Hardware size error detection method and system based on point cloud processing

Abstract

The invention relates to the technical field of hardware size measurement and detection, in particular to a hardware size error detection method and system based on point cloud processing. The method comprises the following steps: firstly, acquiring point cloud data of hardware and carrying out preliminary density clustering on the point cloud data; further analyzing from a plurality of perspectives of distribution similarity characteristics of the target data point and other data points in the initial density clustering cluster to which the target data point belongs, repeated characteristics of curvature and similar characteristics of data density of the target data point to obtain a hardware membership degree of the target data point; further adjusting the distance between the target data point and the central point in the initial density clustering cluster according to the hardware membership degree, carrying out secondary density clustering, and obtaining de-noised point cloud data; and finally, carrying out size error detection on the hardware according to the de-noised point cloud data. According to the method, the possibility that the data points belong to the hardware area is analyzed from multiple angles, the density clustering process is optimized, noise interference in the point cloud data is accurately removed, and a reliable hardware processing product size error detection result is obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of dimensional measurement and inspection, and particularly to a method and system for detecting hardware dimensional errors based on point cloud processing. Background Art

[0002] In construction engineering, the dimensional accuracy of hardware components is crucial for ensuring the safety and stability of building structures. These hardware components, such as steel bars, embedded parts, connectors, gears, etc., need to have high dimensional accuracy and shape and position accuracy in order to achieve precise connection and load-bearing in building structures. Dimensional error detection can discover the deviations of key parameters such as the length, diameter, and angle of these key components, and timely adjust the manufacturing and installation processes to ensure that each component meets the design requirements and ensure its safety and reliability in various key applications.

[0003] Traditional methods for detecting the dimensions of construction engineering hardware mainly rely on manual measurement and simple measuring tools, such as tape measures, calipers, etc. These methods are inefficient, have limited accuracy, and are easily affected by human factors. In recent years, with the rapid development of three-dimensional scanning technology, dimensional error detection methods based on 3D scanning have gradually been applied to the field of construction engineering. By comprehensively scanning the surface of construction hardware with a 3D scanner and then using data processing software to analyze the dimensional errors and shape deviations of the hardware, complex-shaped data can be quickly obtained and comprehensive geometric analysis can be carried out. This method greatly improves the detection efficiency and accuracy, especially when detecting components with complex shapes, the advantages are more obvious.

[0004] However, in practical applications, 3D scanning technology also faces some challenges. Among them, when converting point cloud data into a geometric model, it is necessary to remove noise and useless data from the point cloud data. However, the hardware used in construction engineering usually has complex geometric shapes and fine surface features, such as the rib patterns of steel bars, the anchoring parts of embedded parts, and tiny holes and other features. These features may be misidentified as noise or deleted by the filtering algorithm by mistake. This will lead to incomplete or inaccurate final scanning results, affect the measurement accuracy of the dimensions of construction hardware products, and cause inaccurate error detection results.

[0005] Therefore, for the detection of hardware dimensional errors in construction engineering, more accurate and reliable detection methods and systems need to be developed. These methods need to be able to effectively process complex point cloud data, retain key geometric features, and accurately identify and measure dimensional errors. At the same time, it is also necessary to consider the influence of on-site environments in construction engineering, such as light, dust, etc. on the scanning results to ensure the stability and reliability of the detection system in practical applications. By continuously improving and innovating detection technologies, the quality and safety of construction engineering can be better guaranteed. Summary of the Invention

[0006] To solve the technical problem that the denoising of point cloud data of hardware is inaccurate and affects the detection of hardware size errors, the purpose of the present invention is to provide a method and system for detecting hardware size errors based on point cloud processing. The specific technical solutions are as follows: A method for detecting hardware size errors based on point cloud processing, the method comprising: Obtain the point cloud data of the hardware; perform preliminary density clustering on the point cloud data; select any data point in any preliminary density clustering cluster as the target data point; According to the uniform distribution characteristics of the data points in the local neighborhood of the target data point, obtain the hardware representation degree of the target data point; according to the similarity characteristics of the hardware representation degrees of the target data point and other data points in the preliminary density clustering cluster to which it belongs, and in combination with the repeated characteristics of the curvatures of the target data point and other data points in the point cloud data, obtain the hardware possibility of the target data point; according to the similarity characteristics of the data density of the target data point and the data densities of all data points within the preset neighborhood range of the target data point, obtain the density uniformity parameter of the target data point; Fuse the hardware possibility and the density uniformity parameter of the target data point to obtain the hardware membership degree of the target data point; both the hardware possibility and the density uniformity parameter are positively correlated with the hardware membership degree; adjust the distance between the target data point and the center point in the preliminary density clustering cluster to which it belongs according to the hardware membership degree, and perform secondary density clustering; obtain the denoised point cloud data according to the secondary density clustering result; Perform size error detection on the hardware according to the denoised point cloud data.

