Zipper detection machine based on AI vision and use method

Abstract

The application discloses a zipper detection machine based on AI vision and a use method, relates to the field of zipper quality visual detection equipment, and comprises a base and a transmission part for directional conveying of the zipper. A connecting piece is arranged in the inner cavity of the base through clamping. The zipper detection machine based on AI vision and the use method can compensate the tension fluctuation generated in the transmission process of the zipper in real time through the cooperation of the movable rod and the movable groove in the movable assembly, the butt joint block can move radially along the butt joint plate, the positional deviation of the zipper belt can be corrected, the cooperation of the two can always keep the posture of the zipper detection section stable and the position accurate, the detection area is always kept in the clear imaging range of the AI vision industrial camera, the problem of inaccurate focusing and blurred imaging caused by the floating and swinging of the zipper is avoided, and the identification precision of the micro defect is effectively improved.

Description

Technical Field

[0001] This invention relates to zipper quality visual inspection equipment technology, specifically to an AI vision-based zipper inspection machine and its usage method. Background Technology

[0002] Zippers are widely used connectors in clothing, bags, outdoor equipment, and other fields, and their quality directly affects the user experience and safety of the final product. Defects in zippers, such as missing teeth, misalignment, broken seams, and damaged tape, require rigorous inspection during the production process. Traditional manual visual inspection is inefficient, subjective, and prone to fatigue, and is gradually failing to meet the demands of high-speed, continuous production. AI-based vision-based automated inspection equipment, due to its advantages of high speed, objectivity, and ability to identify minute defects, is gradually becoming an important tool for quality control in the zipper industry.

[0003] Chinese invention patent CN120314316B discloses a fully automatic zipper visual inspection machine. This machine uses a guiding mechanism and a feeding mechanism to ensure that the zipper assembly moves according to the motion logic required by the software in the industrial computer. It continuously feeds zipper assemblies into the machine frame for inspection, marking defective products with numbers for later sorting and collection. The zipper assemblies entering the frame first pass through a lower imaging mechanism. When one zipper in the assembly moves directly in front of the industrial camera, the camera takes a picture and transmits the image to the machine. The industrial computer performs the first analysis and inspection, checking the unclosed part of the zipper and whether the zipper tape and the zipper pull are intact. The zipper is numbered and recorded. Then the zipper continues to be fed forward to the flipping mechanism. The flipping mechanism's push rod reverses the direction of the zipper pull tab. The reversed zipper continues to move to the closing mechanism, passing between the feed plate and the pull plate. The zipper pull passes through the pull groove of the pull plate. During the passage, the pull groove catches the zipper pull and pulls it to the bottom of the zipper tape. Then the zipper head is completely pulled to the bottom of the zipper tape. The zipper head is pulled twice by the two pull plates to double ensure that the zipper head is fully pulled into place.

[0004] When using existing equipment to inspect zippers of different colors, materials, or surface reflectivity, the light source cannot adapt to the rapid changes in material color and surface reflectivity, easily causing local overexposure or dark areas in the image, resulting in the loss of defect features and severely reducing the recognition accuracy of AI vision algorithms, leading to a significant increase in false detection and false negative rates. In addition, during high-speed conveying of zippers, the vibration generated by mechanical transmission causes the zipper belt to float up and down and sway left and right irregularly, causing inaccurate visual focus and blurred imaging, which seriously affects the recognition accuracy of minute defects. Therefore, an AI vision-based zipper inspection machine and its usage method have been developed. Summary of the Invention

[0005] The purpose of this invention is to provide an AI vision-based zipper inspection machine and its usage method to address the aforementioned shortcomings in the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a zipper inspection machine based on AI vision, comprising a base and a transmission component for directional conveying of zippers, wherein a connector is snapped into the inner cavity of the base;

[0007] The movable component, which is assembled on the side of the connector, includes a movable plate, a movable groove and a movable rod for dynamically compensating the tension of the zipper tape, and a docking plate and a docking block for limiting the floating of the zipper tape during operation.

[0008] The movable grooves are symmetrically opened at the ends of the movable plate, and the inner wall of the movable grooves is slidably connected to the outer surface of the movable rod, so that the movable rod can reciprocate along the inner wall of the movable grooves.

