Magnetic ring surface defect intelligent identification system and method based on deep learning

Abstract

The invention discloses an intelligent recognition system and method for magnetic ring surface defects based on deep learning, and relates to the technical field of magnetic ring detection.The intelligent recognition system comprises a first frame body, a controller component, a top dust removal module and a visual recognition module, and the output end of a first electric telescopic rod component is provided with a first lifting plate; a second motor component is installed at the bottom of the first lifting plate, a rotating rod is installed at the output end of the second motor component, a guide rod is installed on the outer side of the rotating rod, a first pipe body is installed on the outer side of the guide rod, a first brush is installed at the bottom of the first pipe body, and the visual recognition module comprises a second right-angle rod, a camera and an illuminating lamp component. The first electric telescopic rod component drives the first brush to move downwards to enable the rotating rod to be inserted into the inner side of the magnetic ring, then the second motor component drives the first brush to rotate, the first brush rotates to remove dust on the upper surface of the magnetic ring, and then the situation that dust on the magnetic ring influences the shooting definition of a camera on the magnetic ring is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of magnetic ring detection, and specifically to an intelligent recognition system and method for magnetic ring surface defects based on deep learning. Background Art

[0002] In the field of magnetic ring detection, image recognition is often used. The defects of magnetic rings are judged by analyzing the patterns on the surface of the magnetic rings through software. When the magnetic rings are transported on the conveyor belt, the surface is easily covered with dust, resulting in misjudgment in image recognition.

[0003] The defects of existing image recognition devices are as follows: 1. The prior art KR101464489B1 discloses a detection method and system based on image recognition. When the surface of the magnetic ring is recognized, it is easily affected by the dust on the surface of the magnetic ring, resulting in incorrect defect recognition. Therefore, an intelligent recognition system for magnetic ring surface defects based on deep learning that can clean the dust on the surface of the magnetic ring is needed to solve this problem.

[0004] 2. The prior art KR101318812B1 discloses a filtering method for detecting the edge direction of an image and its image recognition method. This technology does not have a structure for flipping the product to be photographed and cannot photograph and recognize the bottom of the product. Therefore, an intelligent recognition system for magnetic ring surface defects based on deep learning that can flip the magnetic ring is needed to solve this problem.

[0005] 3. The prior art JP2012063280A discloses an image recognition device and a detector. This technology does not have a structure for cleaning the conveyor belt device. When the conveyor device transports the magnetic ring, the magnetic ring will contact the conveyor device, and the impurities on the conveyor device will adhere to the magnetic ring, affecting the image shooting and image recognition of the magnetic ring. Therefore, an intelligent recognition system for magnetic ring surface defects based on deep learning that can clean the conveyor device is needed to solve this problem.

[0006] 4. The prior art CN207675660U discloses a magnetic ring surface defect detection device. When the conveyor device is cleaned by sprinkling water, when the water is recycled and filtered, the filter component is easily blocked by impurities, resulting in the filter component being unable to effectively filter the cleaning water. Therefore, an intelligent recognition system for magnetic ring surface defects based on deep learning that can sprinkle water to clean the conveyor device of the magnetic ring, recycle and filter the water at the same time, and avoid the filter component from being blocked is needed to solve this problem. Summary of the Invention

[0007] An object of the present application is to provide an intelligent recognition system and method for magnetic ring surface defects based on deep learning, which can solve the technical problems proposed in the prior art.

[0008] To achieve the above object, the present invention provides the following technical solution: An intelligent magnetic ring surface defect recognition system based on deep learning, comprising a first frame, a controller component, a top dust cleaning module and a visual recognition module. A plurality of struts are symmetrically installed at the bottom of the first frame. A controller component is installed on one side of the first frame. The top dust cleaning module includes a first right-angle rod, a first electric telescopic rod component, a first lifting plate, a second motor component, a rotating rod, a guide rod, a first pipe body, a first brush and a bolt. The two first right-angle rods are installed at the top of the first frame. The top of the first right-angle rod is installed with a first electric telescopic rod component, and the first electric telescopic rod component is electrically connected to the controller component. The output end of the first electric telescopic rod component is installed with a first lifting plate. The bottom of the first lifting plate is installed with a second motor component, and the second motor component is electrically connected to the controller component. The output end of the second motor component is installed with a rotating rod. The outside of the rotating rod is installed with a guide rod. The outside of the guide rod is installed with a first pipe body. The bottom of the first pipe body is installed with a first brush. The top of the first pipe body is installed with a bolt through it; The visual recognition module includes a second right-angle rod, a camera and a lighting component. The two second right-angle rods are installed at the top of the first frame. The bottom of the second right-angle rod is installed with a camera, and the camera is electrically connected to the controller component. The lighting component is installed at the bottom of the second right-angle rod, and the lighting component is electrically connected to the controller component.

