Defect detecting and positioning device for building surface based on convolutional neural network

Abstract

The utility model discloses a defect detecting and positioning device for a building surface based on a convolutional neural network, and belongs to the technical field of positioning devices. The utility model discloses a defect detecting and positioning device for a building surface based on a convolutional neural network. According to the defect detecting and positioning device for the building surface based on the convolutional neural network, a user forwards screws an adjusting screw rod to drive a rotating plate to rotate upwards, the rotating plate drives a movable plate not to press a sliding sleeve any more, at the moment, the user can sequentially adjust the positions of four suction cups on an adjusting rod, and after the positions of the suction cups are adjusted, the suction cups are driven to move upwards; a user reversely screws the adjusting screw rod, the rotating plate rotates downwards to drive the rotating plate to simultaneously press the four sliding sleeves at one time, so that the sliding sleeves and the suction cup are positioned, compared with positioning of the suction cup through a fixing bolt and a fixing screw hole, the fixing bolt does not need to be repeatedly screwed, operation is more convenient, labor is saved, and the working efficiency is improved. And the suction cup adjusting efficiency can be improved.

Description

Technical Field

[0001] The utility model relates to the technical field of positioning devices, and more specifically, to a defect detection and positioning device for building surfaces based on a convolutional neural network. Background Art

[0002] Architectural curtain wall refers to the non-load-bearing exterior wall enclosure of a building, which is usually composed of a panel and a supporting structure at the back. Full-glass curtain wall is one of them. Full-glass curtain wall is a fully transparent, full-view glass curtain wall. It uses the transparency of glass to pursue the circulation and integration of the interior and exterior spaces of the building, so that people can clearly see the entire structural system of the glass through the glass, and the structural system is transformed from a simple supporting function to expressing its visibility, thereby showing the artistic sense, layering and three-dimensional sense of architectural decoration. It has the characteristics of light weight, simple material selection, factory processing, fast construction, convenient maintenance and repair, and easy cleaning. Its effect on enriching the facade effect of architectural shape is unmatched by other materials. It is the embodiment of modern technology in architectural decoration. When testing the full-glass curtain wall, it needs to be positioned first.

[0003] The Chinese patent discloses "a detection and positioning device for building curtain walls that is easy to adjust" with patent number CN202121658005.5. The scheme records "a detection and positioning device for building curtain walls that is easy to adjust, including a support mechanism, an adjustment mechanism and an adsorption mechanism, the support mechanism includes a frame, the left and right ends of the inner end of the frame are fixedly installed with slide grooves, the adjustment mechanism includes an adjusting screw, the outer ends of the front and rear ends of the adjusting screw are movably installed with movable rings, the rear end of the movable ring is fixedly installed with a rotating bearing, the rear end of the rotating bearing is fixedly installed with an adjusting nut, the left and right ends of the movable ring are fixedly installed with adjusting rods, and the upper end of the adjusting rod The adjusting rod is fixedly provided with adjusting screw holes, and movable wheels are fixedly installed at the left and right ends of the adjusting rod. "The detection and positioning device for building curtain walls that is easy to adjust has a multi-angle adjustment function, which makes it easier to adjust when positioning building curtain walls of different specifications and shapes, thereby improving convenience. At the same time, the movable wheels can make the movement of the adjusting rod smoother, further improving the activity performance. However, there are still certain problems. The device positions the suction cup after the position adjustment is completed through the fixing bolts and the fixing screw holes. Since there are four suction cups in total, the user needs to repeat the screwing operation of the fixing bolts four times, which is very troublesome and reduces the efficiency of adjusting the suction cup. Utility Model Content

[0004] The purpose of the utility model is to provide a building surface defect detection and positioning device based on a convolutional neural network to solve the problems raised in the above background technology:

[0005] The existing positioning device positions the suction cup after the position adjustment is completed by fixing bolts and fixing screw holes. Since there are four suction cups in total, the user needs to repeat the screwing operation of the fixing bolts four times, which is very troublesome and reduces the efficiency of adjusting the suction cup.

