Point cloud data-oriented flatness error rapid evaluation method, system and device and medium
By decomposing, evaluating and optimizing point cloud data, the problem that traditional measurement equipment is difficult to meet high precision and high efficiency at high sampling speeds is solved, and fast and high-precision flatness error evaluation is achieved.
Patent Information
- Application Number
- CN202410287941.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-03-13
- Publication Date
- 2025-09-16
AI Technical Summary
Traditional measurement equipment finds it difficult to simultaneously meet the requirements of high assessment accuracy and high measurement speed at high sampling speeds, especially on fast full-inspection production lines, and existing data assessment algorithms are inefficient when faced with large amounts of measurement data.
A fast flatness error assessment method for point cloud data is adopted. Through the optimization iteration of the minimum area reference plane, the reference point set update strategy and optimization algorithms such as the SQP algorithm are used to calculate the maximum distance between the point cloud point and the reference plane. The large-scale point cloud data is decomposed into multiple small-scale assessments to optimize the calculation process.
It achieves high-precision flatness error evaluation at high sampling speed, significantly shortens calculation time and improves data processing efficiency.
Smart Images
Figure SMS_21 
Figure SMS_24 
Figure SMS_31
Abstract
Description
Technical Field
[0001] The present invention relates to the field of computer-aided tolerance design, in particular to an evaluation algorithm and system for flatness error. Background Art
[0002] Many new requirements for inspection systems have emerged, such as automated measurement and intelligent assessment. Especially for high-precision parts or high-priced products, in many cases, it is unavoidable to inspect all or even a single product during the production process to control assembly yields. Consequently, the industry's requirements for the sampling speed of measurement equipment and the quality of sampled data processing have increased. Furthermore, the computational speed requirements for assessment methods based on product measurement data have also increased accordingly.
[0003] Traditional measuring equipment, such as contact-type coordinate measuring machines, can still be used to measure some high-precision parts. However, due to limitations in their measurement methods, they struggle to simultaneously meet the requirements of high assessment accuracy and high measurement speed. Therefore, when faced with rapid, full-scale inspection production lines, traditional measuring equipment struggles to meet these requirements. With the advancement of technology, many new measuring equipment and methods, such as blue-light scanning measurement and industrial CT, are now widely used. Advanced instruments can quickly and comprehensively acquire part dimensional information, but this inevitably results in a massive amount of measurement data that needs to be processed and evaluated.
[0004] In the era of slower sampling speeds, the speed of measurement data evaluation algorithms wasn't a major concern. This was because sampling speeds were often slower than the evaluation algorithm's computational speed, and the measurement equipment, rather than data processing, was crucial for the overall measurement process. However, with production lines requiring high-sampling (scanning) speeds for measurement equipment, the efficiency of the corresponding evaluation algorithm becomes even more crucial. Data evaluation algorithms must not only meet computational accuracy requirements but also efficiently process large amounts of measurement data. Furthermore, the evaluation algorithm must be robust enough to accommodate diverse measurement equipment, data sampling methods, and data acquisition paths. Summary of the Invention
[0005] The purpose of the present invention is to provide a method and system for quickly evaluating flatness errors of point cloud data with high precision, high robustness and high processing speed.
[0006] A rapid flatness error assessment method for point cloud data is based on the known conditions of the point cloud coordinates of the object's measured point cloud, the expression of the plane equation of the minimum area reference plane of the point cloud, and the method for calculating the distance from any point in the measured point cloud of the workpiece to the minimum area reference plane of the point cloud.
[0007] The reference point set used to calculate the parameters of the minimum area reference plane is obtained by sampling the measured point cloud. The number of points in the reference point set, m ≥ 5, is much smaller than the number of points in the measured point cloud. The minimum area reference plane is optimized iteratively using the reference point set. After each optimization to obtain the parameters of the current minimum area reference plane, the distance from each point in the measured point cloud to the current minimum area reference plane is traversed and calculated. The purpose of the optimization iteration is to find a target minimum area reference plane that satisfies the maximum distance value from the points in the measured point cloud to the reference plane. The update strategy of the reference point set is: in each iteration, the point in the measured point cloud farthest from the reference plane is added to the reference point set, and the points in the reference point set whose distance to the reference plane is less than a distance constant are eliminated. The twice the absolute value of the minimized maximum distance is used as the flatness error value.
