Splitting and splicing method for reinforcing mesh detection

Abstract

The invention relates to the technical field of reinforcing mesh detection, and discloses a splitting and splicing method for reinforcing mesh detection, which comprises the following steps: step 1, dividing reinforcing meshes; step 2, key intersection point detection; step 3, rebuilding a reinforcing mesh; according to reinforcement mesh division, a super-large reinforcement mesh is divided into a plurality of sub-areas so as to facilitate image acquisition, and image acquisition is performed on the sub-areas by adopting image acquisition equipment and an image acquisition method. Through the automatic shooting, detection and splicing processes, the detection time is greatly shortened, the working efficiency is improved, the labor cost and time cost of manual detection are reduced, the overall detection cost is reduced, the risk that a worker directly makes contact with a large reinforcing mesh is avoided through the automatic detection technology, the working safety is improved, and the detection efficiency is improved. And the detection result and the photo information are stored in the database, so that subsequent analysis and tracing are facilitated, and the credibility of quality detection is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of steel mesh detection, and more particularly to a method for splitting and splicing steel mesh detection. Background Art

[0002] A steel mesh is a building material woven from high-strength steel wires or aluminum wires, which has the advantages of being lightweight, high-strength, corrosion-resistant, anti-aging, etc. It is widely used in various construction projects, and can effectively enhance the stability and bearing capacity of concrete structures, and improve the overall quality and service life of buildings.

[0003] Traditionally, for the detection of ultra-large steel meshes, it often relies on manual visual inspection or simple mechanical measurement. These methods are not only time-consuming and laborious, but also difficult to ensure the accuracy and consistency of detection. With the development of computer vision technology, although there are already detection technologies based on image recognition applied to small steel meshes, for ultra-large components, due to their large size, it is difficult to cover the entire area in a single shot, and the splicing process is complex. Therefore, there is no mature technical solution and it needs to be improved. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides a method for splitting and splicing steel mesh detection to solve the problems existing in the above-mentioned background art.

[0005] The present invention provides the following technical solutions: A method for splitting and splicing steel mesh detection, comprising the following steps:

[0006] Step 1, steel mesh division;

[0007] Step 2, key intersection point detection;

[0008] Step 3, steel mesh reconstruction;

[0009] The steel mesh division is to divide the ultra-large steel mesh into multiple sub-regions for image acquisition, and image acquisition of the sub-regions is performed by using an image acquisition device and an image acquisition method;

[0010] The key intersection point detection uses an improved FS-DETR object detection algorithm as an object detector to identify and extract key intersection points in each sub-region image, so as to obtain information such as the number, diameter and spacing of steel bars in the steel mesh;

[0011] The reconstruction of the steel bar mesh is to splice all sub-region images together to generate the outer contour of an oversized steel bar mesh, thereby obtaining the total length and width of the steel bar mesh. Image splicing is the process of combining images with overlapping regions into an image with a wide field of view and high resolution. The main processes of image splicing include feature matching, registration, and seam removal. Feature matching is the key to image splicing. The ORB algorithm is used in feature matching to extract key feature points in the image and generate descriptors for these feature points. These feature points and descriptors can be used for image matching and target tracking tasks.

[0012] Furthermore, the image acquisition device includes: a gantry with a servo motor and a sliding guide rail, an industrial computer, two detection boxes, and a steel bar mesh welding machine. Each detection box contains an industrial camera and three photoelectric sensors. The industrial camera and photoelectric sensors are installed on the sliding guide rail of the gantry through the detection boxes and can move left and right to adapt to steel bar meshes of different lengths. A certain distance is maintained between the two detection boxes to ensure that the shooting range covers the width of the steel bar mesh. It is equipped with a motor drive system, and the moving speed and position of the industrial camera are controlled by programming to ensure the stability and synchronization of the industrial camera during the shooting process.

[0013] Furthermore, the visual field range of the sub-regions divided by the steel bar mesh is determined by the industrial camera parameters and the working distance. The calculation formula is as follows: The long side size of the chip = the resolution of the long side × the pixel size;

[0014] Furthermore, the workflow of image acquisition is as follows. First, the PLC of the steel bar mesh welding machine sends the device status and the design specifications of the steel bar mesh to the industrial computer. Second, when the servo motor on the gantry receives a movement instruction, it drives the detection box to move along the sliding guide rail. Then, when the photoelectric sensor detects the passage of an object, the industrial camera performs image acquisition and uploads it. Finally, when the production of the steel bar mesh is completed, the image acquisition task is completed. In this article, the industrial cameras inside the two detection boxes are used and recorded as No. 1 and No. 2 respectively. For the photos taken by each camera, the YOLO (You Only Look Once) algorithm is used for feature extraction to identify key information such as steel bar intersection points, diameters, and spacings. According to the preset quality standards, the steel bar meshes in each photo are individually detected to judge their qualification levels, and the detection results and photo information are stored in the database for subsequent analysis and traceability.

