Threaded hole defect intelligent detection system based on AI visual identification

Abstract

The invention relates to a threaded hole defect intelligent detection system based on AI visual identification in the field of defect detection equipment. The threaded hole defect intelligent detection system comprises a workbench, an upper computer, a marking pen, an image acquisition unit and an AI analysis and diagnosis unit, a conveying mechanism is installed on the upper table top of the workbench, a shielding cover box is arranged above the conveying mechanism in a covering mode, and feeding ports are formed in the box walls of the two ends of the shielding cover box; a clamping and positioning mechanism is arranged on the outer side of the conveying mechanism, and a material blocking mechanism is arranged at the output end of the conveying mechanism. The image acquisition unit comprises an endoscopic probe, and a three-dimensional movement mechanism which can drive the endoscopic probe and the marking pen to move in a three-dimensional mode so as to position the threaded hole of the corresponding workpiece to be detected is erected above the transmission mechanism. The upper computer is in signal connection with the AI analysis and diagnosis unit, the image acquisition unit, the transmission mechanism, the clamping and positioning mechanism, the material blocking mechanism and the three-dimensional movement mechanism. The intelligent detection system can perform automatic and intelligent detection, adapts to complex environments, and improves the detection precision and the detection efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of defect detection equipment, in particular to a threaded hole defect intelligent detection system based on AI visual recognition. BACKGROUND

[0002] It is known that the threaded hole machining of a workpiece is prone to defects such as cracking, burr, aluminum chip blockage, etc. If the workpiece is shipped without quality inspection, it will not only cause a large number of returns, increase transportation costs, but also reduce commodity credibility and affect product sales. Traditional manual visual inspection not only consumes manpower, but also is inefficient and prone to missed detection or misjudgment, and the human eye has limited recognition ability for micron-level defects such as scratches and PCB fine particles. With the rise of computer vision detection technology under AI algorithm, image processing and recognition detection using algorithm are more accurate than manual visual inspection, but existing detection equipment still has two shortcomings: first, traditional industrial cameras are easily affected by changes in light and product placement angle offset, resulting in unclear imaging results and affecting computer vision detection results; second, it is difficult to meet the intelligent detection needs of high-speed production lines, especially when there are multiple threaded holes on a workpiece, the detection efficiency is even lower, and it is not easy to distinguish the detection results after detection. SUMMARY

[0003] In order to overcome the shortcomings in the background art and solve the existing technical problems, the present application discloses a threaded hole defect intelligent detection system based on AI visual recognition, which can automatically and intelligently detect, adapt to complex environments, and improve detection accuracy and efficiency.

[0004] To achieve the above-mentioned purpose, the technical scheme adopted by the present application is as follows: A threaded hole defect intelligent detection system based on AI visual recognition, comprising a workbench, an upper computer, a marker pen, an image acquisition unit and an AI analysis and diagnosis unit for identifying threaded hole defects; a transmission mechanism for conveying a workpiece to be inspected is installed on the upper surface of the workbench, a shielding cover box is provided above the transmission mechanism, and a feeding port is provided on the box wall at both ends of the transmission mechanism corresponding to the input end and the output end; a clamping and positioning mechanism corresponding to the workpiece to be inspected is provided on the outer side of the middle part of the transmission mechanism, and a material blocking mechanism capable of temporarily blocking the workpiece to be inspected is provided at the output end of the transmission mechanism; the image acquisition unit comprises an endoscopic probe, a three-dimensional movement mechanism capable of driving the endoscopic probe and the marker pen to move three-dimensionally is provided above the transmission mechanism, and the upper computer is respectively connected to the AI analysis and diagnosis unit, the image acquisition unit, the transmission mechanism, the clamping and positioning mechanism, the material blocking mechanism and the three-dimensional movement mechanism.

[0005] Furthermore, a lifting electric cylinder is fixedly installed on the outer wall of the folding cover box above the feeding port. The telescopic rod end of the lifting electric cylinder is connected to a baffle plate. Slide rails adapted to slide and engage with the corresponding sides of the baffle plate are fixed on both sides of the feeding port.

