A method and system for thickness measurement point planning of an irregular curved surface coating robot

By using an irregular curved surface coating robot to acquire point cloud data for 3D modeling and obstacle avoidance, the problem of controlling the thickness of heat-resistant coatings sprayed on spacecraft surfaces has been solved, and efficient and accurate automated thickness measurement path planning has been achieved.

CN117428757BActive Publication Date: 2026-11-17HUAZHONG UNIV OF SCI & TECH
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202311121086.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-08-31
Publication Date
2026-11-17
Estimated Expiration
2043-08-31

AI Technical Summary

Technical Problem

Existing technologies make it difficult to ensure the thickness and uniformity of heat-resistant coatings when spraying them on spacecraft surfaces. Furthermore, they rely on manual measurement and polishing, resulting in low efficiency, poor quality consistency, and occupational health hazards.

Method used

An irregular curved surface coating robot is used to acquire point cloud data for 3D modeling, calculate curvature matrix and normal vector, plan thickness measurement points, and perform obstacle collision detection and avoidance to achieve automated thickness measurement path planning.

Benefits of technology

It enables accurate location of the normal vector of the thickness measurement point on irregular curved surfaces, automatic detection and obstacle avoidance, improving the accuracy and efficiency of coating thickness control, reducing manual intervention, and enhancing quality consistency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117428757B_ABST
    Figure CN117428757B_ABST
Patent Text Reader

Abstract

This invention discloses a method and system for thickness measurement point planning of an irregular curved surface coating robot. The method includes: acquiring point cloud data of the workpiece to be measured; preprocessing the point cloud data; performing three-dimensional modeling of the workpiece based on the preprocessed point cloud data to generate a three-dimensional model of the workpiece; acquiring the coordinates of sample points of the point cloud data on the curved surface of the three-dimensional model of the workpiece; calculating the curvature matrix of the surface where the sample point coordinates are located, wherein the curvature matrix includes a height difference matrix of the surface; calculating the normal vector of each sample point based on the curvature matrix; taking the normal vector of the sample point closest to the thickness measurement point as the normal vector of the thickness measurement point; the coating robot planning its working path based on the normal vector of the thickness measurement point; and the coating robot performing collision detection and avoidance of obstacles on the planned working path.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of thickness measurement point planning technology, and more specifically, relates to a thickness measurement point planning method and system for an irregular curved surface coating robot. Background Technology

[0002] The surface of the spacecraft needs to be coated with a heat-resistant coating to reduce the impact of high temperatures generated by air friction on the structural strength of the spacecraft during high-speed flight.

[0003] Currently, the spraying process for heat-resistant coatings struggles to guarantee the required thickness accuracy, and the coating thickness uniformity varies significantly. After spraying, manual thickness gauges are required for large-area thickness measurements, followed by repeated manual grinding and polishing to ultimately control the coating thickness within the technical requirements. This process is heavily reliant on manual labor, as measurement accuracy is greatly affected by hand gestures, the amount removed in a single grinding and polishing cycle is unquantifiable, and it is also influenced by various uncontrollable factors such as manual experience and labor intensity. Consequently, the current coating grinding process suffers from low efficiency, poor quality consistency, and difficulty in guaranteeing accuracy.

[0004] Traditional operating methods are no longer sufficient to meet the needs of spacecraft development and have repeatedly become bottleneck processes in manufacturing. In addition, they pose occupational health risks. Therefore, there is an urgent need to conduct research on more advanced automated processing technologies, open up technical pathways, and achieve technological upgrades. Summary of the Invention

[0005] To address the above technical characteristics, this invention proposes a method for thickness measurement point planning for an irregular curved surface coating robot, comprising:

[0006] The point cloud data of the workpiece under test is acquired, the point cloud data is preprocessed, and the workpiece under test is three-dimensionally modeled based on the preprocessed point cloud data to generate a three-dimensional model of the workpiece under test. The coordinates of the sample points of the point cloud data on the surface of the three-dimensional model of the workpiece under test are acquired, and the curvature matrix of the surface where the sample point coordinates are located is calculated, wherein the curvature matrix includes the height difference matrix of the surface.

[0007] The normal vector of each sample point is calculated based on the curvature matrix. The normal vector of the sample point closest to the thickness measurement point is taken as the normal vector of the thickness measurement point. The coating robot plans its working path based on the normal vector of the thickness measurement point.

[0008] The coating robot performs collision detection and avoids obstacles on the planned work path.

[0009] Furthermore, calculating the curvature matrix of the surface where the thickness measurement point is located includes:

[0010]

[0011] Among them, s ij Let n be the curvature matrix on the plane with depth (i, j). ij Let x be the number of sample points in the plane at depth (i, j). k Let k be the coordinate vector of the k-th sample point in three-dimensional space. Let be the mean vector of the coordinates of the sample points on the plane at depth i. Let H be the mean vector of the coordinates of sample points on a plane of depth j. ij Let be the height difference matrix on the plane with depth (i, j).

[0012] Furthermore, the height difference matrix H on the plane at depth (i, j) ij include:

[0013]

[0014] Among them, T ij Let be the shear tensor on the planes of depths i and j.

[0015] Furthermore, the shear tensor T on the planes of depths i and j ij include:

[0016]

[0017] Furthermore, the process of the coating robot performing collision detection and avoidance of obstacles on the planned working path includes: dividing the three-dimensional model of the workpiece under test into multiple Thiessen polygons using Thiessen polygons, and performing collision detection on the Thiessen polygons.

[0018] This invention also proposes a thickness measurement point planning system for an irregular curved surface coating robot, comprising:

[0019] The calculation matrix module is used to acquire point cloud data of the workpiece under test, preprocess the point cloud data, and perform three-dimensional modeling of the workpiece under test based on the preprocessed point cloud data to generate a three-dimensional model of the workpiece under test. It also acquires the coordinates of sample points of the point cloud data of the surface of the three-dimensional model of the workpiece under test, and calculates the curvature matrix of the surface where the sample point coordinates are located, wherein the curvature matrix includes the height difference matrix of the calculated surface.

[0020] The path planning module is used to calculate the normal vector of each sample point based on the curvature matrix, and take the normal vector of the sample point closest to the thickness measurement point as the normal vector of the thickness measurement point. The coating robot plans the working path based on the normal vector of the thickness measurement point.

[0021] The obstacle avoidance module is used by the coating robot to detect and avoid obstacles on the planned working path.

[0022] Furthermore, calculating the curvature matrix of the surface where the thickness measurement point is located includes:

[0023]

[0024] Among them, s ij Let n be the curvature matrix on the plane with depth (i, j). ij Let x be the number of sample points in the plane at depth (i, j). k Let k be the coordinate vector of the k-th sample point in three-dimensional space. Let be the mean vector of the coordinates of the sample points on the plane at depth i. Let H be the mean vector of the coordinates of sample points on a plane of depth j. ij Let be the height difference matrix on the plane with depth (i, j).

[0025] Furthermore, the height difference matrix H on the plane at depth (i, j) ij include:

[0026]

[0027] Among them, T ij Let be the shear tensor on the planes of depths i and j.