[0015] Further, the workflow of the image acquisition can be divided into three stages. In the first stage, driven by the servo motor, the two detection boxes move from the middle position of the gantry sliding rail to the outermost intersection points on both sides of the steel bar mesh. During the movement, the photoelectric sensor and the industrial camera cooperate with each other to complete the image acquisition in the first stage. Whenever the photoelectric sensor is blocked by the steel bars and a preset distance difference is generated, the industrial camera performs the image acquisition operation.

[0016] Further, in the second stage, when the first stage ends, the detection boxes move to the outermost intersection points on both sides of the steel bar mesh and remain fixed in their positions on the gantry. As the production of the steel bar mesh continues, the two detection boxes perform image acquisition on both sides of the steel bar mesh. Similar to the first stage, the photoelectric sensor and the industrial camera cooperate with each other to complete the image acquisition in the second stage.

[0017] Further, in the third stage, the moving direction of the detection is completely opposite to that in the first stage. Driven by the servo motor, the two detection boxes move from the outermost intersection points on both sides of the steel bar mesh to the middle position of the gantry sliding rail. Similar to the first stage, the photoelectric sensor and the industrial camera cooperate with each other to complete the image acquisition in the third stage.

[0018] Further, the main steps of the ORB algorithm are as follows: 1. Construct an image pyramid through downsampling, and regard each pixel on the image as a candidate point; 2. Establish a discrete circle centered on each candidate point, and calculate the difference between the pixel values on the discrete circle and the pixel value of the candidate point; 3. When there are multiple differences exceeding the predetermined threshold, identify the candidate point as a key point, indicating that it has a strong feature representation ability.

[0019] Further, the reconstruction of the steel bar mesh sheet is to use an image stitching algorithm. According to the shooting position and time information, multiple photos are stitched into a complete steel bar mesh sheet image, and the integrity of the stitched image is verified to check for any missing or misaligned parts to ensure that the entire steel bar mesh sheet has been comprehensively detected.

[0020] The technical effects and advantages of the present invention:

[0021] 1. Through the automated shooting, detection, and stitching processes, the present invention greatly shortens the detection time, improves the work efficiency, reduces the labor cost and time cost of manual detection, and lowers the overall detection cost.

[0022] 2. The present invention avoids the risk of direct manual contact with large steel bar mesh sheets through automated detection technology, improving work safety.

[0023] 3. By storing the detection results and photo information in the database, the present invention facilitates subsequent analysis and traceability, improving the credibility of quality detection. Brief Description of the Drawings

[0024] Figure 1 This is a schematic diagram of the overall process of the present invention.

[0025] Figure 2 This is a schematic diagram of the image acquisition device model of the present invention.

[0026] Figure 3 This is a flowchart of the operation of image acquisition of the present invention.

[0027] Figure 4 This is a schematic diagram of the first stage of image acquisition of the present invention.

[0028] Figure 5 This is a schematic diagram of the second stage of image acquisition of the present invention.

[0029] Figure 6 This is a schematic diagram of the third stage of image acquisition of the present invention.

[0030] Figure 7 This is a schematic diagram of the intersection points and key intersection points of the present invention.

[0031] Figure 8 This is a schematic diagram of the improved FS-DETR object detector structure of the present invention.

[0032] Figure 9 This is a schematic diagram of the discrete circle template and tic-tac-toe template of the present invention. Detailed implementation manners

[0033] Next, the technical solutions in the present invention will be clearly and completely described in conjunction with the accompanying drawings in the present invention. In addition, the forms of each structure described in the following embodiments are merely examples, and a method for detecting and splicing a steel bar mesh sheet according to the present invention is not limited to the structures described in the following embodiments. All other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

[0034] Embodiment 1:

[0035] As Figures 1-9 shown, a method for detecting and splicing a steel bar mesh sheet includes the following steps:

[0036] Step 1, division of the steel bar mesh sheet;

[0037] Step 2, detection of key intersection points;

[0038] Step 3, reconstruction of the steel bar mesh sheet;

[0039] The division of the steel bar mesh sheet is to divide an oversized steel bar mesh into multiple sub-regions for image acquisition, and the sub-regions are subjected to image acquisition by using an image acquisition device and an image acquisition method;

