A method for screening noise points based on hole group constraints

Abstract

The present application relates to a noise point screening method based on hole group constraints, including: projecting the three-dimensional point set of the hole group on the bridge hole plate workpiece onto a plane for dimensionality reduction to form a two-dimensional point set of the hole group; establishing a hole group constraint model based on the two-dimensional point set of the hole group; eliminating noise points according to the hole group constraint model; establishing the hole group constraint model based on the two-dimensional point set includes: the hole group distribution forms a parallelogram unit, which is defined as a grid unit, and setting the prior probability distribution of the parameters of the grid unit; under the condition of setting the prior probability distribution of the parameters, calculating the corresponding probability of each point in the two-dimensional point set of the hole group and the nearest corner point in the grid unit, defined as a likelihood function; based on the prior probability distribution and the likelihood function, determining the maximized posterior probability and the parameters of the corresponding grid unit to obtain the hole group constraint model. The present application first performs dimensionality reduction and simplification processing, and then establishes the hole group constraint model based on statistical probability theory, and applies it to noise point screening, so that the screening is accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of noise screening, and in particular to a noise point screening method based on hole group constraints. Background Art

[0002] The bridge orifice plate workpiece has many hole-shaped structures, such as bolt holes, through holes, etc. Each hole is used for bolting positioning and plays an important role in the structural positioning of the workpiece. Therefore, it is necessary to measure the relative position between each hole.

[0003] The holes on a bridge orifice plate form a cluster, distributed in a regular array. Photographic recognition of these points produces a point cloud representing the 3D coordinates of each hole. However, this cloud contains many noise points caused by misidentification. To determine the relative positions of the holes, these noise points must be removed and the relative positions of the remaining valid points extracted. Conventional techniques typically remove these noise points manually, which is inefficient and inaccurate. Summary of the Invention

[0004] The present application provides a noise point screening method based on hole group constraints, which solves the technical problems of low efficiency and inaccuracy caused by manual removal of noise points in related technologies.

[0005] The embodiment of the present application provides a noise point screening method based on hole group constraints, which includes the following steps:

[0006] Projecting a three-dimensional point set of a hole group on a bridge orifice plate workpiece onto a plane for dimensionality reduction to form a two-dimensional point set of the hole group;

[0007] Establishing a hole group constraint model based on the two-dimensional point set of the hole group;

[0008] Eliminate noise points according to the hole cluster constraint model;

[0009] Wherein, establishing a hole group constraint model based on the two-dimensional point set includes:

[0010] The hole clusters are distributed to form parallelogram units, which are defined as grid units, and the prior probability distribution of the parameters of the grid units is set;

[0011] Under the condition of setting the prior probability distribution of the parameters, calculating the corresponding probability between each point in the two-dimensional point set of the hole cluster and the nearest corner point in the grid unit, which is defined as a likelihood function;

[0012] The parameters of the maximized posterior probability and the corresponding grid unit are determined based on the prior probability distribution and the likelihood function to obtain the hole cluster constraint model.

[0013] In one embodiment, by PCAMethod The three-dimensional point set of the hole group on the bridge orifice plate workpiece is projected onto a plane for dimensionality reduction.

[0014] In one embodiment, projecting the three-dimensional point set of the hole group on the bridge orifice plate workpiece onto a plane to perform dimensionality reduction to form the two-dimensional point set of the hole group includes:

[0015] The three-dimensional point set of the hole group is set as: ,in, is a three-dimensional point in the hole cluster, X is a set of three-dimensional points;

[0016] The three-dimensional point set of the hole group is centralized.

[0017] The calculation formula is: ; ;

[0018] Determine the covariance matrix C , the calculation formula is:

[0019] in is the data matrix after data centering, ;

[0020] Pair covariance matrix C Perform eigendecomposition and determine the corresponding eigenvalues and eigenvectors;

[0021] The eigenvectors corresponding to the two largest eigenvalues and Defined as the principal plane, each three-dimensional point is projected onto the principal plane to form a two-dimensional point set of the hole cluster S ;

[0022] The calculation formula is:

[0023] in, are the coordinates of a 3D point on the 2D principal plane, is a two-dimensional real vector space.