[0028] Furthermore, the shear tensor T on the planes of depths i and j ij include:

[0029]

[0030] Furthermore, the process of the coating robot performing collision detection and avoidance of obstacles on the planned working path includes: dividing the three-dimensional model of the workpiece under test into multiple Thiessen polygons using Thiessen polygons, and performing collision detection on the Thiessen polygons.

[0031] Compared with the prior art, the above-described technical solutions conceived in this invention have the following beneficial effects:

[0032] This invention acquires point cloud data of the workpiece under test, preprocesses the point cloud data, and performs 3D modeling of the workpiece based on the preprocessed point cloud data to generate a 3D model of the workpiece. It obtains the coordinates of sample points on the curved surface of the 3D model of the workpiece, calculates the curvature matrix of the surface containing the sample points, where the curvature matrix includes a height difference matrix. Based on the curvature matrix, it calculates the normal vector of each sample point, and uses the normal vector of the sample point closest to the thickness measurement point as the normal vector of the thickness measurement point. The coating robot plans its working path based on the normal vector of the thickness measurement point. The coating robot performs collision detection and avoidance on the planned working path. Through the above technical solution, this invention can accurately find the normal vector of the thickness measurement point on an irregular curved surface, thereby planning the path, and can automatically detect and avoid obstacles. Attached Figure Description

[0033] Figure 1 This is a flowchart of Embodiment 1 of the present invention;

[0034] Figure 2 This is a structural diagram of the system in Embodiment 2 of the present invention. Detailed Implementation

[0035] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0036] The method provided by this invention can be implemented in a terminal environment that may include one or more of the following components: a processor, a storage medium, and a display screen. The storage medium stores at least one instruction, which is loaded and executed by the processor to implement the method described in the following embodiments.

[0037] A processor may include one or more processing cores. The processor uses various interfaces and lines to connect various parts of the terminal, and performs various functions and processes data by running or executing instructions, programs, code sets or instruction sets stored in the storage medium, and by calling data stored in the storage medium.

[0038] Storage media can include random access memory (RAM) or read-only memory (ROM). Storage media can be used to store instructions, programs, code, code sets, or instructions.

[0039] The display screen is used to show the user interface of each application.

[0040] In the formula of this invention, all subscripts are only used to distinguish parameters and have no actual meaning.

[0041] In addition, those skilled in the art will understand that the structure of the terminal described above does not constitute a limitation on the terminal. The terminal may include more or fewer components, or combine certain components, or have different component arrangements. For example, the terminal may also include radio frequency circuits, input units, sensors, audio circuits, power supplies, and other components, which will not be described in detail here.

[0042] Example 1

[0043] like Figure 1 As shown, this embodiment of the invention provides a method for thickness measurement point planning for an irregular curved surface coating robot, including:

[0044] Step 101: Obtain point cloud data of the workpiece under test, preprocess the point cloud data, and perform three-dimensional modeling of the workpiece under test based on the preprocessed point cloud data to generate a three-dimensional model of the workpiece under test. Obtain the coordinates of sample points of the point cloud data of the surface of the three-dimensional model of the workpiece under test, and calculate the curvature matrix of the surface where the sample point coordinates are located, wherein the curvature matrix includes the height difference matrix of the calculated surface.

[0045] Specifically, calculating the curvature matrix of the surface where the thickness measurement point is located includes:

[0046]

[0047] Among them, s ij Let n be the curvature matrix on the plane with depth (i, j). ij Let x be the number of sample points in the plane at depth (i, j). k Let k be the coordinate vector of the k-th sample point in three-dimensional space. Let be the mean vector of the coordinates of the sample points on the plane at depth i. Let H be the mean vector of the coordinates of sample points on a plane of depth j. ij Let be the height difference matrix on the plane with depth (i, j).

[0048] Specifically, the height difference matrix H on the plane with depth (i, j) ij include:

[0049]

[0050] Among them, T ij Let be the shear tensor on the planes of depths i and j.

[0051] Specifically, the shear tensor T on the planes of depths i and j ij include:

[0052]

[0053] Step 102: Calculate the normal vector of each sample point according to the curvature matrix, and take the normal vector of the sample point closest to the thickness measurement point as the normal vector of the thickness measurement point. The coating robot plans the working path according to the normal vector of the thickness measurement point.

[0054] Step 103: The coating robot performs collision detection and avoids obstacles on the planned working path.

[0055] Specifically, the process of the coating robot performing collision detection and avoidance of obstacles on the planned work path includes: dividing the three-dimensional model of the workpiece under test into multiple Thiessen polygons using Thiessen polygons, and performing collision detection on the Thiessen polygons.

[0056] The following is a specific example of obstacle detection in this implementation:

[0057] To determine the coating thickness at various points on an irregular curved surface of a workpiece, it is necessary to plan thickness measurement points on the surface. Besides calculating the accurate normal direction, the thickness measurement points must also avoid holes and unevenness on the irregular curved surface. This section will introduce how to generate uniformly distributed thickness measurement points on the surface based on the coating thickness distribution pattern, while avoiding holes. The positions of measurement points located in areas with holes need to be adjusted, and finally, an interference-free thickness measurement path is generated.

[0058] Since the irregular curved surface design model of the workpiece usually contains multiple NURBS surfaces, it is necessary to stitch the various NURBS surfaces on the surface of the irregular curved surface design model into a complete cylindrical surface, and generate uniformly distributed thickness measurement points on the surface according to the coating thickness distribution law. Let S represent the surface region, which contains all the thickness measurement points:

[0059] S={xi|i=1,2,...,n} (1)

[0060] Where xi is the coordinate of the i-th thickness measurement point, and n is the total number of measurement points.

[0061] However, due to the presence of holes on the surface, the positions of some measuring points need to be fine-tuned to avoid generating measuring points on the holes, which would cause measurement interference. To achieve this, it is necessary to determine whether each measuring point is within the surface region. When a measuring point is not within the surface region, it needs to be offset from the hole by a certain distance to the surrounding area. For example, the measuring point at position xi needs to be offset by a radius of ri to generate a new measuring point xi', i.e.:

[0062] x′ i =x i +r i *n i (2)

[0063] Where ri represents the offset distance and ni represents the offset unit vector.

[0064] Next, it is necessary to determine an interference-free thickness measurement path. Since the robot cannot collide with constraint objects on the surface during the measurement process, an interference-free thickness measurement path needs to be generated. For this purpose, collision detection algorithms from computer graphics can be used.

[0065] Specifically, a marker needs to be generated for each measurement point to indicate whether a measurement needs to be performed at that point. Then, the path between two adjacent marker points is calculated, and obstacles are detected on the path. If obstacles are found, the path needs to be adjusted to ensure that no collisions occur during the measurement process.

[0066] Suppose a measurement needs to be performed at position xi. Then, for the path between two adjacent positions xi and xj, it is necessary to detect whether there is an object intersecting with it. Simply put, the path can be regarded as a line segment, and the object as a geometric shape, and collision detection can be performed between them.

