A vision-based automatic centering method and system for double-sided steel plate shearing

Abstract

This invention discloses a visual detection-based automatic centering method and system for double-sided steel plate shearing. The method includes: installing a combination of area array cameras above each magnetic head position in the double-sided shearing area to detect the roller conveyor area and obtain a rectangular image of the roller conveyor area; calibrating the cameras and, based on the calibration parameters and the information of the steel plate to be sheared, virtually drawing two shearing lines on the left and right sides of the roller conveyor area image; obtaining the pixel coordinates of the left and right contour positions of the steel plate using a preset steel plate foreground extraction model; and controlling the movement of each magnetic head to perform centering operations based on the relative positional relationship between the steel plate contour position pixel coordinates and the coordinates of the virtual shearing lines, until centering is complete. This invention can accurately provide the cutting allowance on both sides of the steel plate, eliminating shearing errors caused by subjective human judgment, and avoiding the use of laser marking equipment, thereby reducing usage and maintenance costs.

Description

Technical Field

[0001] This invention relates to the fields of metal processing and visual inspection technology, and in particular to an automatic centering method and system for double-sided sheared steel plates based on visual inspection. Background Technology

[0002] The centering process for double-sided sheared steel plates currently relies heavily on manual visual laser alignment. Two parallel laser lines emitted from the shears serve as reference lines, and a magnetic centering device is manually controlled to move the steel plate laterally until it aligns with the shears, completing the centering process. This process is entirely manual, physically demanding, and the manually observed allowances on both sides are subjective, lacking actual quantifiable dimensions and thus failing to provide objective data analysis.

[0003] While existing double-sided shear automatic control centering methods have adopted cameras as auxiliary judgment devices, they still rely on the parallel laser line emitted by the shears as a reference. The laser emitter has a short lifespan and suffers from certain width and movement errors. Furthermore, in real-world scenarios, the light emitted by the laser line is almost mirror-reflected on the plate surface, making it difficult to capture in images under fixed viewing angles and natural light conditions. The camera cannot simultaneously analyze the distance between the laser line position and the steel plate contour. Summary of the Invention

[0004] This invention provides a visual detection-based automatic centering method and system for double-sided sheared steel plates, to solve the technical problems of large errors and high costs associated with existing automatic control centering methods for double-sided shears.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0006] On one hand, the present invention provides a visual detection-based automatic centering method for double-sided steel plate shearing, the visual detection-based automatic centering method for double-sided steel plate shearing includes:

[0007] A combination of area scan cameras is installed above each magnetic head position in the double-sided shearing area. The roller conveyor area is detected by the installed area scan camera combination to obtain a rectangular image of the roller conveyor area containing the steel plate to be sheared.

[0008] The cameras in the array camera assembly are calibrated to obtain calibration parameter information. Combining the calibration parameter information with the information of the steel plate to be sheared, two virtual shearing lines are drawn in the image of the roller conveyor area.

[0009] Using a preset steel plate foreground extraction model, the outline of the steel plate in the roller conveyor area image is extracted to obtain the pixel coordinates of the left and right outline positions of the steel plate.

[0010] By utilizing the relative positional relationship between the pixel coordinates of the left and right contours of the steel plate and the coordinates of the two virtual left and right shearing lines, the magnetic heads are controlled to move the steel plate to perform centering operations until centering is completed.

[0011] Furthermore, the area array camera combination includes two or more cameras, each camera is responsible for detecting a portion of the roller conveyor area, and adjacent cameras have 4%-6% overlap in the width direction of the roller conveyor. The camera combination's shooting range can cover the entire width area of ​​the roller conveyor.

[0012] The image captured by the camera is transformed through perspective to obtain a rectangular image of the roller conveyor area.

