Sharpening method and device

Abstract

The application discloses a grinding method and device. If a surface defect exists on a target object in an initial detection area, a target moment when the surface defect is detected in a target area is determined based on a moving speed of the target object, and the target object is controlled to stop running at the target moment. A first real world image of the target area is acquired, the surface defect is identified from the first real world image, and a grinding coordinate of the surface defect in a preset grinding platform coordinate system is determined. The surface defect is ground according to the grinding coordinate. The application can realize coarse positioning of a surface defect, tracking positioning of the surface defect, and precise positioning of the surface defect in a strip steel production process, and finally realize accurate coordinates and efficient grinding of the surface defect of the strip steel.

Description

Technical Field

[0001] This invention relates to the field of surface defect determination technology, and in particular to a grinding method and apparatus. Background Technology

[0002] During strip steel production, surface defects are common: oxide scale, foreign matter indentation, edge cracks, or other defects. These defects may be stretched or deepened during rolling, leading to surface quality issues in the final product. To avoid these defects, grinding is necessary. However, manual grinding presents challenges such as difficulty in detecting defects, inaccurate defect location, limited production line speed, low production efficiency, and high time and labor costs. Summary of the Invention

[0003] In view of the above problems, this application is made in order to provide a grinding method and apparatus that overcomes or at least partially solves the above problems.

[0004] Firstly, a grinding method is provided, including:

[0005] If a surface defect is detected on the target object in the initial detection area, the target time when the surface defect is detected in the target area is determined based on the moving speed of the target object, and the target object is controlled to stop running at the target time.

[0006] Acquire a first real-world image of the target area, identify surface defects from the first real-world image, and determine the grinding coordinates of the surface defects in the preset grinding platform coordinate system;

[0007] The surface defects are repaired by grinding according to the grinding coordinates.

[0008] Optionally, after grinding the surface defects according to the grinding coordinates, the process may also include:

[0009] The image of the target area after grinding is acquired. The image of the target area after grinding is used to identify whether the target object after grinding is qualified. If it is qualified, the target object after grinding is polished according to the grinding coordinates. If not, the surface defects of the target object after grinding are ground according to the grinding coordinates.

[0010] Optionally, acquire a first real-world image of the target area, including:

[0011] By using each of the multiple cameras, images of the portion of the target object located in the target area are acquired, resulting in multiple first images. The multiple cameras have different perspectives.

[0012] Based on the set perspective transformation parameters and motion stitching parameters, multiple first images are subjected to perspective transformation and motion stitching to obtain a first real-world image.

[0013] Optionally, the perspective transformation parameters and the motion stitching parameters are obtained through the following steps:

[0014] Multiple second images are obtained by capturing images of the calibration board located in the target area using each of the multiple cameras;

[0015] For each of the multiple second images, a perspective transformation is performed on the second image to obtain the corresponding third image, and the perspective transformation sub-parameters corresponding to the second image are determined based on the pixel displacement information between the second image and the corresponding third image.

[0016] Based on the perspective transformation sub-parameters corresponding to each second image, the perspective transformation parameters are obtained;

[0017] Multiple third images are moved and stitched together to obtain a second real-world image. Based on the pixel displacement information between each third image and the second real-world image, the motion stitching sub-parameters corresponding to that third image are determined.

[0018] Based on the motion stitching sub-parameters corresponding to each third image, the motion stitching parameters are obtained.

[0019] Optionally, for each of the multiple second images, a perspective transformation is performed on the second image to obtain a corresponding third image, and the perspective transformation sub-parameters corresponding to the second image are determined based on the pixel displacement information between the second image and the corresponding third image, including:

[0020] For each second image, a three-dimensional world coordinate system is created based on all corner points of the second image. Based on the boundary of the calibration plate area corresponding to the second image, the pixels of the second image are moved in the three-dimensional world coordinate system to obtain a third image corresponding to the second image.

[0021] Based on the movement information of corresponding pixels between the second and third images, the perspective transformation matrix is ​​obtained, and the perspective transformation matrix is ​​used as the perspective transformation sub-parameter.

