Convex curved surface workpiece laser processing planning method and device based on grid distortion

Abstract

The invention discloses a convex surface workpiece laser processing planning method and device based on grid distortion, and the method comprises the steps: capturing distortion grid images of a laser grid on a convex surface workpiece through four industrial cameras (X positive direction, X negative direction, Y positive direction and Y negative direction) at the same time, extracting grid node information in combination with an edge detection algorithm, and rapidly calculating the local gradient of the surface of the workpiece. A gradient vector cross product method is adopted to calculate a normal vector, and automatic direct calculation is carried out through a Python program, so that efficient estimation of the curved surface normal vector is realized, and time and cost of a traditional acquisition point cloud calculation normal vector method are reduced; by setting a threshold value delta, performing sorting and iterative calculation on deviations of grid normal vectors, automatically identifying areas with small local curvature changes, performing grid merging and forming block processing areas, the adaptive block processing method reduces the complexity of a processing path, improves the processing efficiency and reduces the processing cost. Meanwhile, it is ensured that the laser incident angle is aligned with the average normal vector of the machining area, and the laser machining quality is optimized.

Description

Technical Field

[0001] The present invention relates to the technical field of laser processing, and particularly relates to a laser processing planning method and device for convex surface workpieces based on grid distortion. Background Art

[0002] Perovskite solar cells, as the third-generation solar cell technology, have the characteristics of high efficiency, low cost, weak light effect, environmental friendliness, etc., and show excellent application potential in the field of photovoltaic power generation. The scribing of perovskite cells is one of the key steps in the manufacturing process of perovskite cells. It uses a laser device to precisely scribe the perovskite thin film to achieve the series connection and performance optimization of the cells.

[0003] The curved surface characteristics of perovskite cells usually come from the flexibility of their substrate materials or preformed curved surface structures (such as PET, metal foils, etc.). If the flat material is scribed first and then bent, the following problems may occur: 1. The scribed groove lines may deform or break during bending, destroying the electrical isolation structure and causing series connection failure between sub-cells (such as P2 connection misalignment); 2. Different film layers (such as FTO, perovskite, metal electrodes) may show delamination or cracks due to bending stress, affecting the stability of the cell. Traditional scribing equipment and methods cannot adapt to the changes of the curved surface, resulting in difficulty in ensuring the scribing accuracy and straightness. The scribing accuracy and straightness are crucial for the performance of the cell. If the scribing is inaccurate, it may lead to uneven current distribution inside the cell, thereby affecting the photoelectric conversion efficiency and stability of the cell. Therefore, directly scribing on the curved surface requires the laser device to have three-dimensional dynamic focusing and curved surface path planning capabilities to ensure the scribing accuracy and straightness.

[0004] In many engineering cases, the actual shape of the curved surface perovskite cell does not conform to the CAD model. Therefore, it is necessary to reversely obtain the shape and normal vector of the workpiece surface and automatically plan the scribing processing path based on the reverse model.

[0005] Due to the following five reasons, the block machining planning based on the surface normal vector has become a necessary pre-step for the automatic programming-free path planning: 1) The stroke of the motion mechanism is limited, making it difficult to cover large-sized surfaces in one go; 2) The curvatures of different regions of the curved surface workpiece may vary greatly, and unified machining parameters cannot adapt to the whole; 3) When continuously machining the curved surface workpiece, the mechanical motion error gradually amplifies; 4) The global path planning of the curved surface workpiece needs to process a large amount of data, with a heavy computational burden, and the data processing can be accelerated through parallel computing; 5) The failure of a single global path planning may lead to the scrapping of the entire curved surface workpiece. However, the methods for obtaining the surface normal vector in the past were less efficient and required more computing resources, increasing the cost of obtaining the normal vector. For example, the implicit function partial derivative calculation method requires accurate parsing of the surface equation, with high computational complexity for high-order NURBS or B-spline surfaces, and it is difficult to meet the real-time requirements of block machining planning; the covariance matrix analysis method based on PCA needs to process a large amount of neighbor point data, with a large amount of calculation, and its efficiency significantly decreases especially in large-sized surfaces; the main purpose of traditional surface reconstruction technology is to obtain the point cloud position, which is time-consuming and costly. Summary of the Invention

[0006] To solve the problems existing in the above-mentioned background technology, the present invention provides a laser machining planning method and device for convex surface workpieces based on grid distortion, which realizes fast calculation of the surface normal vector with high efficiency and low cost and adaptive block machining planning through grid distortion analysis.

