Self-planning high-efficiency laser measurement method for discrete key points in complex assembly scene

Abstract

A self-planning, high-efficiency laser measurement method for discrete key points in complex assembly scenarios is proposed. First, based on a point-line-surface intersection algorithm, pre-identification of laser measurement line-of-sight occlusion is performed. Then, a distributed laser measurement station optimization layout function model is constructed, with the objective function being the minimum weighted sum of 3D measurement uncertainty and key point measurement uncertainty. Constraints such as ranging range, angle measurement range, and feasible region are set. A genetic algorithm is used to solve for the coordinates of the distributed laser measurement stations and the measurement tasks, achieving self-planning of the number and location of measurement stations. Finally, the measurement sequence of discrete key points over a large area is planned, based on the principle of minimizing the measurement path within the sequence, to solve for the measurement sequence of several key points to be measured within each measurement station. This method can be applied to the assembly process of large mechanical products such as aircraft in the aerospace field, greatly improving the efficiency of key point measurement during assembly and showing broad application prospects.

Description

Technical Field

[0001] This invention belongs to the field of large-size / digital measurement and relates to a self-planning, high-efficiency laser measurement method for discrete key points in complex scenarios. Background Technology

[0002] Digital measurement-driven in-situ assembly has become one of the mainstream manufacturing modes for large aircraft components. Its measurement accuracy, efficiency, and reliability directly affect assembly quality and efficiency. For example, the accurate transfer and evaluation of geometric measurement data in aircraft assembly is crucial for ensuring assembly accuracy and shortening assembly time. However, key points (positioning points, reference points, etc.) on aircraft components are often widely distributed, spatially discrete, and subject to severe line-of-sight obstruction at the measurement site, posing a significant challenge to the high-precision, high-efficiency, and high-reliability measurement of a series of key points. Conventional on-site measurement techniques often rely on manual experience for setting up measurement stations, making it difficult to predict whether line-of-sight obstructions will occur, easily leading to the need for temporary addition of measurement stations. Furthermore, the measurement sequence of several targets also relies entirely on personal experience, easily resulting in inefficient measurement problems such as missed measurements, repeated measurements, and repetitive measurements. In addition, the measurement accuracy of laser trackers is related to distance and angle, and experience-based layout cannot guarantee measurement accuracy. Therefore, it is necessary to explore a self-planning, high-efficiency measurement method for complex assembly sites to ensure the measurement accuracy, efficiency, and reliability of key points.

[0003] The patent "A Laser Tracker Site Planning Method Based on Combined Measurement" (patent number CN202011243780.4) discloses a laser tracker site planning method based on combined measurement. This method uses the planned T-Scan scanning trajectory to plan the placement and attitude of the laser tracker. The patent "Spatial Measurement Accessibility Simulation Analysis and Layout Method for Laser Trackers and IGPS" (patent number CN202210363899.8) discloses a spatial measurement accessibility simulation analysis and layout method for laser trackers and IGPS. This method determines the accessibility of measurement points by simulating the difference between measured and theoretical distances, and finds laser tracker sites where all measurement points are measurable through gridded search.

[0004] The aforementioned methods are only applicable to combined measurements using laser trackers and T-Scan, and can only find laser tracker stations that make all measurement points measurable. They do not consider the overall measurement accuracy of the laser tracker in multi-station measurements, and do not address the issues of discrete key point measurement task allocation and measurement sequence planning. Therefore, this invention proposes a self-planning, high-efficiency laser measurement method for discrete key points in complex scenarios, achieving self-planning and full-coverage measurement across multiple measurement stations. This method has good versatility and broad application prospects. Summary of the Invention

[0005] This invention presents a highly efficient self-planning laser measurement method for discrete key points in complex scenarios. First, based on a point-line-surface intersection algorithm, it pre-determines line-of-sight occlusion in laser measurement. Second, it constructs a distributed laser measurement station optimization layout function model, with the objective function being the minimum weighted sum of 3D measurement uncertainty and key point measurement uncertainty. Constraints such as distance measurement range, angle measurement range, and feasible region are set, and a genetic algorithm is used to solve for the coordinates of the distributed laser measurement stations and the measurement tasks, achieving self-planning of the number and location of measurement stations. Finally, it plans the measurement sequence for discrete key points over a large area, based on the principle of minimizing the measurement path within the sequence, solving for the measurement sequence of several key points to be measured within each measurement station. This invention changes the traditional mode of manual, experience-based station layout, solving for problems of incomplete or inaccessible measurements and low efficiency, and has broad application prospects.

