the influence of conveying force cannot be eliminated when detecting defects in the dynamic stage, which reduces the accuracy of defect detection. Furthermore, it is not possible to classify the type of defect location, which makes it impossible to perform defect repair in a targeted manner.

[0004] To address the aforementioned technical shortcomings, a solution is proposed. Summary of the Invention

[0005] The purpose of this invention is to solve the problems mentioned above by proposing a vision analysis-based yarn defect detection system for sizing machines.

[0006] The objective of this invention can be achieved through the following technical solutions:

[0007] The vision-based yarn defect detection system for sizing machines includes a defect monitoring and management platform that monitors several processing steps in real time; and the communication connection includes both static and dynamic detection architectures.

[0008] The static detection architecture performs static detection on the yarn, defines the static phase and the relatively static phase during the operation of the sizing machine, and analyzes each phase to infer the location of defects in the yarn under static conditions and sends the results to the defect monitoring and management platform.

[0009] After the yarn static inspection is completed, the yarn enters the dynamic stage when the sizing machine is running, and dynamic inspection is carried out. The influence of yarn conveying is analyzed based on the conveying force detection. After eliminating the influence of conveying, the yarn image is analyzed to obtain the defect location. The defect type is defined based on the location analysis, and the data is sent to the defect monitoring and management platform.

[0010] In a preferred embodiment of the present invention, when the yarn is in a fixed storage state before the sizing machine is running, the current stage is marked as a static stage; and when the yarn leaves the static stage, the yarn conveying speed is in a constant and uniform state when the sizing machine is running, and the current stage is marked as a relatively static stage.

[0011] In a preferred embodiment of the present invention, yarn images are acquired during the static phase, and vertical and horizontal cross-sectional images of the yarn in a fixed storage state are obtained based on the acquired images. The yarn diameter width variation points within the vertical and horizontal cross-sectional images are obtained through the acquired cross-sectional images, i.e., the intersection points of two yarns with different diameter widths. The lengths of each diameter width of the yarn are obtained based on the corresponding diameter width variation. The deviation between the corresponding diameter width length and the set diameter width length of the yarn is acquired in real time during the static phase, as well as the deviation between the diameter width difference of adjacent diameter width variations and the set diameter width difference, and the acquired deviation values ​​are analyzed.

[0012] In a preferred embodiment of the present invention, if either of the two deviation values ​​collected in real time during the static stage exceeds the corresponding set threshold, the corresponding yarn width segment is designated as a defect location; if neither of the two deviation values ​​collected in real time during the static stage exceeds the corresponding set threshold, the corresponding yarn width segment is designated as a defect-free location and uploaded to the static detection architecture and sent to the defect monitoring and management platform.

[0013] In a preferred embodiment of the present invention, a relatively static stage is entered. During the relatively static stage, the sizing coating is set according to the yarn diameter. The coating of the yarn in the relatively static stage is detected by image acquisition. The coating uniformity deviation of each position of the yarn segment with the same diameter in the relatively static stage is obtained. If the coating uniformity deviation exceeds the set coating error threshold, the defect type analysis is performed on the current coating.

[0014] The location of the yarn segment with coating uniformity is obtained; if there is a coating uniformity deviation at a single yarn segment location, the defect type is marked as a single surface defect and uploaded to the static detection architecture, and then transferred to the defect monitoring and management platform. The current location is repaired, and after the repair is completed, the static detection architecture sends the repair control signal back to the defect detection end.

[0015] If there are deviations in the uniformity of coating at multiple yarn segments and the values ​​are inconsistent, the defect type is marked as a continuous surface defect and uploaded to the static detection architecture. It is then transferred to the defect monitoring and management platform, and the current yarn segment is paused for defect repair and recoating. After completion, the coating completion signal is sent back to the defect detection end.