[0007] To achieve the above purpose, the present invention adopts the following technical solutions: The present invention provides a laser machining planning method for convex surface workpieces based on grid distortion, including: Place the convex surface workpiece under the laser grid projector and project the laser grid onto the surface of the convex surface workpiece; Take the beam emission direction of the laser grid projector as the Z-axis direction of the space coordinate system, and four cameras are respectively arranged in the positive and negative directions of the X-axis and the positive and negative directions of the Y-axis, and the optical axes of the four cameras are all perpendicular to the Z-axis direction, and four distorted grid images are respectively captured by the four cameras; Extract the grid node information on the surface of the workpiece from the captured distorted grid images, record the grid point coordinates of each grid, calculate the local gradient of the surface, and calculate the surface normal vector by the gradient cross product method; By setting the normal vector deviation threshold δ, sort and iteratively calculate the deviation of the adjacent grid normal vectors, automatically identify the regions with relatively small local curvature changes, and perform grid merging to generate block machining regions; Sample along the grid lines in the surface to generate laser machining path points, and generate a laser machining path in combination with the average normal vector after block division.

[0008] In the above technical solution: The four cameras are orthogonally arranged, which can ensure that the data in two directions are completely decoupled during gradient calculation, avoid cross-interference caused by the tilted view angle, and the distorted grid captured by the cameras directly corresponds to the local gradients of the curved surface in the X and Y directions. In addition, the optical axes of all cameras are perpendicular to the Z-axis direction of the laser projection, so that the image distortion is only caused by the height change of the workpiece curved surface, and there is no need to additionally compensate for the perspective distortion caused by the camera tilt, improving the accuracy of distortion capture.

[0009] Furthermore, denoise the captured distorted grid image, compensate for lens distortion, and extract the grid node information on the workpiece surface through the edge detection algorithm.

[0010] Furthermore, the local gradient of the curved surface represents the height change rate (i.e., slope) of the curved surface in the X and Y directions at the grid points of the grid, and the calculation formula is as follows: ; ; ; where, represents the slope of the curved surface in the X direction at the point (a positive value indicates an upward slope, and a negative value indicates a downward slope), represents the slope of the curved surface in the Y direction at the point (a positive value indicates an upward slope, and a negative value indicates a downward slope), the coordinates of are and

[0011] is the height at which the laser intersects the curved surface. ; ; ; ; ; Then perform normalization to obtain .

[0012] Furthermore, sort the normal vector deviations between all adjacent grids, and iteratively judge the adjacent grid with the smallest current normal vector deviation by setting the normal vector deviation threshold δ. If it does not exceed δ, merge the grids into the machining blocks α, β, γ..., and store the coordinates and normal vectors of all grid points of the machining blocks α, β, γ... respectively.

[0013] Further, for the generated laser processing path, between blocks, the change in the machining attitude (change in the normal vector) is smoothed by a B-spline curve fitting algorithm.

[0014] Further, adjust the height of the laser grid projector until the laser grid emitted by the laser grid projector completely covers the surface of the convex curved workpiece.

[0015] Further, the two cameras located in the positive and negative directions of the X-axis are on the same central axis, the two cameras located in the positive and negative directions of the Y-axis are on the same central axis, and the heights of the four cameras are adjustable and kept consistent.

[0016] The present invention also provides a laser processing planning device for a convex curved workpiece based on grid distortion, which is used to implement the above-mentioned laser processing planning method for a convex curved workpiece, and includes a laser grid projector, a cross, columns, a work platform and four cameras. A telescopic component that can move along the Z-axis direction is installed at the middle position of the cross. A laser grid projector is installed at the driving end of the telescopic component. The work platform is located directly below the cross. Four vertical columns are respectively arranged on the top of the work platform. The tops of the four columns are respectively connected to the four ends of the cross in a butt joint manner. A camera is respectively installed on each column.