[0013] Furthermore, when calibrating the cameras in the array camera assembly, the plane of the roller guard plate is used as the calibration reference plane;

[0014] The calibrated parameter information includes:

[0015] Camera mounting height h:

[0016]

[0017] Where f is the focal length of the camera lens, w guard The actual range that the camera can illuminate in the width direction of the roller conveyor is below the plane of the roller conveyor guard plate, and u is the target surface size of the camera in the width direction of the roller conveyor;

[0018] Pixel resolution γ on the calibration reference plane:

[0019]

[0020] Where n is the number of pixels in the width direction of the camera;

[0021] The information about the steel plate to be sheared includes: the actual thickness of the steel plate and the target shearing width;

[0022] Two virtual shearing lines are drawn on the left and right sides in the image of the roller conveyor area, including:

[0023] Calculate the pixel coordinate position I of the virtual reference shear line x_base :

[0024]

[0025] Where d is the actual distance from the fixed shears to the right edge of the roller conveyor in the forward direction. offset This represents the offset of the shear line on the projection. d l h is the horizontal distance from the camera center to the fixed shear line. low t represents the height difference between the horizontal plane of the roller conveyor and the horizontal plane of the roller conveyor guard plate, and t represents the actual thickness of the steel plate.

[0026] Calculate the pixel coordinate position I of the virtual moving side shear line. x_move :

[0027]

[0028] Among them, w target γ represents the target width for shearing the steel plate to be sheared. steel This refers to the camera resolution on the plane of the steel plate when the plate is the measurement plane.

[0029] Furthermore, the steel plate foreground extraction model employs a cascaded semantic segmentation model. The first-stage semantic segmentation model takes the original image captured by the camera as input and outputs a rough outline of the steel plate boundary. The second-stage semantic segmentation model takes the left and right outlines of the steel plate output by the first-stage model as the center, and crops a combination of Regions of Interest (ROIs) along the outline direction on the original image captured by the camera at a specified image size. The output of the second-stage semantic segmentation model overwrites the output of the first-stage model according to the cropping position, ultimately yielding the steel plate outline extraction result.

[0030] Furthermore, the step of controlling the movement of each magnetic head to perform centering operations by utilizing the relative positional relationship between the pixel coordinates of the left and right contour positions of the steel plate and the coordinates of the virtual left and right shearing lines, until centering is completed, includes:

[0031] The steel plate outline is identified, and the position of the magnetic head is initialized, where the initial position coordinates of magnetic head i along the width direction of the roller conveyor are... for:

[0032]

[0033] in, Let be the pixel coordinates of the center of magnetic head i along the roller conveyor direction in the image. for The coordinates of the right outline of the steel plate at the location in the image width direction. for The coordinates of the left outline of the steel plate at the location in the image width direction;

[0034] Identify the outline of the steel plate and perform coarse adjustment of its position. Then, move the position of the i-th magnetic head by setting the minimum shearing amount on the fixed shear side. The amount of movement of the i-th magnetic head at this point... for:

[0035]

[0036] Among them, w min_offset The minimum allowable shearing margin set for the site;

[0037] The outline of the steel plate is identified, and its position is fine-tuned. During this fine-tuning, the position of the i-th read / write head is moved based on the balanced allowance on both sides. The amount of movement of the i-th read / write head is... for:

[0038]

[0039] After the steel plate position is finely adjusted, it is determined whether the centering is complete. The conditions for determining whether the centering is complete include: the shearing allowance on both sides of the steel plate is greater than the set minimum shearing allowance; and the difference between the shearing allowances on both sides is less than the preset value. When both conditions are met, the centering is considered complete. If either condition is not met, the centering operation continues until the centering is complete. If the centering still cannot be completed after exceeding the set number of attempts, an alarm will be triggered.

[0040] On the other hand, the present invention also provides a vision-based automatic centering system for double-sided steel plate shearing, the vision-based automatic centering system for double-sided steel plate shearing includes:

[0041] Multiple sets of area scan cameras are installed above each magnetic head position in the double-sided shearing area to detect the roller conveyor area and obtain a rectangular image of the roller conveyor area containing the steel plate to be sheared; wherein, a set of area scan cameras is installed above each magnetic head position.

[0042] The data processing module is used for:

[0043] The cameras in the array camera assembly are calibrated to obtain calibration parameter information. Combining the calibration parameter information with the information of the steel plate to be sheared, two virtual shearing lines are drawn in the image of the roller conveyor area.

[0044] Using a preset steel plate foreground extraction model, the outline of the steel plate in the roller conveyor area image is extracted to obtain the pixel coordinates of the left and right outline positions of the steel plate.

[0045] The centering operation module is used to control each magnetic head to move the steel plate to perform centering operations by using the relative positional relationship between the pixel coordinates of the left and right contour positions of the steel plate obtained by the data processing module and the coordinates of the two virtual left and right shearing lines, until centering is completed.