[0006] Furthermore, the three-dimensional motion mechanism includes two longitudinal guide rails located on both sides of the transmission mechanism and along the transmission direction of the transmission mechanism. The bottom surface of the longitudinal guide rails is fixed to the worktable by a column. A transverse guide rail that can slide and be positioned along the longitudinal guide rails is slidably installed between the two longitudinal guide rails. A vertical guide rail that can slide and be positioned along the transverse guide rails is slidably installed on the transverse guide rail. A slider plate that can slide and be positioned along the vertical guide rail is slidably installed on the vertical guide rail. The endoscopic probe is installed at the lower end of the slider plate.

[0007] Furthermore, the lower end of the slider plate is rotatably connected to a bracket via a rotating shaft parallel to the transverse guide rail. The bracket is rotated and positioned by a flip drive unit, and the endoscope probe and the marker pen are respectively fixed at both ends of the bracket.

[0008] Furthermore, the AI ​​analysis and diagnosis unit identifies threaded hole defects through AI algorithms such as pixel attribution, boundary detection, region segmentation, and convolutional neural networks.

[0009] Furthermore, the clamping and positioning mechanism includes jacking components symmetrically arranged on both sides of the transmission mechanism. The jacking components include a first upright plate and a first telescopic cylinder. The first upright plate is fixed to the workbench surface, and the cylinder body of the first telescopic cylinder is fixed to the outer surface of the first upright plate. The rod head of the first telescopic cylinder moves through the first upright plate and is fixedly connected to a top block.

[0010] Furthermore, two material blocking mechanisms are symmetrically arranged on both sides of the output end of the transmission mechanism. The material blocking mechanism includes a second vertical plate and a second telescopic cylinder. The second vertical plate is fixed to the workbench surface, and the cylinder body of the second telescopic cylinder is fixed to the outer plate surface of the second vertical plate. The rod head of the second telescopic cylinder moves through the second vertical plate and is fixedly connected to a top plate. The inner plate surface of the top plate is equipped with a vertically oriented positioning guide wheel via a wheel frame.

[0011] Furthermore, the workbench includes a tabletop and four legs fixed to the four corners of the lower surface of the tabletop. An adjustment plate is fixed to the bottom surface of each leg, and support casters and adjustable feet are installed on the lower surface of the adjustment plate.

[0012] Furthermore, a control box is provided below the platform, and the control box has a door and a ventilation opening on its wall where a fan is installed.

[0013] Furthermore, the transmission mechanism includes a driven roller and a driving roller rotatably mounted on the frame, and a conveyor belt connected by the driven roller and the driving roller, wherein the driving roller is driven to rotate by a transmission motor.

[0014] By adopting the technical solutions as described above, the present application has the following beneficial effects: The AI vision recognition-based threaded hole defect intelligent detection system disclosed by the present application is under the folding protection of the shielding cover box, automatically conveys the workpiece with threaded holes processed through the transmission mechanism, continuously performs fixed-point online detection, and uses the material blocking mechanism and the clamping positioning mechanism to regularly position the workpiece to be detected, so that the three-dimensional motion mechanism can drive the endoscopic probe to accurately correspond to the threaded hole to collect images, and after the images are transmitted to the upper computer, the AI analysis and diagnosis unit is used to judge defects, and the threaded hole with defects can also be marked by the marker pen; compared with the traditional industrial camera, the endoscopic probe used in the present application is self-provided with a light source and can be stretched into the hole to take pictures, cooperates with the shielding cover box, solves the problem of external environmental influence such as light change, and has a clearer imaging effect on the threaded hole, and the combination of the AI image detection algorithm is beneficial to improve the accuracy of defect detection, the marker pen can also be used to distinguish the detection results, and the whole detection process can not only adapt to complex environments, but also has high intelligence and automation, thereby greatly improving the detection efficiency and reducing the labor and maintenance costs. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 is the structural schematic diagram of the embodiment of the present application; Figure 2 is Figure 1 the structural schematic diagram of the embodiment of the present application without the shielding cover box; Figure 3 is Figure 2 the enlarged structural schematic diagram of part A of the embodiment of the present application; Figure 4 is Figure 2 the enlarged structural schematic diagram of part B of the embodiment of the present application; Figure 5 is Figure 2 the enlarged structural schematic diagram of part C of the embodiment of the present application; Figure 6 is the structural schematic diagram of the embodiment of the present application from another angle.