[0067] Representing the path between two adjacent markers as line segments [xi, xj] and the object as geometry B, a collision exists if and only if the path intersects with the object body. Thiessen polygons (Triangulated Irregular Network) can be used to divide the geometry and then collision detection can be performed on the Thiessen polygons. Specifically, the object body can be represented as a triangular mesh model, and then the model can be further divided into several small regions, each of which can be considered a face in the triangular mesh. Next, the intersection of the path with each small region is calculated to determine if the path intersects with the geometry. If the intersection of the path with a certain small region is not empty, it means that the path intersects with the geometry, and the path needs to be adjusted to avoid collisions.

[0068] The path can be adjusted by shifting it to either side. Specifically, if the path intersects with the workpiece, it can be moved a small distance so that it crosses the workpiece, thus avoiding a collision.

[0069] The following method can be used to determine whether there is a collision between the path and the workpiece body:

[0070] Intersection detection between path segment [xi,xj] and face Bk on the Thiessen polygon:

[0071] min_dist = the shortest distance between line segment [xi,xj] and surface Bk.

[0072] If min_dist < threshold, it indicates a collision and path adjustment is required.

[0073] 2. Adjust paths where collisions occur:

[0074] Choose a safe distance d, and adjust the path by translation so that the distance between the path and the workpiece is greater than d. The translation distance is...

[0075] t=d*n v (3)

[0076] Where, n v This represents the normal vector corresponding to the intersection point of the workpiece body and the path line segment [xi,xj].

[0077] Automatically generate interference-free paths: For all measurement points, execute the above steps sequentially to generate interference-free thickness measurement paths.

[0078] The following is a specific example of calculating the normal vector of the thickness measurement point in this implementation:

[0079] When using thickness measurement point data to measure the surface of a workpiece, it is usually necessary to obtain the normal direction of each thickness measurement point. In this section, we will introduce a specific method for calculating the normal direction of the thickness measurement points on the workpiece surface based on point cloud data.

[0080] Point cloud data needs to be preprocessed before computation to improve accuracy and efficiency. Point cloud data preprocessing includes:

[0081] 1. Smoothing Filtering: Point cloud data is smoothed by averaging the data from adjacent points to remove noise. This is typically achieved using a Gaussian filter or a mean filter. For each point x... i The result after processing is:

[0082]

[0083] Where, x j For adjacent points in point cloud data, w ij For the weights, N i This represents the number of adjacent points.

[0084] 2. Downsampling: Reducing the number of points in a point cloud data to decrease computational complexity. Downsampling typically employs two methods: random sampling and voxel reduction. Random sampling can be achieved by setting a downsampling rate r to reduce the number of points in the point cloud data.

[0085]

[0086] in, X represents the processed point cloud data, and rand(0,1) represents a random number between 0 and 1.

[0087] 3. Moving Center: Point cloud data needs to be moved to a position coinciding with the origin for subsequent calculations. Given a set X = {x1, x2, ..., x...} of all points in the point cloud. n} Calculate the centroid of the point cloud:

[0088]

[0089] Thus, for each point x i We obtain the coordinates of the M movements:

[0090] x′ i =x i -C (7)

[0091] To calculate the normal direction of the thickness measurement point on the workpiece surface, it is necessary to first calculate the normal vector of the point cloud surface. This can be done by calculating the curvature matrix of the surface where the thickness measurement point is located or by using a nearest neighbor search algorithm. Here, we first introduce the curvature matrix of the surface where the thickness measurement point is located, based on the normal vector (n) on the tangent plane. x n y n z ) and curvature K, representing the curvature vector of the tangent plane of the surface geometry. for:

[0092]

[0093] Let the curvature matrix of the point cloud be S = (s ij ),in

[0094]

[0095] Among them, s ij Let n be the curvature matrix on the plane with depth (i, j). ij Let x be the number of sample points in the plane at depth (i, j). k Let k be the coordinate vector of the k-th sample point in three-dimensional space. Let be the mean vector of the coordinates of the sample points on the plane at depth i. Let H be the mean vector of the coordinates of sample points on a plane of depth j. ij Let be the height difference matrix on the plane with depth (i, j).

[0096] n ij =|{(x k ) i ≤x i}∩{(x k )j ≤x j}| (10)

[0097] Let (i,j) represent the number of samples in a 2D plane of depth "i,j". The normal vector at point (i,j) can be obtained by calculating the second discrete derivative.

[0098] The nearest neighbor search method, also known as the normal vector consensus algorithm, can be used to obtain the surface normal vector. It utilizes the relationship between the normal vector and the surface curvature to calculate the surface normal vector. First, the surface normal vector needs to be estimated. Then, the normal vector is updated again by selecting the neighbor point with the strongest normal vector consensus. This update operation is repeated until the normal vector converges. This method is more suitable for surfaces with uneven surfaces than the curvature method.

[0099] Consider the k neighboring points x around point x. i We can calculate the normal vector of each neighboring point and its surrounding neighboring points, and then obtain the normal vector of that point by calculating the similarity between these normal vectors. First, define the curvature covariance matrix c:

[0100]

[0101] Where m is the center point of the nearest neighbor of point x.

[0102] The eigenvector v of matrix C i and eigenvalues ​​λ i This can be obtained by calculating the eigenvalue decomposition of C. The normal vector is the one with the smallest eigenvector, i.e.

[0103] n = v k (12)

[0104] For the newly updated normal vector, recalculate the normal vectors of each neighboring point and the neighboring points around the base, and then update the normal vectors based on the consistency of the normal vectors.

[0105] Next, the normal direction of the thickness measurement point needs to be calculated. For each thickness measurement point, the nearest point in the point cloud needs to be found, and the normal vector of that point is used as the normal direction of the thickness measurement point. For the point cloud data near the thickness measurement point, this can be obtained by calculating the average normal vector of the patch containing that point. The points around the thickness measurement point are labeled as {P}. i The nearest neighbor is labeled P. i The average normal vector is represented as n av Therefore, the normal direction of the thickness measurement point can be expressed as:

[0106]

[0107] Where, n pθ is the normal vector of the nearest neighbor, and θ is a parameter used to control the degree of influence of the selection of neighbor points on the normal vector.

[0108] Example 2

[0109] like Figure 2 As shown, this embodiment of the invention also provides a thickness measurement point planning system for an irregular curved surface coating robot, comprising:

[0110] The calculation matrix module is used to acquire point cloud data of the workpiece under test, preprocess the point cloud data, and perform three-dimensional modeling of the workpiece under test based on the preprocessed point cloud data to generate a three-dimensional model of the workpiece under test. It also acquires the coordinates of sample points of the point cloud data of the surface of the three-dimensional model of the workpiece under test, and calculates the curvature matrix of the surface where the sample point coordinates are located, wherein the curvature matrix includes the height difference matrix of the calculated surface.

[0111] Specifically, calculating the curvature matrix of the surface where the thickness measurement point is located includes:

[0112]

[0113] Among them, s ij Let n be the curvature matrix on the plane with depth (i, j). ij Let x be the number of sample points in the plane at depth (i, j). k Let k be the coordinate vector of the k-th sample point in three-dimensional space. Let be the mean vector of the coordinates of the sample points on the plane at depth i. Let H be the mean vector of the coordinates of sample points on a plane of depth j. ij Let be the height difference matrix on the plane with depth (i, j).

[0114] Specifically, the height difference matrix H on the plane with depth (i, j) ij include:

[0115]

[0116] Among them, T ij Let be the shear tensor on the planes of depths i and j.

