Method and device for three-dimensional defect detection and size measurement of high-temperature continuous casting sheet billet

Abstract

The invention provides a three-dimensional defect detection and size measurement method and device for a high-temperature continuous casting slab, and relates to the technical field of machine vision detection. The method comprises the following steps: acquiring point cloud data of a high-temperature continuous casting slab; calibrating the point cloud data of the high-temperature continuous casting slab by adopting a calibration block to obtain calibrated point cloud data; calculating an actual depth value according to the calibrated point cloud data; the actual depth value is input into a pre-trained Yolov5 target defect detection model through rendering processing for recognition, and a defect position and a defect category are obtained; and according to the defect position and the defect category, extracting the maximum depth and the average depth of the defect area as the severity characterization of the defect. With the adoption of the method, the limitation of a traditional vertical-view-angle camera in the process of detecting tiny defects can be overcome, and a more comprehensive and effective solution is provided for detecting the surface defects of the continuous casting sheet billet.

Description

Technical Field

[0001] The present invention relates to the technical field of machine vision detection, and in particular to a method and device for three-dimensional defect detection and size measurement of a high-temperature continuous casting slab. Background Art

[0002] In the process of slab production, a large number of defects are often generated due to process factors. Slabs with defects as raw materials will have a serious impact on subsequent products, including product degradation, thus causing economic losses.

[0003] Traditional surface defect detection systems usually use cameras and LED light sources installed directly above the slab surface for imaging, thereby realizing the detection of slab defects. However, for common defects, including dents and cracks, when two-dimensional imaging is performed from directly above, no obvious grayscale difference can be formed on the image, so that the morphology of the defect cannot be effectively characterized, resulting in a high rate of missed detection of such defects by the detection system. In addition, two-dimensional images cannot represent changes in the depth direction, and the severity of the defects cannot be determined, making it difficult to complete high-level processing operations for quality grading. Summary of the invention

[0004] In order to solve the problem that the surface defects of slabs cannot be effectively detected by grayscale images in the prior art, and the technical problem that the depth of defects cannot be quantified, the embodiment of the present invention provides a method and device for three-dimensional defect detection and size measurement of high-temperature continuous casting slabs. The technical solution is as follows:

[0005] On the one hand, a method for three-dimensional defect detection and size measurement of a high-temperature continuous casting slab is provided, which is implemented by a three-dimensional defect detection and size measurement device for a high-temperature continuous casting slab, and the method comprises:

[0006] S1. Obtaining point cloud data of high temperature continuous casting slab;

[0007] S2. Calibrate the point cloud data of the high-temperature continuous casting slab using a calibration block to obtain calibrated point cloud data;

[0008] S3, calculating the actual depth value according to the calibrated point cloud data; the actual depth value is processed by rendering to obtain a pseudo-color image, which is input into a pre-trained Yolov5 target defect detection model for recognition to obtain the defect location and defect category;

[0009] S4. Obtaining defect information on the surface of the high-temperature continuous casting slab according to the defect position and defect category;

[0010] S5, adopting an auxiliary detection method, by adding an area array camera to shoot the upper and lower surfaces of the high-temperature continuous casting slab, to obtain a grayscale image of the high-temperature continuous casting slab; inputting the grayscale image into the pre-trained Yolov5 target defect detection model for recognition, to obtain slight defect information on the surface of the high-temperature continuous casting slab;

[0011] S6. Obtain final defect information on the surface of the high-temperature continuous casting slab according to the defect information on the surface of the high-temperature continuous casting slab and the slight defect information on the surface of the high-temperature continuous casting slab.

[0012] Optionally, the step of obtaining point cloud data of a high-temperature continuous casting slab in S1 includes:

[0013] In the cross-sectional direction of the high-temperature continuous casting slab, multiple groups of 3D cameras are arranged around to collect point cloud data of the upper surface, lower surface, left side and right side respectively; among them, the upper and lower surface cameras shoot the same cross-sectional position, the left and right surface cameras shoot the same cross-sectional position, and the interval between the two cross-sectional positions is greater than a certain preset distance. The laser velocimeter output pulse trigger is used for collection to obtain the point cloud data of the high-temperature continuous casting slab.

[0014] Optionally, the 3D camera is composed of a line laser transmitter and an area array camera, and adopts a split laser triangulation imaging solution. The laser line is perpendicular to the measured plane, and the area array camera is tilted at a fixed angle to shoot, and the angle is determined by the actual depth of field range and depth direction accuracy required to be shot; two different colors of lasers are used between adjacent cameras, and the cameras are equipped with color filters of corresponding bands to eliminate interference when the camera extracts the laser line; wherein the two different colors are blue and green, respectively.

[0015] Optionally, the step S3 of calculating the actual depth value according to the calibrated point cloud data includes:

[0016] S31, extracting the first row of valid data from the calibrated point cloud data matrix, splicing the first row of point cloud data and row coordinates into a two-dimensional data matrix; using the RANSAC algorithm to fit the data into a straight line, determining the inclination angle of the straight line, and constructing a rotation matrix;

[0017] S32, translate the first row of valid data along the x direction to the center position, perform angle rotation to make the z direction data close to the horizontal, and then translate it in the opposite direction along the x direction to obtain a new matrix, and use the new matrix as the target point cloud;

[0018] S33, along the row direction of the point cloud, take the next row of data, splice the corresponding point cloud data with the row coordinates to form a two-dimensional data matrix, align the two-dimensional data matrix with the target point cloud through the ICP algorithm, calculate the rigid body matrix including rotation and translation, and transform the two-dimensional data matrix using the obtained rigid body matrix to obtain a transformed point cloud;

[0019] S34, taking the converted point cloud as the target point cloud, repeating steps S33-S34 until all the point cloud data collected by the camera are processed and combined row by row to form new point cloud data;

[0020] S35, taking the median of the new point cloud data as the reference point, comparing the new point cloud data with the median and taking the difference as the actual depth value of the coordinate position of each point cloud data.