[0006] In order to achieve the above purpose, the utility model provides the following technical solutions:

[0007] A defect detection and positioning device for building surfaces based on convolutional neural networks comprises a frame, the internal rotation connection of the frame is provided with a bidirectional screw rod, the external sleeve of the bidirectional screw rod is provided with an adjusting rod, the adjusting rod is threadedly connected to the bidirectional screw rod, the adjusting rod is slidably connected to the frame, the external sleeve of the adjusting rod is provided with a sliding sleeve, the sliding sleeve can slide freely on the adjusting rod to achieve the purpose of adjusting the position of the suction cup, the sliding sleeve is slidably connected to the adjusting rod, the bottom of the sliding sleeve is fixedly connected to a connecting rod, the bottom of the connecting rod is fixedly connected to a suction cup, a first groove is provided inside the adjusting rod, and in each first groove, a group of rotating plates, a first sliding groove and a first sliding rod are respectively provided on both sides of the second sliding rod, and when the second sliding rod slides, it will drive the first sliding rod to squeeze The rotatable plate is connected to the rotation axis by pressing the inner wall of the first slide groove, thereby driving the rotating plate to rotate, and the rotating plate then drives the movable plate to move, and there are two first grooves in total, and the two first grooves are symmetrically distributed, and the rotating plate is rotatably connected inside the first groove, and a first slide groove is provided inside the rotating plate, and a first slide rod is slidably connected inside the first slide groove, and one end of the first slide rod is fixedly connected to the second slide rod, and the second slide rod passes through the adjusting rod and extends to the outside of the adjusting rod, and the second slide rod is slidably connected to the adjusting rod, and the movable plate is rotatably connected to the outside of the rotating plate, and a second groove is provided on the lower surface of the adjusting rod, and the second groove is communicated with the first groove, and the second groove is used in conjunction with the movable plate, so that the movable plate can be completely received in the second groove.

[0008] Preferably, a second slide groove is provided on one side of the frame, the second slide rod passes through the second slide groove and extends to the outside of the frame, the second slide rod is slidably connected to the frame, the second slide rod is internally slidably connected to a push-pull rod, the internal thread of the push-pull rod is connected to an adjusting screw, the adjusting screw is rotatably connected to the frame, and the bidirectional screw rod is screwed, the bidirectional screw rod will drive the two adjusting rods in the frame to move simultaneously to achieve the purpose of adjusting the width, when the adjusting rod moves, the second slide rod on the adjusting rod will slide on the push-pull rod, and when the user screws the adjusting screw, the adjusting screw will drive the second slide rod to move through the push-pull rod.

[0009] Preferably, a guide groove is provided inside the frame, and a guide rod is slidably connected inside the guide groove. The guide rod is fixedly connected to the push-pull rod, and two guide grooves are symmetrically distributed at both ends of the push-pull rod. The push-pull rod is limited by the guide groove and the guide rod. When the user screws the adjusting screw, the push-pull rod can be driven to move, so that the push-pull rod drives the second sliding rod to move.

[0010] Preferably, the cross-sections of the adjusting rod and the sliding sleeve are both rectangular structures, and the outer wall of the adjusting rod fits with the inner wall of the sliding sleeve to limit the sliding sleeve, so that the sliding sleeve can only slide back and forth on the adjusting rod without twisting and dislocation.

[0011] Preferably, a third groove is provided inside the sliding sleeve, and each sliding sleeve has two third grooves located on both sides of the adjusting rod. A ball is slidably connected inside the third groove, and a spring is provided between the ball and the third groove. Through the elastic force of the spring, the ball in the third groove is always pressed against the adjusting rod, which can increase the friction between the sliding sleeve and the adjusting rod and prevent the sliding sleeve from moving automatically on the adjusting rod.

[0012] Preferably, a scale line is fixedly connected to the top of the adjustment rod, and the provision of the scale line helps the user to adjust the position of the suction cup more accurately.

[0013] Compared with the prior art, the beneficial effects of the utility model are:

[0014] 1) When the defect detection and positioning device for building surfaces based on convolutional neural network is in use, the user rotates the adjusting screw in the forward direction, and the push-pull rod drives the second sliding rod to move toward the inside of the adjusting rod. The second sliding rod drives the rotating plate to rotate upward through the first sliding rod and the first sliding groove, and the rotating plate drives the movable plate to no longer press the sliding sleeve. At this time, the user can adjust the positions of the four suction cups on the adjusting rod in turn. After the positions of the suction cups are adjusted, the user rotates the adjusting screw in the reverse direction. At this time, the rotating plate rotates downward, and the rotating plate drives the rotating plate to press the four sliding sleeves at the same time, thereby positioning the sliding sleeve and the suction cup. Compared with positioning the suction cup by fixing bolts and fixing screw holes, there is no need to repeatedly screw the fixing bolts, and the operation is more convenient and labor-saving, which can improve the efficiency of adjusting the suction cup.