[0009] Preferably, the intelligent magnetic ring surface defect recognition system based on deep learning further includes a conveying module. The conveying module includes a rotating wheel, a conveyor belt and a first motor component. A plurality of rotating wheels are movably installed through the inside of the first frame. The outside of the rotating wheel is installed with a conveyor belt. Two first motor components are symmetrically installed on one side of the first frame, and the output end of the first motor component is connected to one end of the rotating wheel. The first motor component is electrically connected to the controller component.

[0010] Preferably, the intelligent surface defect identification system for magnetic rings based on deep learning further includes a flipping module. The flipping module includes a right-angle rod four, an electric telescopic rod component two, a piston rod two, a lifting plate two, a motor component four, a frame three, an electric telescopic rod component three, a push rod, a guide ring, a clamping rod, and a rubber pad. The right-angle rod four is installed on the top of the frame one. An electric telescopic rod component two is installed on the top of the right-angle rod four and is electrically connected to the controller component. The output end of the electric telescopic rod component two is installed with the piston rod two. One end of the piston rod two is installed with the lifting plate two. A motor component four is installed on one side of the lifting plate two and is electrically connected to the controller component. The output end of the motor component four is installed with the frame three. Electric telescopic rod components three are symmetrically installed on both sides of the frame three and are electrically connected to the controller component. The output end of the electric telescopic rod component three is installed with the push rod. One end of the push rod is installed with the guide ring. A clamping rod is installed on the outer side of the guide ring. A rubber pad is installed on the back of the clamping rod.

[0011] Preferably, the intelligent surface defect identification system for magnetic rings based on deep learning further includes a front cleaning module. The front cleaning module includes a hydraulic cylinder component one, a frame two, a brush roller, and a motor component three. Two hydraulic cylinder components one are symmetrically installed on both sides of the frame one and are electrically connected to the controller component. The output end of the hydraulic cylinder component one is installed with the frame two, and the frame two is located in front of the frame one. The brush roller is movably installed through the inner side of the frame two. The motor component three is installed on one side of the frame two, and the output end of the motor component three is connected to one end of the brush roller. The motor component three is electrically connected to the controller component.

[0012] Preferably, the intelligent surface defect identification system for magnetic rings based on deep learning further includes a watering module. The watering module includes a water sprinkling pipe, a water accumulation frame, a pipe body two, a water pump component, and a hose. The water sprinkling pipe is installed on the top of the frame two and is located above the brush roller. A plurality of water sprinkling openings are formed through the bottom of the water sprinkling pipe. The water accumulation frame is installed on the bottom of the frame one. The front output end of the water accumulation frame is installed with the pipe body two. A water pump component is installed on the front of the water accumulation frame, and the input end of the water pump component is connected to the output end of the pipe body two. The water pump component is electrically connected to the controller component. The output end of the water pump component is installed with the hose, and the output end of the hose is connected to the input end of the water sprinkling pipe.

[0013] Preferably, a plurality of support blocks are symmetrically installed on the inner wall of the water accumulation frame, and a filter plate is arranged on the top of the support blocks.

[0014] Preferably, hydraulic cylinder components two are symmetrically installed on the back of the water accumulation frame and are electrically connected to the controller component. The output end of the hydraulic cylinder component two is installed with a piston rod one, and one end of the piston rod one is installed with a brush two, and the brush two is located above the filter plate.

[0015] Preferably, the intelligent magnetic ring surface defect recognition system based on deep learning further includes a drying module, which includes a right-angle rod three and a hot air blower component. The right-angle rod three is installed on the top of the first frame body, and a plurality of hot air blower components are installed at the bottom of the right-angle rod three, and the hot air blower components are electrically connected to the controller component.

[0016] Preferably, the recognition method of the intelligent magnetic ring surface defect recognition system based on deep learning is as follows: S1. Place the magnetic ring on the conveyor belt, and then drive the magnetic ring backward by the conveyor belt to below the second motor component; S2. The front electric telescopic rod component one drives the brush one to move downward so that the rotating rod inserts into the inner side of the magnetic ring, and then drives the brush one to rotate by the second motor component. The brush one rotates to clean the upper surface of the magnetic ring; S3. Then the brush one moves upward, and then the conveyor belt drives the magnetic ring backward to below the front camera, and then the magnetic ring is photographed by the camera; S4. Then the conveyor belt drives the magnetic ring backward to below the third frame body. The second electric telescopic rod component drives the third frame body to move downward, and then the clamping rod moves inward to clamp the magnetic ring. Then the third frame body moves upward, and then the third frame body rotates 180 degrees to drive the magnetic ring to flip 180 degrees; S5. Then the conveyor belt continues to drive the magnetic ring backward. The rear brush one cleans the magnetic ring, and then the rear camera photographs the magnetic ring. The analysis software in the controller component analyzes the surface defects of the magnetic ring.