[0009] The end of the docking plate engages with the end of the docking block, causing the docking block 272 to move radially along the docking plate, and in conjunction with the movable rod, dynamically adjusts the detection end of the zipper tape;

[0010] An adjustment assembly, which is fitted to the end of the connector, includes an annular plate, an annular groove, a swing ring, and a swing rod for responding to the current reflective properties of the zipper;

[0011] The annular groove is formed on the inner wall of the annular plate and is slidably connected to the outer surface of the swing rod. At the same time, the end of the swing rod is engaged with the outer surface of the swing ring, so that the swing rod on the swing ring can be adjusted along the annular groove, and dynamically adapted to the zipper tape detection position in conjunction with the docking block.

[0012] As a further optimization of the present invention, the movable component further includes a base plate that is snapped into the connector, the end of the base plate being slidably connected to the end of the movable plate, and a movable block being slidably disposed on the side of the movable plate, and a protective plate being fixedly disposed at the end of the movable block, the end of the protective plate being snapped into the base plate.

[0013] As a further optimization of the present invention, a crankshaft is rotatably provided at the end of the base plate, and a drive plate is rotatably provided on the outer surface of the crankshaft;

[0014] The drive plate is rotatably provided with a positioning plate at its end, and positioning blocks are fixedly provided at both ends of the positioning plate.

[0015] As a further optimization of the present invention, a protective block is snapped onto the end of the base plate and on both sides of the crankshaft, and a protective cylinder is fixedly provided at the end of the protective block, with the inner wall of the protective cylinder slidably connected to the outer surface of the lower end of the docking block.

[0016] The inner wall of the protective block is fitted with a limiting block, and the end of the limiting block is attached to the end of the docking plate.

[0017] As a further optimization of the present invention, the adjustment component further includes a locking plate that engages with the connector, a power component is fixedly provided at the middle position of the end of the locking plate, and a telescopic component is engaged at the end of the power component.

[0018] As a further optimization of the present invention, multiple sets of support blocks are evenly arranged at the end of the locking plate, and the end of the support block is fixedly connected to the end of the annular plate.

[0019] As a further optimization of the present invention, a connecting block is snapped onto the end of the telescopic member, and a connecting rod is rotatably provided on the inner wall of the connecting block, the end of the connecting rod being snapped onto the inner wall of the swing ring.

[0020] As a further optimization of the present invention, a limiting block is fixedly provided on the outer surface of the annular plate, a limiting rod is slidably fitted on the inner wall of the limiting block, and an adjusting block is rotatably provided at the end of the limiting rod.

[0021] As a further optimization of the present invention, a support plate is snapped onto the outer surface of the adjusting block, and a pressure plate is snapped onto the end of the adjusting block.

[0022] A zipper detection method based on AI vision, employing any of the detection devices described above, the detection method comprising the following steps:

[0023] S1. First, based on the specifications of the zipper to be inspected, the AI ​​vision inspection controller automatically generates light source parameters, preset values ​​for zipper posture, and target values ​​for tension through the built-in material-optical-mechanical parameter mapping model.

[0024] S2. Drive the crankshaft to run, and drive the positioning plate and positioning block to move back and forth through the drive plate, thereby adjusting the docking plate and docking block to the preset position and completing the initial tension posture setting of the zipper detection section;

[0025] S3. Drive the power component to move the telescopic component, and drive the swing ring to rotate along the annular groove through the connecting block and connecting rod. Adjust the light source to the preset irradiation angle. Then, the limiting rod in the limiting block drives the support plate and pressure plate to lock the swing ring, completing the equipment debugging before testing.

[0026] S4. When the zipper belt experiences tension fluctuations or positional deviations due to transmission, the movable rod slides along the movable groove in real time to compensate for the tension, and the docking block moves radially along the docking plate to correct the positional deviation. This ensures that the zipper detection section remains stable and accurately positioned, so that the detection area is always within the clear imaging range of the AI ​​vision industrial camera.

[0027] After acquiring the zipper image, the S5 AI vision controller analyzes the image using a trained AI detection algorithm to identify defects such as misaligned teeth, missing teeth, and damaged fabric tape.