[0046] Furthermore, the area array camera combination includes two or more cameras, each camera is responsible for detecting a portion of the roller conveyor area, and adjacent cameras have 4%-6% overlap in the width direction of the roller conveyor. The camera combination's shooting range can cover the entire width area of ​​the roller conveyor.

[0047] The image captured by the camera is transformed through perspective to obtain a rectangular image of the roller conveyor area.

[0048] Furthermore, when the data processing module calibrates the cameras in the array camera assembly, it uses the roller guard plate plane as the calibration reference plane.

[0049] The calibrated parameter information includes:

[0050] Camera mounting height h:

[0051]

[0052] Where f is the focal length of the camera lens, w guard The actual range that the camera can illuminate in the width direction of the roller conveyor is below the plane of the roller conveyor guard plate, and u is the target surface size of the camera in the width direction of the roller conveyor;

[0053] Pixel resolution γ on the calibration reference plane:

[0054]

[0055] Where n is the number of pixels in the width direction of the camera;

[0056] The information about the steel plate to be sheared includes: the actual thickness of the steel plate and the target shearing width;

[0057] The data processing module virtually generates two shear lines, left and right, in the image of the roller conveyor area, including:

[0058] Calculate the pixel coordinate position I of the virtual reference shear line x_base :

[0059]

[0060] Where d is the actual distance from the fixed shears to the right edge of the roller conveyor in the forward direction. offset This represents the offset of the shear line on the projection. d l h is the horizontal distance from the camera center to the fixed shear line. low t represents the height difference between the horizontal plane of the roller conveyor and the horizontal plane of the roller conveyor guard plate, and t represents the actual thickness of the steel plate.

[0061] Calculate the pixel coordinate position I of the virtual moving side shear line. x_move :

[0062]

[0063] Among them, w target γ represents the target width for shearing the steel plate to be sheared. steel This refers to the camera resolution on the plane of the steel plate when the plate is the measurement plane.

[0064] Furthermore, the data processing module employs a cascaded semantic segmentation model for steel plate foreground extraction. The first-stage semantic segmentation model takes the original image captured by the camera as input and outputs a rough outline of the steel plate boundary. The second-stage semantic segmentation model takes the left and right outlines of the steel plate output by the first-stage model as the center, and crops a combination of Regions of Interest (ROIs) along the outlines at a specified image size on the original image captured by the camera. The output of the second-stage semantic segmentation model overwrites the output of the first-stage model according to the cropping position, ultimately yielding the steel plate outline extraction result.

[0065] Furthermore, the centering module is specifically used for:

[0066] The steel plate outline is identified, and the position of the magnetic head is initialized, where the initial position coordinates of magnetic head i along the width direction of the roller conveyor are... for:

[0067]

[0068] in, Let be the pixel coordinates of the center of magnetic head i along the roller conveyor direction in the image. for The coordinates of the right outline of the steel plate at the location in the image width direction. for The coordinates of the left outline of the steel plate at the location in the image width direction;

[0069] Identify the outline of the steel plate and perform coarse adjustment of its position. Then, move the position of the i-th magnetic head by setting the minimum shearing amount on the fixed shear side. The amount of movement of the i-th magnetic head at this point... for:

[0070]

[0071] Among them, w min_offset The minimum allowable shearing margin set for the site;

[0072] The outline of the steel plate is identified, and its position is fine-tuned. During this fine-tuning, the position of the i-th read / write head is moved based on the balanced allowance on both sides. The amount of movement of the i-th read / write head is... for:

[0073]

[0074] After the steel plate position is finely adjusted, it is determined whether the centering is complete. The conditions for determining whether the centering is complete include: the shearing allowance on both sides of the steel plate is greater than the set minimum shearing allowance; and the difference between the shearing allowances on both sides is less than the preset value. When both conditions are met, the centering is considered complete. If either condition is not met, the centering operation continues until the centering is complete. If the centering still cannot be completed after exceeding the set number of attempts, an alarm will be triggered.

[0075] In another aspect, the present invention also provides an electronic device comprising a processor and a memory; wherein the memory stores at least one instruction, which is loaded and executed by the processor to implement the above-described method.