[0006] The technical solution of the present invention:

[0007] A self-planning, high-efficiency laser measurement method for discrete key points in complex assembly scenarios, comprising the following steps:

[0008] The first step is to predict line-of-sight occlusion based on the intersection of points, lines, and surfaces.

[0009] First, models of aircraft components, tooling, and other obstructions that could easily block the line of sight in the assembly area are imported into SOLIDWORKS modeling software. The models are then exported in .obj format, completing the triangular facetization of the model, obtaining one triangular facet at a time. Next, a line segment is constructed between the key point to be measured and the measurement origin of the laser tracker. The sum of the areas of the three triangles formed by the intersection of this line segment and the plane containing the triangular facet of the component, along with the vertices of the triangular facet, is calculated to determine whether the line segment intersects with the triangular facet (i.e., whether the line of sight is blocked). Let the coordinates of the three vertices of a triangular facet be q1 = (x1, y1, z1), q2 = (x2, y2, z2), and q3 = (x3, y3, z3). The coordinates of the key point to be measured and the measurement origin of the laser tracker are p... a =(x a ,y a ,z a ), p b =(x b ,y b ,z b The parametric equation of the line segment is expressed as equation (1):

[0010]

[0011] Where λ is a parameter;

[0012] The plane equation of the triangular facet is expressed by the determinant as equation (2):

[0013]

[0014] Combining equations (1) and (2), we calculate the value of parameter λ using equation (3), where C and D are process parameters, and m1 = x b -x a m2 = y b -y a m3 = z b -z a If λ∈[0,1], then the triangular facet on the surface intersects with the line segment, and the coordinates of the intersection point are obtained by equation (1):

[0015]

[0016] The area of ​​the model triangular facet and the total area of ​​the three triangles formed by the intersection of the line and the three vertices of the model triangular facet are calculated using equation (4).

[0017]

[0018] Among them, S 原 S represents the area of ​​the triangular facet of the model. 新 The total area of ​​the three triangles formed by the intersection of the line and the plane and the vertices of the model's triangular facets is a1 = x2 - x1, a2 = x3 - x1, b1 = y2 - y1, b2 = y3 - y1, c1 = z2 - z1 and c2 = z3 - z1;

[0019] At this point, the area of ​​the original model's triangular facets and the total area of ​​the three triangles formed by the intersection of the line and the vertices of the model's triangular facets can be calculated. If S... 原 =S 新 If the line of sight is obstructed, it indicates that the intersection of the measurement line of sight and the triangular plane is inside the triangle, i.e., they intersect, and the measurement line of sight is blocked; otherwise, the measurement line of sight is not obstructed.

[0020] The second step is to construct an optimization function for the distributed laser tracker measurement stations.

[0021] First, the deployable range of laser tracker stations, the ranging range of the laser tracker, and the angular measurement range of the laser tracker are used as constraints. The objective function is to minimize the weighted sum of the uncertainties of the three-dimensional measurement (x, y, z directions in the measurement coordinate system) of the laser tracker and the uncertainties of the key point measurement (ux, uy, uz). The parameters to be optimized are the number and location of laser tracker stations. A measurement point allocation factor θ(i) is introduced to allocate measurement tasks. In summary, the optimization function of the distributed laser tracker measurement stations is constructed, where the objective function is shown in Equation (5), the measurement point allocation factor is shown in Equation (6) (others are shown), and the constraint function is shown in Equation (7).

[0022]

[0023] Where m is the number of laser tracker measurement stations, n is the number of key points, and P LT,j Let p be the coordinates of the j-th measurement station to be measured. i Let i be the coordinates of the i-th key point to be measured. The minimum and maximum values ​​of the vertical angle (the angle between the line connecting the key point to be measured and the origin of the laser tracker and the Z-axis in the coordinate system of the tracker) are given by l. max ε is the maximum permissible measurement distance. x ,ε y ,ε z These represent the feasible regions where stations can be deployed in the x, y, and z directions, respectively. It is the line-of-sight occlusion penalty factor for the i-th control (it can be 1e8 when occluded, and 0 when not occluded).