[0007] Further, the method for obtaining the hardware representation degree includes: Take the surface formed by the data points within the intersection of the first preset neighborhood of the target data point and the preliminary density clustering cluster to which it belongs as the target surface of the target data point; according to the uniform distribution characteristics of the data points within the target surface, obtain the hardware representation degree of the target data point.

[0008] Further, the method for obtaining the hardware representation degree of the target data point according to the uniform distribution characteristics of the data points within the target surface includes: Within the target surface, with the target data point as the origin, select data points to be analyzed in a preset direction away from the origin; according to the distance consistency between adjacent data points to be analyzed in the same preset direction, obtain the hardware representation degree of the target data point; the distance consistency and the hardware representation degree are positively correlated.

[0009] Further, the method for obtaining the curvature includes: Take the maximum curvature of the target data point in all the preset directions as the curvature of the target data point.

[0010] Further, the method for obtaining the hardware possibility includes: Obtain a first possibility parameter of the target data point according to the similarity features of the hardware performance degree between the target data point and other data points within the preliminary density clustering cluster to which it belongs; the similarity features of the hardware performance degree are positively correlated with the first possibility parameter; In the point cloud data, obtain a second possibility parameter of the target data point according to the number of other data points with the same curvature as the curvature of the target data point; the number of other data points with the same curvature is positively correlated with the second possibility parameter; Obtain the hardware possibility of the target data point according to the first possibility parameter and the second possibility parameter of the target data point; both the first possibility parameter and the second possibility parameter are positively correlated with the hardware possibility.

[0011] Further, the method for obtaining the density uniformity parameter includes: Obtain the data density of each data point according to the quantity characteristics of the data points within the second preset neighborhood of each data point; Obtain the density uniformity parameter of the target data point according to the similarity features of the data density between the data density of the target data point and the data densities of all the data points within the preset neighborhood range of the target data point; the similarity features of the data density are positively correlated with the density uniformity parameter.

[0012] Further, the method for performing secondary density clustering by adjusting the distance between the target data point and the center point within the preliminary density clustering cluster to which it belongs according to the hardware membership degree includes: Perform a negative correlation mapping and normalization on the hardware membership degree of the target data point to obtain a distance correction coefficient of the target data point; Take the product of the distance between the target data point and the center point within the preliminary density clustering cluster to which it belongs and the distance correction coefficient as the corrected distance between the target data point and the center point within the preliminary density clustering cluster to which it belongs, and perform secondary density clustering according to the corrected distance.

[0013] Further, the method for detecting the dimensional error of the hardware according to the denoised point cloud data includes: Convert the denoised point cloud data into a hardware surface using Poisson surface reconstruction, convert the surface into a mesh model using the triangular meshing method for the reconstructed hardware surface, measure at least data including the diameter, number of teeth, tooth profile, and tooth pitch of the hardware, and obtain the dimensional error of the hardware by comparing the measured data and the design specification data.

[0014] Further, both the initial density clustering and the secondary density clustering use the DBSCAN clustering algorithm.

[0015] The present invention also provides a hardware size error detection system based on point cloud processing. The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of any one of the hardware size error detection methods based on point cloud processing are implemented.

[0016] The present invention has the following beneficial effects: The present invention first obtains the point cloud data of the hardware and performs initial density clustering on the point cloud data, which is convenient for subsequent analysis of the initial density clustering results and optimization of the clustering process; further obtains the hardware performance degree from the perspective of uniform distribution to evaluate the possibility that the target data point belongs to the hardware, and prepares for the hardware possibility of the subsequent target data point; further makes full use of the structural symmetry and local structural consistency characteristics of the hardware to obtain the hardware possibility of the target data point, providing more basis for subsequent optimization of clustering to improve the denoising effect; further obtains the density uniformity parameter of the target data point according to the similarity characteristics between the data density of the target data point and the data density of all data points within the preset neighborhood range of the target data point, providing a basis for distinguishing defective data points from noise data points; further fuses the hardware possibility and density uniformity parameter of the target data point to obtain the hardware membership degree of the target data point as the point cloud data of the hardware from the perspectives of distribution similarity, hardware structure symmetry, and density similarity, providing a reliable basis for subsequent optimization of clustering and accurate denoising of the point cloud data; further optimizes the clustering process according to the hardware membership degree, performs secondary density clustering, and obtains the denoised point cloud data, which better reflects the true local density change, improves the accuracy of the clustering result, and enhances the accuracy of size error detection; finally, performs size error detection on the hardware according to the denoised point cloud data. The present invention analyzes the possibility that the data point belongs to the hardware area from multiple angles, optimizes the density clustering process, accurately removes the noise interference in the point cloud data, and obtains a reliable size error detection result of the hardware processed product. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for description in the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0018] Figure 1 It is a flowchart of a hardware size error detection method based on point cloud processing provided by an embodiment of the present invention; Figure 2 Flowchart of a method for obtaining gear possibilities provided by an embodiment of the present invention. Detailed implementation manners