[0076] In another aspect, the present invention also provides a computer-readable storage medium storing at least one instruction that is loaded and executed by a processor to implement the above-described method.

[0077] The beneficial effects of the technical solution provided by this invention include at least the following:

[0078] This invention provides a vision-based automatic centering scheme for double-sided steel plate shearing. It utilizes a camera to pinpoint the virtual shearing line position and identify the steel plate contour data. Through numerical comparison, the required movement distance of each magnetic head is determined, ultimately achieving automatic centering of the steel plate. This vision-based automatic centering scheme offers high centering accuracy, overcoming the subjectivity inherent in manual automatic centering processes. It effectively controls the shearing allowance of the steel plate, reducing material waste. Furthermore, this scheme eliminates the need for actual laser line auxiliary equipment, making installation and use more convenient and reducing equipment usage and maintenance costs. Attached Figure Description

[0079] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0080] Figure 1 This is a schematic diagram of the execution flow of the automatic centering method for double-sided steel plate shearing based on visual detection provided in an embodiment of the present invention;

[0081] Figure 2 This is a diagram showing the installation location of the double-sided shearing area device provided in an embodiment of the present invention;

[0082] Figure 3 This is a schematic diagram of calibration data parameters provided in an embodiment of the present invention;

[0083] Figure 4This is a schematic diagram of the virtual shear line offset effect provided in an embodiment of the present invention. Detailed Implementation

[0084] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0085] First Embodiment

[0086] This embodiment provides a vision-based automatic centering method for double-sided sheared steel plates. This method detects the contour position of the steel plate using machine vision and uses a calibrated virtual shearing line as a reference to control the magnetic head for automatic centering of the steel plate. This method can be implemented by an electronic device, such as a terminal or a server. The execution flow of this method is as follows: Figure 1 As shown, it includes the following steps:

[0087] S1, Install area array camera combination above each magnetic head position in the double-sided shearing area, and use area array camera combination to detect the roller conveyor area to obtain a rectangular roller conveyor area image containing the steel plate to be sheared;

[0088] The area array camera assembly installed in the centering area contains at least two cameras, each responsible for detecting a portion of the roller conveyor area. Adjacent cameras have 4%-6% overlap in the width direction of the roller conveyor, and the camera assembly's shooting range can cover the entire width of the roller conveyor.

[0089] Specifically, such as Figure 2 As shown, in this embodiment, an area array camera assembly is installed above the four magnetic head positions in the double-sided shearing area. The area array camera assembly contains two cameras, each with a pixel size of 5400*3600 and a 16mm lens. Each camera is responsible for detecting half of the roller conveyor area, that is, the detection of the 2.5m wide roller conveyor area. The two cameras overlap by 5% in the width direction of the roller conveyor.

[0090] The image captured by the camera is transformed through perspective to obtain a rectangular image of the roller conveyor area.

[0091] S2, calibrate the cameras in the array camera combination to obtain the calibration parameter information, and combine the calibration parameter information with the information of the steel plate to be sheared to virtually draw two shearing lines on the left and right sides in the image of the roller conveyor area.

[0092] Among them, such as Figure 3 As shown, the calibration reference surface selected during the camera calibration process is the roller conveyor guard plate plane;

[0093] The calibrated parameter information includes:

[0094] Camera mounting height h:

[0095]

[0096] Where f is the focal length of the camera lens, w guard The actual range that the camera can illuminate in the width direction of the roller conveyor is below the plane of the roller conveyor guard plate, and u is the target surface size of the camera in the width direction of the roller conveyor;

[0097] Pixel resolution γ on the calibration reference plane:

[0098]

[0099] Where n is the number of pixels in the width direction of the camera;

[0100] The information about the steel plate to be sheared includes: the actual thickness of the steel plate and the target shearing width;

[0101] The measurement plane is a steel plate, and the camera resolution γ on the steel plate plane is... steel for:

[0102]

[0103] Among them, h low The height difference between the horizontal plane of the roller conveyor and the horizontal plane of the roller conveyor guard plate is t, the actual thickness of the steel plate is h, and the vertical height of the camera from the horizontal plane of the guard plate is h, which is also the installation height of the camera.