[0008] The optimization of the minimum area reference plane can be performed using existing algorithms, such as particle swarm optimization and genetic algorithm. Preferably, the optimization iteration of the minimum area reference plane adopts SQP (Sequential Quadratic Programming) algorithm.
[0009] The national standard defines an assessment base surface, which is the minimum area reference plane of a flat surface obtained by a prescribed method. It serves as the basis for evaluating flatness deviation and flatness parameters. The minimum area assessment base surface refers to two parallel planes that encompass the flat surface and have the minimum distance between them. The present invention transforms the problem of solving the flatness error value into the problem of solving an optimal minimum area reference plane. The maximum distance from the points in the measured point cloud to this optimal minimum area reference plane is used to find a pair of planes parallel to the optimal minimum area reference plane. All points in the measured point cloud are within this pair of parallel planes. Therefore, twice the absolute value of the maximum distance can be used as the flatness error value of the measured point cloud.
[0010] This method divides the massive measured point cloud into several optimization and update steps for the reference point set, significantly reducing the computational effort required for each optimization. By incorporating the distance between each point in the measured point cloud and the minimum area reference plane, the method performs a traversal calculation for each point, using this distance as the basis for updating the reference point set, thus achieving iterative optimization of the minimum area reference plane. Even with massive point cloud data, the iterative calculation remains highly efficient.
[0011] Furthermore, the optimal number of reference point sets is 5. The 5-point set has the highest computational efficiency and the calculation accuracy of the flatness error value is good.
[0012] Furthermore, the expression of the minimum area reference plane is: p n x+q n y+z+A n =0, p,q,A are optimization variables, p,q are coefficients of the reference plane, A is a constant, p n ,q n ,An represents the optimization variable result obtained at the nth iteration. (x, y, z) are the coordinates of the point in the measurement point cloud.
[0013] Furthermore, the goal of the optimization iteration is expressed as: minε=2D max (ε is the optimization target), D max is the optimization variable, and D obtained from the last optimization iteration is max *2 is the flatness error value; the constraint condition of the optimization iteration is expressed as: |d i |≤D max ;d i Indicates the distance from point i of the measured point cloud to the reference plane. A current D is calculated in each iteration of the optimization problem of the minimum area reference plane. max value Current p(p n ), current q(q n ), and the current A(A n ).
[0014] Optimize the results based on the structure of the constrained optimization problem Numerically, it can represent the maximum value of the distance between the point in the reference point set and the minimum area reference plane of the reference point set in each iteration. Therefore, it can be considered that after being given the final optimization result, D max , which numerically represents the maximum distance from a point in the target point cloud to the minimum area reference plane.
[0015] Furthermore, the distance expression from the point of the measured point cloud to the reference plane is: p,q are the coefficients of the reference plane, A is a constant, x i ,y i , z i Indicates the coordinates of any point in the measurement point cloud; d i There is a distinction between positive and negative.
[0016] Furthermore, the update strategy of the reference point set is expressed as:
[0017] G i =(x i ,y i , z i ) is a measurement data point;
[0018] S n Represents the reference point set of the current iteration, K represents the distance from the reference plane less than the current The initial value of n is 0; The value is higher than that of the previous iteration big;
[0019] M={G i |i=arg max (|d i (p n ,q n , A n )|), G i ∈S}; M represents the point with the largest distance from the reference plane in the current iteration;
[0020] At the reference point set S n Delete K, add M, and update the reference point set to S n+1 .
[0021] Furthermore, the end condition of iterative optimization is expressed as: That is, after each reference point set is updated, it is determined whether the distance between all points in the measured point cloud S and the reference plane is no greater than If so, the iterative calculation ends and the current As the flatness error value; if it is not satisfied, the iteration is entered again with the updated reference point set to solve the optimization parameters of the minimum area reference plane.