[0028] Compared with existing technologies, the zipper inspection machine and its usage method based on AI vision provided by this invention have the following beneficial effects:

[0029] By cooperating with the movable rod and movable groove inside the active component, the tension fluctuations generated during the zipper transmission can be compensated in real time. At the same time, the docking block can move radially along the docking plate to correct the positional deviation of the zipper belt. The combination of the two can always keep the zipper detection section in a stable posture and accurate position, ensuring that the detection area is always within the clear imaging range of the AI ​​vision industrial camera. This avoids problems such as inaccurate focusing and blurred imaging caused by zipper floating and swaying, and effectively improves the recognition accuracy of minute defects.

[0030] By adjusting the coordination of the power and telescopic components inside the assembly, the swing ring can be driven to rotate along the annular groove, thereby quickly adjusting the illumination angle of the light source. The optimal illumination angle is adapted to zippers with different colors, materials, and surface reflective properties, effectively avoiding the problem of local overexposure or loss of defect features in dark areas of the image. This significantly improves the defect recognition accuracy of the AI ​​vision algorithm and reduces the false detection rate and false negative rate.

[0031] Through the coordinated operation of the moving and adjusting components, dynamic posture adjustment and optical imaging parameter adaptation of zippers of different specifications can be achieved simultaneously. This not only solves the imaging blurring problem caused by mechanical vibration in traditional equipment, but also makes up for the inability of the light source in traditional equipment to adapt to zippers with different reflective properties. It can meet the testing requirements of high-speed continuous production and effectively improve the quality control efficiency of the zipper production process. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0033] Figure 1 This is a schematic diagram of the overall structure provided for an embodiment of the present invention;

[0034] Figure 2 This is a schematic diagram of the connector structure provided in an embodiment of the present invention;

[0035] Figure 3 This is a schematic diagram of the active component structure provided in an embodiment of the present invention;

[0036] Figure 4 An exploded view of the active component structure provided in an embodiment of the present invention;

[0037] Figure 5 An exploded view of the docking block connection structure provided in an embodiment of the present invention;

[0038] Figure 6 This is a schematic diagram of the adjustment component structure provided in an embodiment of the present invention;

[0039] Figure 7 An exploded view of the adjustment component structure provided in an embodiment of the present invention;

[0040] Figure 8 This is an exploded cross-sectional view of the adjustment component provided in an embodiment of the present invention.

[0041] Explanation of reference numerals in the attached figures:

[0042] 1. Base; 2. Movable component; 3. Adjusting component; 11. Connector; 12. Transmission component; 21. Base plate; 22. Moving plate; 221. Moving block; 222. Protective plate; 23. Movable groove; 231. Movable rod; 24. Crankshaft; 241. Drive plate; 25. Positioning plate; 251. Positioning block; 26. Protective block; 261. Protective cylinder; 27. Limiting block; 271. Connecting plate; 272. Connecting block; 31. Locking plate; 32. Power component; 33. Telescopic component; 34. Support block; 35. Annular plate; 351. Annular groove; 36. Limiting block; 361. Limiting rod; 362. Adjusting block; 363. Support plate; 364. Pressure plate; 37. Connecting block; 371. Connecting rod; 38. Swing ring; 381. Swing rod; 39. Snap-fit ​​plate. Detailed Implementation

[0043] To enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0044] Example: Please refer to Figure 1 - Figure 8 A zipper inspection machine based on AI vision includes a base 1 and a transmission component 12 for directional conveying of zippers. A connector 11 is snapped into the inner cavity of the base 1.

[0045] In this solution, the transmission component 12 is a set of conveying rollers for directional conveying of the zipper tape. The rotation of the transmission component 12 can drive the zipper tape to be tested to be conveyed along the specified direction, ensuring the stable progress of the zipper testing process.

[0046] The connector 11 is assembled inside the base 1 using a snap-fit ​​structure, which facilitates subsequent installation, disassembly and maintenance. At the same time, it can quickly replace the appropriate components according to different testing requirements, effectively improving the flexibility of equipment adjustment.

[0047] The AI ​​vision inspection controller is connected to the industrial camera, the adaptive light source adjustment component 3, and the posture movement component 2. The controller has a built-in material-optical-mechanical parameter mapping model and is configured with the following functions: When changing to zippers of different specifications, the light source parameters, posture preset values, and tension target values ​​can be adjusted synchronously according to the identified zipper characteristics; during the inspection process, the posture compensation amount and light source tracking parameters can be adjusted according to the tension fluctuation and vibration signal feedforward, and the AI ​​vision algorithm can be used to analyze the zipper image to complete the defect detection.