[0022] Optionally, multiple third images are moved and stitched together to obtain a second real-world image. Based on the pixel displacement information between each third image and the second real-world image, the moving and stitching sub-parameters corresponding to that third image are determined, including:

[0023] Arrange multiple third images according to a set position, and for each third image, determine the initial coordinates of that third image in the real-world image coordinate system of the calibration board;

[0024] Each third image is moved to stitch together a second real-world image, and the final coordinates of each third image in the calibration board's real-world image coordinate system are determined.

[0025] Determine the moving splicing sub-parameters based on the initial and final coordinates.

[0026] Optionally, based on the set perspective transformation parameters and motion stitching parameters, perspective transformation and motion stitching are performed on multiple first images to obtain a first real-world image, including:

[0027] Based on the perspective transformation parameters, a perspective transformation is performed on each of the multiple first images to obtain a fourth image corresponding to that first image.

[0028] Based on the initial coordinates, arrange multiple fourth images in the first real-world image coordinate system;

[0029] Based on the final coordinates, each fourth image is moved and stitched together to obtain the first real-world image.

[0030] Optionally, surface defects are identified from the first real-world image, and the grinding coordinates of the surface defects in the preset grinding platform coordinate system are determined, including:

[0031] Surface defects are identified from a first real-world image, and the first coordinates of the surface defects in the coordinate system of the first real-world image are determined.

[0032] Determine the coordinate transformation relationship between the first real-world image coordinate system and the preset grinding platform coordinate system, and based on the coordinate transformation relationship, transform the first coordinates to the preset grinding platform coordinate system to obtain the grinding coordinates.

[0033] Optionally, determine the coordinate transformation relationship between the first real-world image coordinate system and the preset grinding platform coordinate system, and based on the coordinate transformation relationship, transform the first coordinates to the preset grinding platform coordinate system to obtain the grinding coordinates, including:

[0034] Determine the first coordinate transformation relationship between the first real-world image coordinate system and the target area coordinate system, and based on the first coordinate transformation relationship, transform the first coordinates to the target area coordinate system to obtain the second coordinates;

[0035] Determine the second coordinate transformation relationship between the target area coordinate system and the preset grinding platform coordinate system, and based on the second coordinate transformation relationship, transform the second coordinates to the preset grinding platform coordinate system to obtain the grinding coordinates.

[0036] Secondly, a grinding device is provided, comprising:

[0037] The operation control unit is used to determine the target time when the surface defect is detected in the target area based on the moving speed of the target object if a surface defect is detected on the target object in the initial detection area, and to control the target object to stop running at the target time.

[0038] The coordinate determination unit is used to acquire a first real-world image of the target area, identify surface defects from the first real-world image, and determine the grinding coordinates of the surface defects in the preset grinding platform coordinate system.

[0039] The grinding unit is used to grind surface defects according to grinding coordinates.

[0040] Thirdly, this application also provides a server, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the server performs the method provided in the first aspect.

[0041] Fourthly, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, causes the computer to perform the method provided in the first aspect.

[0042] Fifthly, this application also provides a computer program product, including a computer program that, when run, causes a computer to perform the method provided in the first aspect.

[0043] The technical solution provided in this application has at least the following technical effects or advantages:

[0044] The grinding method and apparatus provided in this application detect whether a surface defect exists in a target object in an initial detection area. When a surface defect is determined to exist, since the distance between the initial detection area and the target area is fixed, the real-time position of the surface defect can be tracked based on the speed of the target object's movement during its operation. This allows the target object to be stopped at a target time, ensuring the surface defect is located within the target area. The surface defect is identified and its grinding coordinates are determined in a first real-world image of the target area. Based on these coordinates, the surface grinding system can be controlled to grind and polish the surface defect. This achieves coarse positioning, tracking positioning, and fine positioning of surface defects during strip steel production, ultimately enabling precise coordinates and efficient grinding of surface defects in the strip steel.

[0045] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0046] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0047] Figure 1 The grinding method flow in the embodiments of this application Figure 1 ;

[0048] Figure 2 This is a schematic diagram of a camera capturing images of a steel strip in an embodiment of this application;

[0049] Figure 3 The grinding method flow in the embodiments of this application Figure 2 ;

[0050] Figure 4 This is a schematic diagram of the second image in the embodiments of this application;

[0051] Figure 5 This is a schematic diagram of the third image in the embodiments of this application;

[0052] Figure 6 This is a schematic diagram of a real-world image of a chessboard pattern in an embodiment of this application;

[0053] Figure 7 The grinding method flow in the embodiments of this application Figure 3 ;

[0054] Figure 8 The grinding method flow in the embodiments of this application Figure 4 ;

[0055] Figure 9 This application includes an embodiment of the grinding method implementation scenario diagram;

[0056] Figure 10 This is a flowchart illustrating the strip grinding operation in the embodiments of this application;

[0057] Figure 11 This is a structural block diagram of the grinding device in an embodiment of this application. Detailed Implementation

[0058] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings.