[0024] In one embodiment, the hole clusters are distributed to form parallelogram units, which are defined as grid units. Setting the prior probability distribution of the parameters of the grid units includes:

[0025] The hole groups are distributed in triangular units AOB Continuously repeating to form parallelogram units AOBC , the parallelogram unit AOBC is defined as said grid cell;

[0026] The grid cell parameters are set to length a ,width b , angle θ and angle , where the length a Main axis OA Length and width b Main axis OB Length, angle θ=∠BOA , angle Main axis OB With the coordinate axis x The angle between the axes;

[0027] Set the parameters of the grid cells a 、 b 、 θ and The prior probability distribution of for:

[0028] .

[0029] In one embodiment, the prior probability distribution of the parameters is set according to the machining accuracy and measurement accuracy of the bridge hole plate workpiece.

[0030] In one embodiment, under the condition of setting the prior probability distribution of the parameters, the corresponding probability of each point in the two-dimensional point set of the hole cluster and the nearest corner point in the grid unit is calculated, which is defined as the likelihood function including:

[0031] The two-dimensional point set of the hole group S Each point in Convert to O In the local grid coordinate system with point as the origin, the calculation formula is:

[0032] in, and are the two-dimensional point sets of the hole group S The point relative to O Point on the main axis OA and spindle OB The projection distance in the direction, MOD is a modular operation, and It is the coordinate of the point after the projection distance in two directions is localized to a grid cell;

[0033] In setting the prior probability distribution of the parameters In the case of S Each point in Euclidean distance to the corner of the grid cell ,

[0034] The calculation formula is: ;

[0035] in, Yes i The corresponding j The coordinates of the corner points, where j=1,2,3,4 , corresponding to ;

[0036] Evaluate the point by the Gaussian function i Belongs to the corresponding j The probability of a corner point ,

[0037] The calculation formula is: ;

[0038] in, is the standard deviation of the prior probability distribution;

[0039] The maximum probability corner point is selected as its corresponding grid point, and its Gaussian function value is the degree of conformity of the point with respect to the grid unit determined by the current parameters, expressed as: ;

[0040] The sum of the Gaussian function values of all points relative to the grid cells determined by the current parameters is defined as the likelihood function, which is expressed as: .

[0041] In one embodiment, determining the maximized posterior probability and the corresponding grid unit parameters based on the prior probability distribution and the likelihood function to obtain the hole cluster constraint model includes:

[0042] Determine the posterior probability based on the prior probability distribution and the likelihood function ,

[0043] The calculation formula is: ;

[0044] Determine the maximum posterior probability, the calculation formula is: , to obtain the parameters of the grid unit corresponding to the maximum posterior probability, generate the overall grid unit, and obtain the hole group constraint model.

[0045] In one embodiment, when generating the entire grid unit, the two-dimensional hole set of the hole group is determined. S Each point The grid cell index to which it belongs and ,in round is the nearest rounding function, which represents its position in the grid unit.

[0046] In one embodiment, removing noise points according to the hole cluster constraint model includes:

[0047] Based on the hole group constraint model, set the two-dimensional point set of the hole group S Points in Satisfy the normal distribution with mean 0 and standard deviation σ;

[0048] Its Euclidean distance to the corner point of the grid cell The minimum value does not exceed 3 When , it is the true point;

[0049] Its Euclidean distance to the corner point of the grid cell Minimum value exceeds 3 When , it is a noise point;

[0050] The calculation formula is: ;

[0051] .

[0052] In one embodiment, the noise point screening method based on hole cluster constraints further includes: automatically measuring the relative positions between adjacent points according to the parameters of the grid unit.