[0117] Specifically, the shear tensor T on the planes of depths i and j ij include:

[0118]

[0119] The path planning module is used to calculate the normal vector of each sample point based on the curvature matrix, and take the normal vector of the sample point closest to the thickness measurement point as the normal vector of the thickness measurement point. The coating robot plans the working path based on the normal vector of the thickness measurement point.

[0120] The obstacle avoidance module is used by the coating robot to detect and avoid obstacles on the planned working path.

[0121] Specifically, the process of the coating robot performing collision detection and avoidance of obstacles on the planned work path includes: dividing the three-dimensional model of the workpiece under test into multiple Thiessen polygons using Thiessen polygons, and performing collision detection on the Thiessen polygons.

[0122] The following is a specific example of obstacle detection in this implementation:

[0123] To determine the coating thickness at various points on an irregular curved surface of a workpiece, it is necessary to plan thickness measurement points on the surface. Besides calculating the accurate normal direction, the measurement points must also avoid holes and irregularities on the surface. This section will introduce how to generate uniformly distributed thickness measurement points on the surface based on the coating thickness distribution pattern, while avoiding holes. The positions of measurement points located in areas with holes need to be adjusted, and finally, an interference-free thickness measurement path is generated.

[0124] Since the irregular curved surface design model of the workpiece usually contains multiple NURBS surfaces, it is necessary to stitch the various NURBS surfaces on the surface of the irregular curved surface design model into a complete cylindrical surface, and generate uniformly distributed thickness measurement points on the surface according to the coating thickness distribution law. Let S represent the surface region, which contains all the thickness measurement points:

[0125] S={xi|i=1, 2,..., n} (1)

[0126] Where xi is the coordinate of the i-th thickness measurement point, and n is the total number of measurement points.

[0127] However, due to the presence of holes on the surface, the positions of some measuring points need to be fine-tuned to avoid generating measuring points on the holes, which would cause measurement interference. To achieve this, it is necessary to determine whether each measuring point is within the surface region. When a measuring point is not within the surface region, it needs to be offset from the hole by a certain distance to the surrounding area. For example, the measuring point at position xi needs to be offset by a radius of ri to generate a new measuring point xi', i.e.:

[0128] x′ i =x i +r i *n i (2)

[0129] Where ri represents the offset distance and ni represents the offset unit vector.

[0130] Next, it is necessary to determine an interference-free thickness measurement path. Since the robot cannot collide with constraint objects on the surface during the measurement process, an interference-free thickness measurement path needs to be generated. For this purpose, collision detection algorithms from computer graphics can be used.

[0131] Specifically, a marker needs to be generated for each measurement point to indicate whether a measurement needs to be performed at that point. Then, the path between two adjacent marker points is calculated, and obstacles are detected on the path. If obstacles are found, the path needs to be adjusted to ensure that no collisions occur during the measurement process.

[0132] Suppose a measurement needs to be performed at position xi. Then, for the path between two adjacent positions xi and xj, it is necessary to detect whether there is an object intersecting with it. Simply put, the path can be regarded as a line segment, and the object as a geometric shape, and collision detection can be performed between them.

[0133] Representing the path between two adjacent markers as line segments [xi, xj] and the object as geometry B, a collision exists if and only if the path intersects with the object body. Thiessen polygons (Triangulated Irregular Network) can be used to divide the geometry and then collision detection can be performed on the Thiessen polygons. Specifically, the object body can be represented as a triangular mesh model, and then the model can be further divided into several small regions, each of which can be considered a face in the triangular mesh. Next, the intersection of the path with each small region is calculated to determine if the path intersects with the geometry. If the intersection of the path with a certain small region is not empty, it means that the path intersects with the geometry, and the path needs to be adjusted to avoid collisions.

[0134] The path can be adjusted by shifting it to either side. Specifically, if the path intersects with the workpiece, it can be moved a small distance so that it crosses the workpiece, thus avoiding a collision.

[0135] The following method can be used to determine whether there is a collision between the path and the workpiece body:

[0136] Intersection detection between path segment [xi,xj] and face Bk on the Thiessen polygon:

[0137] min_dist = the shortest distance between line segment [xi,xj] and surface Bk.

[0138] If min_dist < threshold, it indicates a collision and path adjustment is required.

[0139] 2. Adjust paths where collisions occur:

[0140] Choose a safe distance d, and adjust the path by translation so that the distance between the path and the workpiece is greater than d. The translation distance is...

[0141] t=d*n v (3)

[0142] Where, n v This represents the normal vector corresponding to the intersection point of the workpiece body and the path line segment [xi,xj].

[0143] Automatically generate interference-free paths: For all measurement points, execute the above steps sequentially to generate interference-free thickness measurement paths.

[0144] The following is a specific example of calculating the normal vector of the thickness measurement point in this implementation:

[0145] When using thickness measurement point data to measure the surface of a workpiece, it is usually necessary to obtain the normal direction of each thickness measurement point. In this section, we will introduce a specific method for calculating the normal direction of the thickness measurement points on the workpiece surface based on point cloud data.

[0146] Point cloud data needs to be preprocessed before computation to improve accuracy and efficiency. Point cloud data preprocessing includes:

[0147] 1. Smoothing Filtering: Point cloud data is smoothed by averaging the data from adjacent points to remove noise. This is typically achieved using a Gaussian filter or a mean filter. For each point x... i The result after processing is:

[0148]

[0149] Where, x j For adjacent points in point cloud data, w ij For the weights, N i This represents the number of adjacent points.

[0150] 2. Downsampling: Reducing the number of points in a point cloud data to decrease computational complexity. Downsampling typically employs two methods: random sampling and voxel reduction. Random sampling can be achieved by setting a downsampling rate r to reduce the number of points in the point cloud data.

[0151]

[0152] in, X represents the processed point cloud data, and rand(0,1) represents a random number between 0 and 1.

[0153] 3. Moving Center: Point cloud data needs to be moved to a position coinciding with the origin for subsequent calculations. Given a set X = {x1, x2, ..., x...} of all points in the point cloud. n} Calculate the centroid of the point cloud:

[0154]

[0155] Thus, for each point x i We obtain the coordinates of the M movements:

[0156] x′ i =x i -C (7)

[0157] To calculate the normal direction of the thickness measurement point on the workpiece surface, it is necessary to first calculate the normal vector of the point cloud surface. This can be done by calculating the curvature matrix of the surface where the thickness measurement point is located or by using a nearest neighbor search algorithm. Here, we first introduce the curvature matrix of the surface where the thickness measurement point is located, based on the normal vector (n) on the tangent plane. x n y n z ) and curvature K, representing the curvature vector of the tangent plane of the surface geometry. for:

[0158]

[0159] Let the curvature matrix of the point cloud be S = (s ij ),in

[0160]

[0161] Among them, s ij Let n be the curvature matrix on the plane with depth (i, j). ij Let x be the number of sample points in the plane at depth (i, j). k Let k be the coordinate vector of the k-th sample point in three-dimensional space. Let be the mean vector of the coordinates of the sample points on the plane at depth i. Let H be the mean vector of the coordinates of sample points on a plane of depth j. ij Let be the height difference matrix on the plane with depth (i, j).

