Glass fiber cloth flaw detection method and device based on AI vision
By introducing a conveyor belt, guide assembly, and flattening assembly into the fiberglass cloth defect detection device, combined with an intelligent vision inspection head driven by an electric slide rail, the effects of strong light and fabric wrinkles on the detection have been resolved, achieving higher detection accuracy and stability.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHANDONG SHUOYUAN NEW MATERIALS CO LTD
- Filing Date
- 2026-03-26
- Publication Date
- 2026-05-29
AI Technical Summary
Traditional AI vision-based fiberglass cloth defect detection devices are easily affected by strong light and wrinkles caused by the fabric not being fully stretched during the detection process, which affects the accuracy of the detection.
A device comprising an inspection box, a conveyor belt, a guide assembly, a flattening assembly, and a protective arc plate was designed. The intelligent vision inspection head is driven by an electric slide rail. Combined with the flattening assembly and the guide assembly, it achieves flattening and protection of the fabric and avoids interference from strong light.
It improves the accuracy of detection, ensures the fabric is flat, prevents strong light interference, and enhances the stability and reliability of detection.
Smart Images

Figure CN122109128A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of fiberglass cloth defect detection technology, specifically to a fiberglass cloth defect detection method and apparatus based on AI vision. Background Technology
[0002] Fiberglass cloth, or simply fiberglass cloth, is a substrate woven from extremely fine glass fibers. It is one of the core raw materials for manufacturing printed circuit boards. Its final quality and reliability depend to a large extent on the quality of its substrate, namely the fiberglass cloth. Therefore, the detection of defects in fiberglass cloth is a crucial quality checkpoint. Common methods for detecting defects in fiberglass cloth involve using optical means in conjunction with AI vision inspection. AI vision-based fiberglass cloth defect detection is an automated inspection device that integrates machine vision and artificial intelligence technologies, which can improve the quality of fiberglass cloth inspection. The defect detection device for fiberglass cloth based on AI vision requires optical illumination followed by processing by an internal module. This process can be affected by strong ambient light, which can cause unstable signal reception by the internal signal receiving components, thus affecting the detection results. Furthermore, the device may be affected by wrinkles caused by incomplete stretching of the fiberglass cloth during the detection process, which can also affect the accuracy of the detection.
[0003] To address the aforementioned issues, we propose an AI-based vision-based defect detection device for fiberglass cloth. Summary of the Invention
[0004] To address the shortcomings of existing technologies, this invention provides a method and apparatus for detecting defects in fiberglass cloth based on AI vision. This solves the problems of traditional devices lacking leveling and protection against the influence of other light sources during the detection process, ensuring the accuracy of the device during use.
[0005] To achieve the above objectives, the present invention provides the following technical solution: a fiberglass cloth defect detection device based on AI vision, comprising a detection box, a detection component for detection is provided on one side of the lower inner wall of the detection box, and a conveyor belt for conveying is embedded on the lower side of the inner wall of the detection box to the left of the detection component. A plurality of guide components for guiding the conveying are provided on the lower side of the inner wall of the detection box around the conveyor belt, and leveling components for pulling and leveling are provided on both the left and right sides of the surface of the conveyor belt. The detection assembly includes an L-shaped fixing block fixedly installed on one side of the inner wall of the detection box. The upper end of the L-shaped fixing block is fixedly mounted with an intelligent vision inspection head for defect detection via a fixed electric slide rail. A display screen is fixedly installed on the upper side of the inner wall of the detection box in front of the intelligent vision inspection head. Protective arc plates for anti-slip are fixedly installed on both the front and rear sides of the bottom surface of the intelligent vision inspection head.
[0006] Furthermore, the guide assembly includes a connecting groove fixedly installed on the lower side of the inner wall of the detection box, and a connecting block for connection is fixedly installed on the inner wall of the connecting groove by bolts. A first connecting frame is fixedly installed on the side of the connecting block away from the connecting groove, and a sliding groove for sliding is opened on the opposite side of the inner wall of the first connecting frame. A slider is fixedly installed on the upper side of the inner wall of the two sliding grooves by two damping springs. A pressing roller for pressing is rotatably installed on the opposite sides of the two corresponding sliders.