[0053] The beneficial effects of the technical solutions provided in the embodiments of the present application include:

[0054] The present application provides a noise point screening method based on hole group constraints. By utilizing the hole group distribution law on the bridge orifice plate workpiece, dimensionality reduction is first performed to simplify the processing. Then, a hole group constraint model is established based on statistical probability theory and applied to noise point screening. The parameters are optimized to achieve noise removal and data classification at the same time. When establishing the hole group constraint model, a priori probability distribution is first set for the parameters of the grid unit to provide a statistically based starting point, ensuring that the parameter search process starts from a reasonable range and increasing the possibility of finding the global optimal solution. The likelihood function is then used to determine the maximum posterior probability to optimize the parameters of the grid unit. This is not only based on minimizing the geometric error, but also uses statistical probability theory to guide the search process. It more comprehensively considers the probabilistic characteristics and uncertainty of the data, thereby combining the noise point screening process with the data classification and grid construction process. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0056] Figure 1 Flowchart of a noise point screening method based on hole group constraints in one embodiment of the present invention.

[0057] Figure 2 This is a schematic diagram of a bridge orifice plate workpiece in one embodiment of the present invention.

[0058] Figure 3 for Figure 2 Schematic diagram of the three-dimensional point cloud of the hole cluster on the bridge orifice plate workpiece shown.

[0059] Figure 4 for Figure 3 Schematic diagram of the two-dimensional point set of the hole cluster on the bridge orifice plate workpiece shown.

[0060] Figure 5 FIG. 4 is a schematic diagram of a grid unit according to an embodiment of the present invention.

[0061] Figure 6 This is the calculation result of maximizing the posterior probability in one embodiment of the present invention.

[0062] Figure 7 FIG. 1 is a schematic diagram of removing noise points in one embodiment of the present invention. DETAILED DESCRIPTION

[0063] In order to enable those skilled in the art to better understand the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0064] The embodiment of the present application provides a noise point screening method based on hole group constraints, which can solve the technical problems of low efficiency and inaccuracy caused by manual removal of noise points in related technologies.

[0065] like Figure 1 As shown, Figure 1 Flowchart of a noise point screening method based on hole group constraints in one embodiment of the present invention.

[0066] This embodiment provides a noise point screening method based on hole group constraints, which includes the following steps:

[0067] Step S1: Projecting a three-dimensional point set of a hole group on a bridge orifice plate workpiece onto a plane for dimensionality reduction to form a two-dimensional point set of the hole group;

[0068] Step S2: establishing a hole group constraint model based on the two-dimensional point set of the hole group;

[0069] Step S3, removing noise points according to the hole cluster constraint model;

[0070] Wherein, step S2, establishing a hole cluster constraint model based on a two-dimensional point set, includes:

[0071] Step S21: The hole clusters are distributed to form parallelogram units, which are defined as grid units, and the prior probability distribution of the parameters of the grid units is set;

[0072] Step S22: Under the condition of setting the prior probability distribution of the parameters, the corresponding probability between each point in the two-dimensional point set of the hole cluster and the nearest corner point in the grid cell is calculated, which is defined as the likelihood function;

[0073] Step S23: Determine the parameters of the maximized posterior probability and its corresponding grid unit based on the prior probability distribution and the likelihood function to obtain a hole cluster constraint model.

[0074] This embodiment provides a noise point screening method based on hole group constraints. It utilizes the hole group distribution law on the bridge orifice plate workpiece, first performs dimensionality reduction to simplify the processing, then establishes a hole group constraint model based on statistical probability theory, applies it to noise point screening, optimizes parameters and simultaneously achieves noise removal and data classification. When establishing the hole group constraint model, a priori probability distribution is first set for the parameters of the grid unit to provide a statistically based starting point, ensuring that the parameter search process starts from a reasonable range and increasing the possibility of finding the global optimal solution; then, the likelihood function is used to determine the maximum posterior probability to optimize the parameters of the grid unit. This method is not only based on minimizing geometric errors, but also uses statistical probability theory to guide the search process, and more comprehensively considers the probabilistic characteristics and uncertainty of the data, thereby combining the noise point screening process with the data classification and grid construction process.

[0075] Each step is described and explained in detail below.

[0076] like Figure 2 As shown, Figure 2 This is a schematic diagram of a bridge orifice plate workpiece in one embodiment of the present invention.

[0077] Taking a bridge orifice plate workpiece as an example, the holes on the workpiece are divided into two hole groups, upper and lower. Every four adjacent points in the hole group can form a rectangle with a certain length and width.