[0162] n ij =|{(x k ) i ≤x i}∩{(x k j≤xj}| (10)

[0163] Let (i,j) represent the number of samples in a 2D plane of depth "i,j". The normal vector at point (i,j) can be obtained by calculating the second discrete derivative.

[0164] The nearest neighbor search method, also known as the normal vector consensus algorithm, can be used to obtain the surface normal vector. It utilizes the relationship between the normal vector and the surface curvature to calculate the surface normal vector. First, the surface normal vector needs to be estimated. Then, the normal vector is updated again by selecting the neighbor point with the strongest normal vector consensus. This update operation is repeated until the normal vector converges. This method is more suitable for surfaces with uneven surfaces than the curvature method.

[0165] Consider the k neighboring points x around point x. i We can calculate the normal vector of each neighboring point and its surrounding neighboring points, and then obtain the normal vector of that point by calculating the similarity between these normal vectors. First, define the curvature covariance matrix C:

[0166]

[0167] Where m is the center point of the nearest neighbor of point x.

[0168] The eigenvector v of matrix C i and eigenvalues ​​λ i This can be obtained by calculating the eigenvalue decomposition of C. The normal vector is the one with the smallest eigenvector, i.e.

[0169] n = v k (12)

[0170] For the newly updated normal vector, recalculate the normal vectors of each neighboring point and the neighboring points around the base, and then update the normal vectors based on the consistency of the normal vectors.

[0171] Next, the normal direction of the thickness measurement point needs to be calculated. For each thickness measurement point, the nearest point in the point cloud needs to be found, and the normal vector of that point is used as the normal direction of the thickness measurement point. For the point cloud data near the thickness measurement point, this can be obtained by calculating the average normal vector of the patch containing that point. The points around the thickness measurement point are labeled as {P}. i The nearest neighbor is labeled P. i The average normal vector is represented as n av Therefore, the normal direction of the thickness measurement point can be expressed as:

[0172]

[0173] Where, n p θ is the normal vector of the nearest neighbor, and θ is a parameter used to control the degree of influence of the selection of neighbor points on the normal vector.

[0174] Example 3

[0175] This invention also proposes a storage medium storing multiple instructions for implementing the thickness measurement point planning method for an irregular curved surface coating robot.

[0176] Optionally, in this embodiment, the storage medium may be located in any computer terminal in a group of computer terminals in a computer network, or in any mobile terminal in a group of mobile terminals.

[0177] Optionally, in this embodiment, the storage medium is configured to store program code for performing the following steps: Step 101, acquiring point cloud data of the workpiece under test, preprocessing the point cloud data, and performing three-dimensional modeling of the workpiece under test based on the preprocessed point cloud data to generate a three-dimensional model of the workpiece under test, acquiring the coordinates of sample points of the point cloud data of the surface of the three-dimensional model of the workpiece under test, and calculating the curvature matrix of the surface where the coordinates of the sample points are located, wherein the curvature matrix includes the calculation of the height difference matrix of the surface.

[0178] Specifically, calculating the curvature matrix of the surface where the thickness measurement point is located includes:

[0179]

[0180] Among them, s ij Let n be the curvature matrix on the plane with depth (i, j). ij Let x be the number of sample points in the plane at depth (i, j). k Let k be the coordinate vector of the k-th sample point in three-dimensional space. Let be the mean vector of the coordinates of the sample points on the plane at depth i. Let H be the mean vector of the coordinates of sample points on a plane of depth j. ij Let be the height difference matrix on the plane with depth (i, j).

[0181] Specifically, the height difference matrix H on the plane with depth (i, j) ij include:

[0182]

[0183] Among them, T ij Let be the shear tensor on the planes of depths i and j.

[0184] Specifically, the shear tensor T on the planes of depths i and j ij include:

[0185]

[0186] Step 102: Calculate the normal vector of each sample point according to the curvature matrix, and take the normal vector of the sample point closest to the thickness measurement point as the normal vector of the thickness measurement point. The coating robot plans the working path according to the normal vector of the thickness measurement point.

[0187] Step 103: The coating robot performs collision detection and avoids obstacles on the planned working path.

[0188] Specifically, the process of the coating robot performing collision detection and avoidance of obstacles on the planned work path includes: dividing the three-dimensional model of the workpiece under test into multiple Thiessen polygons using Thiessen polygons, and performing collision detection on the Thiessen polygons.

[0189] The following is a specific example of obstacle detection in this implementation:

[0190] To determine the coating thickness at various points on an irregular curved surface of a workpiece, it is necessary to plan thickness measurement points on the surface. Besides calculating the accurate normal direction, the measurement points must also avoid holes and irregularities on the surface. This section will introduce how to generate uniformly distributed thickness measurement points on the surface based on the coating thickness distribution pattern, while avoiding holes. The positions of measurement points located in areas with holes need to be adjusted, and finally, an interference-free thickness measurement path is generated.

[0191] Since the irregular curved surface design model of the workpiece usually contains multiple NURBS surfaces, it is necessary to stitch the various NURBS surfaces on the surface of the irregular curved surface design model into a complete cylindrical surface, and generate uniformly distributed thickness measurement points on the surface according to the coating thickness distribution law. Let S represent the surface region, which contains all the thickness measurement points:

[0192] S={xi|i=1, 2,..., n} (1)

[0193] Where xi is the coordinate of the i-th thickness measurement point, and n is the total number of measurement points.

[0194] However, due to the presence of holes on the surface, the positions of some measuring points need to be fine-tuned to avoid generating measuring points on the holes, which would cause measurement interference. To achieve this, it is necessary to determine whether each measuring point is within the surface region. When a measuring point is not within the surface region, it needs to be offset from the hole by a certain distance to the surrounding area. For example, the measuring point at position xi needs to be offset by a radius of ri to generate a new measuring point xi', i.e.:

[0195] x′ i =x i +r i *n i (2)

[0196] Where ri represents the offset distance and ni represents the offset unit vector.

[0197] Next, it is necessary to determine an interference-free thickness measurement path. Since the robot cannot collide with constraint objects on the surface during the measurement process, an interference-free thickness measurement path needs to be generated. For this purpose, collision detection algorithms from computer graphics can be used.

[0198] Specifically, a marker needs to be generated for each measurement point to indicate whether a measurement needs to be performed at that point. Then, the path between two adjacent marker points is calculated, and obstacles are detected on the path. If obstacles are found, the path needs to be adjusted to ensure that no collisions occur during the measurement process.

[0199] Suppose a measurement needs to be performed at position xi. Then, for the path between two adjacent positions xi and xj, it is necessary to detect whether there is an object intersecting with it. Simply put, the path can be regarded as a line segment, and the object as a geometric shape, and collision detection can be performed between them.

[0200] Representing the path between two adjacent markers as line segments [xi, xj] and the object as geometry B, a collision exists if and only if the path intersects with the object body. Thiessen polygons (Triangulated Irregular Network) can be used to divide the geometry and then collision detection can be performed on the Thiessen polygons. Specifically, the object body can be represented as a triangular mesh model, and then the model can be further divided into several small regions, each of which can be considered a face in the triangular mesh. Next, the intersection of the path with each small region is calculated to determine if the path intersects with the geometry. If the intersection of the path with a certain small region is not empty, it means that the path intersects with the geometry, and the path needs to be adjusted to avoid collisions.