[0059] The accompanying drawings illustrate various structural schematics according to embodiments of this application. These drawings are not to scale, and some details have been enlarged for clarity, and some details may have been omitted. The shapes of the various regions and layers shown in the drawings, as well as their relative sizes and positional relationships, are merely exemplary and may deviate from reality due to manufacturing tolerances or technical limitations. Furthermore, those skilled in the art can design regions / layers with different shapes, sizes, and relative positions as needed.

[0060] To better understand the above technical solutions, the following will describe the above technical solutions in detail with reference to specific implementation methods. It should be understood that the embodiments of this disclosure and the specific features in the embodiments are detailed descriptions of the technical solutions of this application, rather than limitations on the technical solutions of this application. In the absence of conflict, the embodiments of this application and the technical features in the embodiments can be combined with each other.

[0061] During strip steel production, surface defects need to be ground. Current technology reduces the production line speed to the minimum crawling speed, and workers have to visually inspect for defects. This severely limits production efficiency and poses a great threat to the personal safety of workers. There are even cases where defects are missed, making it impossible to control product quality.

[0062] In the grinding system, defect location technology is the core technology. If the defect cannot be accurately located, the grinding task cannot be completed and the purpose of improving the surface quality of the strip cannot be achieved. Therefore, it is imperative to be able to scientifically calculate the defect location information.

[0063] In view of this, this application provides a grinding method, please refer to it. Figure 1 , Figure 1 The following is a flowchart of the grinding method in the embodiments of this application, including:

[0064] S101, if a surface defect is detected on the target object in the initial detection area, then based on the moving speed of the target object, determine the target time when the surface defect is detected in the target area, and at the target time, control the target object to stop running.

[0065] S102, acquire the first real-world image of the target area, identify surface defects from the first real-world image, and determine the grinding coordinates of the surface defects in the preset grinding platform coordinate system.

[0066] S103, according to the grinding coordinates, the surface defects are ground.

[0067] The initial inspection area is equipped with a surface inspection device. During the strip rolling process, the surface inspection device captures images of the strip surface to preliminarily detect the presence of surface defects. When a surface defect is confirmed from the images captured by the surface inspection device, its position on the strip is recorded. The distance between the initial inspection area and the target area is fixed. Therefore, during the strip's operation, the position of the surface defect in the strip production line can be tracked in real time based on the strip's running speed, allowing the strip production line to be stopped at a target time so that the surface defect is located in the target area.

[0068] By identifying surface defects in the target area from a first real-world image and determining their grinding coordinates, the surface grinding system can be controlled to grind and polish these defects based on these coordinates. This process achieves coarse positioning, tracking positioning, and fine positioning of surface defects during strip steel production, ultimately enabling precise coordinates and efficient grinding of strip steel surface defects.

[0069] In some optional implementations, after performing the operation S103, the following operations are also included:

[0070] The image of the target area after grinding is acquired. The image of the target area after grinding is used to identify whether the target object after grinding is qualified. If it is qualified, the target object after grinding is polished according to the grinding coordinates. If not, the surface defects of the target object after grinding are ground according to the grinding coordinates.

[0071] After grinding, photos are taken again for verification. If the surface defects of the strip steel are removed, the target object is deemed qualified after grinding, and the surface grinding system is then used for polishing. If surface defects of the strip steel still exist after grinding, grinding is performed again to ensure the quality of grinding.

[0072] The following example illustrates how to precisely locate surface defects in strip steel using the S102 operation.

[0073] In some optional implementations, the specific operations for acquiring the first real-world image of the target area in S102 include:

[0074] The first step involves capturing images of the portion of the target object located in the target area using each of the multiple cameras, resulting in multiple first images with different perspectives from each camera.