[0015] 2) When the present building surface defect detection and positioning device based on convolutional neural network is in use, the user can release the positioning of the sliding sleeve by the movable plate after turning the adjusting screw in the forward direction. At this time, the user can adjust the position of the suction cup. The ball in the third groove will always be pressed against the adjusting rod by the elastic force of the spring, thereby increasing the force required by the user to move the sliding sleeve, thereby preventing the sliding sleeve from sliding by itself, and avoiding the situation where the other adjusted suction cups move out of place by themselves when the user adjusts one of the suction cups. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic diagram of the overall structure of the utility model;

[0017] Figure 2 It is a cross-sectional schematic diagram of the adjustment rod of the utility model;

[0018] Figure 3 It is a structural schematic diagram of the rotating plate of the utility model;

[0019] Figure 4 It is a cross-sectional schematic diagram of the sliding sleeve of the utility model.

[0020] Explanation of the numbers in the figure: 1. Frame; 2. Bidirectional screw; 3. Adjusting rod; 4. Sliding sleeve; 5. Connecting rod; 6. Suction cup; 7. First groove; 8. Rotating plate; 9. First slide groove; 10. First slide rod; 11. Second slide rod; 12. Movable plate; 13. Second groove; 14. Push-pull rod; 15. Adjusting screw; 16. Second slide groove; 17. Guide groove; 18. Guide rod; 19. Third groove; 20. Ball; 21. Spring; 22. Scale line. DETAILED DESCRIPTION

[0021] Embodiment 1

[0022] The existing positioning device positions the suction cup 6 after the position adjustment is completed by fixing bolts and fixing screw holes. Since there are four suction cups 6 in total, the user needs to repeat the screwing operation of the fixing bolts four times, which is very troublesome and reduces the efficiency of adjusting the suction cup 6. The following solution is used to solve this problem.

[0023] See also Figures 1 to 4A defect detection and positioning device for building surface based on convolutional neural network includes a frame 1, the frame 1 is internally rotatably connected with a bidirectional screw rod 2, the bidirectional screw rod 2 is externally sleeved with an adjusting rod 3, the adjusting rod 3 is threadedly connected to the bidirectional screw rod 2, the adjusting rod 3 is slidably connected to the frame 1, the adjusting rod 3 is externally sleeved with a sliding sleeve 4, the sliding sleeve 4 can slide freely on the adjusting rod 3 to achieve the purpose of adjusting the position of the suction cup 6, the sliding sleeve 4 is slidably connected to the adjusting rod 3, the bottom of the sliding sleeve 4 is fixedly connected with a connecting rod 5, the bottom of the connecting rod 5 is fixedly connected with a suction cup 6, the inside of the adjusting rod 3 is provided with a first groove 7, each first groove 7 is located on both sides of the second sliding rod 11, and a group of rotating plates 8, first sliding grooves 9 and first sliding rods 10 are respectively arranged, and when the second sliding rod 11 slides, it will drive the first sliding rod 10 to squeeze The inner wall of the first slide groove 9 drives the rotating plate 8 to rotate, and the rotating plate 8 drives the movable plate 12 to move. There are two first grooves 7, and the two first grooves 7 are symmetrically distributed. The rotating plate 8 is rotatably connected inside the first groove 7, and the rotating plate 8 is provided with a first slide groove 9 inside. The first slide groove 9 is slidably connected inside the first slide groove 9, and one end of the first slide rod 10 is fixedly connected to the second slide rod 11, the second slide rod 11 penetrates the adjusting rod 3 and extends to the outside of the adjusting rod 3, the second slide rod 11 is slidably connected to the adjusting rod 3, and the outer part of the rotating plate 8 is rotatably connected to the movable plate 12, and the lower surface of the adjusting rod 3 is provided with a second groove 13, the second groove 13 is communicated with the first groove 7, and the second groove 13 is used in conjunction with the movable plate 12, and the movable plate 12 can be completely received in the second groove 13.

[0024] A second slide groove 16 is provided on one side of the frame 1, and the second slide rod 11 passes through the second slide groove 16 and extends to the outside of the frame 1. The second slide rod 11 is slidably connected to the frame 1, and the inside of the second slide rod 11 is slidably connected to the push-pull rod 14, and the internal thread of the push-pull rod 14 is connected to the adjusting screw 15, and the adjusting screw 15 is rotatably connected to the frame 1. When the bidirectional screw rod 2 is screwed, the bidirectional screw rod 2 will drive the two adjusting rods 3 in the frame 1 to move at the same time, so that the two adjusting rods 3 are close to or away from each other, so as to achieve the purpose of adjusting the width. When the adjusting rod 3 moves, the second slide rod 11 on the adjusting rod 3 will slide on the push-pull rod 14, and when the user screws the adjusting screw 15, the adjusting screw 15 will drive the second slide rod 11 to move through the push-pull rod 14.