[0201] The path can be adjusted by shifting it to either side. Specifically, if the path intersects with the workpiece, it can be moved a small distance so that it crosses the workpiece, thus avoiding a collision.

[0202] The following method can be used to determine whether there is a collision between the path and the workpiece body:

[0203] Intersection detection between path segment [xi,xj] and face Bk on the Thiessen polygon:

[0204] min_dist = the shortest distance between line segment [xi,xj] and surface Bk.

[0205] If min_dist < threshold, it indicates a collision and path adjustment is required.

[0206] 2. Adjust paths where collisions occur:

[0207] Choose a safe distance d, and adjust the path by translation so that the distance between the path and the workpiece is greater than d. The translation distance is...

[0208] t=d*n v (3)

[0209] Where, n v This represents the normal vector corresponding to the intersection point of the workpiece body and the path line segment [xi,xj].

[0210] Automatically generate interference-free paths: For all measurement points, execute the above steps sequentially to generate interference-free thickness measurement paths.

[0211] The following is a specific example of calculating the normal vector of the thickness measurement point in this implementation:

[0212] When using thickness measurement point data to measure the surface of a workpiece, it is usually necessary to obtain the normal direction of each thickness measurement point. In this section, we will introduce a specific method for calculating the normal direction of the thickness measurement points on the workpiece surface based on point cloud data.

[0213] Point cloud data needs to be preprocessed before computation to improve accuracy and efficiency. Point cloud data preprocessing includes:

[0214] 1. Smoothing Filtering: Point cloud data is smoothed by averaging the data from adjacent points to remove noise. This is typically achieved using a Gaussian filter or a mean filter. For each point x... i The result after processing is:

[0215]

[0216] Where, x j For adjacent points in point cloud data, w ij For the weights, N i This represents the number of adjacent points.

[0217] 2. Downsampling: Reducing the number of points in a point cloud data to decrease computational complexity. Downsampling typically employs two methods: random sampling and voxel reduction. Random sampling can be achieved by setting a downsampling rate r to reduce the number of points in the point cloud data.

[0218]

[0219] in, X represents the processed point cloud data, and rand(0,1) represents a random number between 0 and 1.

[0220] 3. Moving Center: Point cloud data needs to be moved to a position that coincides with the origin for subsequent calculations.

[0221] Given a set X = {x1, x2, ..., xn} of all points in a point cloud. n} Calculate the centroid of the point cloud:

[0222]

[0223] Thus, for each point x i We obtain the coordinates of the M movements:

[0224] x′ i =x i -C (7)

[0225] To calculate the normal direction of the thickness measurement point on the workpiece surface, it is necessary to first calculate the normal vector of the point cloud surface. This can be done by calculating the curvature matrix of the surface where the thickness measurement point is located or by using a nearest neighbor search algorithm. Here, we first introduce the curvature matrix of the surface where the thickness measurement point is located, based on the normal vector (n) on the tangent plane. x n y n z ) and curvature K, representing the curvature vector of the tangent plane of the surface geometry. for:

[0226]

[0227] Let the curvature matrix of the point cloud be S = (s ij ),in

[0228]

[0229] Among them, s ij Let n be the curvature matrix on the plane with depth (i, j). ij Let x be the number of sample points in the plane at depth (i, j). k Let k be the coordinate vector of the k-th sample point in three-dimensional space. Let be the mean vector of the coordinates of the sample points on the plane at depth i. Let H be the mean vector of the coordinates of sample points on a plane of depth j. ij Let be the height difference matrix on the plane with depth (i, j).

[0230] n ij =|{(x k ) i ≤x i}∩{(x k ) j ≤x j}| (10)

[0231] Let (i,j) represent the number of samples in a 2D plane of depth "i,j". The normal vector at point (i,j) can be obtained by calculating the second discrete derivative.

[0232] The nearest neighbor search method, also known as the normal vector consensus algorithm, can be used to obtain the surface normal vector. It utilizes the relationship between the normal vector and the surface curvature to calculate the surface normal vector. First, the surface normal vector needs to be estimated. Then, the normal vector is updated again by selecting the neighbor point with the strongest normal vector consensus. This update operation is repeated until the normal vector converges. This method is more suitable for surfaces with uneven surfaces than the curvature method.

[0233] Consider the k neighboring points x around point x. i We can calculate the normal vector of each neighboring point and its surrounding neighboring points, and then obtain the normal vector of that point by calculating the similarity between these normal vectors. First, define the curvature covariance matrix C:

[0234]

[0235] Where m is the center point of the nearest neighbor of point x.

[0236] The eigenvector v of matrix C i and eigenvalues ​​λ i This can be obtained by calculating the eigenvalue decomposition of C. The normal vector is the one with the smallest eigenvector, i.e.

[0237] n = v k (12)

[0238] For the newly updated normal vector, recalculate the normal vectors of each neighboring point and the neighboring points around the base, and then update the normal vectors based on the consistency of the normal vectors.

[0239] Next, the normal direction of the thickness measurement point needs to be calculated. For each thickness measurement point, the nearest point in the point cloud needs to be found, and the normal vector of that point is used as the normal direction of the thickness measurement point. For the point cloud data near the thickness measurement point, this can be obtained by calculating the average normal vector of the patch containing that point. The points around the thickness measurement point are labeled as {P}. i The nearest neighbor is labeled P. i The average normal vector is represented as n av Therefore, the normal direction of the thickness measurement point can be expressed as:

[0240]

[0241] Where, n pθ is the normal vector of the nearest neighbor, and θ is a parameter used to control the degree of influence of the selection of neighbor points on the normal vector.

[0242] Example 4

[0243] This invention also proposes an electronic device, including a processor and a storage medium connected to the processor. The storage medium stores multiple instructions, which can be loaded and executed by the processor to enable the processor to execute a thickness measurement point planning method for an irregular curved surface coating robot.

[0244] Specifically, the electronic device in this embodiment can be a computer terminal, which may include one or more processors and a storage medium.

[0245] The storage medium can be used to store software programs and modules, such as the thickness measurement point planning method for an irregular curved surface coating robot in this embodiment of the invention. The corresponding program instructions / modules are executed by the processor through running the software programs and modules stored in the storage medium, thereby performing various functional applications and data processing, thus realizing the aforementioned thickness measurement point planning method for an irregular curved surface coating robot. The storage medium may include high-speed random access storage media, and may also include non-volatile storage media, such as one or more magnetic storage systems, flash memory, or other non-volatile solid-state storage media. In some instances, the storage medium may further include storage media remotely configured relative to the processor, which can be connected to the terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0246] The processor can call the information and application stored in the storage medium through the transmission system to perform the following steps: Step 101, acquire the point cloud data of the workpiece under test, preprocess the point cloud data, and perform three-dimensional modeling of the workpiece under test based on the preprocessed point cloud data to generate a three-dimensional model of the workpiece under test, acquire the coordinates of the sample points of the point cloud data of the surface of the three-dimensional model of the workpiece under test, and calculate the curvature matrix of the surface where the sample point coordinates are located, wherein the curvature matrix includes the height difference matrix of the calculated surface.