[0007] Furthermore, the leveling assembly includes a fixed vertical block fixedly installed on the lower side of the inner wall of the detection box, and two support hole blocks for support are fixedly installed on the side of the fixed vertical block near the conveyor belt. An N-type push rod for pushing is slidably installed on the inner wall of the two support hole blocks, and a second connecting frame is fixedly installed on the lower end of the N-type push rod. Springs for elastic pushing are sleeved on both opposite sides of the N-type push rod above the support hole blocks. On opposite sides of the inner wall of the second connecting frame, electric rollers for power transmission are fixedly installed. A pulling roller for pulling is fitted onto the surface of the electric roller, and the pulling roller contains several locking components for locking and limiting. On opposite sides of the inner wall of the detection box, two support blocks for support are fixedly installed. Reinforcing plates for reinforcement are fixedly installed on opposite sides of the two corresponding support blocks. Connecting shafts are fixedly installed on opposite sides of the two corresponding support blocks on one side of the reinforcing plates. A pressing and actuating plate for squeezing and actuating is rotatably installed on opposite sides of the two corresponding connecting shafts. A rotary spring is fixedly installed on the surface of the connecting shaft, and one end of the rotary spring is fixedly connected to the surface of the actuating plate. Shock-absorbing rubber blocks for vibration damping are fixedly installed on the top surfaces of the two reinforcing plates, and the top surface of the shock-absorbing rubber blocks contacts the bottom surface of the actuating plate.
[0008] Furthermore, the snap-fit assembly includes a push groove formed on one side of the inner wall of the pull roller. A top block is fixedly installed on the side of the inner wall of the push groove away from the electric roller by a fixed micro damping spring. An arc-shaped top rod for pressing is fixedly installed on the side of the top block close to the electric roller, and an embedding groove is formed on the surface of the electric roller on the side of the arc-shaped top rod.
[0009] Furthermore, the conveyor belt is located at the center of the lower side of the inner wall of the detection box, the two leveling components are located directly below the detection components, and the several guide components are arranged symmetrically from left to right, and the guide components are inclined relative to the conveyor belt.
[0010] Furthermore, both of the protective arc plates are U-shaped plates made of PVC plastic, and the bottom of the protective arc plate has an arc-shaped warp that protrudes downwards.
[0011] Furthermore, the surfaces of several sliders are in close contact with the inner walls of the corresponding sliding grooves, and the bottom surfaces of several extrusion rollers are in contact with the surface of the conveyor belt.
[0012] Furthermore, the vertical cross-section of several of the supporting hole blocks is trapezoidal, the surface of several of the top blocks is in close contact with the inner wall of the pushing groove, and the inner wall of several of the embedding grooves is in contact with the surface of the corresponding arc-head push rod.
[0013] The method for detecting defects in fiberglass cloth based on AI vision includes the following steps; Step 1: Fabric conveying. The fabric conveying involves placing the fabric to be tested under several guide components, and then moving the fabric to be tested under the testing components through the conveyor belt and guide components, thus completing the fabric conveying. Step 2, Fabric Inspection: The fabric inspection is performed by stopping the conveyor belt when the fabric is delivered to the inspection component, and then conducting the fabric inspection through the cooperation of the flattening component and the inspection component.
[0014] In step two, the flattening component and the detection component work together to transport the fabric to be tested to the area below the intelligent vision inspection head via a conveyor belt. Then, driven by the electric slide rail, the intelligent vision inspection head moves downward. During the downward movement of the intelligent vision inspection head, the protective arc plate will squeeze the actuating plate to rotate around the connecting shaft. During the rotation of the actuating plate, it will squeeze the N-type push rod to move downward. At this time, the N-type push rod will drive the second connecting frame to move downward. During the downward movement of the second connecting frame, the pulling roller on the electric roller will first contact the fabric to be tested. Then, through the rotation and locking of the electric roller and the locking component, the pulling roller can rotate. The rotation of the pulling roller ensures the flattening effect of the tested fabric. Under the continuous downward pressure of the actuating plate, the friction between the pulling roller and the conveyor belt will be higher than the pushing force of the micro damping spring in the locking component. At this time, the pulling roller is stationary, and the electric roller is idling, thus ending the pulling effect on the fabric. In this way, the intelligent vision inspection head begins to detect defects in the fabric. After the inspection is completed, the electric slide rail drives the intelligent vision inspection head to move upward. Under the elastic force of the spring and the rotation of the return spring, the pull roller is lifted upward, and the actuating plate will reset and contact the shock-absorbing rubber block on the reinforcing plate. After the fabric inspection is completed, the fabric is transported to one side by the conveyor belt, thus ending the entire fabric inspection process.
