Wheel hub roundness detection method and system based on machine vision

Abstract

The application relates to the technical field of hub detection, in particular to a hub roundness detection method and system based on machine vision, wherein the method comprises the following steps: acquiring a standard point cloud, a radial section and a standard height of a standard hub, and dividing a height interval according to the standard height; acquiring edge point sets of the radial section according to the standard point cloud, and calculating a global curvature complexity; acquiring a sampling section and a contour point set thereof according to the height interval, and calculating a section space curvature and a reference radius; calculating the total slice quantity by using the standard height and the global curvature complexity, and obtaining a slice and a slice position of the height interval in combination with the section space curvature; acquiring actual point clouds and a plurality of actual radii of a hub to be detected, comparing the actual point clouds and the actual radii with corresponding reference radii, completing roundness detection and outputting a detection result. The technical scheme of the application solves the problem that high-precision detection cannot be realized due to the fixed slices which cannot adapt to the complex surface of the hub in the prior art.

Description

Technical Field

[0001] This application relates to the field of wheel hub inspection technology, and in particular to a wheel hub roundness inspection method and system based on machine vision. Background Technology

[0002] As a core load-bearing component of a vehicle's driving system, the roundness accuracy of the wheel hub directly affects driving stability, tire lifespan, and driving safety. With the automotive industry's ever-increasing demands for component precision, 3D laser scanning technology, with its advantages of non-contact and full-size measurement, has become the mainstream technology for three-dimensional shape inspection of wheel hubs. Existing 3D laser scanning-based wheel hub roundness detection methods suffer from the following technical problems: The use of a global slicing strategy with a fixed slice density fails to adapt to the non-uniform characteristics of the wheel hub's axial morphology, making it difficult to balance detection accuracy and efficiency. Specifically, the axial morphology of wheel hubs exhibits significant variations. For example, the rim area typically contains multiple sealing grooves, mounting steps, and rounded corner transitions—structures with abrupt curvature changes—resulting in high morphological complexity. In contrast, the spoke area is mostly straight or gently curved, with low morphological complexity. Ignoring the relationship between wheel hub morphological features and slice density makes it difficult to balance detection accuracy and efficiency.

[0003] Therefore, existing technologies suffer from the problem of failing to achieve high-precision detection due to the use of fixed slices that cannot adapt to the complex curved surface of the wheel hub. Summary of the Invention

[0004] This application provides a machine vision-based wheel hub roundness detection method and system, which solves the problem that the existing technology uses fixed slices that cannot adapt to the complex curved surface of the wheel hub, resulting in the inability to achieve high-precision detection, so as to achieve the purpose of high-precision and high-efficiency detection of wheel hub roundness.

[0005] In the first aspect, a machine vision-based wheel hub roundness detection method includes: Obtain the standard point cloud, standard height, and multiple radial sections of a standard wheel hub, and determine multiple height intervals with equal heights based on the standard height; The edge point set located on each radial section is obtained based on the standard point cloud and the radial section respectively. The global curvature complexity of the standard wheel hub is calculated based on the edge point set. The total number of slices of the standard wheel hub is calculated based on the global curvature complexity and the standard height. Multiple sampling sections of the standard wheel hub in each height range are obtained. Contour point sets located on each sampling section are obtained according to the standard point cloud. The cross-sectional space curvature of the corresponding sampling section is calculated according to each contour point set. Based on the curvature of the cross-sectional space and the total number of slices, slices for each height interval are obtained, and the slice position of each slice is determined. Based on the standard point cloud, obtain the slice point set located in each slice, and determine the reference radius of the corresponding slice based on each slice point set; Obtain the actual point cloud of the wheel hub to be detected, and obtain the actual radius of each slice in the wheel hub to be detected based on the actual point cloud and the slice position; The roundness of the wheel hub to be tested is measured based on the actual radius and reference radius of each slice to obtain the test result of the wheel hub to be tested.

[0006] By adopting the above technical solution, an adaptive match between the total number of slices and the overall morphological complexity of the wheel hub was achieved, and the roundness detection result of the wheel hub to be tested was obtained.