[0040] The image acquisition device includes a gantry with a servo motor and a sliding guide rail, an industrial computer, two detection boxes, and a steel bar mesh welding machine. Each detection box contains an industrial camera and three photoelectric sensors. The industrial camera and the photoelectric sensors are installed on the sliding guide rail of the gantry through the detection boxes and can move left and right to adapt to steel bar meshes of different lengths. A certain distance is maintained between the two detection boxes to ensure that the shooting range covers the width of the steel bar mesh. It is equipped with a motor drive system, and the movement speed and position of the industrial camera are controlled by programming to ensure the stability and synchronization of the industrial camera during the shooting process. The visual field range of the sub-regions divided by the steel bar mesh is determined by the industrial camera parameters and the working distance. The calculation formula is as follows: The long side dimension of the chip = the resolution of the long side × the pixel size; The working process of image acquisition is as follows. First, the PLC of the steel bar mesh welding machine sends the device status and the design specifications of the steel bar mesh to the industrial computer. Second, when the servo motor on the gantry receives a movement instruction, it drives the detection box to move along the sliding guide rail. Then, when the photoelectric sensor detects that an object has passed, the industrial camera performs image acquisition and uploads it. Finally, when the production of the steel bar mesh is completed, the image acquisition task is completed. In this article, the industrial cameras inside the two detection boxes are used and recorded as No. 1 and No. 2 respectively. For the photos taken by each camera, the YOLO (You Only Look Once) algorithm is used for feature extraction to identify key information such as steel bar intersection points, diameters, and spacings. According to the preset quality standards, the steel bar meshes in each photo are individually detected to judge their qualification levels, and the detection results and photo information are stored in the database for subsequent analysis and traceability. The working process of image acquisition can be divided into three stages. The first stage is that, driven by the servo motor, the two detection boxes move from the middle position of the sliding guide rail of the gantry to the outermost intersection points on both sides of the steel bar mesh. During the movement, the photoelectric sensor and the industrial camera cooperate with each other to complete the image acquisition in the first stage. Whenever the photoelectric sensor is blocked by a steel bar and a preset distance difference is generated, the industrial camera performs the image acquisition operation. The second stage is that when the first stage ends, the detection boxes move to the outermost intersection points on both sides of the steel bar mesh and remain fixed in their positions on the gantry. As the production of the steel bar mesh continues, the two detection boxes perform image acquisition on both sides of the steel bar mesh. The same as the first stage, the photoelectric sensor and the industrial camera cooperate with each other to complete the image acquisition in the second stage. The third stage is that the moving direction is completely opposite to that of the first stage. Driven by the servo motor, the two detection boxes move from the outermost intersection points on both sides of the steel bar mesh to the middle position of the sliding guide rail of the gantry. The same as the first stage, the photoelectric sensor and the industrial camera cooperate with each other to complete the image acquisition in the third stage.

[0042] Embodiment 2:

[0043] As Figures 1-9 shown, the critical intersection point detection uses an improved FS-DETR object detection algorithm as the object detector to identify and extract critical intersection points in each sub-region image, thereby obtaining information such as the number, diameter, and spacing of steel bars in the steel bar mesh; further, these critical intersection points are used for steel bar mesh reconstruction to generate the total length and width of the steel bar mesh. The intersection point refers to the rectangular overlapping area formed by the intersection of the horizontal and vertical steel bars in the steel bar mesh, and the critical intersection point refers to the intersection point located at the center of the sub-region image among all intersection points, as well as the intersection point closest to this intersection point and in the moving direction of the detection box.

[0044] Embodiment 3:

[0045] As Figures 1-9 shown, the steel bar mesh sheet reconstruction is to splice all sub-region images together to generate the outer contour of the super-large steel bar mesh, thereby obtaining the total length and width of the steel bar mesh. Image stitching is the process of combining images with overlapping areas into an image with a wide field of view and high resolution. The main processes of image stitching include feature matching, registration, and seam removal. Feature matching is the key to image stitching. The ORB algorithm is used to extract key feature points in the image and generate descriptors for these feature points. These feature points and descriptors can be used for image matching and object tracking tasks. The main steps of the ORB algorithm are as follows: 1. Build an image pyramid by downsampling, and regard each pixel on the image as a candidate point; 2. Establish a discrete circle centered on each candidate point, and calculate the difference between the pixel values on the discrete circle and the pixel value of the candidate point; 3. When there are multiple differences exceeding a predetermined threshold, the candidate point is identified as a key point, indicating that it has strong feature representation ability. The steel bar mesh sheet reconstruction uses the image stitching algorithm to stitch multiple photos into a complete steel bar mesh sheet image according to the shooting position and time information, and perform integrity verification on the stitched image to check for missing or misaligned parts to ensure that the entire steel bar mesh sheet has been comprehensively detected. However, in the steel bar mesh image in this application, since the steel bar intersection point is a rectangular overlapping area, most pixel values are almost indistinguishable from the pixel value of the candidate point using the traditional discrete circle template construction method. Therefore, this application designs a key point detection method based on a "tic-tac-toe" template in the ORB algorithm to facilitate the distinction between pixel values and the pixel value of the candidate point. As shown in Figure 9 shown in the appendix Figure 9 The left side is the discrete circle template, and the right side is the "tic-tac-toe" template.