[0015] Compared with existing technologies, the present invention provides a method and apparatus for detecting defects in fiberglass cloth based on AI vision, which has the following beneficial effects: 1. This device can flatten the fiberglass cloth during the testing process, avoiding surface wrinkles caused by the fiberglass cloth not being fully stretched, thus ensuring the accuracy of the device's testing. In addition, the device has a protective structure to protect the testing end during the testing process, preventing light sources from other equipment from affecting the signal reception of the testing device, thereby ensuring the normal operation of the device.
[0016] 2. The device utilizes the centered setting of the conveyor belt to ensure the quality of fabric conveying and avoid deviation during the fabric conveying process. The flattening component is located below the detection component, which ensures that the detection component can better detect fabric defects and increase the detection accuracy of the device.
[0017] 3. The device uses an inclined guide component to perform preliminary stretching and flattening of the fabric, which facilitates better flattening of the fabric in the later stage. The material of the protective arc plate and the warping at the bottom can better ensure the squeezing and rotation of the actuating plate and ensure the normal operation of the internal structure of the device.
[0018] 4. This device utilizes the interlocking of the arc-head top rod and the embedded groove to ensure the normal rotation of the electric roller and the pull roller. When the extrusion pressure exceeds the interlocking force of the arc-head top rod and the embedded groove, the electric roller and the pull roller will disengage, thereby ensuring the stability of the fabric detection. Attached Figure Description
[0019] Figure 1 This is a perspective view of the entire invention; Figure 2 This is a three-dimensional view of the entire invention. Figure 3 This is a cross-sectional perspective view of the testing box of the present invention; Figure 4 This is a perspective view of the detection component, guiding component, and leveling component of the present invention; Figure 5 This is a perspective view of the detection component, guiding component, and flattening component of the present invention. Figure 6 This is a vertical sectional perspective view of the guide component of the present invention; Figure 7 This is a perspective view of the support block of the present invention; Figure 8 This is a vertical sectional perspective view of the flattening component of the present invention; Figure 9 for Figure 8 A magnified structural diagram of structure A is shown below; Figure 10 This is a perspective view of the top block of the present invention; Figure 11 This is a perspective view of the L-shaped fixing block of the present invention.
[0020] In the diagram: 1. Inspection box; 2. Inspection component; 201. L-shaped fixing block; 202. Intelligent vision inspection head; 203. Display screen body; 204. Protective arc plate; 2041. Arc-shaped warped part; 3. Conveyor belt; 4. Guide component; 401. Connecting groove; 402. Connecting block; 403. First connecting frame; 404. Sliding groove; 405. Damping spring; 406. Slider; 407. Extrusion roller; 5. Leveling component; 501. Fixing vertical block; 50 2. Supporting hole block; 503. N-type push rod; 504. Second connecting frame; 505. Spring; 506. Electric roller; 507. Pulling roller; 508. Support block; 509. Reinforcing plate; 510. Connecting shaft; 511. Actuating plate; 512. Rotary spring; 513. Shock-absorbing rubber block; 6. Snap-fit assembly; 601. Push groove; 602. Miniature damping spring; 603. Top block; 604. Arc-head top rod; 605. Embedded groove; 7. Electric slide rail. Detailed Implementation
[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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.
[0022] Please see Figures 1 to 11 The fiberglass cloth defect detection device based on AI vision in this embodiment includes a detection box 1. A detection component 2 for detection is provided on one side of the lower inner wall of the detection box 1. A conveyor belt 3 for conveying is embedded on the lower inner wall of the detection box 1 to the left of the detection component 2. Several guide components 4 for guiding the conveying are provided on the lower inner wall of the detection box 1 around the conveyor belt 3. Flattening components 5 for pulling and leveling are provided on both the left and right sides of the surface of the conveyor belt 3. The conveyor belt 3 is located at the center of the lower inner wall of the detection box 1. The center setting of the conveyor belt 3 can ensure better conveying of the cloth. The two flattening components 5 are located directly below the detection component 2. The several guide components 4 are symmetrically arranged on the left and right, and the guide components 4 are inclined relative to the conveyor belt 3. The detection component 2 includes an L-shaped fixing block 201 fixedly installed on one side of the inner wall of the detection box 1. The upper end of the L-shaped fixing block 201 is fixedly installed with an intelligent vision inspection head 202 for defect detection via a fixed electric slide rail 7. The upper side of the inner wall of the detection box 1, in front of the intelligent vision inspection head 202, is fixedly installed with a display screen 203 for display. The front and rear sides of the bottom surface of the intelligent vision inspection head 202 are fixedly installed with anti-slip protective arc plates 204. Both protective arc plates 204 are U-shaped plates made of PVC plastic. The bottom of the protective arc plate 204 has an arc-shaped warp 2041 that protrudes downward. The protective arc plate 204 with the warp structure can better push the actuating plate 511 to rotate.