[0022] Since the flatness error is small, D max The initial optimization value can be set to 0. The three parameters p, q, and A are related to the pose of the minimum area reference plane, so they can be determined by the state of the measured surface when it is measured. For example, when evaluating the point cloud data of a horizontally placed measured plane near the XOY plane of the global coordinate system, the initial values of p, q, and A can be set to [0, 0, 0], that is, the minimum area reference plane is set to z = 0. The optimization of the reference point set in each iteration will be [D max , p, q, A] = [0, 0, 0, 0] is used as the initial value for calculation.
[0023] This method is to evaluate the flatness error of the entire measurement point data by searching for the minimum area reference plane of the reference point set multiple times. In each iteration, the process of searching for the minimum area reference plane of the reference point set is achieved by solving a constrained optimization problem. The purpose of solving the optimization problem is to calculate the relevant parameters of the minimum area reference plane of the reference point set in this iteration, that is, p n ,q n ,A n , to determine this plane.
[0024] A second aspect of the present invention provides a system for rapid flatness error assessment of point cloud data, the system comprising the following modules:
[0025] A preset condition module is used to input and store the plane equation of the minimum area reference plane of the point cloud and the distance calculation formula from any point of the measured point cloud of the workpiece to the point cloud reference plane; the coordinates of the point cloud to be evaluated are also stored in the preset condition module;
[0026] The reference point set module is used to obtain and store the reference point set obtained by sampling the measured point cloud. The number of points in the reference point set is m ≥ 5 and is much smaller than the number of points in the measured point cloud.
[0027] The optimization iteration module is used to find a target minimum area reference plane that satisfies the maximum distance between the points in the measured point cloud and the reference plane. The optimization iteration of the minimum area reference plane is performed using the reference point set. After each optimization, the distance from each point in the measured point cloud to the current minimum area reference plane is calculated.
[0028] The updating strategy of the reference point set during the optimization process is: in each iteration, the point in the measured point cloud that is farthest from the reference plane is added to the reference point set, and the points in the reference point set whose distance to the reference plane is less than a distance constant are eliminated;
[0029] The flatness calculation module takes twice the absolute value of the minimum maximum distance obtained by the optimization iteration module as the flatness error value.
[0030] The reference point set expression of the optimization iteration module, the update strategy of the reference point set, the expression of the end condition of the optimization iteration, the boundary condition of the optimization iteration, the end condition of the optimization iteration, the distance expression from the point of the measurement point cloud to the reference plane, and other related calculations are the same as the above-mentioned fast assessment method of flatness error.
[0031] A third aspect of the present invention provides a device for quickly assessing flatness errors of point cloud data, the device comprising: at least one processor, at least one memory for storing at least one computer program;
[0032] The processor executes the computer program in the memory to implement the above-mentioned flatness error rapid evaluation method steps.
[0033] According to a fourth aspect of the present invention, a computer storage medium is provided, on which a computer program is stored. The computer program is executed by a computer to implement the above-mentioned method steps for rapid flatness error assessment.
[0034] The advantages of the present invention are that it can quickly process huge amounts of point cloud data and has good assessment accuracy. DETAILED DESCRIPTION
[0035] The measured point clouds and point cloud data described in the present invention include, but are not limited to, point clouds obtained through visual measurement methods (such as structured light, binocular cameras, depth cameras, etc.), point clouds obtained through CT and other radiographic methods, computer-generated point clouds, point clouds obtained through lidar scanning, and point clouds obtained through sonar detection. Furthermore, this method does not impose any requirements on the strategy (sampling path) for sampling the point clouds.
[0036] A rapid flatness error assessment method for point cloud data is based on the known conditions of the point cloud coordinates of the object's measured point cloud, the expression of the plane equation of the minimum area reference plane of the point cloud, and the method for calculating the distance from any point in the measured point cloud of the workpiece to the minimum area reference plane of the point cloud.