[0048] The controller's built-in material-optical-mechanical parameter mapping model is a mature model that has already been trained in the existing AI visual inspection field. It can directly output the corresponding light source angle, illumination intensity, and zipper posture preset values ​​based on the input zipper material and color parameters, eliminating the need for repeated manual adjustments and significantly reducing the time cost of model changeover and debugging.

[0049] Furthermore, the movable component 2, which is assembled on the side of the connector 11, includes a movable plate 22, a movable groove 23, and a movable rod 231 for dynamically compensating the tension of the zipper tape, and a docking plate 271 and a docking block 272 for limiting the floating of the zipper tape during operation; the movable groove 23 is symmetrically opened at the end of the movable plate 22, and the inner wall of the movable groove 23 is slidably connected to the outer surface of the movable rod 231, so that the movable rod 231 reciprocates along the inner wall of the movable groove 23; the end of the docking plate 271 is engaged with the end of the docking block 272, so that the docking block 272 moves radially along the docking plate 271, and dynamically adjusts the detection end of the zipper tape in conjunction with the movable rod 231.

[0050] In this embodiment, the movable plate 22 is mounted on the side of the connector 11 and drives the movable rod 231 to slide adaptively through the movable groove 23. The end of the movable plate 22 is provided with a component with telescopic function, such as an electric telescopic rod, to drive the movable plate 22 to move along the side of the connector 11, and the movable groove 23 is inclined.

[0051] When the zipper belt experiences tension fluctuations due to transmission vibration, the movable rod 231 can move along the movable groove 23 in real time to follow the tension changes. In conjunction with the radially moving docking block 272 on the docking plate 271, it dynamically compensates for the tension and position of the zipper belt detection section. This avoids the zipper belt from being too loose and sagging, which would affect imaging, and also prevents it from being too tight and breaking the belt. This ensures the stability of the zipper belt detection section's posture from a tension perspective.

[0052] During the transmission and conveying of the zipper, the zipper belt passes between the movable rod 231 and the docking block 272. When the zipper belt floats up and down or shifts left and right due to transmission vibration, the docking block 272 will move radially synchronously along the docking plate 271 to limit the zipper belt. In conjunction with the tension adjustment of the movable rod 231, the positional deviation of the zipper belt detection section is corrected in real time, avoiding the problem of AI visual imaging blur and inaccurate focus caused by the irregular shaking of the zipper belt, and ensuring that the detection area is always within the clear imaging range of the AI ​​visual camera.

[0053] At the same time, the pressure of the zipper docking end is adjusted by the symmetrically set docking blocks 272, so that the meshing state of the docking end can be fully presented when the zipper is under force. This makes it easy for AI vision to clearly capture the feature information of defects such as tooth misalignment and missing teeth, and will not cause local defects to be obscured due to the zipper tape posture being skewed.

[0054] Furthermore, the active component 2 also includes a base plate 21 that is snapped into the connector 11. The end of the base plate 21 is slidably connected to the end of the movable plate 22, and a movable block 221 is slidably provided on the side of the movable plate 22. A protective plate 222 is fixedly provided at the end of the movable block 221, and the end of the protective plate 222 is snapped into the base plate 21.

[0055] Specifically, the base plate 21 is fixedly connected to the connector 11 through a snap-fit ​​structure. The movable rod 231 is limited by the movable block 221 and the protective plate 222. Since the protective plate 222 is snapped onto the base plate 21, it can limit the travel of the movable plate 22 when it slides at the end of the base plate 21, thus preventing the movable plate 22 from moving beyond its range and causing the zipper belt tension to run out of control.

[0056] Furthermore, a crankshaft 24 is rotatably provided at the end of the base plate 21, and a drive plate 241 is rotatably provided on the outer surface of the crankshaft 24; a positioning plate 25 is rotatably provided at the end of the drive plate 241, and positioning blocks 251 are fixedly provided at both ends of the positioning plate 25.