[0023] The guide assembly 4 includes a connecting groove 401 fixedly installed on the lower side of the inner wall of the detection box 1. A connecting block 402 for connection is fixedly installed on the inner wall of the connecting groove 401 by bolts. A first connecting frame 403 is fixedly installed on the side of the connecting block 402 away from the connecting groove 401. A sliding groove 404 for sliding is opened on the opposite side of the inner wall of the first connecting frame 403. A slider 406 is fixedly installed on the upper side of the inner wall of the two sliding grooves 404 by two damping springs 405. The surfaces of several sliders 406 are tightly fitted with the inner walls of the corresponding sliding grooves 404. The bottom surfaces of several extrusion rollers 407 are in contact with the surface of the conveyor belt 3. Extrusion rollers 407 for extrusion are rotatably installed on the opposite sides of the two corresponding sliders 406.
[0024] The leveling component 5 includes a fixed vertical block 501 fixedly installed on the lower side of the inner wall of the detection box 1. Two support holes 502 for support are fixedly installed on the side of the fixed vertical block 501 near the conveyor belt 3. An N-type push rod 503 for pushing is slidably installed on the inner wall of the two support holes 502. A second connecting frame 504 is fixedly installed on the lower end of the N-type push rod 503. Springs 505 for elastic pushing are sleeved on both opposite sides of the N-type push rod 503 above the support holes 502. On opposite sides of the inner wall of the second connecting frame 504, an electric roller 506 for power transmission is fixedly installed. A pulling roller 507 for pulling is fitted onto the surface of the electric roller 506. The pulling roller 507 has several locking components 6 for locking and limiting inside. On opposite sides of the inner wall of the detection box 1, two support blocks 508 for support are fixedly installed. Reinforcing plates 509 for reinforcement are fixedly installed on opposite sides of the two corresponding support blocks 508. Two connecting shafts 510 are fixedly installed on opposite sides of the reinforcing plate 509. A pressing plate 511 for pressing and actuating is rotatably installed on the opposite sides of the two corresponding connecting shafts 510. A rotary spring 512 is fixedly installed on the surface of the connecting shaft 510, and one end of the rotary spring 512 is fixedly connected to the surface of the pressing plate 511. A shock-absorbing rubber block 513 for shock absorption is fixedly installed on the top surface of both reinforcing plates 509, and the top surface of the shock-absorbing rubber block 513 is in contact with the bottom surface of the pressing plate 511.
[0025] The snap-fit assembly 6 includes a push groove 601 formed on one side of the inner wall of the pull roller 507. A top block 603 is fixedly installed on the side of the inner wall of the push groove 601 away from the electric roller 506 by a fixed miniature damping spring 602. An arc-shaped push rod 604 for pressing is fixedly installed on the side of the top block 603 close to the electric roller 506. An embedding groove 605 is formed on the surface of the electric roller 506 on the side of the arc-shaped push rod 604. The vertical cross-section of the several support hole blocks 502 is trapezoidal. The surface of the several top blocks 603 is tightly fitted with the inner wall of the push groove 601. The inner wall of the several embedding grooves 605 is in contact with the surface of the corresponding arc-shaped push rod 604.
[0026] The method for detecting defects in fiberglass cloth based on AI vision includes the following steps; Step 1: Fabric conveying. Fabric conveying involves placing the fabric to be tested under several guide components 4, and then moving the fabric to be tested under the detection component 2 through the conveyor belt 3 and the guide components 4, thus completing the fabric conveying. Step 2, Fabric Inspection: When the fabric is delivered to the inspection component 2, the conveyor belt 3 stops, and then the flattening component 5 and the inspection component 2 work together to inspect the fabric.