[0078] In one embodiment, by PCA Methods The three-dimensional point set of the hole group on the bridge orifice plate workpiece is projected onto a plane for dimensionality reduction.

[0079] In practice, if the workpiece is flat, each hole group is distributed on the same plane (i.e., the workpiece surface). Therefore, projecting the three-dimensional point set of all holes onto the plane for dimensionality reduction can simplify the operation processing. For the case where most points are on the same plane, use PCAIt is more convenient to perform dimensionality reduction using a method called principal component analysis (PCA). Its purpose is to reduce the dimensionality of the data by retaining the main features in the data while minimizing the loss of information. PCA An orthogonal transformation transforms a set of possibly correlated variables into a set of linearly uncorrelated variables. These new variables are called principal components. Each principal component is a linear combination of the original data, but each combination has a different weight, so that the first principal component contains the most variance, the second principal component contains the second most variance, and so on.

[0080] like Figure 3 and Figure 4 As shown, Figure 3 for Figure 2 Schematic diagram of the three-dimensional point cloud of the hole cluster on the bridge orifice plate workpiece shown. Figure 4 for Figure 3 Schematic diagram of the two-dimensional point set of the hole cluster on the bridge orifice plate workpiece shown. Figure 4 The pore group is divided into two groups, the upper and lower pore groups. The lower pore group is taken as an example as the two-dimensional point set to be analyzed.

[0081] In one embodiment, step S1, projecting a three-dimensional point set of a hole group on a bridge orifice plate workpiece onto a plane for dimensionality reduction to form a two-dimensional point set of the hole group, includes:

[0082] Step S11: The three-dimensional point set of the hole group is set as: ,in, is a three-dimensional point in the hole group, X is a set of three-dimensional points.

[0083] Step S12: centralize the three-dimensional point set of the hole cluster, that is, subtract the average value of each dimension so that the mean value of the data in each dimension is 0, which makes the result more accurate;

[0084] The calculation formula is: ; ;

[0085] Step S13: Determine the covariance matrix C , the calculation formula is: ;

[0086] in is the data matrix after data centering, .

[0087] Step S14: covariance matrix C Perform eigendecomposition and determine the corresponding eigenvalues and eigenvectors;

[0088] Specifically, these eigenvectors represent the principal component directions of the data, and the eigenvalues represent the variance of the data in each direction.

[0089] Step S15: The eigenvectors corresponding to the two largest eigenvalues and Defined as the principal plane, each 3D point is projected onto the principal plane to form a 2D point set of the hole cluster S ;

[0090] The calculation formula is:

[0091] in, are the coordinates of a 3D point on the 2D principal plane, is a two-dimensional real vector space.

[0092] The principal plane is the plane with the largest data variance, which is also the plane of the workpiece where the hole is located. For points that deviate significantly from the principal plane, the distance between them and the plane is calculated, and a threshold distance is set. Points with excessive distance are filtered out.

[0093] Through the above scheme, PCA The method is used to reduce the dimension, compress the data space, and intuitively express the characteristics of multivariate data in the low-dimensional space. If there are more noise points with more chaotic distribution, you can use RANSAC Method to fit a plane.

[0094] like Figure 5 As shown, Figure 5 FIG. 4 is a schematic diagram of a grid unit according to an embodiment of the present invention.

[0095] In one embodiment, in step S21, the hole clusters are distributed to form parallelogram units, which are defined as grid units. The prior probability distribution of the parameters of the grid units is set to include:

[0096] Step S211: hole clusters are distributed in triangular units. AOB Continuously repeating to form parallelogram units AOBC , the parallelogram element AOBC Defined as a grid cell.

[0097] Specifically, if Figure 2 The center position of each four adjacent holes forms a bridge hole workpiece with a length of , width is According to the hole group distribution law, it is regarded as a rectangle composed of a triangular unit. AOB Continuously repeating to form parallelogram units AOBC , the parallelogram element AOBCDefined as a grid cell, each real data point, i.e., a hole, should fall on a grid point (or corner point) of the grid cell.