[0017] Preferably, the following steps are further included in the S5: S51. The second frame body and the brush roller move backward. The brush roller rotates to clean the surface of the conveyor belt. At the same time, the water in the water sprinkling pipe is sprinkled on the brush roller through the water sprinkling port for wetting. The water on the brush roller falls into the water accumulation frame. The filter plate filters the water, and at the same time, the brush two cleans the filter plate.

[0018] Compared with the prior art, the beneficial effects of the present invention are: 1. In the present invention, the electric telescopic rod component one drives the brush one to move downward so that the rotating rod inserts into the inner side of the magnetic ring, and then drives the brush one to rotate by the second motor component. The brush one rotates to clean the upper surface of the magnetic ring, thereby avoiding the influence of dust on the magnetic ring on the shooting clarity of the camera for the magnetic ring.

[0019] 2. The present invention is provided with a flipping module. The electric telescopic rod component two drives the frame three to move downward, then the clamping rod moves inward to clamp the magnetic ring, then the frame three moves upward, and then the frame three rotates 180 degrees to drive the magnetic ring to flip 180 degrees, so that the bottom of the magnetic ring can be flipped upward, facilitating dust cleaning and shooting identification of the top and bottom of the magnetic ring respectively. At the same time, the clamping rod can move downward after flipping, so that the magnetic ring clamped by the clamping rod can stably fall on the conveyor belt.

[0020] 3. The present invention moves the frame two and the brush roller backward. The rotation of the brush roller can clean the surface of the conveyor belt. At the same time, the water in the water sprinkling pipe is sprinkled on the brush roller through the water sprinkling port for wetting. The water on the brush roller falls into the water accumulation frame, and the filter plate filters the water to ensure the cleanliness of the water falling on the brush roller.

[0021] 4. The present invention can drive the brush two to move back and forth through the hydraulic cylinder component two, so that the brush two can clean the top of the filter plate and prevent the holes on the filter plate from being blocked. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 is a perspective view of the present invention; Figure 2 is a schematic structural view of the right-angle rod one of the present invention; Figure 3 is a schematic structural view of the guide rod of the present invention; Figure 4 is a schematic structural view of the frame two and the water accumulation frame of the present invention; Figure 5 is a schematic structural view of the water accumulation frame of the present invention; Figure 6 is a schematic structural view of the water sprinkling pipe of the present invention; Figure 7 is a schematic front sectional view structural view of the water accumulation frame of the present invention; Figure 8 is a schematic structural view of the right-angle rod four of the present invention; Figure 9 is a schematic structural view of the frame three of the present invention; Figure 10 is a schematic structural view of the guide ring of the present invention; Figure 11 is a module flow chart of the present invention; Figure 12 is a flow chart of the usage method of the present invention.

[0023] In the figure: 1. First frame; 2. Support pillar; 3. Rotating wheel; 4. Conveyor belt; 5. First motor component; 6. First right-angle rod; 7. First electric telescopic rod component; 8. First lifting plate; 9. Second motor component; 10. Rotating rod; 11. Guide rod; 12. First pipe body; 13. First brush; 14. Bolt; 15. Second right-angle rod; 16. Camera; 17. First hydraulic cylinder component; 18. Second frame; 19. Brush roller; 20. Third motor component; 21. Water sprinkling pipe; 22. Water accumulation frame; 23. Second pipe body; 24. Water pump component; 25. Hose; 26. Support block; 27. Filter plate; 28. Second hydraulic cylinder component; 29. First piston rod; 30. Second brush; 31. Water sprinkling port; 32. Third right-angle rod; 33. Hot air blower component; 34. Fourth right-angle rod; 35. Second electric telescopic rod component; 36. Second piston rod; 37. Second lifting plate; 38. Fourth motor component; 39. Third frame; 40. Third electric telescopic rod component; 41. Push rod; 42. Guide ring; 43. Clamping rod; 44. Rubber pad; 45. Controller component; 46. Lighting component. Detailed implementation manner

[0024] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0025] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "upper", "lower", "inner", "outer", "front end", "rear end", "both ends", "one end", "the other end", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance.