[0016] In the figure: 1, workbench; 101, table plate; 102, supporting leg; 103, supporting caster; 104, adjustable supporting leg; 2, transmission mechanism; 3, clamping positioning mechanism; 301, first vertical plate; 302, first telescopic cylinder; 303, top block; 4, three-dimensional motion mechanism; 401, longitudinal guide rail; 402, transverse guide rail; 403, vertical guide rail; 404, sliding block plate; 5, endoscopic probe; 6, rotating shaft; 7, marker pen; 8, clamping frame; 9, material blocking mechanism; 901, second vertical plate; 902, second telescopic cylinder; 903, top plate; 904, wheel frame; 905, positioning guide wheel; 10, shielding cover box; 11, lifting cylinder; 12, shielding plate; 13, sliding groove track; 14, control box; 15, fan. Detailed Implementation

[0017] The technical solution of the present invention will be described below with reference to the accompanying drawings of the embodiments of the present invention. In the description, it should be understood that if there are terms such as "upper", "lower", "front", "rear", "left", "right" indicating the orientation or positional relationship, they are only corresponding to the accompanying drawings of the present invention for the convenience of describing the present invention, and are not intended to indicate or imply that the device or element referred to must have a specific orientation.

[0018] Combined with appendix Figure 1 , 2 The intelligent thread hole defect detection system based on AI visual recognition, as described in section 6, includes a workbench 1, a host computer, a marker pen 7, an image acquisition unit, and an AI analysis and diagnosis unit for identifying thread hole defects. The image acquisition unit typically uses a camera for image capture, followed by image transmission and backend imaging. As needed, the AI ​​analysis and diagnosis unit uses mature detection methods, employing AI algorithms such as pixel attribution, boundary detection, region segmentation, and convolutional neural networks to identify defects in thread hole images. Pixel attribution involves analyzing the visual features of each pixel, such as color, texture, brightness, and depth, as well as its contextual relationship with neighboring pixels. Boundary detection identifies the edges of regions where features change significantly, determining the boundaries of different objects or regions. Region segmentation aggregates images into continuous, well-defined regions based on the similarity within pixel features and the differences between regions. Meaningful fragments: Convolutional neural networks can automatically learn complex features and contextual information, efficiently simulating the human visual understanding process of a scene and achieving pixel-level precise segmentation. As needed, the workbench 1 includes a platform 101 and four legs 102 fixed to the four corners of the lower surface of the platform 101. The legs 102 are connected by horizontal and vertical struts to ensure stable support. An adjustment plate is fixed to the bottom of each leg 102, and support casters 103 and adjustable feet 104 are installed on the lower surface of the adjustment plate. By raising and lowering the adjustable feet 104, stable support can be ensured, and they can be switched with the support casters 103 for transfer, transportation, or positioning. Furthermore, a control box 14 is located below the platform 101, where relevant control components can be placed. The control box 14 has a door and a ventilation opening with a fan 15 for ventilation and cooling.

[0019] The upper table of the workbench 1 is provided with a conveying mechanism 2 for conveying the workpieces to be detected. The workpieces with threaded holes machined are automatically conveyed by the conveying mechanism 2 for continuous point online detection. According to the need, the conveying mechanism 2 is generally provided with a belt conveyor, which comprises driven rollers, driving rollers rotatably installed on the frame, and a belt loop drivingly connected by the driven rollers and the driving rollers. The driving rollers are driven to rotate by a transmission motor. The table plate 101 of the workbench 1 can be opened to hide the transmission motor, thereby improving the space utilization. The upper part of the conveying mechanism 2 is provided with a shielding cover box 10. The shielding cover box 10 is provided with feeding ports on the two end walls corresponding to the input end and the output end of the conveying mechanism 2. While not affecting the feeding and discharging of the workpieces, the shielding cover box 10 can shield light, dust and the like, thereby preventing external environmental interference. The outer wall of the folding cover box above the feeding port is fixedly provided with a lifting electric cylinder 11. The lifting electric cylinder 11 is connected with a shielding plate 12 at the end of the telescopic rod. The two sides of the feeding port are fixedly provided with sliding groove tracks 13 matched with the corresponding side edges of the shielding plate 12. The shielding plate 12 is driven to move up and down along the sliding groove tracks 13 by controlling the lifting electric cylinder 11, so that the feeding port can be shielded when detecting the defects of the threaded holes, thereby ensuring more rigorous shielding and better detection effect.