[0098] Step S212: The parameter of the grid unit is set to length a ,width b , angle θ and angle , where the length a Main axis OA Length and width b Main axis OB Length, angle θ=∠BOA , angle Main axis OB With the coordinate axis x The angle between the axes.

[0099] Specifically, the shape of the grid cell involves two side lengths a 、 b and angle θ Three parameters, whose positions involve the coordinates of the vertex O and one of its sides and the coordinate axis x Axis angle Parameters, the angle is not fixed here θ It is a right angle to accommodate more complex distribution situations.

[0100] Step S213: Set the parameters of the grid unit a 、 b 、 θ and The prior probability distribution of for: .

[0101] in, is the uniform distribution function, It obeys the normal distribution with mean u and standard deviation σ.

[0102] Through the above scheme, the parameter initialization of the grid cells is guided by the probability distribution (uniform and normal distribution), and the set prior probability distribution is used as the starting point of the statistical basis. Describes the beliefs about various parameters of the grid cell shape before observing the data.

[0103] In one embodiment, the prior probability distribution of the parameters is set according to the machining accuracy and measurement accuracy of the bridge hole plate workpiece.

[0104] Specifically, the prior probability distribution of the set parameters needs to be determined based on actual conditions, that is, it is related to the actual horizontal and vertical spacing of the holes. For example, the design size distribution of the hole group is a positive grid with both horizontal and vertical hole spacing of 100 mm. However, the machining process has large errors, and the actual hole spacing may fluctuate within 100 ± 3 mm. In this case, the expected actual hole spacing distribution can be set as: a and b All have a mean of 100 and σ =1 Gaussian distribution, that is, N(100,1); because 3 σ =3, or it can be set to a uniform distribution of 97~103, that is U [97,103]. The same is true for measurement accuracy. If the actual hole spacing is 100mm, but the measurement is inaccurate, it may be measured as 100±3mm. Then, assuming that the actual hole distribution is N (100,1) or U [97,103].

[0105] In one embodiment, step S22, under the condition of setting the prior probability distribution of the parameters, calculate the corresponding probability of each point in the two-dimensional point set of the hole cluster and the nearest corner point in the grid unit, which is defined as the likelihood function including:

[0106] Step S221: The two-dimensional point set of the hole group S Each point in Convert to O In the local grid coordinate system with point as the origin, the calculation formula is:

[0107]

[0108] in, and are the two-dimensional point sets of the hole group S The point relative to O Point on the main axis OA and spindle OB The projection distance in the direction, MOD is a modular operation, and It is the coordinate of the point after localizing the projected distance in both directions to a grid cell.

[0109] Put all points in a local cell and measure the contribution of each point to the shape of the grid cell. Specifically, ;

[0110] in, ;

[0111] ;

[0112] .

[0113] Compute the cell index of a point in a grid cell: and ,in round It is a nearest rounding function that represents its position in the grid unit. The nearest rounding function avoids positioning grid errors caused by small errors.

[0114] Step S222: Setting the prior probability distribution of parameters In the case of , calculate the two-dimensional point set of the hole group S Each point in Euclidean distance to the corner of the grid cell ,

[0115] The calculation formula is: ;

[0116] in, Yes i The corresponding j The coordinates of the corner points, where j=1,2,3,4 , corresponding to .

[0117] Specifically, for each grid cell, the parallelogram cell There are four corner points: O,A,B,C . O is the origin, A yes O Point Edge a The vertex of the axis, B yes O Point Edge b The vertex of the axis, the remaining corner points are C .

[0118] Step S223: Evaluate the point using the Gaussian function i Belongs to the corresponding j The probability of a corner point ,

[0119]

[0120] in, is the standard deviation of the prior probability distribution;

[0121] Step S224: Select the corner point with the maximum probability as its corresponding grid point, and use its Gaussian function value as the degree of conformity of the point with respect to the grid unit determined by the current parameters.

[0122] Expressed as: .

[0123] Specifically, the two-dimensional point set of the hole group S Each point in The likelihood contribution of is determined by its probability to the nearest corner point, so the probability of the four corner points is , the largest point is selected as its corresponding grid point, and its Gaussian function value is used to represent the degree of conformity of that point with the grid cell determined by the current parameters. Selecting the maximum value means that each point is most likely to belong to the nearest corner. This is a simplistic assumption, but it is widely used in practice due to its computational simplicity and intuitive rationality.

