A method for three-dimensional measurement of refractory bricks based on principal component analysis
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HEBEI UNIV OF TECH
- Filing Date
- 2023-04-13
- Publication Date
- 2026-06-23
Smart Images
Figure CN116576777B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of machine vision technology, specifically to a three-dimensional measurement method for refractory bricks based on principal component analysis. Background Technology
[0002] With the continuous development of industry, construction, and related sectors, my country has become the world's largest producer, consumer, and exporter of refractory materials, holding a significant position in the international refractory materials field. Refractory bricks are refractory materials with specific shapes and dimensions, made from refractory clay and other refractory raw materials. They can withstand various mechanical actions and physicochemical changes at high temperatures and are widely used in metallurgy, chemical industry, petroleum, power generation, and other industrial fields. The high-temperature operating environment places special demands on the quality of refractory bricks; their dimensional errors, surface defects, and internal structure must all meet corresponding quality standards. However, for a long time, most refractory brick manufacturers in my country have relied on simple tools such as measuring tapes for manual inspection, resulting in low efficiency, low precision, susceptibility to subjective bias, and difficulty in establishing unified evaluation standards. In recent years, with the rapid development of machine vision technology, methods for applying machine vision to the surface quality inspection of refractory bricks have gradually increased. However, most of these methods focus on feature extraction and classification of two-dimensional images, neglecting the measurement and detection of three-dimensional parameters and defects such as length, width, height, flatness, and missing corners. However, directly using and processing the three-dimensional point cloud data of refractory bricks presents problems such as excessive computation, low efficiency, and high cost of the detection system. Summary of the Invention
[0003] To address the shortcomings of existing technologies, this invention provides a three-dimensional measurement method for refractory bricks based on principal component analysis, which can accurately and quickly detect the three-dimensional parameters of refractory bricks.
[0004] This invention provides a three-dimensional measurement method for refractory bricks based on principal component analysis, comprising:
[0005] Obtain point cloud data of the refractory brick to be tested;
[0006] The point cloud data is filtered according to height to remove the point cloud data outside the refractory brick to be tested, and the point cloud data on the surface of the refractory brick to be tested is filtered to obtain the processed upper surface point cloud.
[0007] The point cloud on the upper surface is subjected to dimensionality reduction processing using principal component analysis to transform the three-dimensional point cloud into a two-dimensional point cloud.
[0008] The length and width of the refractory brick to be measured are calculated based on the two-dimensional point cloud computing.
[0009] A perspective projection is performed on the 2D point cloud and converted into a grayscale image. Corner detection is then performed in the grayscale image to obtain the number of corner points.
[0010] The three-dimensional vertex of the refractory brick corresponding to the image corner point is found based on the number of corner points, the projection relationship between the image corner points and the two-dimensional point cloud, and the correspondence between the two-dimensional point cloud and the three-dimensional point cloud in the principal component analysis method. The thickness of the refractory brick to be measured is calculated based on the distance from the three-dimensional vertex to the reference plane.
[0011] Output the length, width, and thickness of the refractory brick to be tested.
[0012] Optionally, the method further includes: determining whether the number of corner points is greater than 4;
[0013] If the value is greater than 4, it is determined that the refractory brick under test has a missing corner;
[0014] If the value is less than or equal to 4, then the refractory brick under test is determined to have no missing corners.
[0015] Optionally, the method further includes: fitting a three-dimensional plane equation based on the three-dimensional vertices, calculating the distances from all three-dimensional points in the point cloud on the upper surface of the refractory brick to the fitted plane, calculating the mean and standard deviation of all the obtained distances, and obtaining the flatness of the refractory brick to be tested.
[0016] Optionally, a specific method for filtering the point cloud data based on height includes:
[0017] Filter points according to the set height range;
[0018] Perform connected component processing on the selected points;
[0019] Preserve point clouds with the same connected components and remove interfering point clouds.
[0020] Optionally, the specific method for performing data dimensionality reduction processing on the upper surface point cloud using principal component analysis includes: aligning the XOY plane of the coordinate system with the brick surface, and aligning the origin of the coordinate system with the center of the brick surface, while retaining the two principal components in the three-dimensional data.