[0021] Optionally, the auxiliary detection method of S5 is adopted, by adding an area array camera to shoot the upper and lower surfaces of the high-temperature continuous casting slab to obtain a grayscale image of the high-temperature continuous casting slab; the grayscale image is input into the pre-trained Yolov5 target defect detection model for recognition to obtain slight defect information on the surface of the high-temperature continuous casting slab, including:

[0022] S51, perform auxiliary detection on the slight defects on the upper and lower surfaces by adding an additional area array camera to shoot the upper and lower surfaces of the slab at an angle less than 15 degrees to obtain a grayscale image;

[0023] S52, preprocessing the grayscale image to obtain a preprocessed image;

[0024] S53, inputting the preprocessed image into the pre-trained Yolov5 target defect detection model for recognition, and obtaining the defect pixel coordinates;

[0025] S54, obtaining the position coordinates and size of the defect on the actual slab according to the defect pixel coordinates and the mapping relationship between the image coordinate system and the world coordinate system;

[0026] S55. According to the actual world coordinates of the defect, the maximum depth and the average depth are extracted at the corresponding positions in the actual depth data as representations of the severity of the defect, and the minor defect information on the surface of the high-temperature continuous casting slab is obtained.

[0027] Optionally, the process of the auxiliary detection method includes:

[0028] Collecting slab images according to the auxiliary detection and acquisition device;

[0029] According to the acquired slab images, the actual movement distance of the slab between two adjacent images and the world coordinates corresponding to the pixel coordinate origin on the images are calculated;

[0030] According to the actual movement distance of the slab and the world coordinates corresponding to the pixel coordinate origin on the image, the world coordinates corresponding to the pixel coordinate origin on all the collected images are calculated;

[0031] According to the world coordinates corresponding to the origin of the pixel coordinates on the image, the position coordinates of the defect in the length direction and the width direction of the slab are calculated.

[0032] On the other hand, a high-temperature continuous casting slab three-dimensional defect detection and size measurement device is provided, and the device is applied to a high-temperature continuous casting slab three-dimensional defect detection and size measurement method, and the device comprises:

[0033] A first acquisition unit is used to acquire point cloud data of a high-temperature continuous casting slab;

[0034] A calibration unit, used for calibrating the point cloud data of the high-temperature continuous casting slab using a calibration block to obtain calibrated point cloud data;

[0035] The recognition and calculation unit is used to calculate the actual depth value according to the calibrated point cloud data; the actual depth value is processed by rendering to obtain a pseudo-color image which is input into a pre-trained Yolov5 target defect detection model for recognition to obtain the defect position and defect category;

[0036] A second acquisition unit is used to obtain defect information on the surface of the high-temperature continuous casting slab according to the defect position and defect category;

[0037] The auxiliary detection unit is used to adopt an auxiliary detection method, by adding an area array camera to shoot the upper and lower surfaces of the high-temperature continuous casting slab, to obtain a grayscale image of the high-temperature continuous casting slab; the grayscale image is input into the pre-trained Yolov5 target defect detection model for recognition, to obtain slight defect information on the surface of the high-temperature continuous casting slab;

[0038] The third acquisition unit is used to obtain the final defect information of the high-temperature continuous casting slab surface according to the defect information of the high-temperature continuous casting slab surface and the slight defect information of the high-temperature continuous casting slab surface. Optionally, the first acquisition unit is used to:

[0039] In the cross-sectional direction of the high-temperature continuous casting slab, multiple groups of 3D cameras are arranged around to collect point cloud data of the upper surface, lower surface, left side and right side respectively; among them, the upper and lower surface cameras shoot the same cross-sectional position, the left and right surface cameras shoot the same cross-sectional position, and the interval between the two cross-sectional positions is greater than a certain preset distance. The laser velocimeter output pulse trigger is used for collection to obtain the point cloud data of the high-temperature continuous casting slab.

[0040] Optionally, the 3D camera is composed of a line laser transmitter and an area array camera, and adopts a split laser triangulation imaging solution. The laser line is perpendicular to the measured plane, and the area array camera is tilted at a fixed angle to shoot, and the angle is determined by the actual depth of field range and depth direction accuracy required to be shot; two different colors of lasers are used between adjacent cameras, and the cameras are equipped with color filters of corresponding bands to eliminate interference when the camera extracts the laser line; wherein the two different colors are blue and green, respectively.

[0041] Optionally, the identification and calculation unit is used to:

[0042] Extract the first row of valid data from the calibrated point cloud data matrix, and splice the first row of point cloud data and row coordinates into a two-dimensional data matrix; use the RANSAC algorithm to fit the data into a straight line, determine the inclination angle of the straight line, and construct a rotation matrix;

[0043] The first row of valid data is translated to the center along the x direction, and the angle rotation is performed to make the z direction data close to the horizontal, and then translated in the opposite direction along the x direction to obtain a new matrix, and the new matrix is ​​used as the target point cloud;

[0044] Take the next row of data along the row direction of the point cloud, splice the corresponding point cloud data with the row coordinates to form a two-dimensional data matrix, align the two-dimensional data matrix with the target point cloud through the ICP algorithm, calculate the rigid body matrix including rotation and translation, and use the obtained rigid body matrix to transform the two-dimensional data matrix to obtain the transformed point cloud;

[0045] The converted point cloud is used as the target point cloud, and steps S33-S34 are repeated until all the point cloud data collected by the camera are processed and combined row by row to form new point cloud data;

[0046] The median of the new point cloud data is taken as the reference point, and the difference between the new point cloud data and the median is taken as the actual depth value of the coordinate position of each point cloud data.

[0047] Optionally, the auxiliary detection unit is used to:

[0048] To assist in the detection of minor defects on the upper and lower surfaces, an additional area array camera is added to capture the upper and lower surfaces of the slab at an angle of less than 15 degrees to obtain a grayscale image.

[0049] Preprocessing the grayscale image to obtain a preprocessed grayscale image;

[0050] The preprocessed grayscale image is input into the pre-trained Yolov5 target defect detection model for recognition to obtain the defect pixel coordinates;

[0051] According to the defect pixel coordinates and the mapping relationship between the image coordinate system and the world coordinate system, the position coordinates and size of the defect on the actual slab are obtained;

[0052] According to the actual world coordinates of the defect, the maximum depth and average depth are extracted at the corresponding position in the actual depth data as the representation of the severity of the defect, and the minor defect information on the surface of the high-temperature continuous casting slab is obtained.

[0053] Optionally, the process of the auxiliary detection method includes:

[0054] Collecting slab images according to the auxiliary detection and acquisition device;

[0055] According to the acquired slab images, the actual movement distance of the slab between two adjacent images and the world coordinates corresponding to the pixel coordinate origin on the images are calculated;

[0056] According to the actual movement distance of the slab and the world coordinates corresponding to the pixel coordinate origin on the image, the world coordinates corresponding to the pixel coordinate origin on all the collected images are calculated;

[0057] According to the world coordinates corresponding to the origin of the pixel coordinates on the image, the position coordinates of the defect in the length direction and the width direction of the slab are calculated.