[0020] The middle part of the conveying mechanism 2 is provided with a clamping and positioning mechanism 3 corresponding to the workpieces to be detected. When the workpieces stop conveying in position, the clamping and positioning mechanism 3 is used to clamp and position the workpieces, so as to realize regular correction and facilitate image acquisition of the threaded holes. Figure 4 As shown in the drawings, according to the need, the clamping and positioning mechanism 3 comprises push components symmetrically arranged on the outer sides of the conveying mechanism 2. The push components comprise a first vertical plate 301 and a first telescopic cylinder 302. The first vertical plate 301 is fixed to the upper table of the workbench 1. The cylinder body of the first telescopic cylinder 302 is fixed to the outer plate surface of the first vertical plate 301. The rod head of the first telescopic cylinder 302 is movably inserted through the first vertical plate 301 and is fixedly connected with a top block 303. The top block 303 can be provided as a long block to increase the contact length with the workpieces, thereby facilitating better clamping and correction of the position. The two ends of the top block 303 are further connected with guide rods movably inserted through the first vertical plate 301 for guiding, so that the first telescopic cylinder 302 drives the top block 303 to move more stably. The output end of the conveying mechanism 2 is provided with a material blocking mechanism 9 capable of temporarily blocking the workpieces to be detected. Only by starting and stopping the conveying mechanism 2, the workpieces can be transported to the vicinity of the clamping and positioning mechanism 3 for posture adjustment. However, position deviation may occur. Therefore, according to the need, as shown in the drawings, Figure 5As shown, two material blocking mechanisms 9 are symmetrically arranged outside the output end of the conveying mechanism 2, and the material blocking mechanism 9 comprises a second vertical plate 901 and a second telescopic cylinder 902. The second vertical plate 901 is fixed to the tabletop of the workbench 1. The cylinder body of the second telescopic cylinder 902 is fixed to the outer plate surface of the second vertical plate 901. The rod head of the second telescopic cylinder 902 is movably arranged through the second vertical plate 901 and is fixedly connected with a top plate 903. The inner plate surface of the top plate 903 is rotatably connected with a wheel shaft vertical positioning guide wheel 905 through a wheel frame 904. The positioning guide wheel 905 can be in contact with the front end surface of the workpiece to assist in positioning the workpiece and facilitate subsequent accurate detection.