[0037] The reference point set used to calculate the parameters of the minimum area reference plane is obtained by sampling the measured point cloud. The number of points in the reference point set, m ≥ 5, is much smaller than the number of points in the measured point cloud. The minimum area reference plane is optimized iteratively using the reference point set. After each optimization to obtain the parameters of the current minimum area reference plane, the distance from each point in the measured point cloud to the current minimum area reference plane is traversed and calculated. The purpose of the optimization iteration is to find a target minimum area reference plane that satisfies the maximum distance value from the points in the measured point cloud to the reference plane. The update strategy of the reference point set is: in each iteration, the point in the measured point cloud farthest from the reference plane is added to the reference point set, and the points in the reference point set whose distance to the reference plane is less than a distance constant are eliminated. The twice the absolute value of the minimized maximum distance is used as the flatness error value.
[0038] In some embodiments, the optimization of the minimum area reference plane can be performed using existing algorithms, such as particle swarm optimization and genetic algorithm. In some embodiments, the optimization iteration of the minimum area reference plane can be performed using SQP (Sequential Quadratic Programming) algorithm.
[0039] In some embodiments, the optimal number of reference point sets is 5. A 5-point set has the highest computational efficiency and good calculation accuracy of the flatness error value.
[0040] In some embodiments, the expression for the minimum area reference plane is: n x+q n y+z+A n =0, p,q,A are optimization variables, p,q are coefficients of the reference plane, A is a constant, p n ,q n ,A n Indicates the optimization variable result obtained at the nth iteration.
[0041] In some embodiments, the goal of the optimization iteration is expressed as: minε=2D max (ε is the optimization target), Dmax is the optimization variable, and D obtained from the last optimization iteration is max *2 is the flatness error value; the constraint condition of the optimization iteration is expressed as: |d i |≤D max ;d i Indicates the distance from the i-th point of the measurement point cloud to the reference plane.
[0042] In some embodiments, the distance from a point in the measurement point cloud to the reference plane is expressed as: p,q are the coefficients of the reference plane, A is a constant, x i ,y i , z i Indicates the coordinates of any point in the measurement point cloud; d i There is a distinction between positive and negative.
[0043] In some embodiments, the update strategy of the reference point set is expressed as:
[0044] G i =(x i ,y i ,z i ) is a measurement data point;
[0045] S n Represents the reference point set of the current iteration, K represents the distance from the reference plane less than the current The initial value of n is 0; The value is higher than that of the previous iteration big;
[0046] M={G i |i=arg max(|d i (p n ,q n , A n )|),G i ∈S}; M represents the point with the largest distance from the reference plane in the current iteration;
[0047] At the reference point set S n Delete K, add M, and update the reference point set to S n+1 .
[0048] In some embodiments, the termination condition of the iterative optimization is expressed as: That is, after each reference point set is updated, it is determined whether the distance between all points in the measured point cloud S and the reference plane is no greater than If so, the iterative calculation ends and the current As the flatness error value; if it is not satisfied, the iteration is entered again with the updated reference point set to solve the optimization parameters of the minimum area reference plane.
[0049] Example 1
[0050] The fast evaluation method of flatness error for point cloud data is based on the following pre-conditions: the plane equation of the reference plane of the minimum area of the point cloud is: px+qy+rz+A=0;
[0051] Where p, q, r, and A are known constants, and p, q, and r are not zero at the same time.
[0052] Let r = 1, that is, the reference plane intersects the z-axis.
[0053] The plane equation becomes
[0054] px+qy+z+A=0
[0055] The plane normal vector is
[0056] A point on the plane is W = (0, 0, -A)
[0057] Measure any point G in the point cloud S i =(x i ,y i , z i ), (i=1,2,3,4..., is the counting subscript) the distance d to the reference plane i for:
[0058]
[0059] (d i There are positive and negative distinctions)
[0060] To evaluate point cloud data, perform the following operations:
[0061] Step 1: Evenly select five points from the measured point cloud S to construct the initial five-point set S n , n=0, (n=1, 2, 3... is the iteration count subscript, the first iteration n=0).