[0026] In the description of the present invention, it should be noted that unless otherwise clearly defined and limited, the terms "installed", "provided with", "connected", etc. should be understood in a broad sense. For example, "connected" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the internal communication of two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0027] Please refer to Figure 1 ,Figure 2 and Figure 3 , an embodiment provided by the present invention: an intelligent magnetic ring surface defect recognition system based on deep learning; It includes a first frame 1, a controller component 45, a top dust cleaning module and a visual recognition module. A plurality of struts 2 are symmetrically installed at the bottom of the first frame 1. A controller component 45 is installed on one side of the first frame 1. The top dust cleaning module includes a first right-angle rod 6, a first electric telescopic rod component 7, a first lifting plate 8, a second motor component 9, a rotating rod 10, a guide rod 11, a first pipe body 12, a first brush 13 and a bolt 14. Two first right-angle rods 6 are installed on the top of the first frame 1. A first electric telescopic rod component 7 is installed at the top of the first right-angle rod 6, and the first electric telescopic rod component 7 is electrically connected to the controller component 45. The output end of the first electric telescopic rod component 7 is installed with a first lifting plate 8. A second motor component 9 is installed at the bottom of the first lifting plate 8, and the second motor component 9 is electrically connected to the controller component 45. The output end of the second motor component 9 is installed with a rotating rod 10. A guide rod 11 is installed on the outer side of the rotating rod 10. A first pipe body 12 is installed on the outer side of the guide rod 11. A first brush 13 is installed at the bottom of the first pipe body 12. A bolt 14 is installed through the top of the first pipe body 12. The visual recognition module includes a second right-angle rod 15, a camera 16 and a lighting component 46. Two second right-angle rods 15 are installed on the top of the first frame 1. A camera 16 is installed at the bottom of the second right-angle rod 15, and the camera 16 is electrically connected to the controller component 45. The lighting component 46 is installed at the bottom of the second right-angle rod 15, and the lighting component 46 is electrically connected to the controller component 45. The first frame 1 can provide an installation position for other components of the device, enabling other components of the device to have a position for installation. The struts 2 can provide support for the first frame 1. The controller component 45 can receive the signal from the camera 16 and can control the first motor component 5, the first electric telescopic rod component 7, the second motor component 9, the first hydraulic cylinder component 17, the third motor component 20, the water pump component 24, the second hydraulic cylinder component 28, the hot air blower component 33, the second electric telescopic rod component 35, the fourth motor component 38, the third electric telescopic rod component 40 and the lighting component 46. The controller component 45 integrates software inside, analyzes the defects of the magnetic ring according to the image captured by the camera 16, and saves the analysis result. It continuously learns according to the result and the picture to facilitate a better comparative analysis of subsequent magnetic rings. The first right-angle rod 6 can provide an installation position for the first electric telescopic rod component 7. The first electric telescopic rod component 7 can convert electrical energy into kinetic energy, thereby driving the first lifting plate 8 to move up and down. The first lifting plate 8 can drive the second motor component 9 and the first brush 13 to move up and down through the up and down movement. The second motor component 9 can convert electrical energy into kinetic energy, thereby driving the rotating rod 10 to rotate. The rotating rod 10 can drive the guide rod 11 and the first brush 13 to rotate through the rotation. The guide rod 11 can provide guidance for the first pipe body 12, enabling the first pipe body 12 to move left and right. The first pipe body 12 can drive the first brush 13 to move left and right through the left and right movement. The first brush 13 can clean the upper surface of magnetic rings with different diameters through the left and right movement. The bolt 14 can squeeze the guide rod 11 through rotation.Thus, the stability of the first pipe body 12 outside the guide rod 11 can be ensured. The second right-angle rod 15 can provide an installation position for the camera 16. The camera 16 can take pictures of the surface of the magnetic ring. The lighting component 46 can convert electrical energy into light energy to illuminate the surface of the magnetic ring.

[0028] Please refer to Figure 1 , an embodiment provided by the present invention: an intelligent magnetic ring surface defect identification system based on deep learning; The intelligent magnetic ring surface defect identification system based on deep learning further includes a conveying module. The conveying module includes a rotating wheel 3, a conveyor belt 4, and a first motor component 5. A plurality of rotating wheels 3 are movably installed through the inside of the first frame 1. The conveyor belt 4 is installed on the outside of the rotating wheel 3. Two first motor components 5 are symmetrically installed on one side of the first frame 1, and the output end of the first motor component 5 is connected to one end of the rotating wheel 3. The first motor component 5 is electrically connected to the controller component 45. The rotating wheel 3 can drive the conveyor belt 4 to move by rotating. The conveyor belt 4 can drive the magnetic ring to move backward by moving backward. The first motor component 5 can convert electrical energy into kinetic energy to drive the conveyor belt 4 to move.