[0007] Preferably, obtaining the set of edge points located on each radial cross section based on the standard point cloud and the radial cross section includes: Based on the standard point cloud, obtain the planar points on each radial section, and obtain the set of planar points located on each radial section based on the planar points; The α-shape algorithm is used to obtain the edge points of the corresponding radial cross section from each of the planar point sets, so as to obtain the edge point set of each radial cross section.

[0008] By adopting the above technical solution, the edge point set of each radial section is accurately extracted using the α-shape algorithm, providing reliable contour feature data for subsequent curvature calculation.

[0009] Preferably, the step of calculating the global curvature complexity of the standard wheel hub based on the set of edge points includes: Calculate the edge space curvature of each edge point in the edge point set, and calculate the curvature change of each edge point based on the edge space curvature; Calculate the standard deviation of curvature change for each edge point set, and take the standard deviation of the largest curvature change as the global curvature complexity.

[0010] By adopting the above technical solution, the complexity of the overall shape of the wheel hub is effectively quantified by using the standard deviation of the maximum curvature change as the global curvature complexity.

[0011] Preferably, the step of calculating the sectional spatial curvature of the corresponding sampling section for each of the contour point sets includes: Obtain the three-dimensional neighborhood point set of each contour point in the contour point set, wherein the three-dimensional neighborhood point set includes a portion of the sampled points in the standard point cloud; Based on the coordinates of the sampling points in the three-dimensional neighborhood point set, obtain the covariance matrix corresponding to each three-dimensional neighborhood point set, and calculate the eigenvalues ​​of each covariance matrix; The contour space curvature of the corresponding contour point is calculated based on the feature value, and the cross-sectional space curvature is calculated based on the contour space curvature.

[0012] By adopting the above technical solution, the curvature of the contour space is calculated using the eigenvalues ​​of the covariance matrix, providing an accurate curvature basis for differentiated slice allocation.

[0013] Preferably, obtaining slices for each height interval based on the cross-sectional spatial curvature and the total number of slices includes: Based on the cross-sectional spatial curvature of the sampling section within each height interval, calculate the interval spatial curvature of each height interval. The initial number of slices for each height interval is calculated based on the interval space curvature and the total number of slices. The remaining number of slices for the standard wheel hub is obtained based on the total number of slices and the initial number of slices. The remaining number of slices is then distributed in descending order to each height interval based on the interval space curvature to obtain the number of slices for each height interval. Finally, the number of slices for each height interval is determined based on the number of slices.

[0014] By adopting the above technical solution, the remaining slices are redistributed in descending order of curvature, so that high curvature areas can obtain denser slice resources.

[0015] Preferably, obtaining the actual radius of each slice in the wheel hub to be detected based on the actual point cloud and the slice position includes: Based on the slice position and the actual point cloud, obtain the set of points to be detected on each slice; Perform circle fitting on each of the points to be detected, and obtain the actual radius based on the fitted circular region.

[0016] By adopting the above technical solution, a circle fitting was performed on the set of points to be detected in each slice, and the actual radius at each slice position of the wheel hub to be detected was accurately obtained.

[0017] Preferably, the step of performing roundness detection on the wheel hub to be tested based on the actual radius and reference radius of each slice to obtain the detection result of the wheel hub to be tested includes: Based on the actual radius and the reference radius, calculate the radial deviation of each slice in the hub to be tested; The detection result is obtained based on the radial deviation using a preset qualified threshold.

[0018] By adopting the above technical solution, the radial deviation of each slice is compared with the qualified threshold, and the roundness test result of the wheel hub to be tested is obtained objectively and accurately.

[0019] Secondly, a machine vision-based wheel hub roundness detection system includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the machine vision-based wheel hub roundness detection method described above is implemented.

[0020] This application has the following advantages: The machine vision-based wheel hub roundness detection method of this application calculates the global curvature complexity based on the radial cross-sectional edge point set of a standard wheel hub. Since the global curvature complexity quantifies the complexity of the overall shape of the wheel hub and determines the total number of slices required to meet the detection accuracy, it achieves adaptive matching between the total number of slices and the overall shape complexity of the wheel hub, avoiding too many or too few slices. It calculates the cross-sectional spatial curvature according to the height interval and dynamically allocates the total number of slices to each height interval according to the spatial curvature ratio of the interval. Since the area with greater spatial curvature has more drastic shape changes and requires more dense slices, it achieves differentiated slice density of dense slices in high curvature areas and sparse slices in low curvature areas, reducing data redundancy while ensuring accuracy. Because the above two-level adaptive strategy prioritizes the allocation of limited slice resources to areas with complex shapes such as the rim, while only allocating necessary slices to the flat spoke areas, it ultimately achieves high-precision and high-efficiency wheel hub roundness detection.