[0046] In summary, as Figures 1-9As shown in the figure, for the splitting and splicing method of steel bar mesh detection, during use, a gantry with adjustable height is used to carry two detection boxes. Each detection box contains a high-resolution industrial camera and three photoelectric sensors. A certain distance is maintained between the two detection boxes to ensure that the shooting range covers the width of the steel bar mesh. The detection boxes are installed on the gantry through precision slide rails to achieve left and right movement to adapt to steel bar meshes of different lengths. An electric motor drive system is equipped, and the movement speed and position of the detection boxes are controlled through programming to ensure the stability and synchronization of the industrial camera during the shooting process. The industrial camera in the middle uses three stages to collect images. In the first stage, the PLC of the steel bar welding machine sends the equipment status and design dimensions of the steel bar mesh to be processed to the industrial computer. After receiving the signal, the industrial computer sends a movement instruction to the servo motor. Driven by the servo motor, the two detection boxes move from the middle position of the sliding guide of the gantry to the outermost intersection points on both sides of the steel bar mesh. During the movement, the photoelectric sensors and the industrial camera cooperate with each other to complete the image collection in the first stage. Whenever the photoelectric sensor is blocked by the steel bar and a preset distance difference is generated, the industrial camera performs the image collection operation. In the second stage, the detection boxes move to the outermost intersection points on both sides of the steel bar mesh and remain fixed in position on the gantry. As the production of the steel bar mesh continues, the two detection boxes perform image collection on both sides of the steel bar mesh. The movement direction in the third stage is completely opposite to that in the first stage. Driven by the servo motor, the two detection boxes move from the outermost intersection points on both sides of the steel bar mesh to the middle position of the sliding guide of the gantry. The second stage and the third stage are the same as the first stage, and the photoelectric sensors and the industrial camera cooperate with each other to complete the image collection. For the photos taken by each camera, the YOLO (You Only Look Once) algorithm is used for feature extraction to identify key information such as steel bar intersection points, diameters, and spacings. According to the preset quality standards, the steel bar meshes in each photo are individually detected to determine their qualification levels. The detection results and photo information are stored in the database for subsequent analysis and traceability. Using the image stitching algorithm, according to the shooting position and time information, after collecting the pictures, multiple photos are stitched into a complete steel bar mesh image, and the integrity of the stitched image is verified to check for any missing or misaligned parts to ensure that the entire steel bar mesh has been comprehensively detected.

[0047] Finally, several points should be noted: First, in the description of this application, it should be noted that unless otherwise specified and limited, the terms "installation", "connection", and "coupling" should be understood in a broad sense. It can be a mechanical connection or an electrical connection, or it can be the internal communication of two components. It can be directly connected. "Up", "down", "left", "right", etc. are only used to represent relative position relationships. When the absolute position of the object being described changes, the relative position relationship may change;

[0048] Secondly: In the attached drawings of the disclosed embodiments of the present invention, only the structures related to the disclosed embodiments are involved. For other structures, reference may be made to the general design. Without conflict, the same embodiment and different embodiments of the present invention can be combined with each other;

[0049] Finally: The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for splitting and splicing the detection of a steel bar mesh, characterized in that, It includes the following steps: Step 1, reinforcement mesh division; Step 2, key intersection detection; Step 3, reinforcement mesh reconstruction; The reinforcement mesh division is to divide the super-large reinforcement mesh into multiple sub-regions for image acquisition, and the image acquisition device and image acquisition method are used to acquire images of the sub-regions; The key intersection detection uses an improved FS-DETR object detection algorithm as the object detector to identify and extract key intersections in each sub-region image, so as to obtain information such as the number, diameter and spacing of steel bars in the reinforcement mesh; The reinforcement mesh reconstruction is to splice all sub-region images together to generate the outer contour of the super-large reinforcement mesh, so as to obtain the total length and width of the reinforcement mesh. Image stitching is the process of combining images with overlapping regions into an image with a wide field of view and high resolution. The main processes of image stitching include feature matching, registration and seam removal. Feature matching is the key to image stitching. The ORB algorithm is used to extract key feature points in the image and generate descriptors of these feature points. These feature points and descriptors can be used for image matching and object tracking tasks.