[0104] like Figure 4 As shown, the shear position on the fixed blade side is a specified distance from the right side of the roller conveyor in the forward direction. When a steel plate is present, due to the influence of thickness, the pixel position marked by the shear line in the image will shift laterally during camera imaging. The specific formula for calculating the offset is as follows:

[0105]

[0106] Where, d l The horizontal distance from the camera center to the fixed shear line;

[0107] Based on the above, the pixel coordinate position I of the virtual reference shear line x_base The calculation formula is:

[0108]

[0109] Where d is the actual distance from the fixed shears to the right edge of the roller conveyor in the forward direction. offset γ is the offset of the shear line on the projection, and γ is the camera resolution in the reference plane;

[0110] Pixel coordinate position I of the virtual moving side shear line x_move The calculation formula is:

[0111]

[0112] Among them, w target The target width for shearing the steel plate to be sheared;

[0113] S3. Using a preset steel plate foreground extraction model, the outline of the steel plate in the roller conveyor area image is extracted to obtain the pixel coordinates of the left and right outline positions of the steel plate.

[0114] The steel plate foreground extraction model employs a cascaded semantic segmentation model. The first-stage semantic segmentation model uses a BiSeNet network. The network input is an original image of 5400*3600 pixels resized to 512*512 pixels, and the output image is also resized to 5400*3600 pixels to obtain a rough outline of the steel plate boundary. The second-stage semantic segmentation model uses the left and right outlines of the steel plate output from the first-stage model as the center. Regions of interest (ROIs) are cropped along the outlines in the original image captured by the camera, using a 512*512 image size. These cropped ROIs are then input into another BiSeNet network. The output of this combined image overwrites the output of the first-stage semantic segmentation model according to the cropped position, ultimately yielding the steel plate outline extraction result.

[0115] S4 uses the relative positional relationship between the pixel coordinates of the left and right contours of the steel plate and the coordinates of the two virtual left and right shearing lines to control each magnetic head to move the steel plate to perform centering operations until centering is completed.

[0116] The process of aligning the steel plates includes:

[0117] S41, Identify the steel plate outline and initialize the position of the magnetic head, where the initial position coordinates of magnetic head i along the width direction of the roller conveyor are... for:

[0118]

[0119] in, Let be the pixel coordinates of the center of magnetic head i along the roller conveyor direction in the image. for The coordinates of the right outline of the steel plate at the location in the image width direction. for The coordinates of the left outline of the steel plate at the location in the image width direction;

[0120] S42, identify the steel plate outline, perform coarse adjustment of the steel plate position, and move the position of the i-th magnetic head by setting the minimum shearing amount on the fixed shear side. At this time, the movement amount of the i-th magnetic head is... for:

[0121]

[0122] Among them, w min_offset The minimum allowable shearing margin set for the site;

[0123] S43, identify the outline of the steel plate and perform fine-tuning of the steel plate position. During fine-tuning, the position of the i-th read / write head is moved primarily based on balancing the allowance on the left and right sides. The amount of movement of the i-th read / write head at this time... for:

[0124]

[0125] S44, after completing the fine adjustment of the steel plate position, determine whether the alignment is complete;

[0126] The conditions for determining whether the alignment is complete include: 1) the shearing allowance on both sides of the steel plate is greater than the set minimum shearing allowance (which can be set to 35mm depending on the specific site conditions); 2) the difference between the shearing allowances on both sides is less than a certain set value (which can be set to 5mm depending on the specific site conditions).

[0127] When both of the above conditions are met, the centering is complete. If either condition is not met, the centering operation continues until it is completed. If the centering fails to be completed after the set number of attempts, an alarm will be triggered.

[0128] In summary, this embodiment provides a vision-based automatic centering method for double-sided steel plate shearing. This method involves adding area array cameras to each magnetic head position within the steel plate centering area to capture images of the steel plate and obtain its contour position. Based on parameter information obtained from on-site calibration and received steel plate shearing information, the pixel positions of two virtual shearing lines are generated in the image. By comparing the steel plate contour with the virtual shearing lines, the movement distance of each centering magnetic head is obtained, and the magnetic head initialization, coarse adjustment, and fine adjustment processes are performed, ultimately achieving alignment between the steel plate and the shears. This method can accurately provide the cutting allowance on both sides of the steel plate, eliminating shearing errors caused by subjective human judgment, and avoiding the use of laser marking equipment, thereby reducing usage and maintenance costs and realizing automatic control centering of double-sided steel plate shearing.