[0021] The above and other objects, advantages and features of this application will become more apparent to those skilled in the art from the following detailed description of specific embodiments of this application in conjunction with the accompanying drawings. Attached Figure Description

[0022] The following sections will describe some specific embodiments of this application in detail by way of example and not limitation, with reference to the accompanying drawings. The same reference numerals in the drawings denote the same or similar parts or components. Those skilled in the art should understand that these drawings are not necessarily drawn to scale. In the drawings: Figure 1 This is a schematic flowchart of a machine vision-based wheel hub roundness detection method according to an embodiment of this application; Figure 2 This is a schematic flowchart illustrating the process of obtaining slices for each height interval based on the cross-sectional spatial curvature and the total number of slices, according to an embodiment of this application. Figure 3 This is a schematic diagram of a machine vision-based wheel hub roundness detection system according to an embodiment of this application. Detailed Implementation

[0023] The following reference Figures 1 to 3This application describes a machine vision-based wheel hub roundness detection method and system according to embodiments of the present application. In the description of this embodiment, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined with "first" and "second" may explicitly or implicitly include at least one of that feature, that is, include one or more of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified. When a feature "includes or contains" one or more of the features it encompasses, unless otherwise specifically described, this indicates that other features are not excluded and may be further included.

[0024] In the description of this embodiment, the terms "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0025] This embodiment provides a machine vision-based wheel hub roundness detection method, which can achieve high-precision and high-efficiency wheel hub roundness detection even when there are significant differences in the axial morphology of the wheel hub.

[0026] Please see Figure 1 , Figure 1 This is a schematic flowchart of a machine vision-based wheel hub roundness detection method according to an embodiment of this application. The method generally includes: Step S101: Obtain the standard point cloud, standard height, and multiple radial sections of the standard wheel hub, and determine multiple height ranges with equal heights based on the standard height; Step S102: Obtain the edge point set on each radial section according to the standard point cloud and radial section of the standard wheel hub. Calculate the global curvature complexity of the standard wheel hub according to the edge point set. Calculate the total number of slices of the standard wheel hub according to the global curvature complexity and the standard height of the standard wheel hub. Step S103: Obtain multiple sampling sections of the standard wheel hub in each height range, obtain the contour point set on each sampling section according to the standard point cloud of the standard wheel hub, and calculate the cross-sectional space curvature of the corresponding sampling section according to each contour point set. Step S104: Obtain slices for each height range based on the cross-sectional spatial curvature of the sampling section and the total number of slices for the standard wheel hub, and determine the slice position of each slice. Step S105: Obtain the slice point set located in each slice according to the standard point cloud of the standard wheel hub, and determine the reference radius of the corresponding slice according to each slice point set. Step S106: Obtain the actual point cloud of the wheel hub to be detected, and obtain the actual radius of each slice in the wheel hub to be detected based on the actual point cloud and the slice position; Step S107: Perform roundness detection on the wheel hub to be inspected based on the actual radius and reference radius of each slice to obtain the inspection result of the wheel hub to be inspected.

[0027] In step S101 above, a three-dimensional reference coordinate system is established using a laser tracker to cover the full-size measurement range of a standard wheel hub. A standard wheel hub refers to a wheel hub without defects and whose roundness meets actual engineering requirements. A full-size scan of the standard wheel hub is then performed using a 3D laser scanner to obtain the standard point cloud of the standard wheel hub in the three-dimensional coordinate system.

[0028] In this embodiment, the three-dimensional coordinate system uses the central axis of the standard wheel hub as the vertical axis, i.e., the Z-axis, and the directions of a pair of mutually perpendicular radii in the standard wheel hub as the horizontal axis and the vertical axis, respectively. The horizontal axis is the X-axis, and the vertical axis is the Y-axis.

