An Evaluation Method for Standard Surface Feature Parameters Evaluation Algorithm
By performing point-group sampling and random perturbation processing on the standard surface of the measurement system, a standard evaluation method is constructed, which solves the inconsistency problem of software algorithm evaluation and improves the accuracy and consistency of the measurement system.
Patent Information
- Application Number
- CN202210026036.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-01-11
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2042-01-11
AI Technical Summary
The existing technology lacks a unified evaluation method for supporting software algorithms for measuring systems, resulting in inconsistent evaluation results, affecting the accuracy and consistency of geometric parameter evaluation.
By performing point-set sampling of typical standard surfaces and adding random perturbations, a standard input data set is constructed, a standard evaluation method is constructed, the output results of the evaluated algorithm are evaluated, and its output is corrected based on the evaluation results to strip away the equipment hardware error and improve the measurement system accuracy.
It realizes the unified evaluation object of the software algorithm, independently evaluates the accuracy of geometric parameters, and improves the accuracy of the measurement system and the consistency of the results.
Smart Images

Figure CN114511197B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to an evaluation method for a standard curved surface characteristic parameter evaluation algorithm, and belongs to the field of metrology. Background Art
[0002] In the traditional field of geometric measurement, the research object often focuses on the comprehensive evaluation of the metrological characteristics of the measuring equipment, such as the evaluation of the measurement results of the three-dimensional coordinate measuring machine according to relevant standards such as ISO10360 and JJF1064. With the development of digital and intelligent measurement technology, more and more measurement results directly rely on the measurement software supporting the measurement system, so the role of the evaluation algorithm in the software in the measurement is becoming increasingly prominent. In many metrological fields, the metrological evaluation results rely on the software calculation after the measurement. When such software is used to provide the final measurement results, the accuracy of its value needs to be traced. However, the traditional metrological evaluation of the measurement equipment as the research object is a comprehensive evaluation of the conformity of the equipment's metrological characteristics. The evaluation results not only include the differences in software algorithms, but also couple the influence of factors such as equipment hardware and environment, and cannot reflect the impact of software algorithms on the accuracy of measurement results alone. At present, some high-level metrological technical institutions are conducting relevant research on the verification, comparison, and evaluation of the software algorithms supporting the measurement system, but they are not comprehensive, and the evaluation objects and evaluation methods are not unified. At the same time, there is a lack of research on this part in China.
[0003] At present, high-level metrology technology institutions abroad have conducted research on the testing, comparison and verification of Gaussian fitting algorithms and Chebyshev fitting algorithms for basic geometric elements such as lines, planes, circles, cylinders, cones, and spheres in measurement software. They have also developed mature online verification software and provided commercial services. However, there are few publicly published patent inventions and related research on core issues such as the definition of test data sets, generation criteria, addition of random perturbations, and fixed values of test data sets.
[0004] The ISO standard and my country's national standards give clear definitions for geometric parameters such as sphericity, taper and cylindricity, but do not provide a unified standard evaluation algorithm that meets the definitions. This directly leads to large divergence and inconsistent values between the evaluation results obtained using different evaluation algorithms, which seriously affects the accuracy and consistency of geometric parameter evaluation. This has been confirmed in many international measurement comparisons. Summary of the invention
[0005] In view of the lack of evaluation algorithm for measurement system supporting software in the prior art, the present invention discloses an evaluation method for standard surface feature parameter evaluation algorithm, the purpose of which is to decouple the influence of factors such as device hardware on software evaluation results, and to improve the accuracy of the measurement system by correcting the output results of the evaluated algorithm.
[0006] The objective of the present invention is achieved through the following technical solutions:
[0007] The present invention discloses an evaluation method for a standard surface characteristic parameter evaluation algorithm, which performs point set sampling on a typical standard surface with a standard analytical equation, adds disturbance to the discrete point set, and constructs a series of standard input data sets; constructs a standard evaluation method for the series of standard input data sets; uses the shape error, standard deviation, and fitted characteristic parameter values obtained by the standard evaluation method as evaluation criteria for evaluating other algorithms; and uses the deviation between the output result of the evaluated algorithm and the evaluation criteria as a correction reference for the evaluated algorithm to improve the accuracy of the measurement system.