[0027] In step two, the leveling component 5 and the detection component 2 work together to transport the fabric to be tested to below the intelligent vision inspection head 202 via the conveyor belt 3. Then, driven by the electric slide rail 7, the intelligent vision inspection head 202 moves downwards. During this downward movement, the protective arc plate 204 presses the actuating plate 511 to rotate around the connecting shaft 510. As the actuating plate 511 rotates, it presses the N-type push rod 503 downwards. At this time, the N-type push rod 503 drives the second connecting frame 504 downwards. During the movement, the pull roller 507 on the electric roller 506 will first contact the fabric to be tested. Then, through the rotation and locking of the electric roller 506 and the locking component 6, the pull roller 507 can rotate. The rotation of the pull roller 507 can ensure the pulling and leveling effect of the tested fabric. Under the continuous pressure of the actuating plate 511, the friction between the pull roller 507 and the conveyor belt 3 will be higher than the pushing force of the miniature damping spring 602 in the locking component 6. At this time, the pull roller 507 is stationary and the electric roller 506 is idling, thus ending the pulling effect on the fabric. The intelligent vision inspection head 202 then begins to inspect the fabric for defects. After inspection, the electric slide rail 7 moves the intelligent vision inspection head 202 upward. Under the elastic force of the spring 505 and the rotation of the return spring 512, the pull roller 507 is lifted upward, and the actuating plate 511 resets and contacts the shock-absorbing rubber block 513 on the reinforcing plate 509. After the fabric inspection is completed, the fabric is conveyed to one side by the conveyor belt 3, thus ending the entire fabric inspection process. The working principle of the above embodiments is as follows: Before the device is installed, the fabric to be tested needs to be placed under several guide components 4. Then, the fabric to be tested is moved towards the detection component 2 by the conveyor belt 3 and the guide components 4. The guide components 4 will push the extrusion roller 407 on the slider 406 to better fit the fabric to be tested through the elastic force of the damping spring 405. This facilitates the normal conveying of the fabric with the conveyor belt 3. In addition, the guide components 4 are set at an angle. This angled guide components 4 can generate a certain lateral tension, which has a preliminary improvement effect on slight wrinkles and facilitates better flattening of the fabric in the later stage, ensuring the accuracy of the detection process. When the fabric is conveyed to the detection component 2, the conveyor belt 3 stops. Then, the flattening component 5 and the detection component 2 work together to detect the fabric. The fabric to be tested is conveyed to the area below the intelligent vision inspection head 202. Then, driven by the electric slide rail 7, the intelligent vision inspection head 202 moves downwards. During this downward movement, the protective arc plate 204 presses against the actuating plate 511, causing it to rotate around the connecting shaft 510. The protective arc plate 204 not only flattens the fabric in conjunction with the detection, but also, due to its U-shaped structure, protects the detection end of the intelligent vision inspection head 202, preventing other light sources from affecting the signal during the detection process. During the rotation of the actuating plate 511, it presses against N... When the N-type push rod 503 moves downward, it will drive the second connecting frame 504 to move downward. During the downward movement of the second connecting frame 504, the pulling roller 507 on the electric roller 506 will first contact the fabric to be tested. Then, through the rotation and locking of the electric roller 506 and the locking assembly 6, the pulling roller 507 can rotate. The rotation of the pulling roller 507 can ensure the pulling and leveling effect of the tested fabric. Under the continuous downward pressure of the actuating plate 511, the friction between the pulling roller 507 and the conveyor belt 3 will be higher than the pushing force of the micro damping spring 602 in the locking assembly 6. At this time, the pulling roller 507 is stationary and the electric roller 506 is idling, thus ending the pulling effect on the fabric. In this way, the intelligent vision inspection head 202 begins to detect defects in the fabric. After the inspection is completed, the electric slide rail 7 drives the intelligent vision inspection head 202 to move upward. Under the elastic force of the spring 505 and the rotation of the return spring 512, the pull roller 507 is lifted upward, and the actuating plate 511 will reset and contact the shock-absorbing rubber block 513 on the reinforcing plate 509. After the fabric inspection is completed, the fabric is conveyed to one side by the conveyor belt 3, thus ending the entire fabric inspection process. The intelligent vision inspection head 202 in this device has the effect of AI control and can transmit the inspection data to the display screen 203 in real time. This technology is existing and mature. This application highlights the innovative structure and does not elaborate on the existing technology. The device performs inspection after the fabric is flattened to avoid the vibration during the flattening process from affecting the accuracy of the inspection.