[0017] Further, the installation height of the camera on the column is adjustable. Four feet are arranged at the bottom of the work platform, and the height of the feet is adjustable. A leveling component is arranged between the column and the cross end bracket.

[0018] In this device, the laser grid projector can be adjusted so that the projection grid can completely cover the workpiece, and the projection range and the camera view angle can be adjusted according to the size and shape of the convex curved workpiece. The height of the camera can be adjusted to adapt to convex curved workpieces of different specifications, improving the versatility and adaptability of the system.

[0019] Compared with the prior art, the present invention has the following beneficial effects: (1) Four industrial cameras (X positive, X negative, Y positive, Y negative) simultaneously capture the distorted grid images of the laser grid on the convex curved workpiece, and combine the edge detection algorithm to extract the grid node information, quickly calculate the local gradient of the workpiece surface, calculate the normal vector by the cross product method of the gradient vector, and directly calculate it automatically through a Python program, realizing the efficient estimation of the surface normal vector and reducing the time and cost of the traditional method of calculating the normal vector by collecting cloud points.

[0020] (2) By setting a threshold value δ, the deviations of the grid normal vectors are sorted and iteratively calculated to automatically identify the regions with relatively small local curvature changes, and grid merging is performed to form machining blocks α, β, γ... This adaptive block division method reduces the complexity of the machining path, improves the machining efficiency, and at the same time ensures that the laser incident angle is aligned with the average normal vector of the machining region, optimizing the laser machining quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0022] Figure 1 It is a schematic structural diagram of a laser machining planning device for a convex surface workpiece in Embodiment 1; Figure 2 It is a laser grid image after surface distortion of a convex surface workpiece in Embodiment 2; Figure 3 It is to generate a laser machining path according to grid lines and normal vectors in Embodiment 2; Figure 4 It is the original image collected by the camera in Embodiment 2; Figure 5 It is to denoise the original image collected by the camera using the Gaussian filtering algorithm in Embodiment 2; Figure 6 It is to obtain grid point information using the Harris corner detection in Embodiment 2; Figure 7 It is a diagram showing the method for calculating normal vectors by traditional point cloud; Figure 8 It is a schematic flow diagram of step (8) in Embodiment 2; Among them, the specific reference numerals are: Laser grid projector 1, telescopic assembly 2, cross 3, leveling assembly 4, column 5, waist-shaped hole 6, camera 7, workbench 8, hydraulic leveling support rod 9. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0023] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0024] Embodiment 1 This embodiment provides a laser machining planning device for a convex surface workpiece based on grid distortion, as Figure 1As shown in the figure, it includes a laser grid projector 1, a cross 3, a column 5, a working platform 8, and four cameras 7 (using black and white industrial cameras with a resolution of 20 million pixels). A telescopic component 2 that can move along the Z-axis direction is installed at the middle position of the cross 3. The driving end of the telescopic component 2 is installed with a laser grid projector 1. The telescopic component 2 can be selected from an electric telescopic adjustment structure (such as an electric push rod) and a manual mechanical telescopic adjustment structure (lead screw + scale). The working platform 8 is located directly below the cross 3. Four vertical columns 5 are respectively arranged on the top of the working platform 8. The tops of the four columns 5 are respectively connected to the four ends of the cross 3 through leveling components 4. A height-adjustable camera 7 is installed on each column 5. Specifically, a vertically arranged waist-shaped hole 6 can be opened on the column 5, and the camera 7 is fixed in the waist-shaped hole 6 through bolts, realizing the adjustment of the height of the camera 7, or adopting a structure of a linear module + a stepping motor to automatically adjust the height of the camera 7, realizing the autofocus and angle adjustment of the camera 7. Multiple hydraulic leveling support rods 9 are arranged at the bottom of the working platform 8 to realize the horizontal angle adjustment of the working platform 8.