[0062] Step 2: Use the SQP method to perform constrained optimization on this five-point set:
[0063] obj.minε=2D max (ε is the optimization target);
[0064] st|d i |≤D max ;
[0065] D max ,p,q,A are optimization variables, according to the optimization p n ,qn ,A n , we can get the minimum area reference plane p of these five points n x+q n y+z+A n =0;
[0066] Step 3: Calculate the distance d from each point in G to this reference plane i ; Add the point M in S that is the largest distance from the reference plane to the point set S0, and remove S o The distance from this reference plane is less than Point K forms a new five-point set S1:
[0067] M={G i |i=arg max(|d i (p n ,q n , A n )|),G i ∈S};
[0068] Step 4: Verify the end condition and determine whether the distance between all points in the measured point cloud S and the reference plane is less than Right now
[0069] If the end condition is met, the flatness error is calculated
[0070] If the end condition is not met, return to step 2.
[0071] Traditional flatness error assessment algorithms often treat the entire point cloud as the calculation object. During the optimization process, the complexity of the objective function and constraints is directly related to the data volume of the point cloud file. As the amount of point cloud data increases, the computational effort and computational time increase significantly.
[0072] The present invention divides the evaluation of an entire point cloud into several evaluations of point clouds composed of five points. Each optimization process is only performed on five points, which greatly reduces the complexity of a single optimization.
[0073] In this invention, the flatness assessment algorithm is affected by the size of the point cloud data in two aspects: the distance between all points and a single reference plane is calculated at each iteration; and the number of iterations. The following explains them in order:
[0074] 1. Calculate the distance from all points to a single reference plane at each iteration: This distance is di. As the formula shows, this is a common numerical calculation. Computers can process this type of numerical calculation very quickly, so even with massive point cloud data, the calculation of di can still maintain good efficiency.
[0075] 2. Number of iterations: The present invention designs an update strategy for the reference point set after each iteration, which introduces a point that is farthest from the reference plane. The introduction of the farthest point M after each iteration rapidly expands the search range, and the entire point cloud can be included in the error band after several iterations. (It can also be understood that the amount of data in the point cloud file will not have a significant impact on the search for the farthest point). Therefore, even if the number of point clouds increases, the number of iterations can be controlled within a range with little fluctuation. The present invention decomposes the one-time optimization of a large amount of data into multiple optimizations of a small amount of data.
[0076] As shown in Table 1, the point cloud data scale, flatness error and time consumption of Example 1.
[0077] Table 1
[0078] Measure point cloud points <![CDATA[D max ]]> p q A Flatness error Time(s) 100 0.094644 -0.00061 2.46E-05 -0.03457 0.189289 0.073838 400 0.116434 0.00048 0.00030 -0.04924 0.232867 0.084227 1600 0.116261 -0.00037 -0.00097 0.01651 0.232522 0.060219 6400 0.123751 0.00014 -0.00011 -0.04487 0.247501 0.077201 25600 0.146833 -0.00040 -0.00053 -0.02769 0.293665 0.070542 102400 0.160902 0.00109 0.00020 -0.07129 0.321803 0.138195 409600 0.165595 -0.00018 -5.86E-06 -0.03549 0.331191 0.260250 1638400 0.167550 -4.99E-05 -2.22E-05 -0.03776 0.335100 1.007726
[0079] As shown in Table 2, the point cloud scale, flatness error and time consumption of Comparative Example 1. V, M, B, et al. Evaluating minimum zone flatness error using new method-Bundle of plains through one point[J]. Precision Engineering, 2016, 43: 554-562. The method described in the paper.
[0080] Table 2
[0081] Measure point cloud points Flatness error Time(s) 100 0.189437 0.96293 400 0.232963 3.425623 1600 0.232663 12.42968 6400 0.247724 47.74412 25600 0.293709 186.8513 102400 0.32192 774.9981 409600 0.331425 2933.536 1638400 0.335454 11777.94
[0082] It can be seen from the above comparative examples that, compared with the traditional method, the method of the present invention greatly shortens the calculation time and improves the calculation efficiency while ensuring the evaluation accuracy of the flatness error.