[0021] Optionally, the specific method for calculating the length and width of the refractory brick to be measured based on the two-dimensional point cloud includes:
[0022] Find the maximum and minimum values of the x-coordinate in a 2D point cloud;
[0023] Find the maximum and minimum values of the ordinate in a 2D point cloud;
[0024] The length of the refractory brick to be measured is obtained by subtracting the minimum value of the x-axis from the maximum value of the x-axis.
[0025] The width of the refractory brick to be measured is obtained by subtracting the minimum value of the ordinate from the maximum value of the ordinate.
[0026] Optionally, the specific method for obtaining the point cloud data of the refractory brick to be tested includes:
[0027] Place the refractory brick to be tested on a reference surface, and place the 3D smart sensor at a preset height above the reference surface. Use the 3D smart sensor to collect point cloud data of the refractory brick to be tested.
[0028] Optionally, the preset height is 765mm.
[0029] Optionally, the set height range is 650mm-725m from the distance to the 3D sensor.
[0030] The beneficial effects of this invention are:
[0031] This invention provides a three-dimensional measurement method for refractory bricks based on principal component analysis. The method reduces the dimensionality of the three-dimensional point cloud data of refractory bricks, removes redundant information in the three-dimensional data, makes the data processing process simpler and more effective, and improves the data processing speed. This method can automatically detect the three-dimensional parameters of refractory bricks with high measurement accuracy and fast detection speed. Attached Figure Description
[0032] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.
[0033] Figure 1 The diagram shows a flowchart of a three-dimensional measurement method for refractory bricks based on principal component analysis provided in the first embodiment of the present invention.
[0034] Figure 2 A schematic diagram of the device structure for collecting point cloud data is shown in an embodiment of the present invention.
[0035] Figure 3 This is a schematic diagram of the point cloud on the upper surface of the refractory brick after filtering in an embodiment of the present invention.
[0036] Figure 4 This is a two-dimensional point cloud image of the point cloud data of the upper surface of the refractory brick in this embodiment of the invention.
[0037] Figure 5 This is a schematic diagram of corner detection using grayscale images in an embodiment of the present invention.
[0038] Labels in the diagram: 1: 3D intelligent sensor, 2: refractory brick to be tested, 3: experimental table. Detailed Implementation
[0039] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0040] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.
[0041] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.
[0042] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.
[0043] As used in this specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrases "if determined" or "if [described condition or event] is detected" may be interpreted, depending on the context, as "once determined," "in response to determination," "once [described condition or event] is detected," or "in response to detection of [described condition or event]."
[0044] It should be noted that, unless otherwise stated, the technical or scientific terms used in this application should have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.
[0045] Please see Figures 1-5 The first embodiment of the present invention provides a three-dimensional measurement method for refractory bricks based on principal component analysis, which includes the following steps:
[0046] Step S1: Obtain point cloud data of the refractory brick to be tested.
[0047] exist Figure 2In this experiment, the 3D smart sensor 1 is placed at a preset height above a reference surface, and the refractory brick 2 to be tested is placed on the reference surface. In this embodiment, the reference surface is the experimental table 3. The 3D smart sensor 1 is used to collect point cloud data of the refractory brick 2 to be tested.
[0048] Step S2: Filter the point cloud data collected in step S1 according to height, remove the point cloud data outside the refractory brick to be tested, and filter the point cloud on the surface of the refractory brick to be tested to obtain the processed upper surface point cloud.
[0049] Step S3: Principal component analysis (PCA) is used to reduce the dimensionality of the point cloud on the upper surface of the refractory brick obtained in Step S2. This ensures that the XOY plane of the coordinate system coincides with the brick surface, and the origin of the coordinate system coincides with the center of the brick surface. The two most important components of the 3D data are retained, thus transforming the 3D point cloud into a 2D point cloud. This reduces data complexity and improves data processing speed. The coordinate positions of the 2D point cloud image are adjusted so that the origin of the coordinate system coincides with the center of the brick surface, facilitating subsequent calculations.
[0050] Step S4: For the reduced point cloud obtained in step S3, find the maximum and minimum values of the horizontal and vertical coordinates, obtain the coordinates of the four vertices in the reduced point cloud, and obtain the length and width of the refractory brick to be tested by calculating the distance between the vertices.