[0075] The second step involves performing perspective transformation and motion stitching on multiple first images based on the set perspective transformation and motion stitching parameters to obtain the first real-world image. It can be understood that the effective shooting range of multiple cameras constitutes the target area, which is located on a set grinding platform. The grinding platform can be a platform set up on the production line for a surface grinding system to perform grinding operations, and the surface grinding system is used to grind the strip steel on the grinding platform.

[0076] The specific operations in S102, which involve identifying surface defects from the first real-world image and determining the grinding coordinates of the surface defects in the preset grinding platform coordinate system, include:

[0077] The first step is to identify surface defects from the first real-world image and determine the first coordinates of the surface defects in the coordinate system of the first real-world image.

[0078] The second step is to determine the coordinate transformation relationship between the first real-world image coordinate system and the preset grinding platform coordinate system, and based on the coordinate transformation relationship, transform the first coordinates to the preset grinding platform coordinate system to obtain the grinding coordinates.

[0079] In some alternative implementations, such as Figure 8 As shown, the second step is as follows:

[0080] S801, determine the first coordinate transformation relationship between the first real-world image coordinate system and the target area coordinate system, and based on the first coordinate transformation relationship, transform the first coordinates to the target area coordinate system to obtain the second coordinates.

[0081] S802, determine the second coordinate transformation relationship between the target area coordinate system and the preset grinding platform coordinate system, and based on the second coordinate transformation relationship, transform the second coordinates to the preset grinding platform coordinate system to obtain the grinding coordinates.

[0082] The grinding method provided in this application involves performing perspective transformation and motion stitching on first images of a target object captured by multiple cameras to obtain a first real-world image, ensuring that the position of surface defects in the first real-world image matches the actual position. The first real-world image originates from the target area, and the target area and the grinding platform have a fixed positional relationship. Therefore, based on the one-to-one coordinate transformation relationship between corresponding coordinate systems, the coordinates of the strip surface defects in the first real-world image can be converted into the actual coordinates of the strip surface defects in the target area. Since the relative position of the target area on the grinding platform is predetermined, the actual coordinates of the strip surface defects in the target area can be converted into the actual coordinates of the strip surface defects on the grinding platform, i.e., the grinding coordinates. This allows for precise defect location, improves grinding efficiency on the production line, and enhances grinding quality and yield.

[0083] It is understood that the above-described grinding methods can be applied to terminal devices and / or servers. Terminal devices and / or servers can be understood as testing equipment. Servers can be, but are not limited to, web servers, database servers, cloud servers, etc. Terminal devices can be, but are not limited to, office computers.

[0084] The following example illustrates a specific solution for applying the method described above to an actual strip steel production line to obtain the specific coordinates of surface defects on the grinding platform.

[0085] (1) Define the surface inspection device, strip steel production line and grinding robot in the same coordinate system.

[0086] For example, such as Figure 2 As shown, four cameras 201 arranged at intervals can be slidably mounted on the slide rail 202. Each camera 201 takes an image of the surface of the strip steel 203 at a set height and angle from the surface of the strip steel 203. Compared with the real image, the captured image has a certain degree of distortion and deviation, and there is regional overlap between each image. Therefore, it is necessary to convert the image of the surface of the strip steel 203 captured by the camera 201 into a real image, and then stitch the images together in the overlapping area.

[0087] When a camera is replaced, maintained, or serviced, the new camera may have changed parameters, and there may be deviations in the installation angle and height. Therefore, when eliminating distortion and deviations and stitching images taken by the new camera, the parameters used to process images taken by the previous camera cannot be reused. New image processing parameters need to be configured based on the parameters and location of the new camera.

[0088] Therefore, in some alternative implementations, such as Figure 3 As shown, the perspective transformation parameters and the motion stitching parameters are obtained through the following steps:

[0089] S301 acquires images of the calibration board located in the target area using each of the multiple cameras, resulting in multiple second images.

[0090] The calibration board uses a black and white checkerboard pattern. Four cameras 201 each capture an image of the checkerboard pattern, which is the second image. The overall pattern of the checkerboard in the second image is as follows: Figure 4 The trapezoid shown is not the same as the actual image of a checkerboard.

[0091] S302, for each of the multiple second images, perform a perspective transformation on the second image to obtain a corresponding third image, and determine the perspective transformation sub-parameters corresponding to the second image based on the pixel displacement information between the second image and the corresponding third image.