[0058] Optionally, the final defect information on the surface of the high-temperature continuous casting slab includes: defect type, defect position, defect length, defect width, defect area and defect depth.

[0059] On the other hand, a high-temperature continuous casting slab three-dimensional defect detection and size measurement device is provided, and the high-temperature continuous casting slab three-dimensional defect detection and size measurement device comprises: a processor; a memory, and the memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, any one of the above-mentioned high-temperature continuous casting slab three-dimensional defect detection and size measurement methods is implemented.

[0060] On the other hand, a computer-readable storage medium is provided, in which at least one instruction is stored, and the at least one instruction is loaded and executed by a processor to implement any one of the above-mentioned high-temperature continuous casting slab three-dimensional defect detection and size measurement methods.

[0061] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0062] The embodiment of the present invention first obtains point cloud data of a high-temperature continuous casting slab; uses a calibration block to calibrate the point cloud data of the high-temperature continuous casting slab to obtain calibrated point cloud data; secondly, according to the calibrated point cloud data, calculates the actual depth value; the actual depth value is processed by rendering to obtain a pseudo-color image, which is input into a pre-trained Yolov5 target defect detection model for identification to obtain a defect position and a defect category; according to the defect position and the defect category, defect information on the surface of the high-temperature continuous casting slab is obtained; an auxiliary detection method is adopted to obtain a grayscale image of the high-temperature continuous casting slab by adding an array camera to shoot the upper and lower surfaces of the high-temperature continuous casting slab; the grayscale image is input into a pre-trained Yolov5 target defect detection model for identification to obtain slight defect information on the surface of the high-temperature continuous casting slab; finally, according to the defect information on the surface of the high-temperature continuous casting slab and the slight defect information on the surface of the high-temperature continuous casting slab, the final defect information on the surface of the high-temperature continuous casting slab is obtained.

[0063] The embodiment of the present invention collects point cloud data of the upper surface, lower surface, left side and right side of the slab by arranging multiple groups of 3D cameras in the cross-sectional direction of the slab, and after calibration and conversion of the point cloud data, the length, thickness, width, curvature and T value of the slab can be calculated, and the point cloud data can be further processed to realize the representation of the relative depth of the surface, and the position and depth of the defect can be detected by using the relative depth data. In addition, in order to increase the detection capability of minor defects, upper and lower surface auxiliary image acquisition devices are installed on the edge side of the roller, and the acquisition camera and the corresponding light source are designed to be at a lower angle. The viewing angle can significantly enhance the defect characteristics with depth on the surface of the slab, thereby enhancing the detection capability of defects; the embodiment of the present invention can simultaneously realize the slab size measurement and defect detection tasks through point cloud data analysis, and by adding auxiliary detection means, simulating the actual manual detection method, increasing the detection capability of minor defects, and realizing comprehensive detection, thereby achieving the purpose of replacing manual detection of slab surface defects, overcoming the limitations of traditional vertical viewing angle cameras in detecting minor defects, ensuring the quality of slabs, reducing production losses, meeting the quality detection needs of enterprises, and providing a more comprehensive and effective solution for continuous casting slab surface defect detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0065] Figure 1 This is a flow chart of a method for three-dimensional defect detection and size measurement of a high-temperature continuous casting slab provided by an embodiment of the present invention;

[0066] Figure 2 is a schematic diagram of an installation position of a detection device provided by an embodiment of the present invention;

[0067] Figure 3 It is a layout diagram of a 3D camera in a cross-sectional direction of a casting blank provided by an embodiment of the present invention;

[0068] Figure 4 is a schematic diagram of an auxiliary image acquisition device provided by an embodiment of the present invention along a direction of looking down at a roller conveyor;

[0069] Figure 5 is a schematic diagram of an auxiliary image acquisition device provided by an embodiment of the present invention along the running direction of a roller conveyor;

[0070] Figure 6 This is a block diagram of a high-temperature continuous casting slab three-dimensional defect detection and size measurement device provided by an embodiment of the present invention;

[0071] Figure 7 It is a structural schematic diagram of a high-temperature continuous casting slab three-dimensional defect detection and size measurement device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0072] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0073] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0074] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same. "of", "corresponding, relevant" and "corresponding" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same.

[0075] In the embodiments of the present invention, sometimes a subscript such as W1 may be written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0076] In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0077] The embodiment of the present invention provides a method for three-dimensional defect detection and size measurement of high-temperature continuous casting slabs, which can be implemented by a three-dimensional defect detection and size measurement device for high-temperature continuous casting slabs, and the three-dimensional defect detection and size measurement device for high-temperature continuous casting slabs can be a terminal or a server. Figure 1 The flowchart of the method for three-dimensional defect detection and size measurement of high-temperature continuous casting slabs shown in the figure can include the following steps:

[0078] S1. Obtain point cloud data of high temperature continuous casting slab.

[0079] Optionally, S1 acquires point cloud data of high temperature continuous casting slab, including:

[0080] In the cross-sectional direction of the high-temperature continuous casting slab, multiple groups of 3D cameras are arranged around to collect point cloud data of the upper surface, lower surface, left side and right side respectively; wherein the upper and lower surface cameras shoot the same cross-sectional position, and the left and right surface cameras shoot the same cross-sectional position, and the interval between the two cross-sectional positions is greater than a certain preset distance, and the laser velocimeter outputs pulse triggers for collection to obtain the point cloud data of the high-temperature continuous casting slab; wherein the placement position of the detection device is as follows: Figure 2 shown.

[0081] Optionally, the 3D camera is composed of a line laser transmitter and an area array camera, and adopts a split laser triangulation imaging solution. The laser line is perpendicular to the measured plane, and the area array camera is tilted at a fixed angle to shoot, and the angle is determined by the actual depth of field range and depth direction accuracy required to be shot; two different colors of lasers are used between adjacent cameras, and the cameras are equipped with color filters of corresponding bands to eliminate interference when the camera extracts the laser line; wherein the two different colors are blue and green, respectively.

[0082] The area array camera has a fixed tilt angle of 40°; blue and green lasers are used between adjacent cameras and the cameras are equipped with color filters of corresponding bands.

[0083] Among them, the laser velocimeter output pulse trigger is used for collection, which can ensure that each row of data has the same resolution, and the upper and lower surface data and the left and right side data are matched and aligned through the length direction position.