[0247] Specifically, calculating the curvature matrix of the surface where the thickness measurement point is located includes:

[0248]

[0249] Among them, s ij Let n be the curvature matrix on the plane with depth (i, j). ij Let x be the number of sample points in the plane at depth (i, j). k Let k be the coordinate vector of the k-th sample point in three-dimensional space. Let be the mean vector of the coordinates of the sample points on the plane at depth i. Let H be the mean vector of the coordinates of sample points on a plane of depth j. ij Let be the height difference matrix on the plane with depth (i, j).

[0250] Specifically, the height difference matrix H on the plane with depth (i, j) ij include:

[0251]

[0252] Among them, T ij Let be the shear tensor on the planes of depths i and j.

[0253] Specifically, the shear tensor T on the planes of depths i and j ij include:

[0254]

[0255] Step 102: Calculate the normal vector of each sample point according to the curvature matrix, and take the normal vector of the sample point closest to the thickness measurement point as the normal vector of the thickness measurement point. The coating robot plans the working path according to the normal vector of the thickness measurement point.

[0256] Step 103: The coating robot performs collision detection and avoids obstacles on the planned working path.

[0257] Specifically, the process of the coating robot performing collision detection and avoidance of obstacles on the planned work path includes: dividing the three-dimensional model of the workpiece under test into multiple Thiessen polygons using Thiessen polygons, and performing collision detection on the Thiessen polygons.

[0258] The following is a specific example of obstacle detection in this implementation:

[0259] To determine the coating thickness at various points on an irregular curved surface of a workpiece, it is necessary to plan thickness measurement points on the surface. Besides calculating the accurate normal direction, the measurement points must also avoid holes and irregularities on the surface. This section will introduce how to generate uniformly distributed thickness measurement points on the surface based on the coating thickness distribution pattern, while avoiding holes. The positions of measurement points located in areas with holes need to be adjusted, and finally, an interference-free thickness measurement path is generated.

[0260] Since the irregular curved surface design model of the workpiece usually contains multiple NURBS surfaces, it is necessary to stitch the various NURBS surfaces on the surface of the irregular curved surface design model into a complete cylindrical surface, and generate uniformly distributed thickness measurement points on the surface according to the coating thickness distribution law. Let S represent the surface region, which contains all the thickness measurement points:

[0261] S={xi|i=1, 2,..., n} (1)

[0262] Where xi is the coordinate of the i-th thickness measurement point, and n is the total number of measurement points.

[0263] However, due to the presence of holes on the surface, the positions of some measuring points need to be fine-tuned to avoid generating measuring points on the holes, which would cause measurement interference. To achieve this, it is necessary to determine whether each measuring point is within the surface region. When a measuring point is not within the surface region, it needs to be offset from the hole by a certain distance to the surrounding area. For example, the measuring point at position xi needs to be offset by a radius of ri to generate a new measuring point xi', i.e.:

[0264] x′ i =x i +r i *n i (2)

[0265] Where ri represents the offset distance and ni represents the offset unit vector.

[0266] Next, it is necessary to determine an interference-free thickness measurement path. Since the robot cannot collide with constraint objects on the surface during the measurement process, an interference-free thickness measurement path needs to be generated. For this purpose, collision detection algorithms from computer graphics can be used.

[0267] Specifically, a marker needs to be generated for each measurement point to indicate whether a measurement needs to be performed at that point. Then, the path between two adjacent marker points is calculated, and obstacles are detected on the path. If obstacles are found, the path needs to be adjusted to ensure that no collisions occur during the measurement process.

[0268] Suppose a measurement needs to be performed at position xi. Then, for the path between two adjacent positions xi and xj, it is necessary to detect whether there is an object intersecting with it. Simply put, the path can be regarded as a line segment, and the object as a geometric shape, and collision detection can be performed between them.

[0269] Representing the path between two adjacent markers as line segments [xi, xj] and the object as geometry B, a collision exists if and only if the path intersects with the object body. Thiessen polygons (Triangulated Irregular Network) can be used to divide the geometry and then collision detection can be performed on the Thiessen polygons. Specifically, the object body can be represented as a triangular mesh model, and then the model can be further divided into several small regions, each of which can be considered a face in the triangular mesh. Next, the intersection of the path with each small region is calculated to determine if the path intersects with the geometry. If the intersection of the path with a certain small region is not empty, it means that the path intersects with the geometry, and the path needs to be adjusted to avoid collisions.

[0270] The path can be adjusted by shifting it to either side. Specifically, if the path intersects with the workpiece, it can be moved a small distance so that it crosses the workpiece, thus avoiding a collision.

[0271] The following method can be used to determine whether there is a collision between the path and the workpiece body:

[0272] Intersection detection between path segment [xi,xj] and face Bk on the Thiessen polygon:

[0273] min_dist = the shortest distance between line segment [xi,xj] and surface Bk.

[0274] If min_dist < threshold, it indicates a collision and path adjustment is required.

[0275] 2. Adjust paths where collisions occur:

[0276] Choose a safe distance d, and adjust the path by translation so that the distance between the path and the workpiece is greater than d. The translation distance is...

[0277] t=d*n v (3)

[0278] Where, n v This represents the normal vector corresponding to the intersection point of the workpiece body and the path line segment [xi,xj].

[0279] Automatically generate interference-free paths: For all measurement points, execute the above steps sequentially to generate interference-free thickness measurement paths.

[0280] The following is a specific example of calculating the normal vector of the thickness measurement point in this implementation:

[0281] When using thickness measurement point data to measure the surface of a workpiece, it is usually necessary to obtain the normal direction of each thickness measurement point. In this section, we will introduce a specific method for calculating the normal direction of the thickness measurement points on the workpiece surface based on point cloud data.

[0282] Point cloud data needs to be preprocessed before computation to improve accuracy and efficiency. Point cloud data preprocessing includes:

[0283] 1. Smoothing Filtering: Point cloud data is smoothed by averaging the data from adjacent points to remove noise. This is typically achieved using a Gaussian filter or a mean filter. For each point x... i The result after processing is:

[0284]

[0285] Where, x j For adjacent points in point cloud data, w ij For the weights, N i This represents the number of adjacent points.

[0286] 2. Downsampling: Reducing the number of points in a point cloud data to decrease computational complexity. Downsampling typically employs two methods: random sampling and voxel reduction. Random sampling can be achieved by setting a downsampling rate r to reduce the number of points in the point cloud data.

[0287]

[0288] in, X represents the processed point cloud data, and rand(0,1) represents a random number between 0 and 1.