[0016] In a preferred embodiment of the present invention, after the sizing machine is running, the yarn enters a dynamic stage. During the yarn transport process, the real-time status of the yarn is detected, and the maximum width of the yarn in the relaxed state during the dynamic stage is obtained from the maximum width of the coating area set under the current diameter. The width deviation value is obtained by comparing the width and marked as coating missing information. When the yarn is in a taut state during the dynamic stage and the yarn transmission vibrates, the coating thickness deviation value of the yarn corresponding to the coating area during the vibration period is obtained and marked as coating deviation information.

[0017] In a preferred embodiment of the present invention, if the coating missing information exceeds the width deviation threshold, or the coating deviation information exceeds the thickness deviation threshold, it is inferred that the yarn coating state is abnormal under different conveying forces, marked as an abnormal state and uploaded to the dynamic detection architecture. After receiving the abnormal state, the dynamic detection architecture adjusts the dynamic conveying of the yarn to ensure that the yarn force is within the set range and avoids the occurrence of a loose or tight state.

[0018] If the missing coating information does not exceed the width deviation threshold and the coating deviation information does not exceed the thickness deviation threshold, it is inferred that the yarn coating status is normal under different conveying forces. It is marked as normal and uploaded to the dynamic detection architecture. After receiving the normal status, the dynamic detection architecture will send the real-time normal status back to the image acquisition end and analyze the image.

[0019] In a preferred embodiment of the present invention, the real-time acquired images are analyzed to perform defect identification and detection;

[0020] During the coating process, images of the conveyed yarn are acquired, and the yarn is divided into several positions based on the acquired images, which are marked as sub-positions. The color difference of each adjacent sub-position within the yarn is acquired, and the location of defects is obtained by comparing the color difference with the real-time display based on the range of color difference changes after the coating slurry is exposed to air.

[0021] After coating is applied to the defective location, the coating thickness is compared with that of the adjacent coated locations. If the thickness of the defective location is higher than that of the adjacent coated locations, it is inferred that the defect at the defective location is of the adhesion type. During the delivery phase of the defective location, the coating thickness at the defective location is continuously monitored.

[0022] As a preferred embodiment of the present invention, if there is a coating thickness fluctuation and the surface adhesion type of the adhering material is an external adhering material or an adhering material with low adhesion, then the defect type of the defect location is marked as an easily changeable adhesion type and uploaded to the dynamic detection architecture together with the defect location.

[0023] If there is no coating thickness fluctuation, and the surface adhesion type is yarn component adhesion or highly adhesive adhesion, then the defect type at the defect location will be marked as difficult to change connection type, and uploaded to the dynamic detection architecture along with the defect location.

[0024] If the thickness at the defect location is lower than that of the adjacent coating location, the defect at the defect location is inferred to be a defect type, and the defect location and defect type are uploaded together to the dynamic detection architecture.

[0025] Compared with the prior art, the beneficial effects of the present invention are:

[0026] 1. In this invention, static detection is used to infer whether there are defects in the yarn before the sizing machine is started, thereby ensuring the reliability of the current sizing machine operation. At the same time, the changes in yarn defects can be used to infer whether the current operation of the sizing machine affects the yarn and causes defects, so as to make timely adjustments to the sizing machine process or components.

[0027] 2. In this invention, the yarn is subjected to different states by conveying force detection, and the coating of the yarn is inferred from the different states. This helps to overcome the influence of the sizing machine operation when detecting yarn defects, improves the accuracy of yarn defect detection, and at the same time, the location of yarn defects can be more clearly inferred from image analysis, so as to accurately repair and improve the sizing coating efficiency and the high efficiency of yarn use. Attached Figure Description

[0028] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0029] Figure 1 This is a schematic diagram of the overall principle of the present invention;

[0030] Figure 2 This is a principle block diagram of Embodiment 1 of the present invention;

[0031] Figure 3 This is a principle block diagram of Embodiment 2 of the present invention. Detailed Implementation

[0032] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0033] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0034] Please see Figure 1 As shown, the sizing machine yarn defect detection system based on visual analysis includes a defect monitoring and management platform, which monitors several processing steps in real time to detect yarn defects in each step when the sizing machine is running; in addition, the defect monitoring and management platform has a communication connection with a static detection architecture and a dynamic detection architecture.