[0129] Second Embodiment

[0130] This embodiment provides a vision-based automatic centering system for double-sided steel plate shearing, including:

[0131] Multiple sets of area scan cameras are installed above each magnetic head position in the double-sided shearing area to detect the roller conveyor area and obtain a rectangular image of the roller conveyor area containing the steel plate to be sheared; wherein, a set of area scan cameras is installed above each magnetic head position.

[0132] The data processing module is used for:

[0133] The cameras in the array camera assembly are calibrated to obtain calibration parameter information. Combining the calibration parameter information with the information of the steel plate to be sheared, two virtual shearing lines are drawn in the image of the roller conveyor area.

[0134] Using a preset steel plate foreground extraction model, the outline of the steel plate in the roller conveyor area image is extracted to obtain the pixel coordinates of the left and right outline positions of the steel plate.

[0135] The centering operation module is used to control each magnetic head to move the steel plate to perform centering operations by using the relative positional relationship between the pixel coordinates of the left and right contour positions of the steel plate obtained by the data processing module and the coordinates of the two virtual left and right shearing lines, until centering is completed.

[0136] The vision-based automatic centering system for double-sided steel shearing plates in this embodiment corresponds to the vision-based automatic centering method for double-sided steel shearing plates in the first embodiment described above. The functions implemented by each module in the vision-based automatic centering system for double-sided steel shearing plates in this embodiment correspond one-to-one with the process steps in the vision-based automatic centering method for double-sided steel shearing plates described above; therefore, they will not be repeated here.

[0137] Third Embodiment

[0138] This embodiment provides an electronic device, which includes a processor and a memory; wherein the memory stores at least one instruction, which is loaded and executed by the processor to implement the method of the first embodiment.

[0139] The electronic device can vary considerably depending on its configuration or performance, and may include one or more processors (central processing units, CPUs) and one or more memories, wherein the memories store at least one instruction that is loaded by the processor and executed in accordance with the above method.

[0140] Fourth embodiment

[0141] This embodiment provides a computer-readable storage medium storing at least one instruction, which is loaded and executed by a processor to implement the method of the first embodiment described above. The computer-readable storage medium may be a ROM, random access memory, CD-ROM, magnetic tape, floppy disk, or optical data storage device, etc. The instruction stored therein can be loaded and executed by a processor in a terminal.

[0142] Furthermore, it should be noted that the present invention can be provided as a method, apparatus, or computer program product. Therefore, embodiments of the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, embodiments of the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code.

[0143] Embodiments of the present invention are described with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0144] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing terminal equipment to cause a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0145] It should also be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0146] Finally, it should be noted that the above description represents a preferred embodiment of the present invention. It should be pointed out that although preferred embodiments have been described, those skilled in the art, once they understand the basic inventive concept of the present invention, can make various improvements and modifications without departing from the principles described herein. These improvements and modifications should also be considered within the scope of protection of the present invention. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the embodiments of the present invention.