[0019] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following, in combination with the accompanying drawings and preferred embodiments, details the specific implementation manners, structures, features and effects of the hardware size error detection method and system based on point cloud processing proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0021] The following specifically describes the specific solutions of the hardware size error detection method and system based on point cloud processing provided by the present invention with reference to the accompanying drawings.

[0022] Please refer to Figure 1 , which shows a flowchart of a hardware size error detection method based on point cloud processing provided by an embodiment of the present invention, specifically including: In construction engineering, there are many types of hardware components, such as steel bars, embedded parts, connectors, gears, etc., all of which can be detected using the point cloud method. This application takes the gear hardware as an example.

[0023] Step S1: Obtain the point cloud data of the gear; perform preliminary density clustering on the point cloud data; select any data point in any preliminary density clustering cluster as the target data point.

[0024] In an embodiment of the present invention, in the gear error detection link, a light-shielding curtain and a reflector are used to control the ambient light, and a 3D scanner is used to laser scan the gear. The gear is firmly placed on the turntable, and the angle and distance between the scanner and the gear are adjusted to ensure that the entire surface of the gear can be scanned. The point cloud data of the gear is obtained, and the scanned data is imported into a dedicated software for fusion processing to generate a complete gear point cloud model.

[0025] Considering that the density clustering algorithm can preliminarily classify the point cloud data of gears, cluster the points with close distances into clusters based on the density relationship between data points, effectively distinguish potential gear structures from non-gear data, after preliminary density clustering, further analyze various characteristics of the data points, adjust the distribution of the data points, re-perform density clustering, and optimize the cluster division, so as to more accurately capture the structural characteristics of the gears, and at the same time remove those noise points that significantly deviate from the gear model to ensure the quality of the final gear surface reconstruction. Therefore, the point cloud data is first subjected to preliminary density clustering; any data point in any preliminary density clustering cluster is selected as the target data point.

[0026] Preferably, in an embodiment of the present invention, the DBSCAN clustering algorithm is used for preliminary density clustering. In the point cloud data, the Euclidean distance between any two points is obtained, and the Euclidean distances corresponding to each data point are sorted from large to small, and the k-th nearest Euclidean distance corresponding to each data point is taken as the k-distance of each data point; the k-distances of all data points in the point cloud data are sorted from small to large and then the k-distance change curve is fitted; the k-distance corresponding to the maximum inflection point on the k-distance change curve is taken as the neighborhood distance; k + 1 is taken as the minimum number of points for DBSCAN clustering.

[0027] As an example, , the maximum inflection point on the k-distance change curve is directly obtained by the maximum curve inflection point (L-curve method), all of which are prior arts and will not be elaborated here.

[0028] It should be noted that preliminary density clustering may misclassify the data points in the gear area as noise points for isolation. To avoid this situation, the isolated points after initial density clustering are divided into the nearest clustering cluster for further analysis.

[0029] Step S2: According to the uniform distribution characteristics of the data points in the local neighborhood of the target data point, obtain the gear manifestation degree of the target data point; according to the similarity characteristics of the gear manifestation degrees of the target data point and other data points in the preliminary density clustering cluster to which it belongs, and in combination with the repeated characteristics of the curvatures of the target data point and other data points in the point cloud data, obtain the gear possibility of the target data point; according to the similarity characteristics of the data density of the target data point and the data densities of all data points within the preset neighborhood range of the target data point, obtain the density uniformity parameter of the target data point.

[0030] The ideal gear point cloud data has a uniform density distribution, while noise appears as scattered outlier points or low-density regions. At the same time, the tooth surface of the gear should exhibit a smooth and continuous point cloud structure. Therefore, first, according to the uniform distribution characteristics of the data points in the local neighborhood of the target data point, the gear representation degree of the target data point is obtained, and the possibility of the target data point belonging to the gear is evaluated from the perspective of uniform distribution, preparing for the subsequent gear possibility of the target data point and finally adjusting the distance between the target data point and the center point within the preliminary density clustering cluster to which it belongs for secondary density clustering.