[0035] Example 1

[0036] Please see Figure 2 As shown, the static detection architecture performs static detection on the yarn. Through static detection, it infers whether there are defects in the yarn before the sizing machine is running, thereby ensuring the reliability of the current sizing machine operation. At the same time, it can infer whether the current operation of the sizing machine affects the yarn and causes defects by the changes in yarn defects, so as to make timely adjustments to the sizing machine process or components.

[0037] Static definition is made during the operation phase of the sizing machine. When the yarn is in a fixed storage state before the sizing machine starts running, the current phase is marked as a static phase. The fixed storage state refers to the storage state of the yarn, such as the spooled state.

[0038] Furthermore, after the yarn leaves the static stage, the yarn conveying speed is constant and uniform when the sizing machine is running, and the current stage is marked as the relatively static stage.

[0039] The static phase is analyzed by acquiring yarn images and obtaining vertical and horizontal cross-sectional images of the yarn in a fixed storage state. The yarn diameter variation points within the vertical and horizontal cross-sectional images are obtained through the acquired cross-sectional images, i.e., the intersection points of two yarns with different diameters. The length of each diameter of the yarn is obtained based on the corresponding diameter variation.

[0040] The system collects real-time data on the deviation between the corresponding yarn diameter length and the set yarn diameter length during the static phase, as well as the deviation between the yarn diameter width difference for adjacent diameter width changes and the set diameter width difference. The collected deviation values ​​are then analyzed.

[0041] If either of the two deviation values ​​collected in real time during the static phase exceeds the corresponding set threshold, it is inferred that there is a defect in the yarn during the static phase. The corresponding yarn segment with the corresponding diameter is taken as the defect location and uploaded to the static detection architecture and sent to the defect monitoring and management platform. After receiving the defect, the defect monitoring and management platform replaces the current yarn or repairs the defect.

[0042] If the two deviation values ​​collected in real time during the static stage do not exceed the corresponding set threshold, it is inferred that there are no defects in the yarn during the static stage. The corresponding yarn segment with the corresponding diameter is taken as the defect-free position and uploaded to the static detection architecture and sent to the defect monitoring and management platform.

[0043] After the static stage confirms that there are no defects, the sizing machine starts running and the yarn enters a relatively static stage. It should be explained that if the yarn does not move at a uniform speed after the sizing machine starts running, the dynamic detection architecture will intervene. When the yarn moves at a uniform speed, it enters a relatively static stage, where the static detection architecture will detect defects.

[0044] During the relatively static stage, the sizing coating is set according to the yarn diameter. The coating is detected by image acquisition, and the coating uniformity deviation of each position of the yarn segment with the same diameter is obtained. If the coating uniformity deviation exceeds the set coating error threshold, the defect type analysis is performed on the current coating.

[0045] To obtain the location of yarn segments with coating uniformity, it should be noted that the specifications of the yarn segment locations are uniform. If a single yarn segment location has a coating uniformity deviation, the defect type is marked as a single surface defect and uploaded to the static detection architecture. It is then transferred to the defect monitoring and management platform, where defect repair is performed at the current location. After the repair is completed, the static detection architecture sends the repair control signal back to the defect detection end.

[0046] If there are deviations in the uniformity of coating at multiple yarn segments and the values ​​are inconsistent, the defect type is marked as a continuous surface defect and uploaded to the static detection architecture. It is then transferred to the defect monitoring and management platform, and the current yarn segment is paused for defect repair and recoating. After completion, the coating completion signal is sent back to the defect detection end.