[0008] The present invention discloses a method for evaluating a standard surface characteristic parameter evaluation algorithm, comprising the following steps:
[0009] Step 1: Perform point set sampling on the standard analytical equations of several typical standard surfaces such as plane, circular surface, spherical surface, cylindrical surface and conical surface to obtain several discrete point sets that satisfy the standard equations;
[0010] Step 2: Add random perturbations that conform to the normal distribution to the point set obtained in step 1 to obtain standard input data;
[0011] Step 3: Considering the influence of the distribution position and posture of the geometric body in space on the obtained point set, the standard input data in step 2 is rotated and translated to obtain a series of standard input data sets, which serve as the unified evaluation object for software algorithm evaluation;
[0012] Step 4: Construct standard evaluation method;
[0013] For the series of standard input data sets obtained in step 3, the corresponding geometric model is constructed through the fitting algorithm according to the different characterization parameters of the simulation elements, so that the sum of the squares of the offset errors is minimized.
[0014] Step 5: Input the series of standard input data sets in step 3 into the software algorithm to be evaluated, and output the surface parameters to be evaluated;
[0015] Step 6: Using the standard evaluation method obtained in step 4, evaluate the series of standard input data sets in step 3 to obtain the shape error and standard deviation of the geometric surface;
[0016] Step 7: Compare the outputs of step 5 and step 6 to obtain the evaluation results, and modify the results of the evaluated algorithm based on the evaluation results.
[0017] Beneficial effects:
[0018] 1. The present invention discloses an evaluation method for a standard surface feature parameter evaluation algorithm, which realizes the unification of software algorithm evaluation objects by constructing a standard data set.
[0019] 2. The present invention discloses an evaluation method for a standard surface feature parameter evaluation algorithm, which realizes the evaluation of the independent influence of the software algorithm on the geometric parameter evaluation result by fitting the standard data set and stripping the measurement error of the equipment.
[0020] 3. The present invention discloses an evaluation method for a standard surface characteristic parameter evaluation algorithm, which improves the accuracy of the measurement system by correcting the output results of the evaluated algorithm. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] Figure 1 It is a flow chart of an evaluation method of a standard surface characteristic parameter evaluation algorithm;
[0022] Figure 2 An ideal cylindrical surface data set in an evaluation method embodiment of a standard surface feature parameter evaluation algorithm;
[0023] Among them: Figure (a) is a complete cylinder, Figure (b) is a half cylinder;
[0024] Figure 3 It is a data set after adding different Gaussian noises to the complete cylinder and 1 / 2 cylinder in an evaluation method embodiment of a standard surface feature parameter evaluation algorithm;
[0025] Where: Figure (a) shows the noise Δ 1 Cylinder dataset (μ=0,σ=0.01), Figure (b) shows the noise Δ 1 Semi-cylindrical dataset (μ=0,σ=0.01), Figure (c) shows the noise Δ 2 Cylinder dataset (μ=0,σ=0.05) and Figure (d) shows the noise Δ 2 Semi-cylindrical data set (μ=0,σ=0.05);
[0026] Figure 4 The cylindrical surface is obtained by fitting a standard data set in an evaluation method embodiment of a standard curved surface feature parameter evaluation algorithm. DETAILED DESCRIPTION
[0027] The present invention will be described in detail below with reference to the accompanying drawings and embodiments. The technical problems solved by the technical solution of the present invention and the beneficial effects are also described. It should be noted that the described embodiments are only intended to facilitate the understanding of the present invention and do not have any limiting effect on it.