[0029] In this embodiment, the height of a standard wheel hub can be detected in advance using a ruler, laser rangefinder, etc., and this height can be used as the standard height of the standard wheel hub.

[0030] In this embodiment, the total height range of the standard wheel hub is divided into multiple height intervals. The number of height intervals directly affects the test results. If too few intervals are divided, wheel hubs with unacceptable roundness may be mistakenly judged as acceptable; if too many intervals are divided, wheel hubs with acceptable roundness may be mistakenly judged as unacceptable. In this embodiment, the total height range of the standard wheel hub is preferably divided evenly into 20 height intervals.

[0031] In this embodiment, multiple radial sections are extracted in a three-dimensional coordinate system with the Z-axis as the central axis. Each radial section intersects the Z-axis, is perpendicular to the XY plane formed by the X-axis and Y-axis, and is evenly distributed around the Z-axis. The number of radial sections can be freely set by the user. The more radial sections there are, the more accurate the global curvature complexity of the standard wheel hub is calculated, and the more accurate the total number of slices of the standard wheel hub is, resulting in a more accurate roundness detection result for the wheel hub to be tested.

[0032] In step S102 above, all sampling points in the standard point cloud of the standard wheel hub are projected onto each radial section to obtain the planar point set of each radial section.

[0033] In this embodiment, the α-shape algorithm is used to extract features from each set of planar points to obtain the edge point set corresponding to each set of planar points. Each edge point set contains multiple edge points, which refer to the planar points corresponding to the edge positions of the standard wheel hub within the set of planar points. The edge points on each radial section are sorted in ascending order of their Z-axis coordinate values ​​to obtain the edge point sequence for each radial section. Taking a set of planar points for a radial section as an example, the method for obtaining its corresponding edge point set using the α-shape algorithm includes: First, define the α-shape algorithm. value, The value controls the fineness of the extracted edge point set. When The smaller the value, the more refined the extracted edge point set; when The larger the value, the coarser the extracted edge point set. In this embodiment, The value is adaptively determined based on the density of the planar points in the planar point set. Specifically, the Euclidean distance between every two planar points in the planar point set is calculated, and the average of all Euclidean distances is calculated. This average value, which is then multiplied by a preset factor, is used as the... The preset multiplier is set according to the actual detection accuracy requirements. The larger the preset multiplier, the more accurate the identification of edge points. In this embodiment, the preset multiplier is preferably 1.5.

[0034] Next, iterate through all planar points in the planar point set, and process the two planar points encountered as follows: obtain the points that pass through these two planar points and have a radius of... The two candidate circles are selected, and then it is determined whether there are other planar points in these two candidate circles; if there are no other planar points in at least one circle, then these two planar points are determined to be edge points.

[0035] Then, all the plane points marked as edge points are extracted to form the edge point set corresponding to the plane point set.

[0036] Finally, the edge point set of the radial section is deduplicated, that is, the repeated edge points are deleted to ensure that each edge point appears only once in the edge point set, so as to obtain the complete edge point set of the radial section.

[0037] In this embodiment, after obtaining the edge point set of each radial section, the edge points in each edge point set are arranged sequentially according to their positions to obtain the edge point sequence of each radial section. Adjacent edge points in each edge point sequence are also adjacent in their corresponding radial sections. Taking the edge point set of one radial section as an example, since each radial section intersects the Z-axis, in this embodiment, a planar coordinate system for the radial section is established with the extension direction of the Z-axis as the horizontal axis, the direction perpendicular to the Z-axis as the vertical axis, and the position of the zero point on the Z-axis as the origin. The coordinates of each edge point in the edge point set of the radial section in the planar coordinate system are obtained. Then, one edge point is randomly selected as the first edge point, one edge point adjacent to the starting point is selected as the second edge point, and one edge point adjacent to the second edge point is selected as the third edge point; and so on, to construct a sequence containing all edge points of the radial section. This sequence is the edge point sequence of the radial section.