[0084] In a feasible implementation, Figure 3 The width of the slab to be inspected ranges from 1800mm to 2200mm, and the thickness is 200mm. Four 3D cameras are used on the upper and lower surfaces to cover the entire surface area, and the resolution of each camera is 2560 1024, the resolution accuracy of the 3D camera in the z direction is designed to be 0.25mm; among them, the cameras in the upper table are numbered 1, 2, 3, and 4, the cameras in the lower table are numbered 5, 6, 7, and 8, the camera on the left is numbered 9, and the camera on the right is numbered 10. The green laser line used by the cameras numbered 1, 3, 5, 7, and 10, and the blue laser line used by the cameras numbered 2, 4, 7, 8, and 9 can offset the interference of the laser line extraction when the adjacent cameras shoot the repeated areas.

[0085] S2. Use a calibration block to calibrate the point cloud data of the high-temperature continuous casting slab to obtain calibrated point cloud data.

[0086] In a feasible implementation, the point cloud data output by the 3D camera is calibrated to a unified coordinate system through a calibration block. The calibration parameters include the translation value of the point cloud data in the x direction and the translation value in the z direction, as well as the rotation angle along the y direction. The values ​​converted by the calibration parameters can be used to calculate the length, width, thickness, curvature and T value of the slab.

[0087] S3. Calculate the actual depth value based on the calibrated point cloud data; process the actual depth value through rendering to obtain a pseudo-color image, which is input into the pre-trained Yolov5 target defect detection model for recognition to obtain the defect location and defect category.

[0088] Among them, rendering processing is an existing technology for processing images, so the specific implementation process is not described in the present invention.

[0089] Optionally, S3 calculates the actual depth value according to the calibrated point cloud data, including:

[0090] S31, extracting the first row of valid data from the calibrated point cloud data matrix, splicing the first row of point cloud data and row coordinates into a two-dimensional data matrix; using the RANSAC algorithm to fit the data into a straight line, determining the inclination angle of the straight line, and constructing a rotation matrix;

[0091] Among them, the RANSAC algorithm is an iterative algorithm used to estimate from a set of data containing outliers. It belongs to conventional technical means, and the specific implementation process is not described in the present invention.

[0092] The expression of the rotation matrix can be expressed by the following formula (1):

[0093] (1)

[0094] in, represents the rotation matrix; Indicates the inclination angle of the line.

[0095] S32, translate the first row of valid data along the x direction to the center position, perform angle rotation to make the z direction data close to the horizontal, and then translate it in the opposite direction along the x direction to obtain a new matrix, and use the new matrix as the target point cloud;

[0096] The new matrix can be expressed by the following formula (2):

[0097] (2)

[0098] in, represents the new matrix; Indicates that the first row of point cloud data and row coordinates are spliced ​​into a two-dimensional data matrix; Indicates the first row of point cloud data; represents the row coordinate matrix; Represents the row coordinate matrix The mean of .

[0099] S33, along the row direction of the point cloud, take the next row of data, splice the corresponding point cloud data with the row coordinates to form a two-dimensional data matrix, align the two-dimensional data matrix with the target point cloud through the ICP algorithm, obtain a rigid body matrix including rotation and translation through calculation, and transform the two-dimensional data matrix using the obtained rigid body matrix to obtain a transformed point cloud;

[0100] Among them, the ICP (Iterative Closest Point, ICP) algorithm is a classic algorithm for point cloud data registration. It gradually optimizes the correspondence between two point clouds in an iterative manner to minimize the difference between them. The specific implementation process is not described in this invention.

[0101] The converted point cloud can be expressed by the following formula (3):

[0102] (3)

[0103] in, represents the transformed point cloud; Represents the corresponding point cloud data With row coordinates Splicing to form a two-dimensional data matrix; The rigid body matrix representing the rotation; Represents the rigid body matrix of translation; i represents the index value in the row direction of the point cloud data.

[0104] S34, taking the converted point cloud as the target point cloud, repeating steps S33-S34 until all the point cloud data collected by the camera are processed and combined row by row to form new point cloud data;

[0105] S35, taking the median of the new point cloud data as the reference point, comparing the new point cloud data with the median and taking the difference as the actual depth value of the coordinate position of each point cloud data.

[0106] S4. Obtain defect information on the surface of the high-temperature continuous casting slab according to the defect position and defect category.

[0107] Among them, according to the defect location and defect category, the maximum depth and average depth of the defect area are extracted as the representation of the severity of the defect.

[0108] S5. An auxiliary detection method is adopted to obtain a grayscale image of the high-temperature continuous casting slab by adding an area array camera to shoot the upper and lower surfaces of the high-temperature continuous casting slab; the grayscale image is input into a pre-trained Yolov5 target defect detection model for recognition to obtain the slight defect information on the surface of the high-temperature continuous casting slab.

[0109] Optionally, the specific implementation process of S5 may include S51-S55:

[0110] S51, perform auxiliary detection on the slight defects on the upper and lower surfaces by adding an additional area array camera to shoot the upper and lower surfaces of the slab at an angle less than 15 degrees to obtain a grayscale image;

[0111] S52, preprocessing the grayscale image to obtain a preprocessed grayscale image;

[0112] In a feasible implementation, the preprocessing implementation process includes: restoring the original shape of the slab through perspective transformation of the initial image captured by the camera, and perspective transforming the captured trapezoidal area into a rectangular area, which is a part of the area in the length direction of the slab and includes the illumination area of ​​the two light sources.

[0113] S53, inputting the preprocessed grayscale image into the pre-trained Yolov5 target defect detection model for recognition, and obtaining the defect pixel coordinates;

[0114] S54, obtaining the position coordinates and size of the defect on the actual slab according to the defect pixel coordinates and the mapping relationship between the image coordinate system and the world coordinate system;

[0115] S55. According to the actual world coordinates of the defect, the maximum depth and the average depth are extracted at the corresponding positions in the actual depth data as representations of the severity of the defect, and the minor defect information on the surface of the high-temperature continuous casting slab is obtained.

[0116] The actual depth data is obtained through the collection of 3D cameras. Since the auxiliary detection is image data collected by the area array camera, the depth cannot be measured. The maximum depth and average depth are extracted through the actual depth data.