[0028] like Figure 1As shown, this example evaluates the algorithm for evaluating the diameter and cylindricity of a cylinder in Gom and the algorithm for evaluating the diameter and standard deviation of a cylinder in Geomagic, wherein the standard algorithm is compared by using the ZEISS CALYPS software certified by PTB, and discloses an evaluation method for the standard surface feature parameter evaluation algorithm. The specific implementation steps are as follows:
[0029] Step 1: For geometric surfaces with standard analytical equations such as planes, circles, spheres, cylinders, and cones, select sampling intervals that fully describe the shape characteristics while ensuring sampling efficiency based on the characterization parameters of standard surfaces, perform point set sampling on the standard equations, construct original standard input data that has a certain correlation with the standard data output, and obtain several discrete point sets that satisfy the standard equations. For example, the standard cylindrical surface equation is:
[0030]
[0031] According to the standard cylindrical surface equation, different structures are constructed Cylindrical dataset under value constraints, such as Figure 2 Full cylinder and half cylinder data sets shown.
[0032] Step 2: Add random perturbations that conform to the normal distribution to the point set obtained in step 1 to obtain standard input data. Add different random perturbations to the data set obtained in step 1 to obtain cylindrical data sets with different error sequences. Figure 3 The data sets obtained by adding Gaussian noise with mean 0 and standard deviation 0.01 and 0.05 to the full cylinder and half cylinder are given.
[0033] Step 3: Considering the distribution position and posture of the point set in space, the point set is rotated and translated to obtain a series of standard input data sets;
[0034] Step 4: The cylindrical surface can be described by the radius of the cylindrical section and the direction vector of the cylindrical axis. The cylinder equation is
[0035] x 2 +y 2 =R 2 (a≤z≤b)
[0036] Its parameterized characteristic model can be expressed as
[0037] cylinder(l,m,n,r)
[0038] Among them, the vector Expressed as the unit direction vector of the cylinder axis, satisfying l 2 +m 2 +n 2 =1, the radius of the cylindrical section is r.
[0039] For the series of standard input data sets obtained in step 3, according to the different parameterized characteristics of the simulated elements, based on the least squares principle, the distance di from each discrete point Pi (xi, yi, zi) to the axis of the fitting cylinder is the smallest. Assume that the center of gravity of each discrete point is but
[0040]
[0041] E=min(N[(mδ x -lδ y ) 2 +(nδ x -lδ z ) 2 +(mδ x -ln) 2 ]+g(l,m,n))
[0042] in
[0043]
[0044] g(l,m,n)=B 11 +B 22 +B 33 -[l,m,n]B[l,m,n] T
[0045] Where (x 0 ,y 0 , z 0 ) is a point on the cylinder axis (l, m, n); δ x , δ y , δ z is the distance from this point to the centroid of all measured points.
[0046] B is a 3X3 symmetric matrix, where
[0047]
[0048]
[0049]
[0050]
[0051] Therefore, for any l, m, n, we have
[0052]
[0053] Therefore, the cylinder axis must pass through the centroid of each discrete point. When [l, m, n]T is the maximum eigenvalue λ of matrix B max (B) When the corresponding unit eigenvector is equal to the unit direction vector of the cylinder axis, E can achieve the minimum value.
[0054] According to the axis unit direction vector obtained above, the original coordinate system is transformed into the coordinate system. Let the discrete point P i (x i ,y i , z i ) is located in the coordinate system O-XYZ, for the discrete point Pi(x i ,y i , z i ) to perform coordinate transformation, assuming that the transformed coordinate system is OX′Y′Z′, so that the z′ axis coincides with the vector [1, m, n] after transformation, then the transformation matrix φ is:
[0055]
[0056] The transformed cylinder is projected as a circle on the XOY plane, so we can only consider the x-coordinate and y-coordinate. According to the principle of least squares method, we can fit the projection circle to get the radius and center coordinates of the circle. Let the equation of the least squares fitting circle be (x 0 ) 2 +(yy 0 ) 2 =R 2 , can be obtained
[0057]
[0058]
[0059]
[0060] in
[0061]
[0062]
[0063]
[0064]
[0065]
[0066]
[0067] Perform inverse coordinate transformation of the center point coordinates obtained above -1A point on the cylinder axis can be obtained. At this time, according to a point on the cylinder axis and the unit direction vector of the cylinder axis, the equation expression of the cylinder axis and the geometric parameters of the cylinder such as the cylinder radius can be obtained. The fitting effect diagram is shown as follows: Figure 4 shown.