[0057] Specifically, the end of the base plate 21 is equipped with a motor or other power output device, which is connected to an external control device. The output end of the motor is engaged with the end of the crankshaft 24. When the motor runs, it can drive the crankshaft 24 to rotate, and then the rotation of the crankshaft 24 drives the drive plate 241 to swing, thereby pushing the positioning plate 25 to slide back and forth along the end of the base plate 21 through the positioning block 251. With the tension compensation action of the movable rod 231, the initial position of the moving plate 22 can be pre-adjusted according to the specifications of the zipper tape before testing, adapting to the initial tension requirements of zipper tapes of different widths and thicknesses, and further improving the equipment's adaptability to zippers of different specifications.

[0058] Furthermore, protective blocks 26 are snapped onto the ends of the base plate 21 and on both sides of the crankshaft 24. Protective cylinders 261 are fixedly installed at the ends of the protective blocks 26. The inner wall of the protective cylinders 261 is slidably connected to the outer surface of the lower end of the docking block 272. Limiting blocks 27 are snapped onto the inner wall of the protective blocks 26. The end of the limiting blocks 27 is attached to the end of the docking plate 271.

[0059] Specifically, the protective block 26 restricts the sliding stroke of the docking block 272 through the upper protective cylinder 261 and simultaneously performs limit guidance to ensure that the docking block 272 always moves radially in the predetermined direction without deviation or jamming. The outer surface of the docking plate 271 meshes with the end of the positioning block 251.

[0060] When the positioning plate 25 is driven to reciprocate by the crankshaft 24, the positioning block 251 will drive the docking plate 271 to make a small position adjustment along the zipper conveying direction, thereby driving the docking block 272 to adjust its position synchronously. This achieves batch synchronous adjustment of the zipper belt limit position without the need for manual adjustment one by one, which greatly improves the adjustment efficiency when switching specifications. At the same time, the limit block 27 can limit the installation position of the docking plate 271, ensuring the positional accuracy of the docking plate 271 after assembly and avoiding the impact of positioning reference offset on the limit accuracy.

[0061] The limiting block 27 is assembled with the protective block 26 through a snap-fit ​​structure. After the docking plate 271 completes the position calibration, the limiting block 27 can be snapped in and pressed against the docking plate 271 to ensure that the docking plate 271 will not be displaced during operation, maintain the positioning accuracy after long-term operation, and reduce the cumulative error of the equipment.

[0062] Furthermore, the adjustment component 3, which is assembled at the end of the connector 11, includes an annular plate 35, an annular groove 351, a swing ring 38, and a swing rod 381 to address the current reflective characteristics of the zipper. The annular groove 351 is formed on the inner wall of the annular plate 35 and is fitted and slidably connected to the outer surface of the swing rod 381. At the same time, the end of the swing rod 381 is engaged with the outer surface of the swing ring 38, so that the swing rod 381 on the swing ring 38 can be angled along the annular groove 351, and dynamically adapts to the detection position of the zipper tape in conjunction with the docking block 272.

[0063] In this embodiment, the annular plate 35 is fixedly connected to the locking plate 31 via the support block 34 and is positioned below the AI ​​vision industrial camera. The swing ring 38 can be circumferentially swung along the annular groove 351 on the inner wall of the annular plate 35 via the swing rod 381. The light source assembly is embedded at the end of the swing ring 38 and is limited and protected by the snap-fit ​​plate 39. The illumination angle can be adjusted synchronously with the swing ring 38.

[0064] Once the AI ​​vision controller recognizes the color, material, and reflective properties of the current zipper, it can drive the telescopic component 33 via the power component 32 to move the connecting block 37. Then, the connecting rod 371 drives the swing ring 38 to adjust the circumferential angle, adjusting the light source to the optimal illumination angle under the corresponding reflective properties. This avoids overexposure or local dark areas caused by reflection on the zipper surface, ensuring that defect features can be completely and clearly captured by the AI ​​vision.

[0065] Furthermore, the adjustment assembly 3 also includes a locking plate 31 that engages with the connector 11. A power component 32 is fixedly provided at the middle position of the end of the locking plate 31, and a telescopic component 33 is engaged at the end of the power component 32.

[0066] Specifically, the locking plate 31 is fixedly assembled with the connector 11 through a snap-fit ​​structure, which can provide a stable installation base for the power component 32 and the telescopic component 33, avoid the overall shaking of the components during power output, and ensure the accuracy of the angle adjustment of the swing ring 38.