[0117] Optionally, the auxiliary detection process includes:

[0118] Collecting slab images according to the auxiliary detection and acquisition device;

[0119] In a feasible implementation, Figure 4 A schematic diagram of an auxiliary image acquisition device provided by an embodiment of the present invention along a top view of a roller conveyor; Figure 5A schematic diagram of an auxiliary image acquisition device provided in an embodiment of the present invention along the running direction of the roller; wherein, designing and installing an auxiliary detection device comprises: installing a bracket perpendicular to the ground at the side edge position of the running roller of the slab production line, and installing two sets of image acquisition devices on the bracket, which are respectively used to capture image data of the upper surface and the lower surface of the slab; the image acquisition device comprises two point light source emitters and a color array camera, the light source is converted into parallel light after being reflected by a concave mirror, and forms a trapezoidal lighting area after irradiating the slab, the shooting range of the array camera is larger than the lighting area formed by the two light sources, the camera and the light source are at the same horizontal height, the two light sources are distributed on both sides of the camera, and the angle α with the camera is in the range of 15°~30°; the two sets of image acquisition devices are respectively an upper surface acquisition device and a lower surface acquisition device, wherein the installation height of the upper surface acquisition device is higher than the upper surface of the slab, and the installation height of the lower surface acquisition device is lower than the lower surface of the slab, and the shooting direction of the acquisition device and the corresponding surface The angle β ranges from 5°~15°.

[0120] Among them, the initial image captured by the camera needs to undergo perspective transformation to restore the original shape of the slab, and the perspective of the captured trapezoidal area is converted into a rectangular area. The rectangular area is a part of the area in the length direction of the slab, which contains the illumination area of ​​the two light sources; among them, the pre-trained Yolov5 target defect detection model is responsible for defect detection of the rectangular area image.

[0121] According to the acquired slab images, the actual movement distance of the slab between two adjacent images and the world coordinates corresponding to the pixel coordinate origin on the images are calculated;

[0122] The actual movement distance of the slab between two adjacent images captured by the camera can be expressed by the following formula (4):

[0123] (4)

[0124] in, It represents the actual movement distance of the slab between two adjacent images captured by the camera; v represents the running speed of the slab; and F represents the acquisition frame rate of the camera.

[0125] According to the actual movement distance of the slab and the world coordinates corresponding to the pixel coordinate origin on the image, the world coordinates corresponding to the pixel coordinate origin on all the collected images are calculated;

[0126] The images captured by the camera are processed in sequence. After the pixel coordinates of the slab head in the y direction are identified, the slab head is taken as the origin in the world coordinates to obtain the world coordinates corresponding to the origin of the pixel coordinates on the current image. The world coordinates corresponding to the origin of the pixel coordinates in the y direction on the current image can be expressed by the following formula (5):

[0127] (5)

[0128] Where, f represents the pixel resolution of the image, which is obtained by field calibration; Indicates the pixel coordinate of the slab head in the y direction; Indicates the world coordinates corresponding to the origin of the pixel coordinates on the current image.

[0129] The world coordinates corresponding to the origin of the pixel coordinates in the y direction on the subsequently acquired image can be expressed by the following formula (6):

[0130] (6)

[0131] in, The world coordinates corresponding to the origin of the pixel coordinates on the subsequently acquired image; i represents the image index after the slab head is determined.

[0132] According to the world coordinates corresponding to the origin of the pixel coordinates on the image, the position coordinates of the defect in the length direction and the width direction of the slab are calculated.

[0133] Among them, from the time the slab head image is recognized to the beginning, each image is subjected to defect detection, and the position coordinates of the defects are output through the pre-trained Yolov5 target defect detection model. The position coordinates of the defects in the length direction of the slab can be expressed by the following formula (7):

[0134] (7)

[0135] in, Indicates the pixel coordinates of the defect in the image height direction; Indicates the position coordinates of the defect in the length direction of the slab.

[0136] The position coordinates of the defect in the width direction of the slab can be expressed by the following formula (8):

[0137] (8)

[0138] in, Indicates the pixel coordinates of the defect in the width direction of the image; Indicates the position coordinates of the defect in the width direction of the slab.

[0139] S6. Obtain final defect information on the surface of the high-temperature continuous casting slab according to the defect information on the surface of the high-temperature continuous casting slab and the slight defect information on the surface of the high-temperature continuous casting slab.

[0140] Optionally, the final defect information on the surface of the high temperature continuous casting slab includes: defect type, defect location, defect length, defect width, defect area and defect depth.

[0141] The embodiment of the present invention first obtains point cloud data of a high-temperature continuous casting slab; uses a calibration block to calibrate the point cloud data of the high-temperature continuous casting slab to obtain calibrated point cloud data; secondly, according to the calibrated point cloud data, calculates the actual depth value; the actual depth value is processed by rendering to obtain a pseudo-color image, which is input into a pre-trained Yolov5 target defect detection model for identification to obtain a defect position and a defect category; according to the defect position and the defect category, defect information on the surface of the high-temperature continuous casting slab is obtained; an auxiliary detection method is adopted to obtain a grayscale image of the high-temperature continuous casting slab by adding an array camera to shoot the upper and lower surfaces of the high-temperature continuous casting slab; the grayscale image is input into a pre-trained Yolov5 target defect detection model for identification to obtain slight defect information on the surface of the high-temperature continuous casting slab; finally, according to the defect information on the surface of the high-temperature continuous casting slab and the slight defect information on the surface of the high-temperature continuous casting slab, the final defect information on the surface of the high-temperature continuous casting slab is obtained.

[0142] The embodiment of the present invention collects point cloud data of the upper surface, lower surface, left side and right side of the slab by arranging multiple groups of 3D cameras in the cross-sectional direction of the slab, and after calibration and conversion of the point cloud data, the length, thickness, width, curvature and T value of the slab can be calculated, and the point cloud data can be further processed to realize the representation of the relative depth of the surface, and the position and depth of the defect can be detected by using the relative depth data. In addition, in order to increase the detection capability of minor defects, upper and lower surface auxiliary image acquisition devices are installed on the edge side of the roller, and the acquisition camera and the corresponding light source are designed to be at a lower angle. The viewing angle can significantly enhance the defect characteristics with depth on the surface of the slab, thereby enhancing the detection capability of defects; the embodiment of the present invention can simultaneously realize the slab size measurement and defect detection tasks through point cloud data analysis, and by adding auxiliary detection means, simulating the actual manual detection method, increasing the detection capability of minor defects, and realizing comprehensive detection, thereby achieving the purpose of replacing manual detection of slab surface defects, overcoming the limitations of traditional vertical viewing angle cameras in detecting minor defects, ensuring the quality of slabs, reducing production losses, meeting the quality detection needs of enterprises, and providing a more comprehensive and effective solution for continuous casting slab surface defect detection.