[0051] Step S5: Perform perspective projection on the reduced point cloud obtained in step S3 to convert it into a grayscale image, and perform corner detection in the grayscale image to obtain the number of corners.
[0052] Step S6: Based on the corner detection results of the grayscale image in Step S5, the projection relationship between the image corners and the reduced point cloud, as well as the correspondence between the reduced point cloud and the three-dimensional point cloud in the PCA operation, is used to obtain the three-dimensional vertices of the refractory brick to be tested corresponding to the image corners. The thickness of the refractory brick to be tested is calculated based on the distance from the three-dimensional vertex to the reference plane.
[0053] Step S7: Fit the three-dimensional plane equation to the three-dimensional vertices in step S6, calculate the mean and variance of the distances from all three-dimensional points in the point cloud on the upper surface of the refractory brick to the fitted plane, and thus calculate the flatness of the refractory brick.
[0054] Step S8: Save and output the detection results.
[0055] In the above steps, the dimensionality of the refractory brick 3D point cloud data is reduced to remove redundant information in the 3D data, making the data processing process simpler and more effective, and improving the data processing speed. This method can automatically detect the 3D parameters of refractory bricks with high measurement accuracy and fast detection speed.
[0056] Furthermore, to detect whether refractory bricks have missing corners, the presence of missing corners can be determined by checking if the number of detected corner points is greater than four. Specific methods include:
[0057] If the number of corner points is greater than 4, the refractory brick under test is judged to have missing corners;
[0058] If the number of corner points is less than or equal to 4, the refractory brick under test is judged to have no missing corners.
[0059] Specifically, in step S1, when using a 3D smart sensor to collect point cloud data of the refractory brick to be tested, the 3D smart sensor is positioned 765mm away from the reference surface.
[0060] Specifically, in step S2, the specific steps for filtering the collected point cloud according to its height to obtain the point cloud on the surface of the refractory brick to be tested include:
[0061] After placing refractory bricks on the reference surface, points within 650mm-725mm of the 3D smart sensor are first selected. Connectivity processing is performed on the selected points. If the distance between two points is no more than 1mm, they are considered to be the same connected component. Interference point clouds from other connected components are removed.
[0062] In step S2, the specific method for filtering the point cloud data of the surface of the refractory brick to be tested is as follows: calculate the mean value M of the Z coordinate of all point clouds on the upper surface of the refractory brick to be tested, set the range (M-1.2, M+1.2), filter the point cloud, and retain the points within the range, with the unit being mm.
[0063] In step S3, the method of performing data dimensionality reduction processing on the point cloud of the upper surface of the refractory brick obtained in step S2 using principal component analysis specifically includes:
[0064] First, the point cloud data format for the upper surface of the refractory brick is as follows:
[0065]
[0066] In equation (1), It is a point cloud matrix composed of N three-dimensional data points on the upper surface of the refractory brick to be tested. Each point represents a 3D spatial information matrix S, where S is a 3D point cloud set composed of N feature elements. m In the diagram, each row represents a point; further, the zero-mean normalization of each point in the three dimensions is calculated using the following formula (2):
[0067]
[0068] In equation (2), The covariance matrix of the sample features is calculated using the following formula (3):
[0069]
[0070] In equation (3), Let C be the covariance matrix of the sample features; further calculate the covariance matrix C. m eigenvalues λ and eigenvectors υ λ Arrange the eigenvectors into a matrix by row according to their corresponding eigenvalues from largest to smallest, and take the first and second columns to form the transformation matrix K; then use the following formula (4) to calculate the dimensionality-reduced point cloud matrix:
[0071]
[0072] In step S4, the specific method for calculating the length and width of the refractory brick to be measured based on two-dimensional point computing includes:
[0073] Since the origin of the coordinate system in the two-dimensional plane is located at the center of the plane, it is very convenient to find the maximum and minimum values of the horizontal and vertical coordinates. Furthermore, the length and width of the refractory brick can be calculated using the distance formula between the two points. That is, if the point cloud matrix obtained by the dimension reduction in formula (4) is expressed in detail as formula (5):
[0074]
[0075] Therefore, the maximum and minimum values in the x' direction and the maximum and minimum values in the y' direction can be represented as x' max , x' min y' max y' min Therefore, the length and width of the refractory brick can be calculated according to formula (6):
[0076]
[0077] In step S5, the process of obtaining the grayscale image from the dimensionality-reduced point cloud matrix is as follows:
[0078] Set the camera's focal length f and pixel size d. u d v Principal point position c u c v After assigning the rotation matrix R and the translation vector t, the camera's projection matrix can be obtained as follows:
[0079]
[0080] If we assume that the z' coordinate of each point in the reduced-dimensional point cloud matrix is 0, then the camera projection matrix can be transformed into:
[0081] H' = [h1 h2 h4], (8)
[0082] Therefore, the correspondence between the dimensionality-reduced point cloud matrix and the image pixels can be established as follows:
[0083]
[0084] in, Let be the homogeneous coordinates of the image pixels. Let i be the homogeneous coordinates of the point cloud after dimensionality reduction, i = 1, 2, ..., N, where N is a natural number.