[0124] Step S225: The sum of the Gaussian function values of all points relative to the grid unit determined by the current parameters is defined as the likelihood function, which is expressed as: .

[0125] Specifically, the likelihood function quantifies the probability distribution given the model parameters. In the case of , the current data set is observed to be a two-dimensional point set S probability.

[0126] according to , and obtain the likelihood function,

[0127] Expressed as: .

[0128] In one embodiment, step S23, determining the parameters of the grid unit that maximizes the posterior probability and its corresponding posterior probability based on the prior probability distribution and the likelihood function to obtain the hole cluster constraint model includes:

[0129] Step S231: Determine the posterior probability based on the prior probability distribution and the likelihood function. ,

[0130] The calculation formula is: .

[0131] Specifically, the posterior probability is the probability of parameter update after given observation data. The prior belief (prior probability, i.e., the distribution law of the hole group) and the actual observed data (i.e., the position of each hole in the two-dimensional point set) are combined through the likelihood function to update the understanding of the parameters.

[0132] According to Bayes' formula, the posterior probability It is given by the product of the prior and the likelihood function: .

[0133] Step S232: determine the maximum posterior probability, the calculation formula is: , in order to obtain the parameters of the grid cell corresponding to the maximum posterior probability, generate the overall grid cell, and obtain the hole group constraint model.

[0134] Specifically, the process of maximizing the posterior probability is regarded as a minimum optimization problem of the negative natural logarithm of the posterior probability. The goal of parameter optimization is to find the parameter value that maximizes the posterior probability, which is equivalent to the most likely parameter setting after considering the data information to seek the best statistical estimate of the parameter.

[0135] In one embodiment, when generating the overall grid unit, the two-dimensional hole set of the hole group is determined. S Each point The grid cell index to which it belongs and ,in round is the nearest rounding function, which represents its position in the grid unit.

[0136] Specifically, after obtaining the grid parameters corresponding to the maximum a posteriori probability, according to the ellipse parameter value corresponding to the maximum a posteriori probability, that is, , and the initial value (the estimated value is set manually according to the working condition) to generate the overall grid unit; the two-dimensional hole set of the hole group S Each point The grid cell index to which it belongs and , as the position representation of each point in the grid cell, is used to determine whether the point is in an adjacent grid cell. The adjacent points are indexed as adjacent grid points, so that the subsequent calculation of the spacing between adjacent grid points is the side length of the grid cell.

[0137] like Figure 6 As shown, Figure 6 This is the calculation result of maximizing the posterior probability in one embodiment of the present invention.

[0138] Depend on Figure 6 It can be seen that the × in the legend represents Figure 4 The grid cell shape calculated from the points in , the real hole should appear at the position of ×, Figure 6 The corresponding grid cells obtained by calculating the grid cell parameters according to the above steps are shown as the basis for subsequent removal of noise points.

[0139] In one embodiment, step S3, removing noise points according to the hole cluster constraint model, includes:

[0140] Based on the hole cluster constraint model, set the two-dimensional point set of the hole cluster S Points in Satisfy the normal distribution with mean 0 and standard deviation σ;

[0141] Its Euclidean distance to the corner point of the grid cell When the minimum value does not exceed 3σ, it is the true point;

[0142] Its Euclidean distance to the corner point of the grid cell When the minimum value exceeds 3σ, it is a noise point;

[0143] The calculation formula is: ;

[0144] .

[0145] Specifically, after determining the grid cell to which each point belongs, the grid point to which it belongs is the vertex closest to the point among the four vertices of the cell to which it belongs. S Points in Satisfies the normal distribution with mean 0 and standard deviation σ. Since the probability outside 3σ is very small, the points outside 3σ are directly regarded as noise points for filtering, not real points. OABC If the minimum value of the distances between these four points is less than 3σ, it is considered to be a real point, not a noise point.