[0143] Figure 6 The block diagram of a high-temperature continuous casting slab three-dimensional defect detection and size measurement device according to an exemplary embodiment is shown. The device is used in a high-temperature continuous casting slab three-dimensional defect detection and size measurement method. Figure 6 The device includes a first acquisition unit 310, a calibration unit 320, a recognition and calculation unit 330, a second acquisition unit 340, an auxiliary detection unit 350 and a third acquisition unit 360. Wherein:

[0144] A first acquisition unit 310 is used to acquire point cloud data of a high temperature continuous casting slab;

[0145] A calibration unit 320, configured to calibrate the point cloud data of the high temperature continuous casting slab using a calibration block to obtain calibrated point cloud data;

[0146] The recognition and calculation unit 330 is used to calculate the actual depth value according to the calibrated point cloud data; the actual depth value is processed by rendering to obtain a pseudo-color image, which is input into a pre-trained Yolov5 target defect detection model for recognition to obtain the defect position and defect category;

[0147] A second acquisition unit 340 is used to obtain defect information on the surface of the high-temperature continuous casting slab according to the defect position and defect type;

[0148] The auxiliary detection unit 350 is used to adopt an auxiliary detection method, by adding an area array camera to shoot the upper and lower surfaces of the high-temperature continuous casting slab, to obtain a grayscale image of the high-temperature continuous casting slab; the grayscale image is input into the pre-trained Yolov5 target defect detection model for recognition, to obtain slight defect information on the surface of the high-temperature continuous casting slab;

[0149] The third acquisition unit 360 is used to obtain final defect information on the surface of the high-temperature continuous casting slab according to the defect information on the surface of the high-temperature continuous casting slab and the slight defect information on the surface of the high-temperature continuous casting slab.

[0150] Optionally, the first acquiring unit 310 is configured to:

[0151] In the cross-sectional direction of the high-temperature continuous casting slab, multiple groups of 3D cameras are arranged around to collect point cloud data of the upper surface, lower surface, left side and right side respectively; among them, the upper and lower surface cameras shoot the same cross-sectional position, the left and right surface cameras shoot the same cross-sectional position, and the interval between the two cross-sectional positions is greater than a certain preset distance. The laser velocimeter output pulse trigger is used for collection to obtain the point cloud data of the high-temperature continuous casting slab.

[0152] Optionally, the 3D camera is composed of a line laser transmitter and an area array camera, and adopts a split laser triangulation imaging solution. The laser line is perpendicular to the measured plane, and the area array camera is tilted at a fixed angle to shoot, and the angle is determined by the actual depth of field range and depth direction accuracy required to be shot; two different colors of lasers are used between adjacent cameras, and the cameras are equipped with color filters of corresponding bands to eliminate interference when the camera extracts the laser line; wherein the two different colors are blue and green, respectively.

[0153] Optionally, the identification and calculation unit 330 is used to:

[0154] Extract the first row of valid data from the calibrated point cloud data matrix, and splice the first row of point cloud data and row coordinates into a two-dimensional data matrix; use the RANSAC algorithm to fit the data into a straight line, determine the inclination angle of the straight line, and construct a rotation matrix;

[0155] The first row of valid data is translated to the center along the x direction, and the angle rotation is performed to make the z direction data close to the horizontal, and then translated in the opposite direction along the x direction to obtain a new matrix, and the new matrix is ​​used as the target point cloud;

[0156] Take the next row of data along the row direction of the point cloud, splice the corresponding point cloud data with the row coordinates to form a two-dimensional data matrix, align the two-dimensional data matrix with the target point cloud through the ICP algorithm, calculate the rigid body matrix including rotation and translation, and use the obtained rigid body matrix to transform the two-dimensional data matrix to obtain the transformed point cloud;

[0157] The converted point cloud is used as the target point cloud, and steps S33-S34 are repeated until all the point cloud data collected by the camera are processed and combined row by row to form new point cloud data;

[0158] The median of the new point cloud data is taken as the reference point, and the difference between the new point cloud data and the median is taken as the actual depth value of the coordinate position of each point cloud data.

[0159] Optionally, the auxiliary detection unit 350 is used to:

[0160] To assist in the detection of minor defects on the upper and lower surfaces, an additional area array camera is added to capture the upper and lower surfaces of the slab at an angle of less than 15 degrees to obtain a grayscale image.

[0161] Preprocessing the grayscale image to obtain a preprocessed grayscale image;

[0162] The preprocessed grayscale image is input into the pre-trained Yolov5 target defect detection model for recognition to obtain the defect pixel coordinates;

[0163] According to the defect pixel coordinates and the mapping relationship between the image coordinate system and the world coordinate system, the position coordinates and size of the defect on the actual slab are obtained;

[0164] According to the actual world coordinates of the defect, the maximum depth and average depth are extracted at the corresponding position in the actual depth data as the representation of the severity of the defect, and the minor defect information on the surface of the high-temperature continuous casting slab is obtained.

[0165] Optionally, the process of the auxiliary detection method includes:

[0166] Collecting slab images according to the auxiliary detection and acquisition device;

[0167] According to the acquired slab images, the actual movement distance of the slab between two adjacent images and the world coordinates corresponding to the pixel coordinate origin on the images are calculated;

[0168] According to the actual movement distance of the slab and the world coordinates corresponding to the pixel coordinate origin on the image, the world coordinates corresponding to the pixel coordinate origin on all the collected images are calculated;

[0169] According to the world coordinates corresponding to the origin of the pixel coordinates on the image, the position coordinates of the defect in the length direction and the width direction of the slab are calculated.

[0170] Optionally, the final defect information on the surface of the high-temperature continuous casting slab includes: defect type, defect position, defect length, defect width, defect area and defect depth.

[0171] The embodiment of the present invention first obtains point cloud data of a high-temperature continuous casting slab; uses a calibration block to calibrate the point cloud data of the high-temperature continuous casting slab to obtain calibrated point cloud data; secondly, according to the calibrated point cloud data, calculates the actual depth value; the actual depth value is processed by rendering to obtain a pseudo-color image, which is input into a pre-trained Yolov5 target defect detection model for identification to obtain a defect position and a defect category; according to the defect position and the defect category, defect information on the surface of the high-temperature continuous casting slab is obtained; an auxiliary detection method is adopted to obtain a grayscale image of the high-temperature continuous casting slab by adding an array camera to shoot the upper and lower surfaces of the high-temperature continuous casting slab; the grayscale image is input into a pre-trained Yolov5 target defect detection model for identification to obtain slight defect information on the surface of the high-temperature continuous casting slab; finally, according to the defect information on the surface of the high-temperature continuous casting slab and the slight defect information on the surface of the high-temperature continuous casting slab, the final defect information on the surface of the high-temperature continuous casting slab is obtained.