[0038] Then, using the five-point difference method, the edge space curvature of each edge point in each edge point sequence is calculated sequentially. Taking one edge point as an example, this edge point is designated as the target edge point, the radial section containing this edge point is designated as the target radial section, and the edge point sequence containing this edge point is designated as the target edge point sequence. Let the target edge point be the a-th edge point in the target edge point sequence. The method for calculating the edge space curvature of this edge point includes: First, the first and second difference values ​​of the target edge points are calculated using the five-point difference method. Let the first difference value of the target edge points be... The second difference is ,but: in, This represents the ordinate value of the second edge point following the target edge point in the target edge point sequence. This is the ordinate value of the first edge point following the target edge point in the target edge point sequence. This is the ordinate value of the second edge point preceding the target edge point in the target edge point sequence. The ordinate value of the first edge point following the target edge point in the target edge point sequence; This represents the x-coordinate of the second edge point following the target edge point in the target edge point sequence. This is the x-coordinate of the first edge point following the target edge point in the target edge point sequence. The x-coordinate of the second edge point preceding the target edge point in the target edge point sequence. It is the x-coordinate of the first edge point following the target edge point in the target edge point sequence.

[0039] Since the longitudinal section of the target is a closed shape, the target edge point sequence is also a closed loop sequence. That is, the edge point preceding the first edge point in the target edge point sequence is the last edge point in the target edge point sequence, the second edge point preceding the first edge point in the target edge point sequence is the second to last edge point in the target edge point sequence, the first edge point following the last edge point in the target edge point sequence is the first edge point in the target edge point sequence, and the second edge point following the last edge point in the target edge point sequence is the second edge point in the target edge point sequence.

[0040] Then, based on the first-order and second-order difference values ​​of the target edge points, the edge space curvature of the target edge points is calculated. Let the edge space curvature of the target edge points be... ,but: In this embodiment, the curvature change of each edge point is calculated based on the edge space curvature of each edge point in each edge point sequence. Specifically, taking an edge point as an example, the absolute value of the difference between the edge space curvature of that edge point and the edge space curvature of its predecessor is taken as the curvature change of that edge point. If the edge point is the first edge point in its edge point sequence, the edge space curvature of the second edge point in the sequence is obtained, and the edge space curvature of the first edge point is subtracted from it. The absolute value of the difference is then taken to obtain the curvature change of the first edge point.

[0041] In this embodiment, the standard deviation of curvature change for each edge point sequence is calculated. The standard deviation of curvature change represents the curvature complexity of its corresponding radial section. Then, the maximum value of the curvature complexity of all radial sections is obtained and used as the global curvature complexity of the standard wheel hub.

[0042] This embodiment calculates the total number of slices for a standard wheel hub based on its global curvature complexity. The formula used to calculate the total number of slices is as follows: In the above formula, To find the function with the maximum value, This is a rounding function. This represents the total number of slices for a standard wheel hub. The global curvature complexity of a standard wheel hub; This refers to the standard height of a standard wheel hub; The preset adjustment coefficient is used to remove the dimension of the standard height H. The larger the preset adjustment coefficient, the larger the total number of slices of the standard wheel hub, that is, the finer the slices of the standard wheel hub, and therefore the more accurate the roundness detection of the wheel hub to be tested. In this embodiment, the preferred value of the adjustment coefficient is 2.

[0043] In step S103 above, P sampling standard heights are uniformly selected within each height interval. For each sampling standard height, a section perpendicular to the Z-axis and parallel to the XY plane is extracted, i.e., the sampling section. The point set of each sampling section is obtained according to the standard point cloud of the standard wheel hub. Then, the edge points of each sampling section are obtained using the same α-shape algorithm as in step S102. The edge points are used as the contour points of their corresponding sampling sections, and the contour point set of each sampling section is constructed based on these contour points.

[0044] In this embodiment, after obtaining the contour point set for each sampling section, the contour space curvature of each contour point in each contour point set is calculated, and the average value of the contour space curvature of each contour point in the contour point set is taken as the cross-sectional space curvature of its corresponding sampling section. Taking one sampling section as an example, after obtaining the contour space curvature of each contour point in the contour point set of the sampling section, the average value of the contour space curvature of all contour points in the contour point set is calculated, and this average value is taken as the cross-sectional space curvature of the sampling section.