[0092] S303, based on the perspective transformation sub-parameters corresponding to each second image, obtain the perspective transformation parameters.

[0093] The operation of S302 includes:

[0094] The first step is to create a three-dimensional world coordinate system for each second image, based on all the corner points of that second image.

[0095] like Figure 4 As shown, the chessboard in a second image is trapezoidal, and the corner points are specifically the four corner points (A0, B0, C0, D0) of the trapezoidal image.

[0096] The second step involves moving the pixels of the second image in three-dimensional world coordinates according to the boundary of the calibration plate region corresponding to the second image, thereby obtaining a third image corresponding to the second image. The goal of this operation is to perform perspective transformation on the second image to achieve trapezoidal correction. It can be understood that the pixels of multiple third images correspond one-to-one with the pixels of the second real-world image, and the third image can be understood as a real-world image of a sub-region of the calibration plate.

[0097] This operation yields the corner transformation matrix of the second image, converting the corner points from the 3D world coordinate system to the real-world plane. After eliminating trapezoidal distortion through perspective transformation stretching, the result is as follows: Figure 5 The third image shown.

[0098] The third step is to obtain the perspective transformation matrix based on the movement information of corresponding pixels between the second and third images, and use the perspective transformation matrix as the perspective transformation sub-parameter.

[0099] When camera 201 captures images of the surface of strip 203 in the target area, deviations will naturally occur. These deviations are not conducive to analyzing the location of defects on the surface of strip 203 through images. Therefore, perspective transformation is required. However, since the images of strip 203 are irregular, the parameters for perspective transformation are difficult to determine. Therefore, perspective transformation is first performed on checkerboard images captured by the four cameras 201, and the perspective transformation parameters for each image are recorded. These parameters can be directly applied to the perspective transformation operation of the strip 203 image in S102 to obtain real-world images of multiple sub-regions on the surface of strip 203. It is understandable that these sub-regions overlap, and these sub-regions are pieced together to form the complete target area.

[0100] S304, move and stitch multiple third images to obtain a second real-world image, and determine the move and stitch sub-parameters corresponding to the third image based on the pixel displacement information between each third image and the second real-world image.

[0101] S305, based on the motion stitching sub-parameters corresponding to each third image, obtain the motion stitching parameters.

[0102] The operation of S304 includes:

[0103] The first step is to arrange multiple third images according to the set positions, and for each third image, determine the initial coordinates of that third image in the real-world image coordinate system of the calibration board.

[0104] The second step is to move each third image to stitch together the second real-world image and determine the final coordinates of each third image in the real-world image coordinate system of the calibration board.

[0105] The third step is to determine the moving and splicing sub-parameters based on the initial and final coordinates.

[0106] Using the specific operation of S303 described above, multiple such Figure 5 The third image shown is stitched together as follows: Figure 6 The image shown is a real-world image of the checkerboard pattern in the target area. This completes the process of capturing multiple images of sub-regions of the target area from multiple cameras, performing perspective transformation and motion stitching to obtain the real-world image of the checkerboard pattern in the target area.

[0107] Specifically, the movement of the third image can be controlled by a predefined keyboard. During the process of completely stitching together the second real-world image, the keyboard input is recorded to calculate the final coordinates of the corner points of each third image after the stitching is completed.

[0108] In summary, based on the perspective transformation and motion stitching process performed on multiple second images of the target area captured by multiple cameras on a calibration board, standard perspective transformation parameters and motion stitching parameters are generated from the displacement parameters of corresponding pixels. These parameters are then applied to multiple first images of the target object captured by multiple cameras on the target area to obtain the first real-world image. This ensures the integrity and authenticity of the images of the target object captured by multiple cameras, thereby facilitating the accuracy of defect detection of the target object and providing convenience for future equipment maintenance.

[0109] In some alternative implementations, such as Figure 7 As shown, in the specific operation of obtaining the first real-world image of the target area in S102, the second step specifically includes:

[0110] S701, according to the perspective transformation parameters, perform perspective transformation on each of the multiple first images to obtain a fourth image corresponding to the first image.

[0111] S702, based on the initial coordinates, arrange multiple fourth images in the first real-world image coordinate system.

[0112] S703, based on the final coordinates, move and stitch each fourth image together to obtain the first real-world image.