[0067] The power component 32 and the telescopic component 33 are electrically connected to the AI ​​vision inspection controller and can receive angle adjustment commands issued by the controller to drive the telescopic component 33 to move linearly along the output direction of the power component 32.

[0068] In this device, the power component 32 is a rotary motor, whose output end is connected to the telescopic component 33 through a gear set. This allows for precise control of the movement stroke of the telescopic component 33. The telescopic component 33 can be a component with telescopic function, such as an electric telescopic rod, thereby precisely controlling the swing angle of the swing ring 38 to meet the accuracy requirements of angle adjustment for zippers with different reflective properties.

[0069] Furthermore, multiple sets of support blocks 34 are evenly arranged at the end of the locking plate 31, and the ends of the support blocks 34 are fixedly connected to the ends of the annular plate 35.

[0070] Specifically, the support blocks 34 are evenly distributed at the ends of the locking plate 31, and the annular plate 35 is stably fixed to one side of the locking plate 31. This not only ensures the structural strength of the annular plate 35 itself, but also leaves sufficient operating space to facilitate subsequent maintenance and replacement of the swing ring 38 and the light source assembly, thus reducing the maintenance difficulty of the equipment.

[0071] Furthermore, a connecting block 37 is snapped onto the end of the telescopic member 33, and a connecting rod 371 is rotatably mounted on the inner wall of the connecting block 37. The end of the connecting rod 371 is snapped onto the inner wall of the swing ring 38.

[0072] Specifically, the connecting block 37 moves linearly and synchronously with the telescopic component 33, thereby pushing the swing ring 38 to deflect circumferentially along the annular groove 351 through the connecting rod 371. Since the connecting rod 371 is rotatably connected to the inner wall of the connecting block 37, it can adaptively offset the angle deviation generated during the deflection of the swing ring 38, avoid transmission jamming, and ensure a smooth and stable angle adjustment process.

[0073] Furthermore, a limiting block 36 is fixedly provided on the outer surface of the annular plate 35, a limiting rod 361 is slidably fitted into the inner wall of the limiting block 36, and an adjusting block 362 is rotatably provided at the end of the limiting rod 361.

[0074] Specifically, the limiting block 36 is fixedly disposed on the outer side of the annular plate 35, and the limiting rod 361 can slide axially along the inner wall of the limiting block 36, with the sliding stroke corresponding to the rotation range of the swing ring 38.

[0075] Furthermore, a support plate 363 is snapped onto the outer surface of the adjusting block 362, and a pressure plate 364 is snapped onto the end of the adjusting block 362.

[0076] Specifically, the support plate 363 and the limiting rod 361 move synchronously. The support plate 363 and the pressure plate 364 are respectively engaged at the upper and lower positions on both sides of the swing ring 38. After the light source angle is adjusted, the support plate 363 and the pressure plate 364 lock the position of the swing ring 38 to prevent the swing ring 38 from being displaced by vibration during the operation of the equipment, and ensure that the light source illumination angle is always maintained at the set optimal state, thus stably ensuring the AI ​​visual imaging quality.

[0077] The control device can choose a microcontroller as the control terminal. In this embodiment, the microcontroller is a typical embedded microcontroller unit, consisting of an arithmetic logic unit (ALU), a controller, memory, input / output devices, etc., essentially a miniature computer. Compared to general-purpose microprocessors used in personal computers, it emphasizes self-sufficiency (no external hardware required) and cost savings. Its biggest advantage is its small size, allowing it to be placed inside the instrument, but it has limited storage capacity, simple input / output interfaces, and low power consumption.

[0078] Working principle: When using this device, the AI ​​vision inspection controller first generates light source parameters, preset values ​​for zipper posture, and target values ​​for tension based on the specifications of the zipper to be inspected through the built-in material-optical-mechanical parameter mapping model.

[0079] Subsequently, the crankshaft 24 is driven to run, and the positioning plate 25 and positioning block 251 are driven to reciprocate through the drive plate 241, which drives the docking plate 271 and docking block 272 to adjust to the preset position, thus completing the initial tension posture setting of the zipper detection section.