[0172] The embodiment of the present invention collects point cloud data of the upper surface, lower surface, left side and right side of the slab by arranging multiple groups of 3D cameras in the cross-sectional direction of the slab, and after calibration and conversion of the point cloud data, the length, thickness, width, curvature and T value of the slab can be calculated, and the point cloud data can be further processed to realize the representation of the relative depth of the surface, and the position and depth of the defect can be detected by using the relative depth data. In addition, in order to increase the detection capability of minor defects, upper and lower surface auxiliary image acquisition devices are installed on the edge side of the roller, and the acquisition camera and the corresponding light source are designed to be at a lower angle. The viewing angle can significantly enhance the defect characteristics with depth on the surface of the slab, thereby enhancing the detection capability of defects; the embodiment of the present invention can simultaneously realize the slab size measurement and defect detection tasks through point cloud data analysis, and by adding auxiliary detection means, simulating the actual manual detection method, increasing the detection capability of minor defects, and realizing comprehensive detection, thereby achieving the purpose of replacing manual detection of slab surface defects, overcoming the limitations of traditional vertical viewing angle cameras in detecting minor defects, ensuring the quality of slabs, reducing production losses, meeting the quality detection needs of enterprises, and providing a more comprehensive and effective solution for continuous casting slab surface defect detection.

[0173] Figure 7 is a schematic diagram of the structure of a high-temperature continuous casting slab three-dimensional defect detection and size measurement device provided by an embodiment of the present invention, such as Figure 7 As shown, the high temperature continuous casting slab three-dimensional defect detection and size measurement equipment may include the above Figure 6 The high temperature continuous casting slab three-dimensional defect detection and size measurement device shown. Optionally, the high temperature continuous casting slab three-dimensional defect detection and size measurement device 410 may include a first processor 2001.

[0174] Optionally, the high temperature continuous casting slab three-dimensional defect detection and size measurement device 410 may further include a memory 2002 and a transceiver 2003 .

[0175] The first processor 2001, the memory 2002 and the transceiver 2003 may be connected via a communication bus.

[0176] Combine the following Figure 7 The components of the high temperature continuous casting slab three-dimensional defect detection and dimension measurement device 410 are specifically introduced:

[0177] The first processor 2001 is the control center of the high temperature continuous casting slab three-dimensional defect detection and dimension measurement device 410, which can be a processor or a general term for multiple processing elements. For example, the first processor 2001 is one or more central processing units (CPUs), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiment of the present invention, such as one or more microprocessors (digital signal processors, DSPs), or one or more field programmable gate arrays (field programmable gate arrays, FPGAs).

[0178] Optionally, the first processor 2001 can perform various functions of the high-temperature continuous casting slab three-dimensional defect detection and size measurement device 410 by running or executing a software program stored in the memory 2002 and calling data stored in the memory 2002.

[0179] In a specific implementation, as an embodiment, the first processor 2001 may include one or more CPUs, such as Figure 7 CPU0 and CPU1 are shown in FIG.

[0180] In a specific implementation, as an embodiment, the high temperature continuous casting slab three-dimensional defect detection and size measurement device 410 may also include multiple processors, such as Figure 7 The first processor 2001 and the second processor 2004 are shown in FIG. Each of these processors may be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). The processor here may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0181] The memory 2002 is used to store the software program for executing the solution of the present invention, and the first processor 2001 controls the execution. The specific implementation method can refer to the above method embodiment and will not be elaborated here.

[0182] Optionally, the memory 2002 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory 2002 may be integrated with the first processor 2001, or may exist independently, and may be accessed through the interface circuit ( Figure 7 (not shown) is coupled to the first processor 2001, which is not specifically limited in this embodiment of the present invention.

[0183] The transceiver 2003 is used to communicate with a network device or a terminal device.

[0184] Optionally, the transceiver 2003 may include a receiver and a transmitter ( Figure 7 The receiver is used to implement a receiving function, and the transmitter is used to implement a sending function.

[0185] Optionally, the transceiver 2003 may be integrated with the first processor 2001, or may exist independently, and may be connected to the first processor 2001 through the interface circuit ( Figure 7 (not shown) is coupled to the first processor 2001, which is not specifically limited in this embodiment of the present invention.

[0186] It should be noted that Figure 7 The structure of the high-temperature continuous casting slab three-dimensional defect detection and size measurement device 410 shown in the figure does not constitute a limitation on the router. The actual knowledge structure recognition device may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0187] In addition, the technical effects of the high-temperature continuous casting slab three-dimensional defect detection and size measurement equipment 410 can refer to the technical effects of the high-temperature continuous casting slab three-dimensional defect detection and size measurement method described in the above method embodiment, and will not be elaborated here.

[0188] It should be understood that the first processor 2001 in the embodiment of the present invention may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0189] It should also be understood that the memory in the embodiments of the present invention may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic random access memory (DRAM), synchronous DRAM (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0190] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a tape), an optical medium (for example, a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0191] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0192] In the present invention, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0193] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0194] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0195] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described equipment, devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be elaborated here.

[0196] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0197] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0198] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0199] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.

[0200] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for three-dimensional defect detection and size measurement of high-temperature continuous casting slabs, characterized in that: The method comprises: S1. Obtaining point cloud data of high temperature continuous casting slab; S2. Calibrate the point cloud data of the high-temperature continuous casting slab using a calibration block to obtain calibrated point cloud data; S3, calculating the actual depth value according to the calibrated point cloud data; the actual depth value is processed by rendering to obtain a pseudo-color image, which is input into a pre-trained Yolov5 target defect detection model for recognition to obtain the defect location and defect category; S4. Obtaining defect information on the surface of the high-temperature continuous casting slab according to the defect position and defect category; S5, adopting an auxiliary detection method, by adding an area array camera to shoot the upper and lower surfaces of the high-temperature continuous casting slab, to obtain a grayscale image of the high-temperature continuous casting slab; inputting the grayscale image into the pre-trained Yolov5 target defect detection model for recognition, to obtain slight defect information on the surface of the high-temperature continuous casting slab; S6. Obtain final defect information on the surface of the high-temperature continuous casting slab according to the defect information on the surface of the high-temperature continuous casting slab and the slight defect information on the surface of the high-temperature continuous casting slab.