[0025] A guide groove 17 is provided inside the frame 1, and a guide rod 18 is slidably connected inside the guide groove 17. The guide rod 18 is fixedly connected to the push-pull rod 14. Two guide grooves 17 are provided, which are symmetrically distributed at the two ends of the push-pull rod 14. The push-pull rod 14 is limited by the guide groove 17 and the guide rod 18. When the user screws the adjusting screw 15, the push-pull rod 14 can be driven to move, so that the push-pull rod 14 can drive the second sliding rod 11 to move.

[0026] The cross-sections of the adjusting rod 3 and the sliding sleeve 4 are both rectangular structures. The outer wall of the adjusting rod 3 fits with the inner wall of the sliding sleeve 4, which serves to limit the sliding sleeve 4 so that the sliding sleeve 4 can only slide back and forth on the adjusting rod 3 without twisting or dislocation, thereby ensuring that the direction of the suction cup 6 on the connecting rod 5 is fixed.

[0027] The use steps of the utility model are as follows: when using the defect detection and positioning device for building surfaces based on convolutional neural networks, the user screws the bidirectional screw rod 2, and the bidirectional screw rod 2 rotates and then drives the two adjusting rods 3 on the inner side of the frame 1 to move simultaneously, so that the two adjusting rods 3 are close to each other or away from each other, thereby adjusting the width. After the width is adjusted, the user screws the adjusting screw 15 on one side of the frame 1 forward, and at this time the adjusting screw 15 will drive the push-pull rod 14 to move in the direction close to the frame 1, and the push-pull rod 14 will drive the second sliding rod 11 to move toward the inside of the adjusting rod 3, and the second sliding rod 11 will then drive the first sliding rod 10 to squeeze the inner wall of the first sliding groove 9, so that the first sliding rod 10 drives the rotating plate 8 to rotate upward. The rotating plate 8 drives the movable plate 12 to no longer press the sliding sleeve 4. At this time, the user can adjust the positions of the four suction cups 6 on the adjusting rod 3 in turn. After the position of the suction cup 6 is adjusted, the user reverses the adjustment screw 15. At this time, the adjustment screw 15 drives the push-pull rod 14 to move away from the frame 1. The push-pull rod 14 drives the second slide bar 11 to move to the outside of the adjusting rod 3. The second slide bar 11 drives the rotating plate 8 to rotate downward through the first slide bar 10 and the first slide groove 9. The rotating plate 8 drives the movable plate 12 to press the four sliding sleeves 4 at the same time, so as to position the sliding sleeve 4 and prevent the suction cup 6 from moving. After positioning, the user can detect cracks on the curtain wall surface through the following steps:

[0028] Step 1: Collect and obtain multiple images of defective areas on the curtain wall surface as an image training set;

[0029] Step 2: Preprocess the image training set;

[0030] Step 3: Input the preprocessed image training set into the improved convolutional neural network model for training to obtain a trained convolutional neural network-based segmentation model;

[0031] Step 4: Input the image to be detected containing the defective area on the curtain wall surface into the trained segmentation model to obtain the segmentation result, which includes the defective area and other areas.

[0032] Step 5: Based on the segmentation results, calculate the position and size of the defect area in the image.

[0033] The above detection steps are known technologies and will not be described in detail here.

[0034] Embodiment 2

[0035] After the user turns the adjusting screw 15 forward, the movable plate 12 can be controlled to no longer press the sliding sleeve 4, thereby releasing the positioning of the sliding sleeve 4 by the movable plate 12. At this time, the user can directly adjust the position of the suction cup 6 through the sliding sleeve 4. In order to prevent the adjusted suction cup 6 from moving again, the following solution is used to solve this problem.

[0036] See also Figures 1 to 4 The difference between the combined implementation and the base is that a third groove 19 is provided inside the sliding sleeve 4, and each sliding sleeve 4 has two third grooves 19, which are located on both sides of the adjusting rod 3. A ball 20 is slidably connected inside the third groove 19, and a spring 21 is provided between the ball 20 and the third groove 19. Through the elastic force of the spring 21, the ball 20 in the third groove 19 is always pressed against the adjusting rod 3. The two balls 20 cooperate to clamp the sliding sleeve 4 on the adjusting rod 3 to prevent the sliding sleeve 4 from automatically moving on the adjusting rod 3.