[0024] The third step is to solve the coordinates of the measurement station layout based on a genetic algorithm.

[0025] In equation (6), θ(i) is used to allocate the key points to be measured to each measurement station according to the principle of shortest distance, that is: to determine the distance from the key points to be measured to each measurement station, if p i If the distance to station k is closest, then θ(i) = 1; otherwise, it is 0. A genetic algorithm is then used to optimize the calculation of the number and spatial location of measurement stations in the formula, with the fitness function serving as the objective function. Considering that the feasible region in a large aerospace assembly site is not parallel to each coordinate axis, the 3D scene to be measured is transformed into a 3D space parallel to the coordinate axes for solution. The obtained configuration parameters are then inversely transformed back to the original 3D space to obtain the final configuration parameters of the measurement stations.

[0026] The fourth step is the automatic planning of measurement sequences for discrete key points over a large area.

[0027] Suppose there are n key points to be measured within a certain measurement station, and let G = (V p ,d p ) is a weighted graph, V p ={1,2,…,n} is the vertex set, d p Let the set of side lengths be d, and the distance between each vertex be d. i,j (d i,j >0, i,j∈V p The optimized mathematical model is Equation (8).

[0028]

[0029] Among them, D s d is the reciprocal of the total distance for each measurement point sequence. i,j Let a represent the distance between the i and j key points to be tested. i,jLet S be the process parameters, where S is all non-empty subsets of V, and |S| is the total number of vertices in the weighted graph G contained in set V. A keypoint measurement sequence planning model based on an improved genetic algorithm was then established. Heuristic search was used to reduce optimization time, ultimately solving for the measurement sequence.

[0030] The beneficial effects of this invention are that it avoids the problem of incomplete or unmeasurable key geometric features caused by unreasonable layout of measurement stations in traditional empirical methods, greatly improving measurement efficiency and ensuring measurement accuracy. This method not only has good practicality and versatility, but also features good robustness, high accuracy, and high stability. It can be used for digital measurement in the manufacturing field of high-end aerospace equipment, laying a foundation for improving assembly efficiency and demonstrating broad application prospects. Attached Figure Description

[0031] Figure 1 This is a flowchart of a self-planning, high-efficiency laser measurement method for discrete key points in complex scenarios.

[0032] Figure 2 This is a diagram of a measurement line-of-sight occlusion judgment method based on triangle area reconstruction;

[0033] Figure 3 It is based on a genetic algorithm to solve the coordinate map of the measurement station layout; Detailed Implementation

[0034] The specific embodiments of the present invention are described in detail below with reference to the technical solutions and accompanying drawings.

[0035] The laser tracker used in this embodiment is a Leica AT960M model, with a maximum permissible measurement error (MPE) of ±(15μm+6μm / m). The target sphere and target base used as measurement points are 0.5 inches in size. A closed cylinder with a bottom radius of 2100mm and a height of 9000mm was used as the measurement object, and the measurement experiment was conducted within a range of 24m×5m.

[0036] As attached Figure 1 As shown, this method first uses a point-line-plane intersection algorithm to pre-determine whether the laser measurement line of sight is obstructed. Secondly, it constructs a distributed laser measurement station optimization layout function, with the objective function being the minimum weighted sum of 3D uncertainty and point measurement uncertainty, setting constraints such as the deployable range of the measurement system stations, the ranging range of the measurement system, and the angle measurement range. Then, it uses a genetic algorithm to design the configuration parameters such as the number and location of the measurement stations. Finally, it performs measurement sequence planning for large-scale discrete targets, planning a measurement sequence for several key points to be measured within each measurement station, minimizing the measurement path within the sequence, ultimately achieving self-planning and efficient laser measurement of discrete key points in complex scenes. The specific steps of this method are as follows:

[0037] The first step is to predict line-of-sight occlusion based on the intersection of points, lines, and surfaces.

[0038] Fifteen points were randomly selected on the cylinder as key points. A laser tracker was used to measure the coordinates of all key points at several stations, and these coordinates were then standardized to a unified (global) coordinate system. The initial coordinate values ​​of all key points are as follows:

[0039] Table 1 Initial values ​​of key points to be measured

[0040]

[0041]

[0042] Next, the three-dimensional model of the closed cylinder in the laboratory was triangulated to prepare for solving the layout of the measurement stations.