[0028] The installation, connection, or setting methods disclosed in this embodiment are all common mechanical connection methods. Any method that can achieve its beneficial effect can be implemented. In addition, the electrical components in this embodiment are all electrically connected to the main controller and the power supply. The main controller can be a conventional known device such as a computer that plays a control role. Those skilled in the art can control the electrical components through simple programming. Moreover, the existing disclosed power connection technology is also common knowledge in the field. Therefore, the specific structural composition and working principle will not be described in detail in this embodiment.
Claims
1. A fiberglass cloth defect detection device based on AI vision, comprising a detection box (1), characterized in that: The detection box (1) has a detection component (2) for detection on one side of the inner wall below, and a conveyor belt (3) for conveying is embedded on the lower side of the inner wall of the detection box (1) to the left of the detection component (2). The lower side of the inner wall of the detection box (1) is provided with several guide components (4) for guiding the conveying on the periphery of the conveyor belt (3), and leveling components (5) for pulling and leveling are provided on both the left and right sides of the surface of the conveyor belt (3). The detection component (2) includes an L-shaped fixing block (201) fixedly installed on one side of the inner wall of the detection box (1), and an intelligent vision inspection head (202) for defect detection is fixedly installed on the upper end of the L-shaped fixing block (201) via a fixed electric slide rail (7). A display screen (203) for display is fixedly installed on the upper side of the inner wall of the detection box (1) in front of the intelligent vision inspection head (202). Protective arc plates (204) for anti-slip are fixedly installed on both the front and rear sides of the bottom surface of the intelligent vision inspection head (202).
2. The fiberglass cloth defect detection device based on AI vision according to claim 1, characterized in that: The guide assembly (4) includes a connecting groove (401) fixedly installed on the lower side of the inner wall of the detection box (1), and a connecting block (402) for connection is fixedly installed on the inner wall of the connecting groove (401) by bolts. A first connecting frame (403) is fixedly installed on the side of the surface of the connecting block (402) away from the connecting groove (401), and a sliding groove (404) for sliding is opened on the opposite side of the inner wall of the first connecting frame (403). A slider (406) is fixedly installed on the upper side of the inner wall of the two sliding grooves (404) by two damping springs (405). A pressing roller (407) for pressing is rotatably installed on the opposite sides of the two corresponding sliders (406).
3. The fiberglass cloth defect detection device based on AI vision according to claim 2, characterized in that: The leveling assembly (5) includes a fixed vertical block (501) fixedly installed on the lower side of the inner wall of the detection box (1), and two support hole blocks (502) for support are fixedly installed on the side of the fixed vertical block (501) near the conveyor belt (3). The inner walls of the two support hole blocks (502) are slidably installed with an N-type push rod (503) for pushing, and the lower end of the N-type push rod (503) is fixedly installed with a second connecting frame (504). The opposite sides of the N-type push rod (503) above the support hole block (502) are each sleeved with a spring (505) for elastic pushing. On opposite sides of the inner wall of the second connecting frame (504), an electric roller (506) for power transmission is fixedly installed, and a pulling roller (507) for pulling is sleeved on the surface of the electric roller (506). The pulling roller (507) is provided with several locking components (6) for locking and limiting. On opposite sides of the inner wall of the detection box (1), two support blocks (508) for support are fixedly installed. On opposite sides of the two corresponding support blocks (508), a reinforcing plate (509) for reinforcement is fixedly installed. Two connecting shafts (510) are fixedly installed on one side of the reinforcing plate (509) on opposite sides. Two corresponding connecting shafts (510) are rotatably mounted on opposite sides for pressing and actuating. A rotary spring (512) is fixedly installed on the surface of the connecting shaft (510), and one end of the rotary spring (512) is fixedly connected to the surface of the actuating plate (511). A shock-absorbing rubber block (513) for shock absorption is fixedly installed on the top surface of both reinforcing plates (509), and the top surface of the shock-absorbing rubber block (513) is in contact with the bottom surface of the actuating plate (511).