[0146] This approach optimizes parameters while dynamically removing noise and classifying data, providing a mathematically rigorous and practical model optimization framework. Furthermore, for each grid point, if there are multiple points corresponding to it, the points that are closer to the grid point should be selected.

[0147] like Figure 7 As shown, Figure 7 FIG. 1 is a schematic diagram of removing noise points in one embodiment of the present invention.

[0148] Depend on Figure 7 It can be seen that after the noise points are removed by the method provided in this application, points representing real holes are left, the noise points are filtered out, and the screening efficiency is higher and more accurate.

[0149] In one embodiment, the noise point screening method based on hole cluster constraints further includes: step S4, automatically measuring the relative positions between adjacent points according to the parameters of the grid unit.

[0150] The position and shape of the grid cells are obtained through the above steps, and the two-dimensional point set of the hole group corresponding to each grid point is retained. S By calculating the side length of each grid unit composed of the remaining points, the relative position of each hole can be obtained intuitively.

[0151] Depend on Figure 7 As can be seen, the figure shows the spacing between adjacent holes, that is, the relative hole positions are automatically measured.

[0152] It should be noted that the above-mentioned serial numbers of the embodiments of the present application are only for description and do not represent the advantages and disadvantages of the embodiments. The terms "including" and "having" in the specification and claims of this application and the above-mentioned drawings, as well as any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device comprising a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units inherent to these processes, methods, products or devices. The terms "first", "second" and "third" are used to distinguish different objects, etc., and do not represent a sequential order, nor do they limit "first", "second" and "third" to being different types.

[0153] In the description of the embodiments of this application, the words "exemplary," "for example," or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary," "for example," or "for example" in the embodiments of this application should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary," "for example," or "for example" is intended to present the relevant concepts in a concrete manner.

[0154] In the description of the embodiments of the present application, unless otherwise specified, “ / ” means or, for example, A / B can mean A or B; “and / or” in the text is merely a description of the association relationship of associated objects, indicating that three relationships may exist, for example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, in the description of the embodiments of the present application, “multiple” refers to two or more than two.

[0155] In some processes described in the embodiments of this application, multiple operations or steps are included that appear in a specific order. However, it should be understood that these operations or steps may not be executed in the order in which they appear in the embodiments of this application or may be executed in parallel. The sequence numbers of the operations are only used to distinguish different operations, and the sequence numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations or steps may be executed in sequence or in parallel, and these operations or steps may be combined.

[0156] The above are only preferred embodiments of the present application and do not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A noise point screening method based on hole group constraint, characterized in that: It includes the following steps: Projecting a three-dimensional point set of a hole group on a bridge orifice plate workpiece onto a plane for dimensionality reduction to form a two-dimensional point set of the hole group; Establishing a hole group constraint model based on the two-dimensional point set of the hole group; Eliminate noise points according to the hole cluster constraint model; Wherein, establishing a hole group constraint model based on the two-dimensional point set includes: The hole clusters are distributed to form parallelogram units, which are defined as grid units, and the prior probability distribution of the parameters of the grid units is set; Under the condition of setting the prior probability distribution of the parameters, calculating the corresponding probability between each point in the two-dimensional point set of the hole cluster and the nearest corner point in the grid unit, which is defined as a likelihood function; Determining the parameters of the maximized posterior probability and the corresponding grid unit based on the prior probability distribution and the likelihood function to obtain the hole cluster constraint model; Eliminating noise points according to the hole group constraint model includes: Based on the hole group constraint model, set the two-dimensional point set of the hole group S Point in Satisfy the normal distribution with mean 0 and standard deviation σ; Its Euclidean distance to the corner point of the grid cell When the minimum value does not exceed 3σ, it is the true point; Its Euclidean distance to the corner point of the grid cell When the minimum value exceeds 3σ, it is a noise point; The calculation formula is: ; ； in, are the coordinates of a three-dimensional point on the two-dimensional principal plane, and O, A, B, and C are the corner points of the grid unit.

2. The noise point screening method based on hole group constraint according to claim 1, characterized in that: The three-dimensional point set of the hole group on the bridge hole plate workpiece is projected onto a plane by the PCA method to reduce the dimension.