[0113] In some alternative implementations, after obtaining the second real-world image, the edges of the second real-world image are cropped to make the second real-world image reach a set size, and the coordinate range of the pixels to be cropped at the edges of the second real-world image is determined.

[0114] Thus, after obtaining the first real-world image by performing the S703 operation, the process also includes:

[0115] Based on the coordinate range of pixels cropped at the edges of the second real-world image, the edge portion of the first real-world image is cropped so that the size of the first real-world image reaches the set size.

[0116] The following is combined with Figure 9 , 10 This paper introduces a specific operational example of using the method provided in the embodiments of this application to repair defects in steel strip surfaces.

[0117] The first step involves setting an advance deceleration threshold based on the approximate location information of the defect area 906 on the surface of the strip 901 provided by the surface inspection device 905 (i.e., the approximate location information of the sub-region where the surface defect is located). When the tracking position of the strip 901 (the real-time position of the defect area 906) reaches the threshold, the strip 901 automatically decelerates so that the defect position is stopped at the target area 907 of the grinding area 902 after the defect is detected. If the distance between the surface inspection device 905 and the grinding area 902 is far enough (greater than the braking distance of the maximum operating speed of the production line), advance deceleration is not necessary.

[0118] When a defect is detected on the surface, the surface detection device 905 can send the defect coordinates to the first-level automation of the production line and start the defect stop positioning, stopping the defect location in the grinding area 902.

[0119] In the second step, after the defective area 906 stops at the target area 907 of the grinding area 902, the grinding robot 903, located on one side of the target area 907, takes a picture of the defective area 906 using the vision camera 904 installed on its head. After taking the picture, the picture is synthesized and analyzed according to the operation in S102 to calculate the precise coordinate position of the defective area 906 in the grinding area 902, and the grinding coordinate position is sent to the grinding robot.

[0120] The third step involves the grinding robot 903 grinding the defective area 906 on the surface of the strip steel 901 according to precise grinding coordinates. After grinding, camera 904 takes a picture of the defective area 906 for verification. If the verification fails, grinding continues; if the verification is successful, the robot switches to a grinding head to polish the defective area. It is understood that the grinding head on the robot is equipped with a two-way force control device and a floating device to achieve a good grinding effect.

[0121] Based on the same inventive concept, embodiments of this application also provide a grinding device, such as... Figure 11 As shown, the grinding device 1100 includes:

[0122] The operation control unit 1101 is used to determine the target time when the surface defect is detected in the target area based on the moving speed of the target object if a surface defect is detected on the target object in the initial detection area, and to control the target object to stop running at the target time.

[0123] The coordinate determination unit 1102 is used to acquire a first real-world image of the target area, identify surface defects from the first real-world image, and determine the grinding coordinates of the surface defects in the preset grinding platform coordinate system.

[0124] The grinding unit 1103 is used to grind surface defects according to grinding coordinates.

[0125] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the computer to perform the method provided in the above embodiments.

[0126] This application also provides a server, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it causes the server to perform the method provided in the above embodiments.

[0127] This application also provides a computer program product, including a computer program that, when run, causes a computer to perform the methods provided in the above embodiments.

[0128] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0129] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0130] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of this application may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

Claims

1. A grinding method, characterized in that, include: If a surface defect is detected on the target object in the initial detection area, the target time when the surface defect is detected in the target area is determined based on the moving speed of the target object, and the target object is controlled to stop running at the target time. Acquire a first real-world image of the target area, identify the surface defects from the first real-world image, and determine the grinding coordinates of the surface defects in a preset grinding platform coordinate system; The surface defects are repaired by grinding according to the grinding coordinates; The acquisition of the first real-world image of the target area includes: Multiple first images are obtained by capturing images of the portion of the target object located in the target area using each of multiple cameras, with the multiple cameras having different perspectives; Based on the set perspective transformation parameters and motion stitching parameters, perspective transformation and motion stitching are performed on multiple first images to obtain the first real-world image; The perspective transformation parameters and the motion stitching parameters are obtained through the following steps: Each of the plurality of cameras acquires an image of the calibration board located in the target area, resulting in a plurality of second images. For each of the multiple second images, a perspective transformation is performed on the second image to obtain a corresponding third image, and the perspective transformation sub-parameters corresponding to the second image are determined based on the pixel displacement information between the second image and the corresponding third image. The perspective transformation parameters are obtained based on the perspective transformation sub-parameters corresponding to each of the second images; Multiple third images are moved and stitched together to obtain a second real-world image. Based on the pixel displacement information between each third image and the second real-world image, the moving and stitching sub-parameters corresponding to the third image are determined. The motion stitching parameters are obtained based on the motion stitching sub-parameters corresponding to each of the third images; For each of the plurality of second images, a perspective transformation is performed on the second image to obtain a corresponding third image. Based on the pixel displacement information between the second image and the corresponding third image, perspective transformation sub-parameters corresponding to the second image are determined, including: For each second image, a three-dimensional world coordinate system is created based on all corner points of the second image. Based on the boundary of the calibration plate area corresponding to the second image, the pixels of the second image are moved in the three-dimensional world coordinate system to obtain the third image corresponding to the second image. Based on the movement information of corresponding pixels between the second image and the third image, a perspective transformation matrix is ​​obtained, and the perspective transformation matrix is ​​used as the perspective transformation sub-parameter.