[0031] Preferably, in an embodiment of the present invention, considering that the data points in the neighborhood of the target data point may not belong to the same initial density clustering cluster as the target data point, and the distances between the data points in different clustering clusters are relatively far, and their distribution characteristics themselves have differences, which cannot be used as the analysis basis for the gear representation degree of the target data point. Therefore, the surface formed by the data points within the intersection of the first preset neighborhood of the target data point and the preliminary density clustering cluster to which it belongs is used as the target surface of the target data point; the analysis range is restricted within the target surface, and the target surface is regarded as the local neighborhood of the target data point. According to the uniform distribution characteristics of the data points within the target surface, the gear representation degree of the target data point is obtained.

[0032] As an example, the first preset neighborhood is a region formed by a sphere with a radius of 5, the center of the sphere is the target data point, and the unit length is 1 mm.

[0033] In other embodiments of the present invention, the implementer can also set other first preset neighborhoods.

[0034] Preferably, in an embodiment of the present invention, within the target surface, taking the target data point as the origin, select the data points to be analyzed in the preset direction away from the origin; according to the distance consistency between adjacent data points to be analyzed in all the same preset directions, obtain the gear representation degree of the target data point; only by analyzing the data point distribution characteristics in the preset direction to represent the uniform distribution characteristics of the data of the entire surface, simplify the analysis process and improve the gear detection efficiency, where the distance consistency is positively correlated with the gear representation degree.

[0035] As an example, there are a total of 9 preset directions, and the preset directions are evenly distributed; by obtaining the surface normal line passing through the target data point, constructing a plane perpendicular to the normal line on the normal line, and evenly selecting 9 preset directions within the plane and projecting back to the surface, at this time, it is possible to select the data points to be analyzed in the preset direction away from the origin within the target surface. The calculation formula of the gear representation degree includes: ; Among them, represents the The gear performance degree of a target data point; Denote the exponential function with as the base; Denote the total number of preset directions; Denote the serial number of the preset direction; Denote the number of data points to be analyzed in the th preset direction; Denote the serial number of the data point to be analyzed; Denote the th distance parameter corresponding to the th data point to be analyzed in the th preset direction of the th target data point; Denote the average value of all distance parameters in the th preset direction of the th target data point; Denote the variance of all distance parameters in the

[0036] In the calculation formula of the gear performance degree, by means of the overall difference between the distance parameter corresponding to the data point to be analyzed of the target data point in the preset direction and the average value of the distance parameters, and the variance of the distance parameters, the distance consistency between adjacent data points to be analyzed in the same preset direction is represented. The smaller the difference between the distance parameter and the average value of the distance parameters, that is, the smaller, the more consistent the distance parameters are, the stronger the distance consistency between adjacent data points to be analyzed in the preset direction, the more uniform the distribution, and the greater the gear performance degree; the larger, the stronger the volatility of the distance parameters, the weaker the distance consistency between adjacent data points to be analyzed in the preset direction, the more uneven the distribution, and the smaller the gear performance degree.

[0037] It should be noted that the distance parameter of the data point to be analyzed is the Euclidean distance between the data point to be analyzed and the data point close to the origin in the same preset direction. For example, the distance parameter corresponding to the first data point to be analyzed in a certain preset direction is the Euclidean distance from the origin to the first data point to be analyzed; the distance parameter corresponding to the second data point to be analyzed is the Euclidean distance from the first data point to be analyzed to the second data point to be analyzed.

[0038] In other embodiments of the present invention, the implementer can also select the data points to be analyzed on the surface by means of, such as CAD models or finite element analysis networks; the weighted summation method can also be used for and Perform weighted summation. For example, the weighted weights are 0.6 and 0.4 respectively to calculate the gear performance degree. Other numbers of preset directions can also be set, which are all well-known technical means to those skilled in the art and will not be elaborated here.

[0039] In the embodiments of the present invention, in order to ensure that the noise can be recognized as a separate outlier rather than being wrongly classified into the gear area, it is necessary to accurately determine whether each data point in the point cloud data belongs to the gear area. Considering the structural symmetry and local structural consistency of the gear, the possibility of a data point belonging to the gear can be evaluated by analyzing the similarity of the gear performance degrees between data points and the repeatability of the geometric attributes of data points. Therefore, according to the similarity characteristics of the gear performance degree between the target data point and other data points within the preliminary density clustering cluster to which it belongs, and combining the repeat characteristics of the curvature between the target data point and other data points in the point cloud data, the gear possibility of the target data point is obtained, providing more basis for adjusting the distance between the target data point and the center point within the preliminary density clustering cluster to which it belongs.