[0021] The image acquisition unit comprises an endoscopic probe 5, i.e. a camera probe with a light source. The diameter of the endoscopic probe 5 is small, and the endoscopic probe 5 can be inserted into the hole of the threaded hole to take pictures. Even if the distance is very close, there will be no reflection or shadow, which ensures excellent imaging effect. The top of the conveying mechanism 2 is provided with a three-dimensional movement mechanism 4 which can drive the endoscopic probe 5 and the marker pen 7 to move in three dimensions to position the threaded hole of the workpiece to be detected. The three-dimensional movement mechanism 4 can be a mechanical arm, or according to the needs, the three-dimensional movement mechanism 4 comprises two longitudinal guide rails 401 arranged on the two sides of the conveying mechanism 2 and along the conveying direction of the conveying mechanism 2. The bottom surface of the longitudinal guide rail 401 is fixed to the workbench 1 through a stand column. The two longitudinal guide rails 401 are slidably connected with a transverse guide rail 402 which can slide and position along the longitudinal guide rail 401. The rail body of the transverse guide rail 402 is slidably connected with a vertical guide rail 403 which can slide and position along the transverse guide rail 402. The rail body of the vertical guide rail 403 is slidably connected with a sliding block plate 404 which can slide and position along the vertical guide rail 403. The endoscopic probe 5 and the marker pen 7 are arranged at the lower end of the sliding block plate 404. Through X, Y and Z three-direction linear motion control, the endoscopic probe 5 and the marker pen 7 can also move in space, which facilitates the automatic and intelligent detection and defect marking of the threaded hole on the workpiece. Figure 3 As shown, the lower end of the sliding block plate 404 is rotatably connected with a clamping frame 8 through a rotating shaft 6 parallel to the transverse guide rail 402. The clamping frame 8 is rotatably connected with a flip driving unit. The driving unit can be a flip motor. The output shaft of the flip motor is coaxially fixed with the rotating shaft 6. The endoscopic probe 5 is clamped and fixed to one end of the clamping frame 8. The marker pen 7 is clamped and fixed to the other end of the clamping frame 8. By controlling the flip of the clamping frame 8, the endoscopic probe 5 or the marker pen 7 can be controlled to probe or mark downward. The marker pen 7 can be used for marking or carving marking, and is mainly used for marking the threaded hole with defects.

[0022] The host computer is respectively connected with the AI analysis and diagnosis unit, the image acquisition unit, the conveying mechanism 2, the clamping and positioning mechanism 3, the material blocking mechanism 9 and the three-dimensional movement mechanism 4. Through the host computer, the various mechanisms and units are controlled to coordinate the action, so as to realize the automation and intelligence of the system.

[0023] The AI vision recognition-based threaded hole defect intelligent detection system is implemented. The first telescopic cylinder 302 of the clamping positioning mechanism 3 is in a retracted state, and the second telescopic cylinder 902 of the material blocking mechanism 9 is in an extended state. When the workpiece to be detected is conveyed to a position by the conveying mechanism 2 and the workpiece is contacted by the positioning guide wheel 905 of the material blocking mechanism 9, the conveying mechanism 2 is paused, the first telescopic cylinder 302 of the clamping positioning mechanism 3 is controlled to drive the top block 303 to extend, and the workpiece is regularly positioned. After the workpiece is positioned, the telescopic cylinders of the clamping positioning mechanism 3 and the material blocking mechanism 9 can be retracted and returned, or can be retracted and returned after detection. During detection, the endoscopic probe 5 is moved to a position for image acquisition by controlling the three-dimensional motion mechanism 4 according to the relative position of the known threaded hole on the workpiece. The image is displayed on the display screen of the upper computer. After detection by the AI analysis and diagnosis unit, if there is a defect, the marking pen 7 is controlled to flip downward and mark beside the defect threaded hole. If there is no defect, the endoscopic probe 5 continues to check the next threaded hole until the workpiece is detected. Then, the conveying mechanism 2 is started to convey the workpiece to the subsequent process.

[0024] The parts of the present application not described in detail are prior art. It is obvious for those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and the present application can be implemented in other specific forms without departing from the spirit or essential characteristics of the present application. Therefore, the above embodiments should be regarded as exemplary and non-limiting, the scope of the present application is defined by the appended claims rather than the above description, and therefore all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present application, and any reference signs in the claims should not be regarded as limiting the content of the claims involved.

Claims

1. An intelligent inspection system for threaded hole defects based on AI visual recognition, characterized in that: The utility model provides a workbench, host computer, marker pen, image acquisition unit and AI analysis diagnosis unit for identifying thread hole defects are included, the upper surface of workbench is equipped with transmission mechanism for conveying workpiece to be detected, the upper cover of transmission mechanism is equipped with shielding cover box, the both ends of shielding cover box correspond to the two ends of transmission mechanism input end and output end and are equipped with feeding port, the outer side of transmission mechanism middle part is equipped with clamping positioning mechanism corresponding to workpiece to be detected, the output end of transmission mechanism is equipped with material blocking mechanism that can temporarily block workpiece to be detected, image acquisition unit includes endoscopic probe, the upper of transmission mechanism is equipped with three -dimensional movement mechanism that can drive endoscopic probe and marker pen to carry out three -dimensional movement to position corresponding workpiece thread hole to be detected, host computer is connected with AI analysis diagnosis unit, image acquisition unit, transmission mechanism, clamping positioning mechanism, material blocking mechanism and three -dimensional movement mechanism corresponding signal respectively.