3. The noise point screening method based on hole group constraint according to claim 2, characterized in that: The projecting of the three-dimensional point set of the hole group on the bridge orifice plate workpiece onto a plane for dimensionality reduction to form the two-dimensional point set of the hole group includes: The three-dimensional point set of the hole group is set as: ,in, is a three-dimensional point in the hole cluster, X is a set of three-dimensional points; The three-dimensional point set of the hole group is centralized. The calculation formula is: ; ; Determine the covariance matrix C , the calculation formula is: ; in is the data matrix after data centering, ; Pair covariance matrix C Perform eigendecomposition and determine the corresponding eigenvalues and eigenvectors; The eigenvectors corresponding to the two largest eigenvalues and Defined as the principal plane, each three-dimensional point is projected onto the principal plane to form a two-dimensional point set of the hole cluster S ; The calculation formula is: ; in, are the coordinates of a 3D point on the 2D principal plane, is a two-dimensional real vector space.

4. The noise point screening method based on hole group constraint according to claim 1, characterized in that: The hole clusters are distributed to form parallelogram units, which are defined as grid units. The prior probability distribution of the parameters of the grid units is set as follows: The hole groups are distributed in triangular units AOB Continuously repeating to form parallelogram units AOBC , the parallelogram unit AOBC is defined as said grid cell; The grid cell parameters are set to length a ,width b , angle θ and angle , where the length a Main axis OA Length and width b Main axis OB Length, angle θ=∠BOA , angle Main axis OB With the coordinate axis x The angle between the axes; Set the parameters of the grid cells a 、 b 、 θ and The prior probability distribution of for: 。 5. The noise point screening method based on hole group constraint according to claim 4, characterized in that: The prior probability distribution of the parameters is set according to the machining accuracy and measurement accuracy of the bridge hole plate workpiece.

6. The noise point screening method based on hole group constraint according to claim 4, characterized in that: Under the condition of setting the prior probability distribution of the parameters, the corresponding probability of each point in the two-dimensional point set of the hole cluster and the nearest corner point in the grid unit is calculated, which is defined as the likelihood function including: The two-dimensional point set of the hole group S Each point in Convert to O In the local grid coordinate system with point as the origin, the calculation formula is: ; in, and are the two-dimensional point sets of the hole group S The point relative to O Point on the main axis OA and spindle OB The projection distance in the direction, MOD is a modular operation, and It is the coordinate of the point after the projection distance in two directions is localized to a grid cell; In setting the prior probability distribution of the parameters In the case of S Each point in Euclidean distance to the corner of the grid cell , The calculation formula is: ; in, Yes i The corresponding j The coordinates of the corner points, where j=1,2,3,4 , corresponding to ; Evaluate the point by the Gaussian function i Belongs to the corresponding j The probability of a corner point , The calculation formula is: ; in, is the standard deviation of the prior probability distribution; The maximum probability corner point is selected as its corresponding grid point, and its Gaussian function value is the degree of conformity of the point with respect to the grid unit determined by the current parameters, expressed as: ; The sum of the Gaussian function values of all points relative to the grid cells determined by the current parameters is defined as the likelihood function, which is expressed as: .

7. The noise point screening method based on hole group constraint according to claim 6, characterized in that: The method of determining the maximized posterior probability and the parameters of the corresponding grid unit based on the prior probability distribution and the likelihood function to obtain the hole cluster constraint model includes: Determine the posterior probability based on the prior probability distribution and the likelihood function , The calculation formula is: ; Determine the maximum posterior probability, the calculation formula is: , to obtain the parameters of the grid unit corresponding to the maximum posterior probability, generate the overall grid unit, and obtain the hole group constraint model.

8. The noise point screening method based on hole group constraint according to claim 7, characterized in that: When generating the entire grid unit, determine the two-dimensional hole set of the hole group S Each point The grid cell index to which it belongs and ,in round is the nearest rounding function, which represents its position in the grid unit.

9. The noise point screening method based on hole group constraint according to claim 1, characterized in that: The noise point screening method based on hole group constraints further includes: automatically measuring the relative positions between adjacent points according to the parameters of the grid unit.

Images