[0083] For the seventh set of test point data (1,638,400 points), comparative tests were conducted with the reference point set consisting of 5 points (Comparative Example 2), 10 points (Comparative Example 3), 50 points (Comparative Example 4), 100 points (Comparative Example 5), 500 points (Comparative Example 6), and 1,000 points (Comparative Example 7). Comparative Examples 1-6 are the same as Example 1 except for the number of reference point sets. The comparison results are shown in Table 3:
[0084] Table 3
[0085]
[0086] The comparison results in Table 3 show that when m is in the range of 5 to 1000, the accuracy of the flatness error assessment results is not affected, and the algorithm time consumption is also controlled within an acceptable range. This is much lower than the 11777.94 seconds required by the comparison method (OPPBM) to calculate the same amount of point cloud data (1638400 points).
[0087] In some embodiments, a system for rapid flatness error assessment of point cloud data includes the following modules:
[0088] A preset condition module is used to input and store the plane equation of the minimum area reference plane of the point cloud and the distance calculation formula from any point of the measured point cloud of the workpiece to the point cloud fitting plane; the coordinates of the point cloud to be evaluated are also stored in the preset condition module;
[0089] The reference point set module is used to obtain and store the reference point set obtained by sampling the measured point cloud. The number of points in the reference point set is m ≥ 5 and is much smaller than the number of points in the measured point cloud.
[0090] The optimization iteration module is used to find a target minimum area reference plane that satisfies the maximum distance between the points in the measured point cloud and the reference plane. The optimization iteration of the minimum area reference plane is performed using the reference point set. After each optimization, the distance from each point in the measured point cloud to the current minimum area reference plane is calculated.
[0091] The updating strategy of the reference point set during the optimization process is: in each iteration, the point in the measured point cloud that is farthest from the reference plane is added to the reference point set, and the points in the reference point set whose distance to the reference plane is less than a distance constant are eliminated;
[0092] The flatness calculation module takes twice the absolute value of the minimum maximum distance obtained by the optimization iteration module as the flatness error value.
[0093] The reference point set expression of the optimization iteration module, the update strategy of the reference point set, the expression of the end condition of the optimization iteration, the boundary condition of the optimization iteration, the end condition of the optimization iteration, the distance expression from the point of the measurement point cloud to the reference plane, and other related calculations are the same as the above-mentioned fast assessment method of flatness error.
[0094] In some embodiments, a device for rapid flatness error assessment of point cloud data includes: at least one processor, at least one memory for storing at least one computer program;
[0095] The processor executes the computer program in the memory to implement the above-mentioned flatness error rapid evaluation method steps.
[0096] In some embodiments, a computer storage medium stores a computer program, and the computer program is executed by a computer to implement the above-mentioned method steps for rapid flatness error assessment.
[0097] The above specific implementation cases are used to illustrate the present invention and are only preferred embodiments of the present invention, rather than limiting the present invention. Any modifications, equivalent substitutions, improvements, etc. made to the present invention within the spirit of the present invention and the scope of protection of the claims shall fall within the scope of protection of the present invention.
Claims
1. A method for rapid flatness error assessment based on point cloud data, characterized by: The point cloud coordinates of the object's measured point cloud are known conditions, the expression form of the plane equation of the minimum area reference plane of the point cloud is known conditions, and the method for calculating the distance from any point of the workpiece's measured point cloud to the minimum area reference plane of the point cloud is also known conditions; The reference point set used for calculating the minimum area reference plane parameters is obtained by sampling the measured point cloud. The number of points in the reference point set m ≥ 5 and is much smaller than the number of points in the measured point cloud. Perform optimization iterations of the minimum area reference plane using a reference point set; After each optimization to obtain the parameters of the current minimum area reference plane, the distance from each point in the measured point cloud to the current minimum area reference plane is traversed and calculated; the purpose of the optimization iteration is to find a target minimum area reference plane that satisfies the maximum distance value from the points in the measured point cloud to the reference plane to be minimized; the update strategy of the reference point set is: in each iteration, the point in the measured point cloud farthest from the reference plane is added to the reference point set, and the points in the reference point set whose distance to the reference plane is less than the distance constant are eliminated; twice the absolute value of the minimized maximum distance is used as the flatness error value.
2. The method for rapid flatness error assessment of point cloud data according to claim 1, wherein: The optimization iteration of the minimum area reference plane adopts the SQP (sequential quadratic programming) algorithm.