[0040] Preferably, in an embodiment of the present invention, the maximum curvature of the target data point in all preset directions is used as the curvature of the target data point.

[0041] Preferably, in an embodiment of the present invention, the method for obtaining the gear possibility includes: Please refer to Figure 2 , which shows a flowchart of a method for obtaining the gear possibility provided by an embodiment of the present invention, specifically including: Step S201: Obtain the first possibility parameter of the target data point according to the similarity characteristics of the gear performance degree between the target data point and other data points within the preliminary density clustering cluster to which it belongs.

[0042] Among them, the similarity characteristics of the gear performance degree are positively correlated with the first possibility parameter.

[0043] Considering that the more similar the gear performance degree between the target data point and other data points within the same clustering cluster is, the smaller the difference is, it indicates that the target data point is more likely to be a data point in the gear area and less likely to be a noise point. Therefore, the first possibility parameter of the target data point is obtained according to the similarity characteristics of the gear performance degree between the target data point and other data points within the preliminary density clustering cluster to which it belongs. The similarity characteristics of the gear performance degree are positively correlated with the first possibility parameter.

[0044] As an example, the average value of the absolute value of the difference between the gear performance degree of the target data point and the gear performance degrees of other data points within the same preliminary density clustering cluster is negatively correlated through the negative correlation function and then used as the first possibility parameter of the target data point. Among them, represents the natural constant The exponential function with a base; Represents the independent variable.

[0045] By means of the absolute value of the difference, the difference characteristics between the gear performance degree of the target data point and the gear performance degrees of other data points within the same preliminary density clustering cluster are represented. The larger the absolute value of the difference, the more obvious the difference characteristics. After negative correlation mapping, the similarity characteristics of the gear performance degree are represented, and the first possibility parameter is obtained. The larger the similarity characteristics of the gear performance degree, the larger the first possibility parameter.

[0046] In other embodiments of the present invention, the implementer can also use the method of taking the reciprocal for negative correlation mapping to convert the difference characteristics of the gear performance degree into similarity characteristics and obtain the first possibility parameter.

[0047] Step S202: In the point cloud data, obtain the second possibility parameter of the target data point according to the number of other data points with the same curvature as the target data point.

[0048] The number of other data points with the same curvature is positively correlated with the second possibility parameter.

[0049] Considering that the more data points with the same curvature as the target data point in the point cloud data, it indicates that the symmetry of the gear presented by the target data point is stronger and it is more likely to be a gear data point, and the gear possibility is greater. Therefore, in the point cloud data, obtain the second possibility parameter of the target data point according to the number of other data points with the same curvature as the target data point; the number of other data points with the same curvature is positively correlated with the second possibility parameter.

[0050] As an example, map the number of data points with the same curvature as the target data point in the point cloud data to the positive correlation normalization function In the function, use the mapped value as the second possibility parameter. Through the number of data points with the same curvature as the target data point in the point cloud data, the repeated characteristics of the curvature between the target data point and other data points in the point cloud data are reflected. The more data points with the same curvature as the target data point, the more obvious the repeated characteristics of the curvature, the more likely it is to be a gear data point, and the greater the gear possibility and the second possibility parameter.

[0051] In other embodiments of the present invention, the implementer can also use other positive correlation normalization functions such as In the function, they are all commonly used existing functions and will not be elaborated.

[0052] Step S203: Obtain the gear possibility of the target data point according to the first possibility parameter and the second possibility parameter of the target data point.

[0053] Among them, both the first possibility parameter and the second possibility parameter are positively correlated with the gear possibility.

[0054] After obtaining the first possibility parameter and the second possibility parameter, the first possibility parameter and the second possibility parameter can be fused to obtain the gear possibility of the target data point. Among them, both the first possibility parameter and the second possibility parameter are positively correlated with the gear possibility.

[0055] As an example, the positive correlation is represented by multiplication, and the product of the first possibility parameter and the second possibility parameter is used as the gear possibility of the target data point.

[0056] In other embodiments of the present invention, the implementer can also represent the positive correlation by weighted summation. For example, the weighted weights of 0.6 and 0.4 are respectively assigned to the first possibility parameter and the second possibility parameter, and the summation result is used as the gear possibility of the target data point.

[0057] There may be manufacturing defects in the production process of gear processing products, and these defects will have a significant impact on the products. During the denoising process of gear point cloud data, the defects are part of the gear and need to be retained. Therefore, it is necessary to distinguish defect data points from noise data points; considering that noise data points exist alone and there is a large density gap between noise points and data points in the neighborhood, while the density of defect data points and other data points in the neighborhood changes continuously or uniformly, the defect data points and noise data points can be distinguished by the density change between the data point and the data points in the neighborhood. Therefore, according to the similarity characteristics of the data density of the target data point and the data density of all data points within the preset neighborhood range of the target data point, the density uniformity parameter of the target data point is obtained.