[0045] This embodiment uses a contour point from a set of contour points in one of the sampling sections as an example to describe in detail the method for calculating the contour space curvature of the contour point. In this embodiment, the contour point is taken as the target contour point, and its contour space curvature calculation method includes: First, a spherical spatial region with the target contour point as the center and a preset length as the radius is obtained, and the sampling points in the standard point cloud of the standard wheel hub located in the spherical spatial region are taken as the neighbor sampling points of the target contour point. In this embodiment, the neighbor sampling points are taken as the target neighbor points. Then, based on the coordinates of each target neighborhood point in the three-dimensional spatial coordinate system, the covariance matrix of the target contour points is constructed; Finally, the three eigenvalues ​​of the covariance matrix of the target contour points are calculated, and the contour space curvature of the target contour points is calculated based on these eigenvalues.

[0046] In this embodiment, the covariance matrix of the target contour points is a symmetric positive semi-definite matrix of size 3×3. Let the number of target neighborhood points be m, where the coordinates of the w-th target neighborhood point in the three-dimensional spatial coordinate system are... If the covariance matrix of the target contour points is C, then: in: in, The average coordinates of all neighborhood points of the target on the X-axis. The average Y-axis coordinate of all neighborhood points of the target. The average coordinate value of all target neighborhood points on the Z-axis, i.e.: After obtaining the covariance matrix of the target contour points, the characteristic equation of the target contour points is constructed based on the covariance matrix. The characteristic equation is as follows: In the above formula, 'a' is the coefficient of the quadratic term, 'b' is the coefficient of the linear term, and 'c' is a constant. Let be the real-valued eigenvalues ​​of the target contour points, and: In this embodiment, the characteristic equation of the target contour points is solved to obtain the real-valued feature values ​​of the target contour points. There are three such real-valued feature values: ,in Let the curvature of the contour space of the target contour points be . ,but: In step S104 above, slices for each height interval are obtained based on the curvature of the cross-sectional space and the total number of slices, and the slice position of each slice is determined based on the height value of each height interval. Please refer to [link to relevant documentation]. Figure 2 , Figure 2 This is a schematic flowchart illustrating the process of obtaining slices for each height interval based on the cross-sectional spatial curvature and the total number of slices according to an embodiment of this application. The process corresponding to this flowchart includes: Step S114: Calculate the mean value of the cross-sectional spatial curvature of the sampled cross-section in each height interval to obtain the interval spatial curvature of each height interval. Step S124: Calculate the initial number of slices for each height interval based on the total number of slices for the standard wheel hub and the spatial curvature of each height interval. Step S134: Obtain the sum of the initial slice counts for all height ranges, and use this sum and the total slice count of the standard wheel hub to calculate the remaining slice count of the standard wheel hub. Step S144: Based on the spatial curvature of each height interval, the remaining number of slices of the standard wheel hub are distributed in descending order to each height interval to obtain the slices of each height interval.

[0047] In step S114 above, the mean value of the spatial curvature of all cross sections in each height interval is calculated to obtain the spatial curvature of each height interval.

[0048] In step S124 above, the initial number of slices for each height interval is obtained. Taking the i-th height interval as an example, the formula used to calculate the initial number of slices is: In the above formula, Let be the initial number of slices for the i-th height interval, where i is the index of the height interval; Let be the interval space curvature of the i-th height interval; I is the total number of height intervals. In this embodiment, I = 20, which is the floor sign.

[0049] In step S134 above, the number of remaining slices of the standard wheel hub is calculated, and the formula used to calculate the number of remaining slices is: In the above formula, This represents the number of remaining slices for a standard wheel hub.

[0050] In step S144 above, The slices are then redistributed in descending order of spatial curvature for each height interval. Specifically, all height intervals are first sorted in descending order of their spatial curvature to obtain a height interval sequence. Then, the height interval sequence is iterated through sequentially. For each height interval, the number of slices in that interval is increased by 1, while the number of remaining slices for the standard wheel hub is simultaneously decreased by one, until the number of remaining slices is 0. This yields the final number of slices for each height interval. .