[0037] A scale line 22 is fixedly connected to the top of the adjustment rod 3 , and the user can more accurately adjust the suction cup 6 by observing the position between the sliding sleeve 4 and the scale line 22 .

[0038] The use steps of the utility model are as follows: when the defect detection and positioning device for building surfaces based on convolutional neural network is in use, after the user positively rotates the adjusting screw 15, the positioning of the sliding sleeve 4 by the movable plate 12 can be released. At this time, the user can adjust the position of the suction cup 6, and a third groove 19 is provided inside the sliding sleeve 4. The ball 20 in the third groove 19 will always be pressed against the adjusting rod 3 by the elastic force of the spring 21, thereby increasing the force required by the user to move the sliding sleeve 4, thereby preventing the sliding sleeve 4 from sliding by itself, and avoiding the situation where the user adjusts one of the suction cups 6 and the other suction cups 6 that have been adjusted move out of place by themselves.

[0039] The above shows and describes the basic principle, main features and advantages of the utility model. Those skilled in the art should understand that the utility model is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the utility model and are not used to limit the utility model. Without departing from the spirit and scope of the utility model, the utility model may have various changes and improvements, which fall within the scope of the utility model to be protected. The scope of protection of the utility model is defined by the attached claims and their equivalents.

Claims

1. A defect detection and positioning device for a building surface based on a convolutional neural network, comprising a frame (1), wherein a bidirectional screw rod (2) is rotatably connected to the inside of the frame (1), an adjusting rod (3) is sleeved on the outside of the bidirectional screw rod (2), the adjusting rod (3) is threadedly connected to the bidirectional screw rod (2), and the adjusting rod (3) is slidably connected to the frame (1), characterized in that: The outer part of the adjusting rod (3) is sleeved with a sliding sleeve (4), the sliding sleeve (4) is slidably connected to the adjusting rod (3), the bottom of the sliding sleeve (4) is fixedly connected to a connecting rod (5), the bottom of the connecting rod (5) is fixedly connected to a suction cup (6), the inside of the adjusting rod (3) is provided with a first groove (7), there are two first grooves (7), the two first grooves (7) are symmetrically distributed, the inside of the first groove (7) is rotatably connected to a rotating plate (8), the inside of the rotating plate (8) is provided with a first slide groove (9), the first slide groove (9) ) is slidably connected to the inside of the adjusting rod (3), one end of the first sliding rod (10) is fixedly connected to the second sliding rod (11), the second sliding rod (11) passes through the adjusting rod (3) and extends to the outside of the adjusting rod (3), the second sliding rod (11) is slidably connected to the adjusting rod (3), the outside of the rotating plate (8) is rotatably connected to the movable plate (12), the lower surface of the adjusting rod (3) is provided with a second groove (13), the second groove (13) is communicated with the first groove (7), and the second groove (13) is used in conjunction with the movable plate (12).

2. The device for detecting and locating defects on building surfaces based on a convolutional neural network according to claim 1, characterized in that: A second slide groove (16) is provided on one side of the frame (1), the second slide rod (11) passes through the second slide groove (16) and extends to the outside of the frame (1), the second slide rod (11) is slidably connected to the frame (1), the interior of the second slide rod (11) is slidably connected to a push-pull rod (14), the interior of the push-pull rod (14) is threadedly connected to an adjusting screw (15), and the adjusting screw (15) is rotatably connected to the frame (1).

3. The building surface defect detection and positioning device based on convolutional neural network according to claim 1, characterized in that: A guide groove (17) is provided inside the frame (1), a guide rod (18) is slidably connected inside the guide groove (17), and the guide rod (18) is fixedly connected to the push-pull rod (14).

4. The building surface defect detection and positioning device based on convolutional neural network according to claim 1, characterized in that: The cross-sections of the adjusting rod (3) and the sliding sleeve (4) are both rectangular structures, and the outer side wall of the adjusting rod (3) fits the inner side wall of the sliding sleeve (4).

5. The building surface defect detection and positioning device based on convolutional neural network according to claim 1, characterized in that: A third groove (19) is provided inside the sliding sleeve (4), a ball (20) is slidably connected inside the third groove (19), and a spring (21) is provided between the ball (20) and the third groove (19).

6. The building surface defect detection and positioning device based on convolutional neural network according to claim 1, characterized in that: A scale line (22) is fixedly connected to the top of the adjusting rod (3).

Images