[0058] Preferably, in an embodiment of the present invention, considering that the more data points there are in the fixed neighborhood of a data point, the greater the data point density around the data point, the data density of each data point is obtained according to the number characteristics of the data points in the second preset neighborhood of each data point, and the data density is represented by the number characteristics; considering that noise points appear randomly and the data density of noise points is quite different from the data density of gear data points, the more similar the data density is and the smaller the difference is, it indicates that the target data point is less likely to be a noise point. To limit the analysis range, the target data point is compared with all data points within the preset neighborhood range, and according to the similarity characteristics of the data density of the target data point and the data density of all data points within the preset neighborhood range of the target data point, the density uniformity parameter of the target data point is obtained; the similarity characteristics of the data density are positively correlated with the density uniformity parameter.

[0059] As an example, the second preset neighborhood of the target data point is a region formed by a sphere with a radius of 3, the center of the sphere is the target data point, and the unit length is 1 mm; the preset neighborhood range of the target data point is the same as the second preset neighborhood of the target data point; the data density within the preset neighborhood range of the target data point is represented by the mean value, and the average value of the data densities of all data points within the preset neighborhood range of the target data point is taken as the comparison data density, representing the data density of all data points within the preset neighborhood range of the target data point; the absolute value of the difference between the data density of the target data point and the comparison data density is used with the negative correlation function After performing negative correlation mapping, it is used as the density uniformity parameter of the target data point; the smaller the absolute value of the difference between the data density of the target data point and the comparison data density, the more similar the data density of the target data point is to that of all data points within the preset neighborhood range, indicating that the target data point is less likely to be a noise point, and the larger the density uniformity parameter.

[0060] In other embodiments of the present invention, the implementer can also set other second preset neighborhoods and preset neighborhood ranges, and the second preset neighborhood and the preset neighborhood range can also be different.

[0061] Step S3: Integrate the gear possibility and the density uniformity parameter of the target data point to obtain the gear membership degree of the target data point; both the gear possibility and the density uniformity parameter are positively correlated with the gear membership degree; according to the gear membership degree, adjust the distance between the target data point and the center point within the preliminary density clustering cluster to which it belongs, and perform secondary density clustering; obtain the denoised point cloud data according to the secondary density clustering result.

[0062] By analyzing the gear possibility and the density uniformity parameter of the target data point, the possibility that the target data point is the point cloud data of a gear is analyzed from the perspectives of distribution similarity, gear structure symmetry, and density similarity respectively. Therefore, further integrate the gear possibility and the density uniformity parameter of the target data point, combine multiple analysis angles, and comprehensively obtain the gear membership degree of the target data point, providing a reliable basis for subsequent optimized clustering and accurate denoising of the point cloud data; both the gear possibility and the density uniformity parameter are positively correlated with the gear membership degree.

[0063] Preferably, in an embodiment of the present invention, since both the gear possibility and the density uniformity parameter are in the range of [0, 1], the product of the gear possibility and the density uniformity parameter of the target data point is directly used as the gear membership degree of the target data point, and the gear membership degree is also distributed in the range of [0, 1].

[0064] In other embodiments of the present invention, if the gear possibility or the density uniformity parameter is not in the range of [0, 1], the product of the gear possibility and the density uniformity parameter of the target data point can be linearly normalized and then used as the gear membership degree.

[0065] Considering that the greater the gear membership degree, the more likely the data point belongs to the gear area, and the smaller the gear membership degree, the more likely it is a noise point. Therefore, by adjusting the distance between the target data point and the center point within the preliminary density clustering cluster according to the gear membership degree and performing secondary density clustering, the true local density change can be better reflected, the accuracy of the clustering result can be improved, the noise points and the true gear data points can be better distinguished, and finally the accuracy of the dimension error detection can be enhanced.

[0066] Preferably, in an embodiment of the present invention, considering that the greater the gear membership degree, the more the target data point belongs to the gear area and the more it needs to be clustered into the cluster, and the closer the distance from the center point of the clustering cluster should be; on the contrary, the smaller the gear membership degree, the farther away from the center point of the clustering cluster; based on this, the gear membership degree of the target data point is negatively correlated and normalized to obtain the distance correction coefficient of the target data point; The product of the distance between the target data point and the center point within the preliminary density clustering cluster to which it belongs and the distance correction coefficient is used as the corrected distance between the target data point and the center point within the preliminary density clustering cluster to which it belongs, and secondary density clustering is performed according to the corrected distance.