Claims

1. A method for automatic centering of double-sided sheared steel plates based on vision detection, characterized in that, include: A combination of area scan cameras is installed above each magnetic head position in the double-sided shearing area. The area scan camera combination detects the roller conveyor area and obtains a rectangular image of the roller conveyor area containing the steel plate to be sheared. The area scan camera combination contains at least two cameras, each responsible for detecting a portion of the roller conveyor area. Adjacent cameras have 4%-6% overlap in the width direction of the roller conveyor, and the camera combination's shooting range covers the entire width of the roller conveyor area. The images captured by the cameras are then transformed by perspective to obtain a rectangular image of the roller conveyor area. Using the roller conveyor guard plate plane as the calibration reference plane, the cameras in the array camera assembly are calibrated to obtain calibration parameter information. Combining the calibration parameter information with the information of the steel plate to be sheared, two virtual shearing lines are created in the image of the roller conveyor area. The calibration parameter information includes the camera's installation height and the pixel resolution on the calibration reference plane; the information of the steel plate to be sheared includes the actual thickness of the steel plate and the target shearing width. Creating the two virtual shearing lines in the image of the roller conveyor area includes: calculating the pixel coordinate positions of the virtual reference shearing line; and calculating the pixel coordinate positions of the virtual moving side shearing line. Using a pre-defined steel plate foreground extraction model, the outline of the steel plate in the roller conveyor area image is extracted to obtain the pixel coordinates of the left and right outline positions of the steel plate. The steel plate foreground extraction model adopts a cascaded semantic segmentation model. The first-stage semantic segmentation model takes the original image captured by the camera as input and outputs a rough outline of the steel plate boundary. The second-stage semantic segmentation model takes the left and right outlines of the steel plate output by the first-stage semantic segmentation model as the center and crops the regions of interest (ROIs) along the outline direction of the original image captured by the camera at a specified image size. The output of the second-stage semantic segmentation model covers the output of the first-stage semantic segmentation model according to the cropping position, and finally obtains the steel plate outline extraction result. The system utilizes the relative positional relationship between the pixel coordinates of the left and right contours of the steel plate and the coordinates of the two virtual left and right shearing lines to control the movement of each magnetic head to perform a centering operation until centering is complete. This includes: identifying the steel plate contour and initializing the magnetic head position; identifying the steel plate contour and coarsely adjusting the steel plate position to move the position of magnetic head i while fixing the minimum shearing amount on the shearing side; identifying the steel plate contour and finely adjusting the steel plate position, where the movement of magnetic head i is based on the balanced allowance on the left and right sides; after completing the fine adjustment, determining whether centering is complete; the conditions for determining whether centering is complete include: both shearing allowances on both sides of the steel plate are greater than the set minimum shearing allowance; and the difference between the shearing allowances on both sides is less than a preset value; when both conditions are met simultaneously, centering is complete; if either condition is not met, the centering operation continues until centering is complete; if centering fails after exceeding a set number of attempts, an alarm is triggered.

2. The automatic centering method for double-sided sheared steel plates based on vision detection as described in claim 1, characterized in that, Camera mounting height The expression is: ； in, For camera lens focal length, This represents the actual range that the camera can illuminate along the width of the roller conveyor, located below the plane of the roller conveyor guard plate. The target surface size of the camera is in the width direction of the roller conveyor; Pixel resolution on the calibration reference plane The expression is: ； in, This represents the number of pixels along the width of the camera. Pixel coordinates of the virtual baseline shear line The calculation formula is: ； in, This represents the actual distance from the fixed scissors to the right edge of the roller conveyor in the forward direction. This represents the offset of the shear line on the projection. , The horizontal distance from the camera center to the fixed shear line. This refers to the height difference between the horizontal plane of the roller conveyor and the horizontal plane of the roller conveyor guard plate. This refers to the actual thickness of the steel plate. Pixel coordinates of the virtual moving side shear line The calculation formula is: ； in, The target width for shearing the steel plate to be sheared. This refers to the camera resolution on the plane of the steel plate when the plate is the measurement plane. .