2. The AI vision recognition-based threaded hole defect intelligent detection system according to claim 1, characterized in that: The outer wall of the folding cover box above the feeding port is fixedly installed with a lifting electric cylinder, the end of the telescopic rod of the lifting electric cylinder is connected with a shielding plate, and the both sides of the feeding port are fixedly provided with a sliding groove track matched with the corresponding side of the shielding plate.

3. The AI vision recognition-based threaded hole defect intelligent detection system according to claim 1, characterized in that: The three-dimensional movement mechanism includes two longitudinal guides arranged on the both sides of the transmission mechanism and along the conveying direction of the transmission mechanism, the bottom surface of the longitudinal guide is fixed to the workbench through a stand column, a transverse guide capable of sliding positioning along the longitudinal guide is slidingly installed between the two longitudinal guides, a vertical guide capable of sliding positioning along the transverse guide is slidingly installed on the transverse guide, a sliding block plate capable of sliding positioning along the vertical guide is slidingly installed on the vertical guide, and the endoscopic probe is installed on the lower end of the sliding block plate.

4. The AI vision-identification-based threaded hole defect intelligent detection system according to claim 3, characterized in that: The lower end of the sliding block plate is rotationally connected with a clamping frame through a rotating shaft parallel to the transverse guide, the clamping frame is rotationally positioned through a turnover driving unit, and the endoscopic probe and the marker pen are respectively clamped and fixed on the both ends of the clamping frame.

5. The AI vision recognition-based threaded hole defect intelligent detection system according to claim 1, characterized in that: The AI analysis diagnosis unit identifies thread hole defects through the AI algorithms of pixel attribution, boundary detection, region division and convolutional neural network. 6.The AI vision-identified threaded hole defect intelligent detection system according to claim 1, characterized in that: The clamping positioning mechanism includes a push assembly symmetrically arranged outside the both sides of the transmission mechanism, the push assembly includes a first vertical plate and a first telescopic cylinder, the first vertical plate is fixed to the upper surface of the workbench, the cylinder body of the first telescopic cylinder is fixed to the outer plate surface of the first vertical plate, and the rod head of the first telescopic cylinder is movably inserted through the first vertical plate and is fixedly connected with a top block.

7. The AI vision recognition-based threaded hole defect intelligent detection system according to claim 1, characterized in that: Two material blocking mechanisms are symmetrically arranged outside the both sides of the output end of the transmission mechanism, the material blocking mechanism includes a second vertical plate and a second telescopic cylinder, the second vertical plate is fixed to the upper surface of the workbench, the cylinder body of the second telescopic cylinder is fixed to the outer plate surface of the second vertical plate, the rod head of the second telescopic cylinder is movably inserted through the second vertical plate and is fixedly connected with a top plate, and the inner plate surface of the top plate is rotationally installed with a vertical positioning guide wheel through a wheel frame. 8.The AI vision-identifying-based threaded hole defect intelligent detection system according to claim 1, characterized in that: The workbench includes a table plate and four supporting legs fixed to the lower plate surface of the table plate, the bottom surface of the supporting leg is fixed with an adjusting plate, and the lower plate surface of the adjusting plate is installed with a supporting caster and an adjustable supporting leg. 9.The AI vision-identifying-based threaded hole defect intelligent detection system according to claim 8, characterized in that: A control box is arranged below the table plate, the wall of the control box is provided with a box door and an air vent provided with a fan. 10.The AI vision-identification-based threaded hole defect intelligent detection system according to claim 1, characterized in that: The transmission mechanism comprises a driven roller rotatably mounted on the frame, a driving roller, and a transmission belt loop connected in transmission with the driven roller and the driving roller, the driving roller being driven to rotate by the transmission motor.

Images