[0043] The second step is to construct a distributed laser measurement station optimization function.

[0044] Based on the actual experimental site size and the number of measurement system stations, a feasible domain space for the arrangement of measurement stations is defined. In this example, the number of measurement stations is 3, and the feasible domain is defined as follows:

[0045] X1∈[-7500,7500] Y1∈[-5000,5000] Z1∈[2100,3000]

[0046] X2∈[-7500,7500] Y2∈[-6000,-3000] Z2∈[-1600,1600]

[0047] X3∈[-7500,7500] Y3∈[2500,5000] Z3∈[-1000,2100]

[0048] Based on the characteristics of the laser tracker model and the requirements for measurement range and accuracy, constraints are set, including the farthest distance from the measurement station to the key point and the vertical angle range of the key point relative to the measurement station position. The configuration parameters such as the number and position of the measurement stations are used as parameters to be optimized, and a distributed laser measurement station optimization layout function is constructed.

[0049] In this example, the nonlinear constraints are set as follows: the distance between the measurement station and the key point is less than 10m, and the measurement vertical angle range is less than 145°.

[0050] The third step is to solve the coordinates of the measurement station layout based on a genetic algorithm.

[0051] Based on equation (5), the coordinates of the three stations are obtained. Then, based on equations (3), (4) and (6), each station is assigned to measure key points. The planning results are as follows:

[0052] Table 2. Station Coordinates and Key Responsibilities

[0053]

[0054] Coordinates of the key points to be measured assigned to station 1:

[0055] Table 3 Coordinates of key points measured at station 1

[0056]

[0057]

[0058] Coordinates of the key points to be measured assigned to station 2:

[0059] Table 4. Coordinates of key points measured at station 2

[0060]

[0061] Coordinates of the key points to be measured assigned to station 3:

[0062] Table 5. Coordinates of key measurement points at station 3.

[0063]

[0064] The fourth step is the automatic planning of measurement sequences for discrete key points over a large area.

[0065] Taking station 1 as an example, the objective function is to minimize the measurement path of the key points to be measured. The population size is set to 1000, the maximum number of iterations to 500, the crossover probability to 0.8, the mutation probability to 0.2, the exchange structure probability to 0.2, the reverse structure probability to 0.5, and the insertion structure probability to 0.3. Based on the improved genetic algorithm, the sequence of points to be measured is obtained through iterative optimization. The total distances before and after optimization are 16906.66 mm and 7854.023 mm, respectively.

[0066] This invention uses a point-line-surface algorithm to determine occlusion of the measurement line of sight. The objective function is to minimize the weighted sum of three-dimensional uncertainty and point measurement uncertainty. Constraints such as the deployable range of the measurement system stations, the ranging range of the measurement system, and the angle measurement range are set to construct an optimized layout function for distributed laser measurement stations. Then, a genetic algorithm is used to solve for configuration parameters such as the number and location of measurement stations. Finally, measurement sequence planning is performed for large-scale discrete targets, planning a measurement sequence for several key points to be measured within each measurement station, minimizing the measurement path within the sequence. This avoids the problem of incomplete or unmeasured key geometric features caused by unreasonable station placement in empirical methods, greatly improving measurement efficiency. In summary, this method not only has good practicality and versatility but also features good robustness, high accuracy, and high stability. This method can be used in the assembly process of large mechanical products such as aircraft in the aerospace field, greatly improving the efficiency of key point measurement during assembly, and has broad application prospects.