3. The automatic centering method for double-sided steel plate shearing based on vision detection as described in claim 2, characterized in that, When identifying the steel plate outline and initializing the position of the magnetic head, the initial position coordinates of magnetic head i along the width direction of the roller conveyor are... for: ； in, Let be the pixel coordinates of the center of magnetic head i along the roller conveyor direction in the image. for The coordinates of the right outline of the steel plate at the location in the image width direction. for The coordinates of the left outline of the steel plate at the location in the image width direction; When identifying the steel plate outline and coarsely adjusting the steel plate position, and moving the position of the i-th magnetic head by setting the minimum shearing amount on the fixed shear side, the amount of movement of the i-th magnetic head is... for: ； in, The minimum allowable shearing margin set for the site; When moving the position of read / write head i based on the left and right side margin balance, the amount of movement of read / write head i is... for: 。 4. A vision-based automatic centering system for double-sided steel plate shearing, characterized in that, include: Multiple sets of area scan cameras are mounted above each magnetic head position in the double-sided shearing area to detect the roller conveyor area and obtain a rectangular image of the roller conveyor area containing the steel plate to be sheared. Each magnetic head position is equipped with one set of area scan cameras. Each area scan camera set contains at least two cameras, each responsible for detecting a portion of the roller conveyor area. Adjacent cameras have 4%-6% overlap in the width direction of the roller conveyor, and the camera set's shooting range covers the entire width of the roller conveyor area. The images captured by the cameras are transformed by perspective to obtain a rectangular image of the roller conveyor area. The data processing module is used for: Using the roller conveyor guard plate plane as the calibration reference plane, the cameras in the array camera assembly are calibrated to obtain calibration parameter information. Combining this calibration parameter information with the information of the steel plate to be sheared, two virtual shearing lines are created in the roller conveyor area image. The calibration parameter information includes the camera installation height and pixel resolution on the calibration reference plane; the information of the steel plate to be sheared includes the actual thickness of the steel plate and the target shearing width. Creating the two virtual shearing lines in the roller conveyor area image includes: calculating the pixel coordinates of the virtual reference shearing line; and calculating the pixel coordinates of the virtual moving side shearing line. Using a pre-defined steel plate foreground extraction model, the outline of the steel plate in the roller conveyor area image is extracted to obtain the pixel coordinates of the left and right outline positions of the steel plate. The steel plate foreground extraction model adopts a cascaded semantic segmentation model. The input of the first-stage semantic segmentation model is the original image captured by the camera, and the output is a rough outline of the steel plate boundary. The input of the second-stage semantic segmentation model is a combination of regions of interest (ROIs) traversed and cropped along the outline direction of the original image captured by the camera, based on the left and right outlines of the steel plate output by the first-stage semantic segmentation model. The output of the second-stage semantic segmentation model covers the output of the first-stage semantic segmentation model according to the cropping position, and finally the steel plate outline extraction result is obtained. The centering operation module is used to control the movement of each magnetic head to perform centering operations by utilizing the relative positional relationship between the pixel coordinates of the left and right contour positions of the steel plate and the coordinates of the two virtual left and right shearing lines. This process continues until centering is complete. The module includes: identifying the steel plate contour and initializing the magnetic head position; identifying the steel plate contour and performing coarse adjustment of the steel plate position to move the position of magnetic head i while fixing the minimum shearing amount on the shearing side; identifying the steel plate contour and performing fine adjustment of the steel plate position, where fine adjustment is based on the balanced movement of the left and right side margins; after fine adjustment, determining whether centering is complete; the conditions for determining whether centering is complete include: both sides of the steel plate having shearing margins greater than the set minimum shearing margin; and the difference between the shearing margins on both sides being less than a preset value; when both conditions are met simultaneously, centering is complete. If either condition is not met, the centering operation continues until completion. If centering fails after a set number of attempts, an alarm is triggered.

5. The visual detection-based automatic centering system for double-sided steel plate shearing as described in claim 4, characterized in that, Camera mounting height The expression is: ； in, For camera lens focal length, This represents the actual range that the camera can illuminate along the width of the roller conveyor, located below the plane of the roller conveyor guard plate. The target surface size of the camera is in the width direction of the roller conveyor; Pixel resolution on the calibration reference plane The expression is: ； in, This represents the number of pixels along the width of the camera. Pixel coordinates of the virtual baseline shear line The calculation formula is: ； in, This represents the actual distance from the fixed scissors to the right edge of the roller conveyor in the forward direction. This represents the offset of the shear line on the projection. , The horizontal distance from the camera center to the fixed shear line. This refers to the height difference between the horizontal plane of the roller conveyor and the horizontal plane of the roller conveyor guard plate. This refers to the actual thickness of the steel plate. Pixel coordinates of the virtual moving side shear line The calculation formula is: ； in, The target width for shearing the steel plate to be sheared. This refers to the camera resolution on the plane of the steel plate when the plate is the measurement plane. .

6. The visual detection-based automatic centering system for double-sided steel plate shearing as described in claim 5, characterized in that, When identifying the steel plate outline and initializing the position of the magnetic head, the initial position coordinates of magnetic head i along the width direction of the roller conveyor are... for: ； in, Let be the pixel coordinates of the center of magnetic head i along the roller conveyor direction in the image. for The coordinates of the right outline of the steel plate at the location in the image width direction. for The coordinates of the left outline of the steel plate at the location in the image width direction; When identifying the steel plate outline and coarsely adjusting the steel plate position, and moving the position of the i-th magnetic head by setting the minimum shearing amount on the fixed shear side, the amount of movement of the i-th magnetic head is... for: ； in, The minimum allowable shearing margin set for the site; When moving the position of read / write head i based on the left and right side margin balance, the amount of movement of read / write head i is... for: 。

Images