[0025] The hydraulic system can provide a large force and precise position control. The hydraulic leveling support rods 9 can independently adjust the height of each support point to level the working platform 8. Workpieces with different surface curvatures may require adjusting the level of the working platform 8 at different positions to ensure the accuracy of projection and capture. When the curvature of the workpiece changes greatly, it is necessary to adjust the angle of the working platform 8 so that the laser projection grid can better cover the curved surface, avoiding projection distortion or inaccurate capture by the camera 7 caused by the inclination of the working platform 8. For large-sized workpieces, the camera 7 needs different installation heights to capture the distorted grids on the entire surface. The slide rail design of the waist-shaped hole 6 enables flexible adjustment of the position of the camera 7, and it can move up and down without disassembly, adapting to convex curved workpieces of different size specifications, improving the versatility and adaptability of the system.

[0026] Embodiment 2 This embodiment provides a laser processing planning method for convex curved workpieces based on grid distortion, which is implemented using the device in Embodiment 1, and specifically includes the following steps: (1) As Figure 2 shown, the length of the convex curved workpiece is 950 mm, and the height is 370 mm. The convex curved workpiece is placed below the laser grid projector 1 with a wavelength of 650 nm, and a laser grid pattern is projected onto the surface of the convex curved workpiece. The grid spacing is 100 mm. Adjust the telescopic component 2 until the laser grid emitted by the laser grid projector 1 completely covers the surface of the convex curved workpiece.

[0027] (2) Taking the beam emission direction of the laser grid projector 1 as the Z-axis direction of the spatial coordinate system, four cameras 7 are respectively arranged in the positive and negative directions of the X-axis and the positive and negative directions of the Y-axis to establish a workpiece coordinate system, and the optical axes of the four cameras 7 are all perpendicular to the Z-axis direction. The two cameras 7 located in the positive and negative directions of the X-axis are on the same central axis, and the two cameras 7 located in the positive and negative directions of the Y-axis are on the same central axis. The heights of the four cameras 7 are adjustable and kept consistent, and four distorted grid images are respectively captured by the four cameras 7.

[0028] (3) Denoise and perform lens distortion compensation on the captured distorted grid images, and extract the grid node information on the workpiece surface through the Harris corner detection algorithm.

[0029] When using a 20-megapixel black-and-white industrial camera 7 to capture images, sensor noise, ambient light interference, or electronic noise may cause random noise points in the images. To remove these noise points, Gaussian filtering is used to convolve the images, suppressing high-frequency noise and retaining edge information. This method can improve the accuracy of subsequent algorithms, reduce noise interference, and avoid misjudging noise as intersections or edges in subsequent edge detection. At the same time, it can improve the efficiency of subsequent algorithms, make the preprocessed images cleaner, and reduce the complexity of subsequent algorithms.

[0030] The Harris corner detection algorithm, through non-maximum suppression, only retains the pixels with the largest amplitude in the gradient direction, refining the edges to a single-pixel width. Through double-threshold detection, a high threshold and a low threshold are set. Strong edges are retained, and weak edges (between the two thresholds) are only retained when connecting strong edges. The time complexity of this algorithm is low, which can improve the efficiency of grid point extraction.

[0031] (4) Record the grid point coordinates of each grid in the positive and negative directions of the X-axis, the positive and negative directions of the Y-axis, and its coordinates can be expressed as , where is the height at which the laser intersects the curved surface.

[0032] (5) The local gradient of the curved surface represents the height change rate (i.e., slope) of the curved surface along the X and Y directions at the grid points of the grid, and the calculation formula is as follows: ; ; ; Among them, represents the slope of the curved surface along the X direction at the point (a positive value indicates an upward slope, and a negative value indicates a downward slope), represents the slope of the curved surface along the Y direction at the point (a positive value indicates an upward slope, and a negative value indicates a downward slope).

[0033] (6) Calculate the normal vector N of the surface, which can be calculated by the cross product of the gradient vectors: ; ; ; ; .

[0034] (7) To obtain the unit normal vector, normalization is required: .

[0035] (8) Sort the normal vector deviations between all adjacent meshes, and iteratively judge the adjacent mesh with the smallest current normal vector deviation by setting the normal vector deviation threshold δ. If it does not exceed δ, merge the meshes into the machining blocks α, β, γ..., and store the grid point coordinates and normal vectors of all machining blocks α, β, γ... respectively. The specific process of this step is shown in Figure 8 .

[0036] In the above steps (4) to (8), by using a calculation program written in Python, inputting the mesh node information, obtaining the normal vector, and merging the meshes through the region growing algorithm.