[0067] The calculation formula of the corrected distance includes: ; Wherein, represents the corrected distance of the th target data point; represents the gear membership degree of the th target data point; represents the distance correction coefficient of the th target data point; represents the distance between the th target data point and the center point within the preliminary density clustering cluster to which it belongs.

[0068] In the calculation formula of the corrected distance, the gear membership degree is negatively correlated and normalized in the form of subtracting the gear membership degree from the constant 1; the distance between the target data point and the center point within the preliminary density clustering cluster is the Euclidean distance; the distance between the target data point and the center point within the preliminary density clustering cluster is reduced to different degrees. The greater the gear membership degree, the greater the reduction amplitude, and the smaller the gear membership degree, the smaller the reduction amplitude, so that the data points with smaller gear membership degrees are farther from the clustering center, ensuring that the clustering center can reflect the local characteristics of the data points to the greatest extent and finally improving the clustering effect of the entire point cloud data.

[0069] It should be noted that the secondary density clustering also uses the DBSCAN clustering algorithm, and k is also set to 4, which is the same as the preliminary density clustering process and will not be elaborated here.

[0070] In other embodiments of the present invention, the distance correction coefficient can also be mapped to [0, 2], and then the distance correction is performed; the implementer can also use other density clustering methods such as the OPTICS clustering algorithm for density clustering, which are all well-known technical means to those skilled in the art and will not be elaborated herein.

[0071] Based on various distribution characteristics of the data, after performing optimized secondary density clustering, the denoised point cloud data can be obtained according to the results of the secondary density clustering, providing reliable data support for dimensional error detection.

[0072] In one embodiment of the present invention, the outliers in the results of the secondary density clustering are removed, and the clustering clusters are retained as the denoised point cloud data.

[0073] Step S4: Perform dimensional error detection on the gear according to the denoised point cloud data.

[0074] The denoised point cloud data retains the effective geometric information on the gear surface, so that the actual geometric shape of the gear can be better captured, and finally dimensional error detection is performed on the gear according to the denoised point cloud data.

[0075] Preferably, in one embodiment of the present invention, the denoised point cloud data is converted into the gear surface using Poisson surface reconstruction, and the reconstructed gear surface is converted into a mesh model using the triangular meshing method. The measurements include at least the diameter, number of teeth, tooth profile, and pitch data of the gear. By comparing the measured data with the design specification data, the dimensional error of the gear is obtained. These are all well-known technical means to those skilled in the art and will not be elaborated herein.

[0076] In one embodiment of the present invention, after obtaining the dimensional error of the gear, it further includes storing the detection data, and marking and removing the gears with dimensional errors exceeding the design specifications. The design specifications of the gears are related to the specific production process and can be obtained through production manuals or product design drawings, which will not be elaborated herein.

[0077] One embodiment of the present invention also provides a hardware dimensional error detection system based on point cloud processing. The system includes a memory, a processor, and a computer program. The memory is used to store the corresponding computer program, and the processor is used to run the corresponding computer program. When the computer program runs in the processor, it can implement the hardware dimensional error detection method based on point cloud processing described in steps S1 - S4.

[0078] In summary, to solve the technical problem that inaccurate denoising of the point cloud data of gears affects the detection of gear size errors, the present invention first obtains the point cloud data of gears and performs preliminary density clustering on the point cloud data; further analyzes from multiple angles including the distribution similarity characteristics between the target data points and other data points within the preliminary density clustering cluster, the repeated characteristics of curvature, and the similarity characteristics of the data density of the target data points to obtain the gear membership degree of the target data points; further adjusts the distance between the target data points and the central points within the preliminary density clustering cluster according to the gear membership degree, performs secondary density clustering and obtains the denoised point cloud data; finally, performs size error detection on the gears based on the denoised point cloud data. The present invention analyzes the possibility of data points belonging to the gear area from multiple angles, optimizes the density clustering process, accurately removes the noise interference in the point cloud data, obtains reliable detection results for the size errors of gear processing products, ensures the compliance of gear products, and ensures the reliability of gears in applications.

[0079] It should be noted that the above sequence of embodiments of the present invention is only for description and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0080] Each embodiment in this specification is described in a progressive manner. The same or similar parts between the embodiments can be referred to each other, and the key points of each embodiment are the differences from other embodiments.