[0289] 3. Moving Center: Point cloud data needs to be moved to a position coinciding with the origin for subsequent calculations. Given a set X = {x1, x2, ..., x...} of all points in the point cloud. n} Calculate the centroid of the point cloud:

[0290]

[0291] Thus, for each point x i We obtain the coordinates of the M movements:

[0292] x′ i =x i -C (7)

[0293] To calculate the normal direction of the thickness measurement point on the workpiece surface, it is necessary to first calculate the normal vector of the point cloud surface. This can be done by calculating the curvature matrix of the surface where the thickness measurement point is located or by using a nearest neighbor search algorithm. Here, we first introduce the curvature matrix of the surface where the thickness measurement point is located, based on the normal vector (n) on the tangent plane. x n y n z ) and curvature K, representing the curvature vector of the tangent plane of the surface geometry. for:

[0294]

[0295] Let the curvature matrix of the point cloud be S = (s ij ),in

[0296]

[0297] Among them, s ij Let n be the curvature matrix on the plane with depth (i, j). ij Let x be the number of sample points in the plane at depth (i, j). k Let k be the coordinate vector of the k-th sample point in three-dimensional space. Let be the mean vector of the coordinates of the sample points on the plane at depth i. Let H be the mean vector of the coordinates of sample points on a plane of depth j. ij Let be the height difference matrix on the plane with depth (i, j).

[0298] n ij =|{(x k ) i ≤x i}∩{(x k ) j ≤x j}| (10)

[0299] Let (i,j) represent the number of samples in a 2D plane of depth "i,j". The normal vector at point (i,j) can be obtained by calculating the second discrete derivative.

[0300] The nearest neighbor search method, also known as the normal vector consensus algorithm, can be used to obtain the surface normal vector. It utilizes the relationship between the normal vector and the surface curvature to calculate the surface normal vector. First, the surface normal vector needs to be estimated. Then, the normal vector is updated again by selecting the neighbor point with the strongest normal vector consensus. This update operation is repeated until the normal vector converges. This method is more suitable for surfaces with uneven surfaces than the curvature method.

[0301] Consider the k neighboring points x around point x. iWe can calculate the normal vector of each neighboring point and its surrounding neighboring points, and then obtain the normal vector of that point by calculating the similarity between these normal vectors. First, define the curvature covariance matrix C:

[0302]

[0303] Where m is the center point of the nearest neighbor of point x.

[0304] The eigenvector v of matrix C i and eigenvalues ​​λ i This can be obtained by calculating the eigenvalue decomposition of C. The normal vector is the one with the smallest eigenvector, i.e.

[0305] n = v k (12)

[0306] For the newly updated normal vector, recalculate the normal vectors of each neighboring point and the neighboring points around the base, and then update the normal vectors based on the consistency of the normal vectors.

[0307] Next, the normal direction of the thickness measurement point needs to be calculated. For each thickness measurement point, the nearest point in the point cloud needs to be found, and the normal vector of that point is used as the normal direction of the thickness measurement point. For the point cloud data near the thickness measurement point, this can be obtained by calculating the average normal vector of the patch containing that point. The points around the thickness measurement point are labeled as {P}. i The nearest neighbor is labeled P. i The average normal vector is represented as n av Therefore, the normal direction of the thickness measurement point can be expressed as:

[0308]

[0309] Where, n p θ is the normal vector of the nearest neighbor, and θ is a parameter used to control the degree of influence of the selection of neighbor points on the normal vector.

[0310] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0311] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0312] In the several embodiments provided by this invention, it should be understood that the disclosed technical content can be implemented in other ways. The system embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or indirect coupling or communication connection between units or modules, and may be electrical or other forms.

[0313] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0314] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0315] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes: USB flash drives, read-only storage media (ROM), random access storage media (RAM), portable hard drives, magnetic disks, optical disks, and other media capable of storing program code.

[0316] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for planning thickness measurement points for an irregular curved surface coating robot, characterized in that, include: The point cloud data of the workpiece under test is acquired, the point cloud data is preprocessed, and the workpiece under test is three-dimensionally modeled based on the preprocessed point cloud data to generate a three-dimensional model of the workpiece under test. The coordinates of sample points of the point cloud data on the surface of the three-dimensional model of the workpiece under test are acquired, and the curvature matrix of the surface where the sample point coordinates are located is calculated, wherein the curvature matrix includes the height difference matrix of the surface. The normal vector of each sample point is calculated based on the curvature matrix. The normal vector of the sample point closest to the thickness measurement point is taken as the normal vector of the thickness measurement point. The coating robot plans its working path based on the normal vector of the thickness measurement point. The coating robot performs collision detection and avoids obstacles on the planned work path.

2. The thickness measurement point planning method for an irregular curved surface coating robot as described in claim 1, characterized in that, Calculating the curvature matrix of the surface containing the coordinates of the sample point includes: , in, For a depth of ( , The curvature matrix on the plane, For a depth of ( , The number of sample points on the plane, For the first The coordinate vector of each sample point in three-dimensional space. For a depth of The mean vector of the coordinates of the sample points on the plane. For a depth of The mean vector of the coordinates of the sample points on the plane. For a depth of ( , The height difference matrix on the plane; Depth is ( , Height difference matrix on the plane include: , in, For depth and depth The shear tensor on the plane; Depth and depth shear tensor on the plane include: 。 3. The thickness measurement point planning method for an irregular curved surface coating robot as described in claim 1, characterized in that, The coating robot performs collision detection and avoidance of obstacles on the planned working path by: dividing the three-dimensional model of the workpiece under test into multiple Thiessen polygons, and performing collision detection on the Thiessen polygons.

4. A thickness measurement point planning system for an irregular curved surface coating robot, characterized in that, include: The calculation matrix module is used to acquire point cloud data of the workpiece under test, preprocess the point cloud data, and perform three-dimensional modeling of the workpiece under test based on the preprocessed point cloud data to generate a three-dimensional model of the workpiece under test. It also acquires the coordinates of sample points of the point cloud data on the surface of the three-dimensional model of the workpiece under test, and calculates the curvature matrix of the surface where the sample point coordinates are located, wherein the curvature matrix includes the height difference matrix of the calculated surface. The path planning module is used to calculate the normal vector of each sample point based on the curvature matrix, and take the normal vector of the sample point closest to the thickness measurement point as the normal vector of the thickness measurement point. The coating robot plans the working path based on the normal vector of the thickness measurement point. The obstacle avoidance module is used by the coating robot to detect and avoid obstacles on the planned working path.

5. The thickness measurement point planning system for an irregular curved surface coating robot as described in claim 4, characterized in that, Calculating the curvature matrix of the surface containing the coordinates of the sample point includes:

6. Among them, For a depth of ( , The curvature matrix on the plane, For a depth of ( , The number of sample points on the plane, For the first The coordinate vector of each sample point in three-dimensional space. For a depth of The mean vector of the coordinates of the sample points on the plane. For a depth of The mean vector of the coordinates of the sample points on the plane. For a depth of ( , The height difference matrix on the plane; Depth is ( , Height difference matrix on the plane include: , in, For depth and depth The shear tensor on the plane; Depth and depth shear tensor on the plane include: 。 7. The thickness measurement point planning system for an irregular curved surface coating robot as described in claim 4, characterized in that, The coating robot performs collision detection and avoidance of obstacles on the planned working path by: dividing the three-dimensional model of the workpiece under test into multiple Thiessen polygons, and performing collision detection on the Thiessen polygons.

Citation Information

Patent Citations

  • Layered spraying track planning method used for spraying robot

    CN102831306A

  • Coating thickness detection device, spraying system comprising same and spraying control method

    CN109499796A