2. The grinding method as described in claim 1, characterized in that, After grinding the surface defects according to the grinding coordinates, the process further includes: Acquire an image of the target area after grinding, and identify whether the target object after grinding is qualified from the image of the target area after grinding. If it is qualified, polish the target object after grinding according to the grinding coordinates; if not, grind the surface defects of the target object after grinding according to the grinding coordinates.

3. The grinding method as described in claim 1, characterized in that, The process of moving and stitching together multiple third images to obtain a second real-world image, and determining the moving and stitching sub-parameters corresponding to each third image based on the pixel displacement information between each third image and the second real-world image, includes: Arrange the multiple third images according to a set position, and for each third image, determine the initial coordinates of the third image in the real-world image coordinate system of the calibration board; Each of the third images is moved to stitch together the second real-world image, and the final coordinates of each of the third images in the real-world image coordinate system of the calibration board are determined. The moving splicing sub-parameters are determined based on the initial coordinates and the final coordinates.

4. The grinding method as described in claim 3, characterized in that, The process of performing perspective transformation and motion stitching on multiple first images based on set perspective transformation parameters and motion stitching parameters to obtain the first real-world image includes: According to the perspective transformation parameters, a perspective transformation is performed on each of the multiple first images to obtain a fourth image corresponding to that first image; Based on the initial coordinates, the multiple fourth images are arranged in the first real-world image coordinate system; Based on the final coordinates, each of the fourth images is moved and stitched together to obtain the first real-world image.

5. The grinding method as described in claim 1, characterized in that, The step of identifying the surface defect from the first real-world image and determining the grinding coordinates of the surface defect in the preset grinding platform coordinate system includes: The surface defect is identified from the first real-world image, and the first coordinate of the surface defect in the first real-world image coordinate system is determined. Determine the coordinate transformation relationship between the first real-world image coordinate system and the preset grinding platform coordinate system, and based on the coordinate transformation relationship, transform the first coordinates to the preset grinding platform coordinate system to obtain the grinding coordinates.

6. The grinding method as described in claim 5, characterized in that, The step of determining the coordinate transformation relationship between the first real-world image coordinate system and the preset grinding platform coordinate system, and transforming the first coordinates to the preset grinding platform coordinate system based on the coordinate transformation relationship to obtain the grinding coordinates, includes: Determine the first coordinate transformation relationship between the first real-world image coordinate system and the target area coordinate system, and based on the first coordinate transformation relationship, transform the first coordinates to the target area coordinate system to obtain the second coordinates; A second coordinate transformation relationship is determined between the target area coordinate system and the preset grinding platform coordinate system, and based on the second coordinate transformation relationship, the second coordinates are transformed to the preset grinding platform coordinate system to obtain the grinding coordinates.

7. A grinding apparatus for implementing the method according to any one of claims 1-6, characterized in that, include: The operation control unit is used to determine the target time when the surface defect is detected in the target area based on the moving speed of the target object if a surface defect is detected on the target object in the initial detection area, and to control the target object to stop running at the target time; The coordinate determination unit is used to acquire a first real-world image of the target area, identify the surface defects from the first real-world image, and determine the grinding coordinates of the surface defects in a preset grinding platform coordinate system. The grinding unit is used to grind the surface defects according to the grinding coordinates.

Images