[0029] Please refer to Figure 1 , Figure 8 Figure 9 and Figure 10 , an embodiment provided by the present invention: an intelligent magnetic ring surface defect identification system based on deep learning; Including a flipping module, the intelligent magnetic ring surface defect identification system based on deep learning further includes a flipping module. The flipping module includes a right-angle rod four 34, an electric telescopic rod component two 35, a piston rod two 36, a lifting plate two 37, a motor component four 38, a frame three 39, an electric telescopic rod component three 40, a push rod 41, a guide ring 42, a clamping rod 43, and a rubber pad 44. The right-angle rod four 34 is installed on the top of the frame one 1. An electric telescopic rod component two 35 is installed on the top of the right-angle rod four 34, and the electric telescopic rod component two 35 is electrically connected to the controller component 45. The output end of the electric telescopic rod component two 35 is installed with a piston rod two 36. One end of the piston rod two 36 is installed with a lifting plate two 37. A motor component four 38 is installed on one side of the lifting plate two 37, and the motor component four 38 is electrically connected to the controller component 45. The output end of the motor component four 38 is installed with a frame three 39. Electric telescopic rod components three 40 are symmetrically installed on both sides of the frame three 39, and the electric telescopic rod components three 40 are electrically connected to the controller component 45. The output end of the electric telescopic rod component three 40 is installed with a push rod 41. One end of the push rod 41 is installed with a guide ring 42. A clamping rod 43 is installed on the outer side of the guide ring 42. A rubber pad 44 is installed on the back of the clamping rod 43. The right-angle rod four 34 can provide an installation position for the electric telescopic rod component two 35. The electric telescopic rod component two 35 can convert electrical energy into kinetic energy, thereby driving the piston rod two 36 to move up and down. The piston rod two 36 can drive the lifting plate two 37 to move up and down through the up and down movement. The lifting plate two 37 can drive the motor component four 38 and the frame three 39 to move up and down through the up and down movement. The motor component four 38 can convert electrical energy into kinetic energy, thereby driving the frame three 39 to rotate. The frame three 39 can drive the guide ring 42 and the clamping rod 43 to flip through the rotation, so as to flip the magnetic ring placed inside the clamping rod 43. The electric telescopic rod component three 40 can convert electrical energy into kinetic energy, thereby driving the push rod 41 to move back and forth. The push rod 41 can drive the guide ring 42 and the clamping rod 43 to move back and forth through the back and forth movement. The guide ring 42 can provide guidance for the clamping rod 43, so that the clamping rod 43 can move down under the action of gravity. The front and rear two clamping rods 43 can clamp the magnetic ring placed inside. The rubber pad 44 can prevent the clamping rod 43 from directly contacting the magnetic ring, thereby reducing the damage to the magnetic ring.

[0030] Please refer to Figure 1 and Figure 4 , an embodiment provided by the present invention: an intelligent magnetic ring surface defect identification system based on deep learning; Including a front cleaning module, the intelligent magnetic ring surface defect identification system based on deep learning further includes a front cleaning module. The front cleaning module includes a first hydraulic cylinder component 17, a second frame 18, a brush roller 19, and a third motor component 20. The two first hydraulic cylinder components 17 are symmetrically installed on both sides of the first frame 1, and the first hydraulic cylinder component 17 is electrically connected to the controller component 45. The output end of the first hydraulic cylinder component 17 is installed with a second frame 18, and the second frame 18 is located in front of the first frame 1. The brush roller 19 is movably installed through the inside of the second frame 18. The third motor component 20 is installed on one side of the second frame 18, and the output end of the third motor component 20 is connected to one end of the brush roller 19. The third motor component 20 is electrically connected to the controller component 45. The first hydraulic cylinder component 17 can convert hydraulic energy into kinetic energy, thereby driving the second frame 18 to move back and forth. The second frame 18 can drive the brush roller 19 to move back and forth through the back-and-forth movement. The brush roller 19 can clean the dust on the front of the conveyor belt 4 by rotating. The third motor component 20 can convert electrical energy into kinetic energy, thereby driving the brush roller 19 to rotate.

[0031] Please refer to Figure 4 , Figure 5 , Figure 6 and Figure 7 , an embodiment provided by the present invention: an intelligent magnetic ring surface defect identification system based on deep learning; Including a water sprinkling module, the intelligent magnetic ring surface defect identification system based on deep learning further includes a water sprinkling module. The water sprinkling module includes a water sprinkling pipe 21, a water accumulation frame 22, a second pipe body 23, a water pump component 24, and a hose 25. The water sprinkling pipe 21 is installed on the top of the second frame 18 and is located above the brush roller 19. A plurality of water sprinkling openings 31 are formed through the bottom of the water sprinkling pipe 21. The water accumulation frame 22 is installed at the bottom of the first frame 1. The front output end of the water accumulation frame 22 is installed with the second pipe body 23. The front of the water accumulation frame 22 is installed with the water pump component 24, and the input end of the water pump component 24 is connected to the output end of the second pipe body 23. The water pump component 24 is electrically connected to the controller component 45. The output end of the water pump component 24 is installed with the hose 25, and the output end of the hose 25 is connected to the input end of the water sprinkling pipe 21. A plurality of support blocks 26 are symmetrically installed on the inner wall of the water accumulation frame 22. A filter plate 27 is arranged on the top of the support block 26. The back of the water accumulation frame 22 is symmetrically installed with a second hydraulic cylinder component 28, and the second hydraulic cylinder component 28 is electrically connected to the controller component 45. The output end of the second hydraulic cylinder component 28 is installed with a first piston rod 29. One end of the first piston rod 29 is installed with a second brush 30, and the second brush 30 is located above the filter plate 27. The intelligent magnetic ring surface defect identification system based on deep learning further includes a drying module. The drying module includes a third right-angle rod 32 and a hot air blower component 33. The third right-angle rod 32 is installed on the top of the first frame 1. A plurality of hot air blower components 33 are installed at the bottom of the third right-angle rod 32, and the hot air blower components 33 are electrically connected to the controller component 45. The water sprinkling pipe 21 can provide a transmission path for the water flow in the hose 25 to fall onto the brush roller 19. The water sprinkling openings 31 can provide a transmission path for the water in the water sprinkling pipe 21 to be discharged. The water accumulation frame 22 functions to store water. The second pipe body 23 can provide a transmission path for the water in the water accumulation frame 22 to enter the water pump component 24, and the height of the second pipe body 23 is lower than that of the filter plate 27. The water pump component 24 can convert electrical energy into kinetic energy, thereby pumping the water in the water accumulation frame 22 towards the water sprinkling pipe 21. The hose 25 can provide a transmission path for the water in the water pump component 24 to enter the water sprinkling pipe 21. The support blocks 26 function to support the filter plate 27. The filter plate 27 can filter the water falling into the water accumulation frame 22. The second hydraulic cylinder component 28 can convert hydraulic energy into kinetic energy, thereby driving the first piston rod 29 to move back and forth. The first piston rod 29 can drive the second brush 30 to move back and forth through the back-and-forth movement, and the second brush 30 can clean the top of the filter plate 27 through the back-and-forth movement to prevent the filter plate 27 from being blocked. The third right-angle rod 32 can provide an installation position for the hot air blower components 33. The hot air blower components 33 can convert electrical energy into heat energy to dry the magnetic rings on the conveyor belt 4.