2. The high temperature continuous casting slab three-dimensional defect detection and size measurement method according to claim 1, characterized in that: The step S1 of acquiring point cloud data of a high temperature continuous casting slab includes: In the cross-sectional direction of the high-temperature continuous casting slab, multiple groups of 3D cameras are arranged around to collect point cloud data of the upper surface, lower surface, left side and right side respectively; among them, the upper and lower surface cameras shoot the same cross-sectional position, and the left and right surface cameras shoot the same cross-sectional position. The interval between the two cross-sectional positions is greater than a certain preset distance. The laser velocimeter output pulse trigger is used for collection to obtain the point cloud data of the high-temperature continuous casting slab.

3. The high temperature continuous casting slab three-dimensional defect detection and size measurement method according to claim 2, characterized in that: The 3D camera is composed of a line laser transmitter and an array camera, and adopts a split laser triangulation imaging solution. The laser line is perpendicular to the measured plane, and the array camera is tilted at a fixed angle to shoot, and the angle is determined by the actual depth of field range and depth direction accuracy required for shooting; two different colors of lasers are used between adjacent cameras, and the cameras are equipped with color filters of corresponding bands to eliminate interference when the camera extracts the laser line; wherein the two different colors are blue and green, respectively.

4. The method for three-dimensional defect detection and size measurement of high temperature continuous casting slab according to claim 1, characterized in that: The step S3 calculates the actual depth value according to the calibrated point cloud data, including: S31, extracting the first row of valid data from the calibrated point cloud data matrix, splicing the first row of point cloud data and row coordinates into a two-dimensional data matrix; using the RANSAC algorithm to fit the data into a straight line, determining the inclination angle of the straight line, and constructing a rotation matrix; S32, translate the first row of valid data along the x direction to the center position, perform angle rotation to make the z direction data close to the horizontal, and then translate it in the opposite direction along the x direction to obtain a new matrix, and use the new matrix as the target point cloud; S33, along the row direction of the point cloud, take the next row of data, splice the corresponding point cloud data with the row coordinates to form a two-dimensional data matrix, align the two-dimensional data matrix with the target point cloud through the ICP algorithm, calculate the rigid body matrix including rotation and translation, and transform the two-dimensional data matrix using the obtained rigid body matrix to obtain a transformed point cloud; S34, taking the converted point cloud as the target point cloud, repeating steps S33-S34 until all the point cloud data collected by the camera are processed and combined row by row to form new point cloud data; S35, taking the median of the new point cloud data as the reference point, comparing the new point cloud data with the median and taking the difference as the actual depth value of the coordinate position of each point cloud data.

5. The method for three-dimensional defect detection and size measurement of high temperature continuous casting slab according to claim 1, characterized in that: The S5 adopts an auxiliary detection method, by adding an area array camera to shoot the upper and lower surfaces of the high-temperature continuous casting slab, to obtain a grayscale image of the high-temperature continuous casting slab; the grayscale image is input into the pre-trained Yolov5 target defect detection model for recognition, to obtain slight defect information on the surface of the high-temperature continuous casting slab, including: S51, perform auxiliary detection on the slight defects on the upper and lower surfaces by adding an additional area array camera to shoot the upper and lower surfaces of the slab at an angle less than 15 degrees to obtain a grayscale image; S52, preprocessing the grayscale image to obtain a preprocessed image; S53, inputting the preprocessed image into the pre-trained Yolov5 target defect detection model for recognition, and obtaining the defect pixel coordinates; S54, obtaining the position coordinates and size of the defect on the actual slab according to the defect pixel coordinates and the mapping relationship between the image coordinate system and the world coordinate system; S55. According to the actual world coordinates of the defect, the maximum depth and the average depth are extracted at the corresponding positions in the actual depth data as representations of the severity of the defect, and the minor defect information on the surface of the high-temperature continuous casting slab is obtained.

6. The method for three-dimensional defect detection and size measurement of high temperature continuous casting slab according to claim 5, characterized in that: The process of the auxiliary detection method includes: Collecting slab images according to the auxiliary detection and acquisition device; According to the acquired slab images, the actual movement distance of the slab between two adjacent images and the world coordinates corresponding to the pixel coordinate origin on the images are calculated; According to the actual movement distance of the slab and the world coordinates corresponding to the pixel coordinate origin on the image, the world coordinates corresponding to the pixel coordinate origin on all the collected images are calculated; According to the world coordinates corresponding to the origin of the pixel coordinates on the image, the position coordinates of the defect in the length direction and the width direction of the slab are calculated.

7. The method for three-dimensional defect detection and size measurement of high temperature continuous casting slab according to claim 1, characterized in that: The final defect information of the surface of the high temperature continuous casting slab in S6 includes: defect type, defect position, defect length, defect width, defect area and defect depth.

8. A high-temperature continuous casting slab three-dimensional defect detection and size measurement device, the high-temperature continuous casting slab three-dimensional defect detection and size measurement device is used to implement the high-temperature continuous casting slab three-dimensional defect detection and size measurement method as claimed in any one of claims 1 to 7, characterized in that: The device comprises: A first acquisition unit is used to acquire point cloud data of a high-temperature continuous casting slab; A calibration unit, used for calibrating the point cloud data of the high-temperature continuous casting slab using a calibration block to obtain calibrated point cloud data; The recognition and calculation unit is used to calculate the actual depth value according to the calibrated point cloud data; the actual depth value is processed by rendering to obtain a pseudo-color image which is input into a pre-trained Yolov5 target defect detection model for recognition to obtain the defect position and defect category; A second acquisition unit is used to obtain defect information on the surface of the high-temperature continuous casting slab according to the defect position and defect category; The auxiliary detection unit is used to adopt an auxiliary detection method, by adding an area array camera to shoot the upper and lower surfaces of the high-temperature continuous casting slab, to obtain a grayscale image of the high-temperature continuous casting slab; the grayscale image is input into the pre-trained Yolov5 target defect detection model for recognition, to obtain slight defect information on the surface of the high-temperature continuous casting slab; The third acquisition unit is used to obtain final defect information on the surface of the high-temperature continuous casting slab according to the defect information on the surface of the high-temperature continuous casting slab and the slight defect information on the surface of the high-temperature continuous casting slab.

9. A high temperature continuous casting slab three-dimensional defect detection and size measurement device, characterized in that: The high temperature continuous casting slab three-dimensional defect detection and size measurement equipment comprises: processor; A memory having computer-readable instructions stored thereon, wherein when the computer-readable instructions are executed by the processor, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores program codes, which can be called by a processor to execute the method according to any one of claims 1 to 7.

Images