3. The method for rapid flatness error assessment based on point cloud data according to claim 1, wherein: The optimal number of reference point sets is 5 points.
4. The method for rapid flatness error assessment of point cloud data according to any one of claims 1 to 3, characterized in that: The expression of the minimum area reference plane is: p n x+q n y+z+A n =0, p,q,A are optimization variables, p,q are coefficients of the reference plane, A is a constant, p n ,q n ,A n represents the optimization variable result obtained at the nth iteration, and (x, y, z) is the coordinate of the point in the measurement point cloud.
5. The method for rapid flatness error assessment of point cloud data according to any one of claims 1 to 3, characterized in that: The goal of the optimization iteration is expressed as: minε=2D max (ε is the optimization target), D max is the optimization variable, and D obtained from the last optimization iteration is max *2 is the flatness error value; the constraint condition of the optimization iteration is expressed as: |d i |≤D max ;d i Indicates the distance from the i-th point of the measured point cloud to the reference plane.
6. The method for rapid flatness error assessment of point cloud data according to any one of claims 1 to 3, characterized in that: The distance expression from a point in the measured point cloud to the reference plane is: p,q are the coefficients of the reference plane, A is a constant, x i ,y i , z i Indicates the coordinates of any point in the measurement point cloud; d i There is a distinction between positive and negative.
7. The method for rapid flatness error assessment of point cloud data according to any one of claims 1 to 3, characterized in that: The update strategy of the reference point set is expressed as: G i =(x i ,y i , z i ) is a measurement data point; S n Represents the reference point set of the current iteration, K represents the distance from the reference plane less than the current The initial value of n is 0; The value is higher than that of the previous iteration big; M={G i |i=argmax(|d i (p n ,q n , A n )|), G i ∈S}; M represents the point with the largest distance from the reference plane in the current iteration; At the reference point set S n Delete K, add M, and update the reference point set to S n+1 .
8. The method for rapid flatness error assessment of point cloud data according to any one of claims 1 to 3, characterized in that: The end condition of iterative optimization is expressed as: That is, after each reference point set is updated, it is determined whether the distance between all points in the measured point cloud S and the reference plane is no greater than If so, the iterative calculation ends and the current As the flatness error value; if it is not satisfied, the iteration is entered again with the updated reference point set to solve the optimization parameters of the minimum area reference plane.
9. A system for rapid flatness error assessment of point cloud data, characterized in that: The system includes the following modules: A preset condition module is used to input and store the plane equation of the minimum area reference plane of the point cloud and the distance calculation formula from any point of the measured point cloud of the workpiece to the point cloud reference plane; the coordinates of the point cloud to be evaluated are also stored in the preset condition module; The reference point set module is used to obtain and store the reference point set obtained by sampling the measured point cloud. The number of points in the reference point set is m ≥ 5 and is much smaller than the number of points in the measured point cloud. An optimization iteration module is used to find a target minimum area reference plane, which satisfies the requirement that the maximum distance between the points in the measured point cloud and the reference plane is minimized; Perform optimization iterations of the minimum area reference plane using a reference point set; After each optimization to obtain the parameters of the current minimum area reference plane, the distance from each point of the measured point cloud to the current minimum area reference plane is traversed and calculated; The updating strategy of the reference point set during the optimization process is: in each iteration, the point in the measured point cloud that is farthest from the reference plane is added to the reference point set, and the points in the reference point set whose distance to the reference plane is less than a distance constant are eliminated; The flatness calculation module takes twice the absolute value of the minimized maximum distance obtained by the optimization iteration module as the flatness error value.
10. A device for rapid flatness error assessment based on point cloud data, characterized by: The apparatus comprises: at least one processor, at least one memory for storing at least one computer program; The processor executes the computer program in the memory to implement the steps of the method for rapid flatness error assessment according to any one of claims 1 to 8.
11. A computer storage medium, characterized in that: The computer storage medium stores a computer program, which is executed by a computer to implement the steps of the method for rapid flatness error assessment according to any one of claims 1 to 8.