[0032] The identification method of the intelligent magnetic ring surface defect identification system based on deep learning is as follows: S1. Place the magnetic ring on the conveyor belt 4, and then drive the magnetic ring to move backward under the second motor component 9 through the conveyor belt 4; S2. The first electric telescopic rod component 7 in the front drives the first brush 13 to move downward so that the rotating rod 10 is inserted into the inner side of the magnetic ring, and then the second motor component 9 drives the first brush 13 to rotate, and the first brush 13 rotates to clean the upper surface of the magnetic ring; S3. Then the first brush 13 moves upward, and then the conveyor belt 4 drives the magnetic ring to move backward under the camera 16 in the front, and then the camera 16 takes a picture of the magnetic ring; S4. Then the conveyor belt 4 drives the magnetic ring to move backward under the third frame 39, the second electric telescopic rod component 35 drives the third frame 39 to move downward, then the clamping rod 43 moves inward to clamp the magnetic ring, then the third frame 39 moves upward, and then the third frame 39 rotates 180 degrees to drive the magnetic ring to flip 180 degrees; S5. Then the conveyor belt 4 continues to drive the magnetic ring to move backward, the rear first brush 13 cleans the magnetic ring, and then the rear camera 16 takes a picture of the magnetic ring, and the analysis software in the controller component 45 analyzes the surface defects of the magnetic ring.

[0033] Step S5 also includes the following steps: S51. The second frame 18 and the brush roller 19 move backward, the brush roller 19 rotates to clean the surface of the conveyor belt 4, and at the same time, the water in the water sprinkling pipe 21 is sprinkled on the brush roller 19 through the water sprinkling port 31 for wetting, the water on the brush roller 19 falls into the water accumulation frame 22, the filter plate 27 filters the water, and at the same time, the second brush 30 cleans the filter plate 27.

[0034] Working principle: Before using the intelligent recognition system for surface defects of magnetic rings based on deep learning, it is necessary to check whether there are any problems affecting the use of the intelligent recognition system for surface defects of magnetic rings based on deep learning. Place the magnetic ring on the conveyor belt 4, and then drive the magnetic ring backward by the conveyor belt 4 to the lower part of the motor component two 9. The electric telescopic rod component one 7 in the front drives the brush one 13 to move downward so that the rotating rod 10 is inserted into the inner side of the magnetic ring. Then, drive the brush one 13 to rotate by the motor component two 9. The brush one 13 rotates to clean the upper surface of the magnetic ring. Then, the brush one 13 moves upward. Then, the conveyor belt 4 drives the magnetic ring backward to the lower part of the camera 16 in the front. Then, take a picture of the magnetic ring by the camera 16. Then, the conveyor belt 4 drives the magnetic ring backward to the lower part of the frame three 39. Drive the frame three 39 to move downward by the electric telescopic rod component two 35. Then, move the clamping rod 43 inward to clamp the magnetic ring. Then, the frame three 39 moves upward. Subsequently, the frame three 39 rotates 180 degrees to drive the magnetic ring to flip 180 degrees. Subsequently, the conveyor belt 4 continues to drive the magnetic ring backward. The rear brush one 13 cleans the magnetic ring. Then, the rear camera 16 takes a picture of the magnetic ring. The analysis software in the controller component 45 analyzes the surface defects of the magnetic ring. At the same time, the frame two 18 and the brush roller 19 move backward. The brush roller 19 rotates to clean the surface of the conveyor belt 4. At the same time, the water in the water sprinkling pipe 21 is sprinkled on the brush roller 19 through the water sprinkling port 31 for wetting. The water on the brush roller 19 falls into the water accumulation frame 22. The filter plate 27 filters the water. At the same time, the brush two 30 cleans the filter plate 27.