[0068] Step 5: Input the cylindrical surface data set shown in the figure into the software to be evaluated. It can be obtained that for the standard cylindrical surface with an added mean of 0 and a standard deviation of 0.01, the cylindrical diameter calculated by Gom is 20.00032mm, and the cylindricity is 0.062442mm; the cylindrical diameter calculated by Geomagic is 20.015118mm, and the standard deviation is 0.01037mm; the cylindrical diameter calculated by ZEISSCALYPSO is 20.000320mm, the cylindricity is 0.062442mm, and the standard deviation is 0.010218mm.
[0069] Step 6: For the general equation of the cylindrical surface obtained in step 4, determine the discrete point Pi (x i ,y i , z i ) to the axis and the radius, thus obtaining the discrete point Pi(x i ,y i , z i ) and the position relationship of the cylindrical surface, the discrete point Pi(x i ,y i , z i ) is divided into two categories: inside and outside the cylindrical surface; the difference D_value between the actual displacement Z and the ideal displacement Zs of the two types of discrete points to the fitted cylindrical surface is calculated; the cylindricality of the cylindrical surface is the difference between the maximum and minimum displacements. Inputting the cylindrical surface data set into the above standard cylindrical surface evaluation algorithm, it can be calculated that the cylinder diameter is 20.00033mm and the cylindricality is 0.070703mm.
[0070] cylindricity=max(D_value)-min(D_value)
[0071] Evaluate the shape error and standard deviation of the standard surface, and complete the evaluation of the standard surface geometric parameter evaluation algorithm. By calculating the standard deviation of D_value, the standard deviation of the cylinder can be obtained as 0.010212mm. The calculation formula is as follows
[0072]
[0073] Where μ is D value The mean of .
[0074]
[0075] Step 7: Compare the outputs of step 5 and step 6 to obtain an evaluation result. According to the evaluation result, perform data correction on the Gom algorithm for the diameter and cylindricity of the cylinder and the Geomagic algorithm for the diameter and standard deviation of the cylinder in this embodiment.
[0076] By inputting a series of standard input data sets into the comparison software algorithm, it can be obtained that the difference in cylinder diameter between the standard algorithm provided by this method and the certified high-precision algorithm is 0.000010mm, the difference between the standard deviations is 0.010101mm, and the cylindricity deviation is 0.00011mm, which proves the high accuracy and assessment feasibility of the standard algorithm provided by this method.
[0077] The difference in cylinder diameter between the standard algorithm and the software algorithm to be evaluated is 0.014788mm and -0.000010mm, the difference between the standard deviation is 0.000158mm, and the cylindricality deviation is 0.008261mm. Based on this deviation, the algorithm to be evaluated can be corrected to improve the measurement accuracy of the measurement system where the algorithm to be evaluated is located.
[0078] The specific description above further illustrates the purpose, technical solutions and beneficial effects of the invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. A method for evaluating a standard surface characteristic parameter evaluation algorithm, characterized in that: The following steps are included: Step 1: Perform point set sampling on the standard analytical equations of several typical standard surfaces such as plane, circular surface, spherical surface, cylindrical surface and conical surface to obtain several discrete point sets that satisfy the standard equations; Step 2: Add random perturbations that conform to the normal distribution to the point set obtained in step 1 to obtain standard input data; Step 3: Considering the influence of the distribution position and posture of the geometric body in space on the obtained point set, the standard input data in step 2 is rotated and translated to obtain a series of standard input data sets, which serve as the unified evaluation object for software algorithm evaluation; Step 4: Construct standard evaluation method; In step 4, the standard evaluation method for constructing a standard cylindrical surface through a fitting algorithm is as follows: The standard cylindrical surface equation is: The cylindrical surface is described by the radius of the cylindrical section and the direction vector of the cylindrical axis. The cylinder equation is x 2 +y 2 =R 2 (a≤z≤b) Its parameterized characteristic model is expressed as cylinder(l,m,n,r) Among them, the vector Expressed as the unit direction vector of the cylinder