Claims

1. A self-planning, high-efficiency laser measurement method for discrete key points in complex assembly scenarios, characterized in that, The steps are as follows: The first step is to predict line-of-sight occlusion based on the intersection of points, lines, and surfaces. First, import models of obstructions that could easily block the line of sight into SOLIDWORKS modeling software. Then, export the models in .obj format, forming triangular facets of the model. One triangular facet is obtained at a time. Next, construct a line segment between the key point to be measured and the measurement origin of the laser tracker. Calculate the intersection of this line segment with the plane containing the triangular facet corresponding to the obstruction, and the sum of the areas of the three triangles formed by the vertices of the triangular facet. Determine whether the line segment intersects with the triangular facet. Let the coordinates of the three vertices of a triangular facet be q1 = (x1, y1, z1), q2 = (x2, y2, z2), and q3 = (x3, y3, z3); ​​and the coordinates of the key point to be measured and the measurement origin of the laser tracker be p... a =(x a ,y a ,z a ), p b =(x b ,y b ,z b The parametric equation of the line segment is expressed as equation (1): Where λ is a parameter; The plane equation of the triangular facet is expressed by the determinant as equation (2): Combining equations (1) and (2), we calculate the value of parameter λ using equation (3), where C and D are process parameters, and m1 = x b -x a m2 = y b -y a m3 = z b -z a If λ∈[0,1], then the triangular facet on the surface intersects with the line segment, and the coordinates of the intersection point are obtained by equation (1): The area of ​​the model triangular facet and the total area of ​​the three triangles formed by the intersection of the line and the three vertices of the model triangular facet are calculated using equation (4). Among them, S 原 S represents the area of ​​the triangular facet of the model. 新 The total area of ​​the three triangles formed by the intersection of the line and the plane and the vertices of the model's triangular facets is a1 = x2 - x1, a2 = x3 - x1, b1 = y2 - y1, b2 = y3 - y1, c1 = z2 - z1 and c2 = z3 - z1; Thus, the areas of the model's triangular facets and the total area of ​​the three triangles formed by the intersection of the line and the vertices of the model's triangular facets have been calculated; if S 原 =S 新 If the intersection of the measurement line of sight and the triangular plane is inside the triangle, that is, they intersect, and the measurement line of sight is blocked; otherwise, the measurement line of sight is not blocked. The second step is to construct an optimization function for the distributed laser tracker measurement stations. First, the deployable range of laser tracker stations, the ranging range of the laser tracker, and the angular measurement range are used as constraints. The objective function is to minimize the weighted sum of the three-dimensional measurement uncertainty of the laser tracker and the measurement uncertainty of key points. The parameters to be optimized are the number and location of laser tracker stations. A measurement point allocation factor θ(i) is introduced to allocate measurement tasks. In summary, an optimization function for distributed laser tracker measurement stations is constructed, where the objective function is shown in Equation (5), the measurement point allocation factor is shown in Equation (6), and the constraint function is shown in Equation (7). Where m is the number of laser tracker measurement stations; n is the number of key points; P LT,j p represents the coordinates of the j-th measurement station to be measured; i Let i be the coordinates of the i-th key point to be measured; Let l represent the minimum and maximum values ​​of the vertical angle, which is the angle between the line connecting the key point to be measured and the origin of the laser tracker, and the Z-axis in the tracker's coordinate system. max ε is the maximum permissible measurement distance. x ,ε y ,ε z The feasible regions for station deployment in the x, y, and z directions are respectively. The line-of-sight occlusion penalty factor is 1e8 when occlusion occurs and 0 when there is no occlusion. The third step is to solve the coordinates of the measurement station layout based on a genetic algorithm. In equation (6), θ(i) is used to allocate the key points to be measured to each measurement station according to the principle of shortest distance, that is: to determine the distance from the key points to be measured to each measurement station, if p i If the distance to station k is closest, then θ(i) = 1; otherwise, it is 0. Subsequently, a genetic algorithm is used to optimize the solution for the number and spatial location of measurement stations in the formula. The fitness function in the algorithm is the objective function. Considering that the feasible domain in the large-scale aviation assembly site is not parallel to each coordinate axis, the three-dimensional scene to be measured is transformed into a three-dimensional space parallel to the coordinate axis for solution. The configuration parameters obtained are inversely transformed back to the original three-dimensional space to obtain the final configuration parameters of the measurement station. The fourth step is the automatic planning of measurement sequences for discrete key points over a large area. Suppose there are n key points to be measured within a certain measurement station, and let G = (V p ,d p ) is a weighted graph, V p ={1,2,…,n} is the vertex set, d p Let the set of side lengths be d, and the distance between each vertex be d. i,j (d i,j >0, i,j∈V p The optimized mathematical model is Equation (8); Among them, D s d is the reciprocal of the total distance for each measurement point sequence. i,j Let a represent the distance between the i and j key points to be tested. i,j Here are the process parameters, S is all non-empty subsets of V, and |S| is the total number of vertices in the weighted graph G contained in set V; Then, a key point measurement sequence planning model based on an improved genetic algorithm was established. Heuristic search was used to reduce optimization time, and the measurement sequence was finally solved.

Images