[0035] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and without departing from the spirit or basic characteristics of the present invention, the present invention can be implemented in other specific forms. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, it is intended to embrace all changes within the meaning and scope of the equivalent elements of the claims in the present invention. Any reference signs in the claims should not be regarded as limiting the claimed rights.

Claims

1. A magnetic ring surface defect intelligent identification system based on deep learning, characterized by: The invention comprises a frame body (1), a controller component (45), a top dust cleaning module and a visual recognition module. The bottom of the frame body (1) is symmetrically provided with a plurality of pillars (2). The controller component (45) is installed on one side of the frame body (1). The top dust cleaning module comprises a right-angle rod (6), an electric telescopic rod component (7), a lifting plate (8), a motor component (9), a rotating rod (10), a guide rod (11), a tube body (12), a brush (13) and a bolt (14). The two right-angle rods (6) are installed on the top of the frame body (1). The top of the right-angle rod (6) is installed with an electric telescopic rod component (7). The electric telescopic rod component 1 (7) is electrically connected to the controller component (45); a lifting plate 1 (8) is installed at the output end of the electric telescopic rod component 1 (7); a motor component 2 (9) is installed at the bottom of the lifting plate 1 (8); and the motor component 2 (9) is electrically connected to the controller component (45); a rotating rod (10) is installed at the output end of the motor component 2 (9); a guide rod (11) is installed on the outer side of the rotating rod (10); a tube body 1 (12) is installed on the outer side of the guide rod (11); a brush 1 (13) is installed at the bottom of the tube body 1 (12); and a bolt (14) is installed through the top of the tube body 1 (12); The visual recognition module comprises a right-angle rod (15), a camera (16) and a lighting component (46); the two right-angle rods (15) are mounted on the top of the frame (1); the camera (16) is mounted on the bottom of the right-angle rod (15), and the camera (16) is connected to the controller component (45) via electrical signals; the lighting component (46) is mounted on the bottom of the right-angle rod (15), and the lighting component (46) is connected to the controller component (45) via electrical signals.

2. According to claim 1, a magnetic ring surface defect intelligent identification system based on deep learning is characterized in that: The magnetic ring surface defect intelligent identification system based on deep learning also includes a conveying module, the conveying module includes a rotating wheel (3), a conveying belt (4) and a motor component (5), a plurality of rotating wheels (3) are movably installed on the inner side of the frame (1), the outer side of the rotating wheel (3) is installed with a conveying belt (4), two motor components (5) are symmetrically installed on one side of the frame (1), and the output end of the motor component (5) is connected to one end of the rotating wheel (3), and the motor component (5) is connected to the controller component (45) via an electrical signal.

3. According to claim 1, a magnetic ring surface defect intelligent identification system based on deep learning is characterized in that: The magnetic ring surface defect intelligent identification system based on deep learning also includes a flip module, which includes a right-angle rod four (34), an electric telescopic rod component two (35), a piston rod two (36), a lifting plate two (37), a motor component four (38), a frame three (39), an electric telescopic rod component three (40), a push rod (41), a guide ring (42), a clamping rod (43) and a rubber pad (44), wherein the right-angle rod four (34) is installed on the top of the frame one (1), the electric telescopic rod component two (35) is installed on the top of the right-angle rod four (34), and the electric telescopic rod component two (35) is connected to the controller component (45) by electrical signals, and the output end of the electric telescopic rod component two (35) is installed with the piston rod two (36) ), a lifting plate 2 (37) is installed at one end of the piston rod 2 (36), a motor component 4 (38) is installed on one side of the lifting plate 2 (37), and the motor component 4 (38) is connected to the controller component (45) by electrical signals, a frame body 3 (39) is installed at the output end of the motor component 4 (38), and electric telescopic rod components 3 (40) are symmetrically installed on both sides of the frame body 3 (39), and the electric telescopic rod components 3 (40) are connected to the controller component (45) by electrical signals, a push rod (41) is installed at the output end of the electric telescopic rod component 3 (40), a guide ring (42) is installed at one end of the push rod (41), a clamping rod (43) is installed on the outer side of the guide ring (42), and a rubber pad (44) is installed on the back of the clamping rod (43).