axis, satisfying l 2 +m 2 +n 2 =1, the radius of the cylindrical section is r; For the series of standard input data sets obtained in step 3, according to the different parameterized characteristics of the simulation elements, the discrete points P are calculated based on the least squares principle. i (x i ,y i ,z i ) to the distance d from the axis of the fitting cylinder i Minimum, assuming that the centroid of each discrete point is but E=min(N[(mδ x -lδ y ) 2 +(nδ x -lδ z ) 2 +(mδ x -ln) 2 ]+g(l,m,n)) in g(l,m,n)=B 11 +B 22 +B 33 -[l,m,n]B[l,m,n] T Where (x0, y0, z0) is a point on the cylinder axis (l, m, n), δx, δy, δz are the distances from this point to the centroid of all measuring points; B is a 3X3 symmetric matrix, where Therefore, for any l, m, n, we have Therefore, the cylinder axis must pass through the centroid of each discrete point. When [l,m,n] T is the maximum eigenvalue λ of matrix B max (B) When the corresponding unit eigenvector is equal to the unit direction vector of the cylinder axis, E reaches its minimum value. According to the axis unit direction vector obtained above, the original coordinate system is transformed into the coordinate system. Let the discrete point P i (x i ,y i ,z i ) is located in the coordinate system O-XYZ, for discrete point P i (x i ,y i ,z i ) to perform coordinate transformation, assuming that the transformed coordinate system is O-X'Y'Z', so that the z' axis coincides with the vector [l,m,n] after transformation, then the transformation matrix φ is: The transformed cylinder is projected as a circle on the XOY plane, so only the x coordinate and y coordinate can be considered. The radius and center coordinates of the circle can be obtained by fitting the projection circle according to the principle of least squares method. The equation of the least squares fitting circle is (x-x0) 2 +(y-y0) 2 =R 2 ,have to in Perform inverse coordinate transformation of the center point coordinates obtained above -1 Get a point on the cylinder axis. At this time, according to the point on the cylinder axis and the unit direction vector of the cylinder axis, we can get the equation of the cylinder axis and the geometric parameters of the cylinder such as the cylinder radius. Step 5: Input the series of standard input data sets in step 3 into the software algorithm to be evaluated, and output the surface parameters to be evaluated; Step 6: Using the standard evaluation method obtained in step 4, evaluate the series of standard input data sets in step 3 to obtain the shape error and standard deviation of the geometric surface; Step 7: Compare the outputs of step 5 and step 6 to obtain the evaluation results, and modify the results of the evaluated algorithm based on the evaluation results.
2. The method for evaluating a standard surface characteristic parameter evaluation algorithm according to claim 1, characterized in that: The implementation method of step 4 is For the series of standard input data sets obtained in step 3, the corresponding geometric model is constructed through the fitting algorithm according to the different characterization parameters of the simulation elements, so that the sum of the squares of the offset errors is minimized.
3. The method for evaluating a standard surface characteristic parameter evaluation algorithm according to claim 1, characterized in that: In step 6, the implementation method for the cylindrical surface is For the general equation of the cylindrical surface obtained in step 4, by judging the discrete point P i (x i ,y i ,z i ) to the axis and the radius, thus obtaining the discrete point P i (x i ,y i ,z i ) and the position relationship of the cylindrical surface, the discrete point P i (x i ,y i ,z i ) is divided into two categories: inside and outside the cylindrical surface; the difference D_value between the actual displacement Z and the ideal displacement Zs of the two types of discrete points to the fitted cylindrical surface is calculated; the cylindricity of the cylindrical surface is the difference between the maximum and minimum displacements; cylindricity=max(D_value)-min(D_value) Evaluate the shape error and standard deviation of the standard surface, complete the evaluation of the standard surface geometric parameter evaluation algorithm, and calculate the standard deviation of the cylinder by calculating the standard deviation of D_value. The calculation formula is as follows Where μ is D value The mean of
Citation Information
Patent Citations
Method for automatically evaluating errors of three-dimensional geometrical shapes
CN104050372A
Complex curved surface shape error evaluating method
CN106840022A
Shape evaluation method and shape evaluation apparatus
US20170370714A1