[0085] In step S5, the specific steps of Harris corner detection include:
[0086] First, the two-dimensional refractory brick image is converted to grayscale. Then, the gradients of the image in the u and v directions are calculated using the following formula (10):
[0087]
[0088] In equation (10), I is the image matrix, I u Let I be the gradient of the image in the u direction. v The gradient of the image in the v direction is further calculated using the following formula (11) to multiply the gradients in the two directions:
[0089]
[0090] Furthermore, the three are Gaussian weighted using the following formula (12) to generate the elements A, B, C of matrix M:
[0091]
[0092] In equation (12), M is the cross-correlation matrix, ω(u,v) is a Gaussian weighted function centered at point (u,v), and the corner response function R is further calculated using the following formula (13), and R values less than a certain threshold t are set to zero:
[0093] R = {R:det(M) - k trace(M)} 2 <t},(13)
[0094] Non-maximum suppression is applied to R, and the local maximum points obtained are the corner points of the image; whether there are missing corners is determined by whether the number of edge corner points is greater than 4.
[0095] Furthermore, in step S6, the specific steps for determining the three-dimensional vertices of the refractory brick based on the image corner points are as follows:
[0096] Assume the set of corner points in the image obtained through corner detection is m.corner The corresponding set of corner points in the reduced-dimensional point cloud is represented by S'. corner Therefore, the relationship between the two can be established according to formula (9) as follows:
[0097]
[0098] In equation (14), This is the homogeneous coordinate form of the set of corner points in the image. This represents the homogeneous coordinate form of the reduced-dimensional point cloud set. The reduced-dimensional point cloud corner set S' is obtained from formula (14). corner Then, the three-dimensional vertex set S of the refractory brick can be obtained further according to formulas (2) and (4). corner That is:
[0099]
[0100] In equation (15), K + The pseudo-inverse matrix of the transformation matrix K is represented by the symbol. This indicates that the elements in the corresponding rows of the matrix are added one by one.
[0101] In some embodiments, step S7 specifically includes the following steps:
[0102] First, based on the three-dimensional vertex coordinates in step S6, fit the three-dimensional plane equation;
[0103] Secondly, based on the point cloud of the upper surface of the refractory brick obtained in step S2, the distance from all three-dimensional points to the fitted plane is calculated using the distance formula from the point to the three-dimensional plane equation.
[0104] Next, calculate the mean and standard deviation of all distances obtained in the previous step. Points within three times the standard deviation are considered interiors, denoted as inliers, and points outside this range are considered outliers. The set of interior points is used to fit a plane equation, and based on this, the smoothness of the refractory brick surface is calculated, denoted as smoothness, which can be expressed as:
[0105]
[0106] Finally, determine whether the flatness value obtained from formula (16) exceeds the preset allowable value. If it does not exceed the allowable value, it is deemed qualified; otherwise, it is deemed unqualified.
[0107] In addition, in step S8, saving the test results means storing the length, width, height, whether there are missing corners, and flatness of the refractory bricks in the above steps into a database file; outputting the test results specifically means sending the test results to the receiving end according to the agreed protocol.