2. A method for splitting and splicing the detection of a steel bar mesh sheet according to claim 1, characterized in that: The image acquisition device includes: a gantry with a servo motor and a sliding guide rail, an industrial computer, two detection boxes, and a reinforcement mesh welding machine. Each detection box contains an industrial camera and three photoelectric sensors. The industrial camera and photoelectric sensors are installed on the sliding guide rail of the gantry through the detection boxes and can move left and right to adapt to different lengths of reinforcement meshes. A certain distance is maintained between the two detection boxes to ensure that the shooting range covers the width of the reinforcement mesh. It is equipped with a motor drive system, and the moving speed and position of the industrial camera are controlled by programming to ensure the stability and synchronization of the industrial camera during the shooting process.

3. A method for splitting and splicing the detection of a steel bar mesh sheet according to claim 1, characterized in that: The visual field range of the sub-regions divided by the steel bar mesh is determined by the industrial camera parameters and the working distance, and the calculation formula is as follows: the long side size of the chip = the resolution of the long side × the pixel size; 4. A method for splitting and splicing the detection of a steel bar mesh sheet according to claim 1, characterized in that: The working process of the image acquisition is as follows. First, the PLC of the reinforcement mesh welding machine sends the device status and the design specifications of the reinforcement mesh to the industrial computer. Second, when the servo motor on the gantry receives a movement instruction, it drives the detection box to move along the sliding guide rail. Then, when the photoelectric sensor detects the passage of an object, the industrial camera performs image acquisition and uploads it. Finally, when the production of the reinforcement mesh is completed, the image acquisition task is completed. In this article, the industrial cameras inside the two detection boxes are used and recorded as No. 1 and No. 2 respectively. For the photos taken by each camera, the YOLO (You Only Look Once) algorithm is used for feature extraction to identify key information such as steel bar intersections, diameters and spacings. According to the preset quality standards, each reinforcement mesh in each photo is detected separately to judge its qualification degree, and the detection results and photo information are stored in the database for subsequent analysis and traceability.

5. A method for splitting and splicing the detection of a steel bar mesh sheet according to claim 1, characterized in that: The workflow of the image acquisition can be divided into three stages. In the first stage, driven by the servo motor, the two detection boxes move from the middle position of the gantry sliding rail to the outermost intersection points on both sides of the steel mesh. During the movement, the photoelectric sensor and the industrial camera cooperate with each other to complete the image acquisition in the first stage. Whenever the photoelectric sensor is blocked by the steel bars and a preset distance difference is generated, the industrial camera performs the image acquisition operation.

6. A method for splitting and splicing the detection of a steel bar mesh sheet according to claim 5, characterized in that: In the second stage, when the first stage ends, the detection boxes move to the outermost intersection points on both sides of the steel mesh and remain fixed in their positions on the gantry. As the production of the steel mesh continues, the two detection boxes perform image acquisition on both sides of the steel mesh. Similar to the first stage, the photoelectric sensor and the industrial camera cooperate with each other to complete the image acquisition in the second stage.

7. A method for splitting and splicing the detection of a steel bar mesh sheet according to claim 5, characterized in that: In the third stage, the moving direction is exactly opposite to that in the first stage. Driven by the servo motor, the two detection boxes move from the outermost intersection points on both sides of the steel mesh to the middle position of the gantry sliding rail. Similar to the first stage, the photoelectric sensor and the industrial camera cooperate with each other to complete the image acquisition in the third stage.

8. A method for splitting and splicing the detection of a steel bar mesh sheet according to claim 1, characterized in that: The main steps of the ORB algorithm are as follows:

1. Construct an image pyramid through downsampling, and regard each pixel on the image as a candidate point; 2. Establish a discrete circle centered on each candidate point, and calculate the difference between the pixel values on the discrete circle and the pixel value of the candidate point; 3. When there are multiple differences exceeding a predetermined threshold, the candidate point is identified as a key point, indicating that it has a strong feature representation ability.

9. A method for splitting and splicing the detection of a steel bar mesh sheet according to claim 1, characterized in that: The reconstruction of the steel mesh sheet is to use the image stitching algorithm. According to the shooting position and time information, multiple photos are stitched into a complete steel mesh sheet image, and the integrity of the stitched image is verified to check whether there are missing or misaligned parts to ensure that the entire steel mesh sheet has been comprehensively detected.

Images