[0080] Then, the power component 32 drives the telescopic component 33 to move, and the swing ring 38 rotates along the annular groove 351 through the connecting block 37 and the connecting rod 371, adjusting the light source to the preset illumination angle. Then, the limiting rod 361 in the limiting block 36 drives the support plate 363 and the pressure plate 364 to lock the swing ring 38, completing the equipment debugging before testing.

[0081] After the test begins, the transmission component 12 drives the zipper belt to be conveyed in the specified direction. The zipper belt passes between the movable rod 231 and the docking block 272 and enters the test area.

[0082] When the zipper belt experiences tension fluctuations or positional deviations due to transmission, the movable rod 231 slides along the movable groove 23 in real time to compensate for the tension, and the docking block 272 moves radially along the docking plate 271 to correct the positional deviation, thus maintaining the stability and accuracy of the zipper detection section's posture and position, ensuring that the detection area is always within the clear imaging range of the AI ​​vision industrial camera.

[0083] After acquiring the zipper image, the AI ​​vision controller analyzes the image using a trained AI detection algorithm to identify defects such as tooth misalignment, missing teeth, and tape damage. Simultaneously, based on the tension fluctuations and vibration signals collected during the detection process, it feeds forward to adjust the zipper posture compensation amount and light source tracking parameters to dynamically maintain the optimal detection state, ultimately completing the entire process of automatic defect detection for the zipper.

[0084] Example 2: A zipper detection method based on AI vision, using any of the detection devices described above, the detection method includes the following steps:

[0085] S1. First, based on the specifications of the zipper to be inspected, the AI ​​vision inspection controller automatically generates light source parameters, preset values ​​for zipper posture, and target values ​​for tension through the built-in material-optical-mechanical parameter mapping model.

[0086] S2. Drive the crankshaft 24 to run, and drive the positioning plate 25 and positioning block 251 to reciprocate through the drive plate 241, thereby driving the docking plate 271 and docking block 272 to adjust to the preset position, and complete the setting of the initial tension posture of the zipper detection section.

[0087] S3. Drive the power component 32 to move the telescopic component 33. Through the connecting block 37 and the connecting rod 371, drive the swing ring 38 to rotate along the annular groove 351, adjust the light source to the preset irradiation angle, and then the limiting rod 361 in the limiting block 36 drives the support plate 363 and the pressure plate 364 to lock the swing ring 38, completing the equipment debugging before the test.

[0088] S4. When the zipper belt experiences tension fluctuations or positional deviations due to transmission, the movable rod 231 slides along the movable groove 23 in real time to compensate for the tension, and the docking block 272 moves radially along the docking plate 271 to correct the positional deviation, so as to always maintain the stable posture and accurate position of the zipper detection section, and ensure that the detection area is always within the clear imaging range of the AI ​​vision industrial camera.

[0089] After acquiring the zipper image, the S5 AI vision controller analyzes the image using a trained AI detection algorithm to identify defects such as misaligned teeth, missing teeth, and damaged fabric tape.

[0090] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A zipper inspection machine based on AI vision, comprising a base (1) and a transmission component (12) for directional conveying of zippers, characterized in that, The inner cavity of the base (1) is fitted with a connector (11); The active component (2), which is mounted on the side of the connector (11), includes a movable plate (22), an active groove (23) and an active rod (231) for dynamically compensating the tension of the zipper belt, and a docking plate (271) and a docking block (272) for limiting the floating of the zipper belt during operation. The movable groove (23) is symmetrically opened at the end of the movable plate (22), and the inner wall of the movable groove (23) is slidably connected to the outer surface of the movable rod (231), so that the movable rod (231) moves back and forth along the inner wall of the movable groove (23). The end of the docking plate (271) is engaged with the end of the docking block (272), so that the docking block 272 moves radially along the docking plate (271), and the detection end of the zipper tape is dynamically adjusted in conjunction with the movable rod (231); The adjustment component (3), which is assembled at the end of the connector (11), includes an annular plate (35), an annular groove (351), a swing ring (38), and a swing rod (381) for dealing with the current reflective properties of the zipper. The annular groove (351) is formed on the inner wall of the annular plate (35) and is slidably connected to the outer surface of the swing rod (381). At the same time, the end of the swing rod (381) is engaged with the outer surface of the swing ring (38), so that the swing rod (381) on the swing ring (38) can be adjusted along the annular groove (351) and dynamically adapted to the zipper tape detection position in conjunction with the docking block (272).