4. According to claim 1, a magnetic ring surface defect intelligent identification system based on deep learning is characterized in that: The magnetic ring surface defect intelligent identification system based on deep learning also includes a front cleaning module, which includes a hydraulic cylinder component one (17), a frame two (18), a brush roller (19) and a motor component three (20), wherein the two hydraulic cylinder components one (17) are symmetrically mounted on both sides of the frame one (1), and the hydraulic cylinder component one (17) is electrically connected to the controller component (45), the output end of the hydraulic cylinder component one (17) is mounted with the frame two (18), and the frame two (18) is located in front of the frame one (1), and the brush roller (19) is movably mounted through the inner side of the frame two (18), the motor component three (20) is mounted on one side of the frame two (18), and the output end of the motor component three (20) is connected to one end of the brush roller (19), and the motor component three (20) is electrically connected to the controller component (45).

5. According to claim 1, a magnetic ring surface defect intelligent identification system based on deep learning is characterized in that: The deep learning-based magnetic ring surface defect intelligent identification system further comprises a watering module, wherein the watering module comprises a watering pipe (21), a water accumulation frame (22), a second pipe body (23), a water pump component (24) and a hose (25), wherein the watering pipe (21) is mounted on the top of the second frame body (18), and the watering pipe (21) is located above the brush roller (19), and a plurality of watering ports (31) are provided through the bottom of the watering pipe (21), and the water accumulation frame (22) is mounted on the frame body. The bottom of the first (1), the front output end of the water accumulation frame (22) is equipped with a second pipe body (23), the front of the water accumulation frame (22) is equipped with a water pump component (24), and the input end of the water pump component (24) is connected to the output end of the second pipe body (23), the water pump component (24) is connected to the controller component (45) via an electrical signal, and the output end of the water pump component (24) is equipped with a hose (25), and the output end of the hose (25) is connected to the input end of the sprinkler pipe (21).

6. According to claim 5, a magnetic ring surface defect intelligent identification system based on deep learning is characterized in that: A plurality of support blocks (26) are symmetrically mounted on the inner wall of the water accumulation frame (22), and a filter plate (27) is arranged on the top of the support block (26).

7. The deep learning-based intelligent recognition system for magnetic ring surface defects according to claim 6 is characterized in that: A second hydraulic cylinder component (28) is symmetrically mounted on the back of the water accumulation frame (22), and the second hydraulic cylinder component (28) is electrically connected to the controller component (45). A piston rod (29) is mounted on the output end of the second hydraulic cylinder component (28), and a second brush (30) is mounted on one end of the piston rod (29), and the second brush (30) is located above the filter plate (27).

8. The magnetic ring surface defect intelligent identification system based on deep learning according to claim 1 is characterized by: The deep learning-based magnetic ring surface defect intelligent identification system further comprises a drying module, the drying module comprising a right-angle rod three (32) and a hot air blower component (33), the right-angle rod three (32) being mounted on the top of the frame one (1), a plurality of hot air blower components (33) being mounted on the bottom of the right-angle rod three (32), and the hot air blower components (33) being electrically signal-connected to the controller component (45).

9. The identification method of a magnetic ring surface defect intelligent identification system based on deep learning according to any one of claims 1 to 8, characterized in that: The recognition method of the magnetic ring surface defect intelligent recognition system based on deep learning is as follows: S1, placing the magnetic ring on the conveyor belt (4), and then driving the magnetic ring backwards to below the motor component 2 (9) through the conveyor belt (4); S2, the electric telescopic rod component 1 (7) in the front drives the brush 1 (13) to move downward so that the rotating rod (10) is inserted into the inner side of the magnetic ring, and then the motor component 2 (9) drives the brush 1 (13) to rotate, and the brush 1 (13) rotates to clean the dust on the upper surface of the magnetic ring; S3, the brush 1 (13) then moves upward, and the conveyor belt (4) drives the magnetic ring to move backward to below the camera (16) in front, and then the camera (16) takes a picture of the magnetic ring; S4, the conveyor belt (4) then moves backward with the magnetic ring to the bottom of the frame body three (39), and the frame body three (39) is driven downward by the electric telescopic rod component two (35), and then the clamping rod (43) moves inward to clamp the magnetic ring, and then the frame body three (39) moves upward, and then the frame body three (39) rotates 180 degrees to drive the magnetic ring to flip 180 degrees; S5. Then, the conveyor belt (4) continues to drive the magnetic ring to move backward, and the brush 1 (13) at the rear cleans the magnetic ring. Then, the camera (16) at the rear takes a picture of the magnetic ring, and the analysis software in the controller component (45) analyzes the surface defects of the magnetic ring.

10. The identification method of a magnetic ring surface defect intelligent identification system based on deep learning according to claim 9, characterized in that: The step S5 also includes the following steps: S51, the frame body (18) and the brush roller (19) move backwards, the brush roller (19) rotates to clean the surface of the conveyor belt (4), and at the same time, water in the watering pipe (21) is sprinkled on the brush roller (19) through the watering port (31) to moisten the brush roller (19), and the water on the brush roller (19) falls into the water accumulation frame (22), the filter plate (27) filters the water, and at the same time, the brush (30) cleans the filter plate (27).

Images