Claims

1. A method for detecting hardware size errors based on point cloud processing, characterized in that The method includes: Obtaining the point cloud data of the hardware; performing preliminary density clustering on the point cloud data; selecting any data point in any preliminary density clustering cluster as the target data point; Obtaining the hardware performance degree of the target data point according to the uniform distribution characteristics of the data points in the local neighborhood of the target data point; obtaining the hardware possibility of the target data point according to the similarity characteristics of the hardware performance degrees between the target data point and other data points in the preliminary density clustering cluster to which it belongs, in combination with the repeated characteristics of the curvature between the target data point and other data points in the point cloud data; obtaining the density uniformity parameter of the target data point according to the similarity characteristics of the data density of the target data point and the data density of all data points within the preset neighborhood range of the target data point; Fusing the hardware possibility and the density uniformity parameter of the target data point to obtain the hardware membership degree of the target data point; both the hardware possibility and the density uniformity parameter are positively correlated with the hardware membership degree; adjusting the distance between the target data point and the center point in the preliminary density clustering cluster to which it belongs according to the hardware membership degree, and performing secondary density clustering; obtaining the denoised point cloud data according to the secondary density clustering result; Performing size error detection on the hardware according to the denoised point cloud data.

2. The hardware size error detection method based on point cloud processing according to claim 1, wherein The method for obtaining the hardware performance degree includes: Taking the surface formed by the data points within the intersection of the first preset neighborhood of the target data point and the preliminary density clustering cluster to which it belongs as the target surface of the target data point; obtaining the hardware performance degree of the target data point according to the uniform distribution characteristics of the data points within the target surface.

3. The method for detecting hardware dimension errors based on point cloud processing according to claim 2, wherein The method for obtaining the hardware performance degree of the target data point according to the uniform distribution characteristics of the data points within the target surface includes: Within the target surface, taking the target data point as the origin, selecting the data points to be analyzed in the preset direction away from the origin; obtaining the hardware performance degree of the target data point according to the distance consistency between the adjacent data points to be analyzed in the same preset direction; the distance consistency and the hardware performance degree are positively correlated.

4. The method for detecting hardware size error based on point cloud processing according to claim 3, characterized in that The method for obtaining the curvature includes: Taking the maximum curvature of the target data point in all the preset directions as the curvature of the target data point.

5. The hardware dimension error detection method based on point cloud processing according to claim 4, characterized in that The method for obtaining the hardware possibility includes: Obtaining the first possibility parameter of the target data point according to the similarity characteristics of the hardware performance degrees between the target data point and other data points in the preliminary density clustering cluster to which it belongs; the similarity characteristics of the hardware performance degrees are positively correlated with the first possibility parameter; In the point cloud data, obtaining the second possibility parameter of the target data point according to the number of other data points with the same curvature as the target data point; the number of other data points with the same curvature is positively correlated with the second possibility parameter; Obtaining the hardware possibility of the target data point according to the first possibility parameter and the second possibility parameter of the target data point; both the first possibility parameter and the second possibility parameter are positively correlated with the hardware possibility.

6. The hardware dimension error detection method based on point cloud processing according to claim 1, characterized in that The method for obtaining the density uniformity parameter includes: Obtain the data density of each data point according to the quantity characteristics of the data points within the second preset neighborhood of each data point; Obtain the density uniformity parameter of the target data point according to the similarity characteristics between the data density of the target data point and the data densities of all data points within the preset neighborhood range of the target data point; the similarity characteristics of the data density are positively correlated with the density uniformity parameter.

7. The hardware dimension error detection method based on point cloud processing according to claim 1, characterized in that The method for performing secondary density clustering by adjusting the distance between the target data point and the center point within the preliminary density clustering cluster according to the hardware membership degree includes: Perform a negative correlation mapping and normalization on the hardware membership degree of the target data point to obtain the distance correction coefficient of the target data point; Use the product of the distance between the target data point and the center point within the preliminary density clustering cluster to which it belongs and the distance correction coefficient as the corrected distance between the target data point and the center point within the preliminary density clustering cluster to which it belongs, and perform secondary density clustering according to the corrected distance.

8. The method for detecting hardware size error based on point cloud processing according to claim 1, wherein The method for detecting the dimensional error of the hardware according to the denoised point cloud data includes: Convert the denoised point cloud data into the hardware surface using Poisson surface reconstruction, convert the reconstructed hardware surface into a mesh model using the triangular meshing method, measure at least the data of the diameter, number of teeth, tooth profile, and tooth pitch of the hardware, and obtain the dimensional error of the hardware by comparing the measured data and the design specification data.

9. The method for detecting hardware size error based on point cloud processing according to claim 1, wherein Both the preliminary density clustering and the secondary density clustering use the DBSCAN clustering algorithm.

10. A hardware size error detection system based on point cloud processing, the system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method for detecting the dimensional error of the hardware based on point cloud processing according to any one of claims 1 to 9.

Images