[0047] This embodiment analyzes and detects defects based on a relatively static stage, specifically identifying the type of defects and improving the efficiency of defect repair. It can also detect defects in different static scenarios, ensuring the operating efficiency of the sizing machine and avoiding waste of sizing material.

[0048] Example 2

[0049] After completing the static yarn detection in the previous embodiment, this embodiment performs dynamic yarn detection. Please refer to [link to previous embodiment]. Figure 3As shown, the dynamic detection architecture consists of three stages: conveyor stress detection, image analysis, and defect recognition and detection. By detecting conveyor stress, the yarn is subjected to different states, and the coating of the yarn is inferred from these different states. This helps to overcome the influence of the sizing machine operation when detecting yarn defects, thus improving the accuracy of yarn defect detection. At the same time, image analysis can more clearly infer the location of yarn defects, enabling accurate repair to improve the efficiency of sizing coating and the high efficiency of yarn use.

[0050] After the sizing machine starts running, the yarn goes through a dynamic stage. During the yarn transport process, the real-time status of the yarn is detected. The maximum width of the yarn in the relaxed state during the dynamic stage is obtained and the maximum width of the coating area set under the current diameter is obtained. The width deviation value is marked as missing coating information based on the width comparison. The maximum width is represented by the maximum distance between the two ends of the yarn after relaxation and the formation of an S-shape.

[0051] If the deviation value is negative, it is marked as an over-coating value. It should be explained that when the over-coating value is negative, the comparison is made according to the set threshold, and the coating area is reset according to the actual slurry consumption. This application does not perform scheme analysis.

[0052] When the yarn is under tension during the dynamic phase and the yarn transmission vibrates, the deviation value of the yarn coating thickness corresponding to the coating area during the vibration period is obtained and marked as coating deviation information.

[0053] The missing coating information and coating deviation information are compared with the width deviation threshold and thickness deviation threshold, respectively:

[0054] If the coating information is missing beyond the width deviation threshold, or the coating deviation information exceeds the thickness deviation threshold, it is inferred that the yarn coating state is abnormal under different conveying forces. This is marked as an abnormal state and uploaded to the dynamic detection architecture. After receiving the abnormal state, the dynamic detection architecture adjusts the dynamic conveying of the yarn to ensure that the yarn force is within the set range and to avoid the occurrence of a loose or tight state. At the same time, it adjusts the real-time coating area in combination with the yarn coating process to adapt to the yarn coating.

[0055] If the missing coating information does not exceed the width deviation threshold and the coating deviation information does not exceed the thickness deviation threshold, it is inferred that the yarn coating status is normal under different conveying forces, and it is marked as normal and uploaded to the dynamic detection architecture. After receiving the normal status, the dynamic detection architecture will send the real-time normal status back to the image acquisition end and analyze the image.

[0056] The images are analyzed in real time to detect and identify defects.

[0057] During the coating process, images of the conveyed yarn are acquired, and the yarn is divided into several positions based on the acquired images, which are marked as sub-positions. The color difference of each adjacent sub-position within the yarn is acquired, and the location of defects is obtained by comparing the color difference with the real-time display based on the range of color difference changes after the coating slurry is exposed to air.

[0058] After coating is completed at the defect location, the coating thickness is compared with that of the adjacent coating location. If the thickness of the defect location is higher than that of the adjacent coating location, it is inferred that the defect at the defect location is an adhesion type. During the defect location delivery stage, the coating thickness at the defect location is continuously monitored. If there is a fluctuation in the coating thickness, the surface adhesion type is an external adhesion or an adhesion with low adhesion. In this case, the defect type at the defect location is marked as an easily changeable adhesion type and uploaded to the dynamic detection architecture along with the defect location.

[0059] If there is no coating thickness fluctuation, and the surface adhesion type is yarn component adhesion or highly adhesive adhesion, then the defect type at the defect location will be marked as difficult to change connection type, and uploaded to the dynamic detection architecture along with the defect location.