[0108] This invention employs principal component analysis to reduce the dimensionality of refractory brick 3D point cloud data obtained from a 3D smart sensor, removing redundant information and simplifying the data processing process, thereby increasing processing speed. The method provided in this invention offers high measurement accuracy, fast detection speed, and can measure the 3D geometric dimensions of various refractory bricks (including straight and wedge-shaped refractory bricks).
[0109] In summary, the 3D measurement method for refractory bricks based on principal component analysis provided by this invention uses a 3D intelligent sensor to collect point cloud data of refractory bricks. Principal component analysis is then used to reduce the data dimensionality, removing redundant information from the 3D point cloud data and simplifying the data processing steps. This method can accurately measure the length, width, and height of refractory bricks. Furthermore, it can determine whether refractory bricks have missing corners or unevenness, reducing the workload of workers while improving measurement accuracy and efficiency. This provides a new solution to the existing problems of 3D dimensional measurement and defect detection of refractory bricks.
[0110] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention 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 the present invention, and they should all be covered within the scope of the claims and specification of the present invention.
Claims
1. A method for three-dimensional measurement of refractory bricks based on principal component analysis, characterized by, include: Obtain point cloud data of the refractory brick to be tested; The point cloud data is filtered according to height to remove the point cloud data outside the refractory brick to be tested, and the point cloud data on the surface of the refractory brick to be tested is filtered to obtain the processed upper surface point cloud. The point cloud on the upper surface is subjected to dimensionality reduction processing using principal component analysis to transform the three-dimensional point cloud into a two-dimensional point cloud. The length and width of the refractory brick to be measured are calculated based on the two-dimensional point cloud computing. A perspective projection is performed on the 2D point cloud and converted into a grayscale image. Corner detection is then performed in the grayscale image to obtain the number of corner points. The three-dimensional vertex of the refractory brick corresponding to the image corner point is found based on the number of corner points, the projection relationship between the image corner points and the two-dimensional point cloud, and the correspondence between the two-dimensional point cloud and the three-dimensional point cloud in the principal component analysis method. The thickness of the refractory brick to be measured is calculated based on the distance from the three-dimensional vertex to the reference plane. Output the length, width, and thickness of the refractory brick to be tested; The method further includes: determining whether the number of corner points is greater than 4; If the value is greater than 4, it is determined that the refractory brick under test has a missing corner; If the value is less than or equal to 4, then the refractory brick under test is determined to have no missing corners. Based on the three-dimensional vertices, fit the three-dimensional plane equation, calculate the distance from all three-dimensional points in the point cloud on the upper surface of the refractory brick to the fitted plane, calculate the mean and standard deviation of all the distances, and obtain the flatness of the refractory brick to be tested.
2. The method of claim 1, wherein, Specific methods for filtering the point cloud data based on height include: Filter points according to the set height range; Perform connected component processing on the selected points; Preserve point clouds with the same connected components and remove interfering point clouds.
3. The method as described in claim 1, characterized in that, The specific method for dimensionality reduction processing of the point cloud on the upper surface using principal component analysis includes: aligning the XOY plane of the coordinate system with the brick surface, and aligning the origin of the coordinate system with the center of the brick surface, while retaining the two principal components in the three-dimensional data.
4. The method as described in claim 1, characterized in that, The specific method for calculating the length and width of the refractory brick to be measured based on the two-dimensional point cloud computing includes: Find the maximum and minimum values of the x-coordinate in a 2D point cloud; Find the maximum and minimum values of the ordinate in a 2D point cloud; The length of the refractory brick to be measured is obtained by subtracting the minimum value of the x-axis from the maximum value of the x-axis. The width of the refractory brick to be measured is obtained by subtracting the minimum value of the ordinate from the maximum value of the ordinate.
5. The method as described in claim 2, characterized in that, The specific method for obtaining the point cloud data of the refractory brick to be tested includes: Place the refractory brick to be tested on a reference surface, and place the 3D smart sensor at a preset height above the reference surface. Use the 3D smart sensor to collect point cloud data of the refractory brick to be tested.
6. The method as described in claim 5, characterized in that, The preset height is 765mm.
7. The method as described in claim 5, characterized in that, The set height range is 650mm-725mm from the 3D smart sensor.