2. The zipper inspection machine based on AI vision according to claim 1, characterized in that, The active component (2) also includes a base plate (21) that is snapped into the connector (11). The end of the base plate (21) is slidably connected to the end of the movable plate (22), and a movable block (221) is slidably provided on the side of the movable plate (22). A protective plate (222) is fixedly provided at the end of the movable block (221), and the end of the protective plate (222) is snapped into the base plate (21).

3. The zipper inspection machine based on AI vision according to claim 2, characterized in that, A crankshaft (24) is rotatably provided at the end of the base plate (21), and a drive plate (241) is rotatably provided on the outer surface of the crankshaft (24). The end of the drive plate (241) is rotatably provided with a positioning plate (25), and both ends of the positioning plate (25) are fixedly provided with positioning blocks (251).

4. The zipper inspection machine based on AI vision according to claim 3, characterized in that, Protective blocks (26) are snapped onto the end of the base plate (21) and on both sides of the crankshaft (24). A protective cylinder (261) is fixedly installed at the end of the protective block (26). The inner wall of the protective cylinder (261) is slidably connected to the outer surface of the lower end of the docking block (272). The inner wall of the protective block (26) is fitted with a limiting block (27), and the end of the limiting block (27) is attached to the end of the docking plate (271).

5. A zipper inspection machine based on AI vision according to claim 1, characterized in that, The adjustment component (3) also includes a locking plate (31) that engages with the connector (11). A power component (32) is fixedly provided at the middle position of the end of the locking plate (31), and a telescopic component (33) is engaged at the end of the power component (32).

6. A zipper inspection machine based on AI vision according to claim 5, characterized in that, The locking plate (31) has multiple sets of support blocks (34) evenly arranged at its end, and the end of the support block (34) is fixedly connected to the end of the annular plate (35).

7. A zipper inspection machine based on AI vision according to claim 6, characterized in that, The end of the telescopic member (33) is fitted with a connecting block (37), and the inner wall of the connecting block (37) is rotatably fitted with a connecting rod (371). The end of the connecting rod (371) is fitted with the inner wall of the swing ring (38).

8. A zipper inspection machine based on AI vision according to claim 7, characterized in that, A limiting block (36) is fixedly provided on the outer surface of the annular plate (35), and a limiting rod (361) is slidably fitted on the inner wall of the limiting block (36). An adjusting block (362) is rotatably provided at the end of the limiting rod (361).

9. A zipper inspection machine based on AI vision according to claim 8, characterized in that, The outer surface of the adjusting block (362) is fitted with a support plate (363), and the end of the adjusting block (362) is fitted with a pressure plate (364).

10. A zipper detection method based on AI vision, characterized in that, Using the detection device as described in any one of claims 1-9, the detection method includes the following steps: S1. First, based on the specifications of the zipper to be inspected, the AI ​​vision inspection controller automatically generates light source parameters, preset values ​​for zipper posture, and target values ​​for tension through the built-in material-optical-mechanical parameter mapping model. S2. Drive the crankshaft (24) to run, and drive the positioning plate (25) and positioning block (251) to move back and forth through the drive plate (241), thereby driving the docking plate (271) and docking block (272) to adjust to the preset position, and complete the setting of the initial tension posture of the zipper detection section; S3. Drive the power component (32) to move the telescopic component (33), and drive the swing ring (38) to rotate along the annular groove (351) through the connecting block (37) and the connecting rod (371), adjust the light source to the preset irradiation angle, and then the limiting rod (361) in the limiting block (36) drives the support plate (363) and the pressure plate (364) to lock the swing ring (38), thus completing the equipment debugging before the test; S4. When the zipper belt experiences tension fluctuations or positional deviations due to transmission, the movable rod (231) slides along the movable groove (23) in real time to compensate for the tension, and the docking block (272) moves radially along the docking plate (271) to correct the positional deviation, so as to always maintain the stable posture and accurate position of the zipper detection section, and make the detection area always within the clear imaging range of the AI ​​vision industrial camera. After acquiring the zipper image, the S5 AI vision controller analyzes the image using a trained AI detection algorithm to identify defects such as misaligned teeth, missing teeth, and damaged fabric tape.

Images