[0051] In this embodiment, the number of slices for each height range is obtained. Next, the spatial location of each slice is determined, and multiple slice locations are generated within each height interval using an equally spaced method. All slice locations are then merged to form the optimal slice location sequence. The method for obtaining the optimal slice location sequence includes: First, obtain the length of each height interval. (The rest of the text appears to be incomplete and requires further context.) height range For example, the formula used to calculate the interval length is: In the above formula, For the first The length of each height interval; For the first The height value of the left endpoint of each height interval; For the first The height value of the right endpoint of each height interval.

[0052] Secondly, obtain the slice position of each slice within each height interval. (The sentence is incomplete and requires more context to translate accurately.) Taking the height range as an example, in the first... Generate at equal intervals within each height range The first slice. height range Taking the j-th slice as an example, the formula used to calculate its slice position is: In the above formula, For the first The slice position of the j-th slice in a height interval.

[0053] Finally, based on the height values ​​corresponding to the slice positions, all slices are sorted from smallest to largest to obtain the optimal slice position sequence. Each height value in the optimal slice position sequence determines a plane, and each plane is parallel to the XY plane. Each plane is considered as a slice to obtain all slices.

[0054] In step S105 above, the slice point set of each slice is obtained in the same way as in step S103, and the least squares method is used to perform circle fitting on the slice point set of each slice to obtain the reference radius of each slice.

[0055] In step S106 above, the actual point cloud of the wheel hub to be detected is obtained, and its nominal diameter is obtained. The actual point cloud is projected onto each slice to obtain the set of points to be detected of the wheel hub on each slice. The least squares method is used to perform circle fitting on each set of points to be detected to obtain the actual radius of each slice.

[0056] In step S107 above, a preset qualified threshold is obtained. The smaller the qualified threshold, the more accurate the roundness detection result of the wheel hub to be tested. In this embodiment, the qualified threshold is preferably 0.1% of the nominal diameter of the wheel hub to be tested.

[0057] In this embodiment, the method for performing roundness detection on the wheel hub to be inspected based on the actual radius and reference radius of each slice to obtain the inspection result of the wheel hub to be inspected includes: Calculate the radius difference between the actual radius and the corresponding reference radius for each slice; taking one slice as an example, let the actual radius of the slice be r, and the reference radius be r. The radius difference between the actual radius of the slice and the corresponding reference radius is: ; Determine whether each radius difference is less than the preset qualified threshold; If so, the test result for the roundness of the wheel hub to be tested is deemed qualified; If not, the test result for the roundness of the wheel hub to be tested is deemed unqualified.

[0058] The flowchart provided in this embodiment is not intended to indicate that the operations of the method will be performed in any particular order, or that all operations of the method are included in every case. Furthermore, the method may include additional operations. Within the scope of the technical concept provided by the method in this embodiment, additional variations can be made to the above method.

[0059] It should be understood that in some embodiments, the components may be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods may be implemented using software or firmware stored in memory and executed by a suitable instruction execution system.

[0060] This embodiment also provides a machine vision-based wheel hub roundness detection system, which is as follows: Figure 3 As shown, it includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the computer program is executed by the processor, it implements the steps of the machine vision-based wheel hub roundness detection method of any of the above embodiments.

[0061] The computer program used to perform the operations of this application may be assembly instructions, Instruction Set Architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, integrated circuit configuration data, or source code or object code written in any combination of one or more programming languages ​​and procedural programming languages. The computer program may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter case, the remote computer may be connected to the user's computer via any type of network (including a Local Area Network (LAN) or Wide Area Network (WAN)) or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, to perform aspects of this application, electronic circuits, including, for example, programmable logic circuits, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), may execute computer-readable program instructions using status information from computer-readable program instructions to personalize the electronic circuits.

[0062] Therefore, those skilled in the art should recognize that although many exemplary embodiments of this application have been shown and described in detail herein, many other variations or modifications conforming to the principles of this application can be directly determined or derived from the disclosure of this application without departing from the spirit and scope of this application. Thus, the scope of this application should be understood and construed as covering all such other variations or modifications.