[0060] If the thickness at the defect location is lower than that of the adjacent coating location, the defect at the defect location is inferred to be a defect type, and the defect location and defect type are uploaded to the dynamic detection architecture together;

[0061] The dynamic detection architecture transmits the location and type of defects to the defect monitoring and management platform. After receiving the information, the platform repairs defects according to their corresponding types. If the attachment type is easily changed, the yarn is cleaned; if the connection type is difficult to change, the yarn segment is removed; if it is a defect type, the yarn segment is repaired.

[0062] In use, the static detection architecture performs static detection on the yarn, defining the static stage and the relatively static stage during the operation of the sizing machine. Each stage is analyzed to infer the location of defects in the yarn under static conditions and sent to the defect monitoring and management platform. After the yarn static detection is completed, the yarn enters the dynamic stage when the sizing machine starts operating, and dynamic detection is performed. The influence of conveying force detection on yarn transport is analyzed, and after eliminating the influence of transport, the yarn image is analyzed to obtain the defect location. Based on the location analysis, the defect type is defined and sent to the defect monitoring and management platform.

[0063] Thresholds, preset values, preset ranges, etc. are set for result comparison and analysis to determine whether they are good or bad. The value of these thresholds is determined by a combination of large-scale model analysis of sample data and human experience. They can also be adjusted appropriately based on seasonal or common-sense influences.

[0064] Furthermore, the settings for weighting ratios, influence factors, etc., are based on the magnitude of each parameter's influence on the results. The specific values ​​are allocated to ultimately reflect the impact on the results. The settings for input and storage are also determined by a combination of large-scale model analysis of sample data and human experience. Appropriate adjustments can also be made based on seasonal or rational influence conditions.

[0065] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A vision-based yarn defect detection system for sizing machines, characterized in that, It includes a defect monitoring and management platform that monitors several processing steps in real time; and the communication connection includes both static and dynamic detection architectures. The static detection architecture performs static detection on the yarn, defines the static phase and the relatively static phase during the operation of the sizing machine, and analyzes each phase to infer the location of defects in the yarn under static conditions and sends the results to the defect monitoring and management platform. After completing the static yarn inspection, the yarn enters the dynamic stage when the sizing machine is running, and dynamic inspection is carried out. The influence of yarn conveying is analyzed based on the conveying force detection. After eliminating the influence of conveying, the yarn image is analyzed to obtain the defect location. The defect type is defined based on the location analysis, and the data is sent to the defect monitoring and management platform. Before the sizing machine starts running and the yarn is in a fixed storage state, the current stage is marked as the static stage; and after the yarn leaves the static stage, the yarn conveying speed is constant and uniform when the sizing machine starts running, and the current stage is marked as the relative static stage. If the yarn is not moving at a uniform speed after the sizing machine starts, the dynamic detection architecture will intervene. When the yarn is moving at a uniform speed, it will enter a relatively static stage, and the static detection architecture will detect defects. During the static phase, yarn images are acquired, and vertical and horizontal cross-sectional images of the yarn in a fixed storage state are obtained based on the acquired images. Through the acquired cross-sectional images, the yarn diameter and width variation points in the vertical and horizontal cross-sectional images are obtained, that is, the intersection points of two yarns with different diameters. The length of each diameter of the yarn is obtained according to the corresponding diameter change; the deviation between the corresponding diameter length and the set diameter length of the yarn is collected in real time during the static stage, as well as the deviation between the diameter difference of the yarn with the corresponding adjacent diameter change and the set diameter difference, and the collected deviation values ​​are analyzed. If either of the two deviation values ​​collected in real time during the static phase exceeds the corresponding set threshold, the corresponding yarn width segment will be regarded as the defect location. If the two deviation values ​​collected in real time during the static phase do not exceed the corresponding set threshold, the corresponding yarn width segment will be regarded as the defect-free position and uploaded to the static detection architecture and sent to the defect monitoring and management platform. Entering the relatively static stage, the sizing coating is set according to the yarn diameter and width. The coating is detected by image acquisition in the relatively static stage, and the coating uniformity deviation of each position of the yarn segment with the same diameter and width in the relatively static stage is obtained. If the coating uniformity deviation exceeds the set coating error threshold, the defect type analysis is performed on the current coating. After the sizing machine starts running, the yarn enters a dynamic phase. During the yarn transport process, the real-time status of the yarn is detected. The maximum width of the yarn in the relaxed state during the dynamic phase is obtained from the maximum width of the coating area set under the current diameter. The width deviation value is obtained based on the width comparison and marked as coating missing information. When the yarn is in a taut state during the dynamic phase and the yarn transmission vibrates, the coating thickness deviation value of the yarn in the coating area during the vibration period is obtained and marked as coating deviation information. If the coating information is missing beyond the width deviation threshold, or the coating deviation information exceeds the thickness deviation threshold, it is inferred that the yarn coating state is abnormal under different conveying forces. It is marked as an abnormal state and uploaded to the dynamic detection architecture. After receiving the abnormal state, the dynamic detection architecture adjusts the dynamic conveying of the yarn to ensure that the yarn force is within the set range and to avoid the occurrence of a loose or tight state. If the missing coating information does not exceed the width deviation threshold and the coating deviation information does not exceed the thickness deviation threshold, it is inferred that the yarn coating status is normal under different conveying forces, marked as normal status and uploaded to the dynamic detection architecture. After receiving the normal status, the dynamic detection architecture will send the real-time normal status back to the image acquisition end and analyze the image.