4. The fiberglass cloth defect detection device based on AI vision according to claim 3, characterized in that: The snap-fit assembly (6) includes a push groove (601) formed on one side of the inner wall of the pull roller (507). A top block (603) is fixedly installed on the side of the inner wall of the push groove (601) away from the electric roller (506) by a fixed micro damping spring (602). An arc-shaped top rod (604) for pressing is fixedly installed on the side of the top block (603) close to the electric roller (506). An embedding groove (605) is formed on the surface of the electric roller (506) on the side of the arc-shaped top rod (604).
5. The fiberglass cloth defect detection device based on AI vision according to claim 1, characterized in that: The conveyor belt (3) is located at the center of the lower side of the inner wall of the detection box (1). The two flattening components (5) are located directly below the detection component (2). Several guide components (4) are arranged symmetrically on the left and right, and the guide components (4) are inclined relative to the conveyor belt (3).
6. The fiberglass cloth defect detection device based on AI vision according to claim 1, characterized in that: Both of the protective arc plates (204) are U-shaped plates made of PVC plastic, and the bottom of the protective arc plate (204) has an arc-shaped warped part (2041) that protrudes downward.
7. The fiberglass cloth defect detection device based on AI vision according to claim 2, characterized in that: The surfaces of several sliders (406) are in close contact with the inner walls of the corresponding sliding grooves (404), and the bottom surfaces of several extrusion rollers (407) are in contact with the surface of the conveyor belt (3).
8. The fiberglass cloth defect detection device based on AI vision according to claim 4, characterized in that: The vertical cross-section of several of the support holes (502) is trapezoidal, the surface of several of the top blocks (603) is closely fitted with the inner wall of the push groove (601), and the inner wall of several of the embedding grooves (605) is in contact with the surface of the corresponding arc-head push rod (604).
9. A method for detecting defects in fiberglass cloth based on AI vision, characterized in that: The fiberglass cloth defect detection device based on AI vision as described in any one of claims 1-8, wherein the fiberglass cloth defect detection method based on AI vision includes the following steps; Step 1: Fabric conveying. The fabric conveying involves placing the fabric to be tested under several guide components (4), and then conveying it through the conveyor belt (3) and guide components (4) to move the fabric to be tested under the detection component (2), thus completing the fabric conveying. Step 2, fabric inspection. The fabric inspection is carried out by stopping the conveyor belt (3) when the fabric is delivered to the inspection component (2), and then by the flattening component (5) and the inspection component (2) working together to inspect the fabric.
10. The method for detecting defects in fiberglass cloth based on AI vision according to claim 9, characterized in that: In step two, the flattening component (5) and the detection component (2) work together to transport the fabric to be tested to the underside of the intelligent vision inspection head (202) via the conveyor belt (3). Then, driven by the electric slide rail (7), the intelligent vision inspection head (202) moves downward. During the downward movement of the intelligent vision inspection head (202), the protective arc plate (204) will squeeze the actuating plate (511) to rotate around the connecting shaft (510). During the rotation of the actuating plate (511), it will squeeze the N-type push rod (503) to move downward. At this time, the N-type push rod (503) will drive the second connecting frame (504) to move downward. During the downward movement, the pull roller (507) on the electric roller (506) will first contact the fabric to be tested. Then, through the rotation and locking of the electric roller (506) and the locking assembly (6), the pull roller (507) can be rotated. The rotation of the pull roller (507) can ensure the pulling and leveling effect of the fabric to be tested. Under the continuous pressure of the actuating plate (511), the friction between the pull roller (507) and the conveyor belt (3) will be higher than the pushing force of the micro damping spring (602) in the locking assembly (6). At this time, the pull roller (507) is stationary and the electric roller (506) is idling, thus ending the pulling effect on the fabric. In this way, the intelligent vision inspection head (202) begins to inspect the fabric for defects. After the inspection is completed, the electric slide rail (7) drives the intelligent vision inspection head (202) to move upward. Under the elastic force of the spring (505) and the rotation of the rotary spring (512), the pull roller (507) is lifted upward, and the actuating plate (511) will reset and contact the shock-absorbing rubber block (513) on the reinforcing plate (509). After the fabric inspection is completed, the fabric is conveyed to one side by the conveyor belt (3), thus ending the entire process of fabric inspection.