Claims

1. A machine vision-based method for detecting the roundness of wheel hubs, characterized in that, include: Obtain the standard point cloud, standard height, and multiple radial sections of a standard wheel hub, and determine multiple height intervals with equal heights based on the standard height; The edge point set located on each radial section is obtained based on the standard point cloud and the radial section respectively. The global curvature complexity of the standard wheel hub is calculated based on the edge point set. The total number of slices of the standard wheel hub is calculated based on the global curvature complexity and the standard height. Multiple sampling sections of the standard wheel hub in each height range are obtained. Contour point sets located on each sampling section are obtained according to the standard point cloud. The cross-sectional space curvature of the corresponding sampling section is calculated according to each contour point set. Based on the curvature of the cross-sectional space and the total number of slices, slices for each height interval are obtained, and the slice position of each slice is determined. Based on the standard point cloud, obtain the slice point set located in each slice, and determine the reference radius of the corresponding slice based on each slice point set; Obtain the actual point cloud of the wheel hub to be detected, and obtain the actual radius of each slice in the wheel hub to be detected based on the actual point cloud and the slice position; The roundness of the wheel hub to be tested is measured based on the actual radius and reference radius of each slice to obtain the test result of the wheel hub to be tested.

2. The machine vision-based wheel hub roundness detection method according to claim 1, characterized in that, The step of obtaining the edge point set located on each radial cross section based on the standard point cloud and the radial cross section includes: Based on the standard point cloud, obtain the planar points on each radial section, and obtain the set of planar points located on each radial section based on the planar points; The α-shape algorithm is used to obtain the edge points of the corresponding radial cross section from each of the planar point sets, so as to obtain the edge point set of each radial cross section.

3. The machine vision-based wheel hub roundness detection method according to claim 1, characterized in that, The calculation of the global curvature complexity of the standard wheel hub based on the set of edge points includes: Calculate the edge space curvature of each edge point in the edge point set, and calculate the curvature change of each edge point based on the edge space curvature; Calculate the standard deviation of curvature change for each edge point set, and take the standard deviation of the largest curvature change as the global curvature complexity.

4. The machine vision-based wheel hub roundness detection method according to claim 1, characterized in that, The step of calculating the sectional spatial curvature of the corresponding sampling section based on each of the contour point sets includes: Obtain the three-dimensional neighborhood point set of each contour point in the contour point set, wherein the three-dimensional neighborhood point set includes a portion of the sampled points in the standard point cloud; Based on the coordinates of the sampling points in the three-dimensional neighborhood point set, obtain the covariance matrix corresponding to each three-dimensional neighborhood point set, and calculate the eigenvalues ​​of each covariance matrix; The contour space curvature of the corresponding contour point is calculated based on the feature value, and the cross-sectional space curvature is calculated based on the contour space curvature.

5. The machine vision-based wheel hub roundness detection method according to claim 1, characterized in that, The step of obtaining slices for each height interval based on the curvature of the cross-sectional space and the total number of slices includes: Based on the cross-sectional spatial curvature of the sampling section within each height interval, calculate the interval spatial curvature of each height interval. The initial number of slices for each height interval is calculated based on the interval space curvature and the total number of slices. The remaining number of slices for the standard wheel hub is obtained based on the total number of slices and the initial number of slices. The remaining number of slices is then distributed in descending order to each height interval based on the interval space curvature to obtain the number of slices for each height interval. Finally, the number of slices for each height interval is determined based on the number of slices.

6. The machine vision-based wheel hub roundness detection method according to claim 1, characterized in that, The step of obtaining the actual radius of each slice in the wheel hub to be detected based on the actual point cloud and the slice position includes: Based on the slice position and the actual point cloud, obtain the set of points to be detected on each slice; Perform circle fitting on each of the points to be detected, and obtain the actual radius based on the fitted circular region.

7. The machine vision-based wheel hub roundness detection method according to claim 1, characterized in that, The step of performing roundness detection on the wheel hub to be tested based on the actual radius and reference radius of each slice to obtain the detection result of the wheel hub to be tested includes: Based on the actual radius and the reference radius, calculate the radial deviation of each slice in the hub to be tested; The detection result is obtained based on the radial deviation using a preset qualified threshold.

8. A machine vision-based wheel hub roundness detection system, characterized in that, It includes a processor and a memory, the memory storing computer program instructions, which, when executed by the processor, implement the steps of the machine vision-based wheel hub roundness detection method according to any one of claims 1-7.

Images