2. The vision analysis-based yarn defect detection system for sizing machines according to claim 1, characterized in that, The location of the yarn segment with coating uniformity is obtained; if there is a coating uniformity deviation at a single yarn segment location, the defect type is marked as a single surface defect and uploaded to the static detection architecture, and then transferred to the defect monitoring and management platform. The current location is repaired, and after the repair is completed, the static detection architecture sends the repair control signal back to the defect detection end. If there are deviations in the uniformity of coating at multiple yarn segments and the values ​​are inconsistent, the defect type is marked as a continuous surface defect and uploaded to the static detection architecture. It is then transferred to the defect monitoring and management platform, and the current yarn segment is paused for defect repair and recoating. After completion, the coating completion signal is sent back to the defect detection end.

3. The vision analysis-based yarn defect detection system for sizing machines according to claim 1, characterized in that, The images are analyzed in real time to detect and identify defects. During the coating process, images of the conveyed yarn are acquired, and the yarn is divided into several positions based on the acquired images, which are marked as sub-positions. The color difference of each adjacent sub-position within the yarn is acquired, and the location of defects is obtained by comparing the color difference with the real-time display based on the range of color difference changes after the coating slurry is exposed to air. After coating is applied to the defective location, the coating thickness is compared with that of the adjacent coated locations. If the thickness of the defective location is higher than that of the adjacent coated locations, it is inferred that the defect at the defective location is of the adhesion type. During the delivery phase of the defective location, the coating thickness at the defective location is continuously monitored.

4. The vision analysis-based yarn defect detection system for sizing machines according to claim 3, characterized in that, If there is a fluctuation in coating thickness, and the surface adhesion type of the adhering material is an external adhering material or an adhering material with low adhesion, then the defect type at the defect location will be marked as an easily changeable adhesion type, and uploaded to the dynamic detection architecture along with the defect location. If there is no coating thickness fluctuation, and the surface adhesion type is yarn component adhesion or highly adhesive adhesion, then the defect type at the defect location will be marked as difficult to change connection type, and uploaded to the dynamic detection architecture along with the defect location. If the thickness at the defect location is lower than that of the adjacent coating location, the defect at the defect location is inferred to be a defect type, and the defect location and defect type are uploaded together to the dynamic detection architecture.

Images