[0037] The following is part of the Python code: def region_growing_clustering(mesh_faces, face_normals, threshold_angle_deg): num_faces = len(mesh_faces) threshold_cos = np.cos(np.radians(threshold_angle_deg)) # Convert the angle to cosine value # Initialize the unvisited faces unvisited = set(range(num_faces)) clusters = face_adjacency = build_face_adjacency(mesh_faces) # Need to pre-build the face adjacency relationship while unvisited: # Randomly select a seed face seed = unvisited.pop() cluster = [seed] queue = deque([seed]) while queue: current_face = queue.popleft() # Traverse the adjacent faces of the current face for neighbor in face_adjacency[current_face]: if neighbor in unvisited: # Calculate the cosine of the normal vector angle cos_angle = np.dot(face_normals[current_face], face_normals[neighbor]) # If the similarity meets the threshold, merge if cos_angle >= threshold_cos: unvisited.remove(neighbor) cluster.append(neighbor) queue.append(neighbor) clusters.append(cluster) return clusters

[0038] (9)Uniformly sample the grid lines in the surface to generate discrete laser processing path points, and generate the laser processing path by combining the average normal vectors after block division. As Figure 3 shown, between blocks, the change in the processing attitude is smoothed by the B-spline curve fitting algorithm. First, chord length parameterization is performed, and then the accumulated chord length method is used to generate a uniform knot vector. After substituting the processing point set into the overdetermined system of equations, the control points are solved.

[0039] ; ; ; ; Among them, Q is the matrix of shape value points (n×3), N is the matrix of basis functions (n×m, element N i,p (u k )), and P is the matrix of control points (m×3).

[0040] Through the B-spline curve fitting algorithm, the machining path curve can be effectively smoothed to avoid the jitter of the laser head.

[0041] In order to verify the advantages of the method of the present invention in the actual machining process, a comparative analysis was carried out through experiments with the traditional method in terms of the calculation efficiency of the normal vector. The present invention only needs to collect about 75 grid nodes for calculation (as Figures 4 - 6 shown), and the traditional dense point cloud method needs to collect about 12,000 point clouds (as Figure 7 shown), and based on the PCA (Principal Component Analysis) method, the local neighborhood is fitted, which takes a long time and has a high data redundancy. In contrast, the present invention uses the laser grid projection and the industrial camera 7 to jointly collect the distorted images, extracts the key corner points through image processing to construct a sparse grid, the number of nodes is significantly reduced, and the normal vector is directly solved according to the grid topological relationship, avoiding the neighborhood search for each point and the calculation of the covariance matrix.

[0042] Through experimental comparison, in the same hardware environment (Intel i7 CPU, single-threaded execution), the traditional point cloud method takes about 6.8 seconds to calculate 12,000 point clouds and synthesize the normal vectors of the triangular patches in the region, while the present invention only needs 2.3 seconds under the sparse nodes, and the calculation efficiency is improved by about 66.2%, significantly shortening the preprocessing time, which is especially suitable for the requirements of online machining planning.

[0043] The laser machining planning method for convex surface workpieces based on grid distortion of the present invention has the following advantages: (1) A more efficient normal vector calculation method The prior art with the publication number of CN115761137A adopts a fusion method of point cloud data and normal vector regression. Although it can generate a high-precision surface, due to the involvement of neural network calculations and images in multiple illumination directions, the calculation efficiency is low and it is not suitable for time-sensitive scenarios. The present invention synchronously collects the distorted grid images through four industrial cameras 7, combines edge detection to extract grid nodes, and directly calculates the surface normal vector without point cloud data collection and neural network regression, improving the normal vector calculation efficiency.

[0044] (2) Low calculation amount and low hardware requirements, reducing the usage cost The prior art with the publication number CN115761137A uses point cloud data fusion neural network for surface reconstruction, which has a high computational complexity and requires large memory and high computing power devices, and is not suitable for resource-constrained environments. CN201410647810.6 relies on binocular vision matching and involves precise camera 7 calibration and stereo matching calculations, with relatively high computational resource requirements. The present invention only uses four industrial cameras 7 and edge detection algorithms to extract grid nodes, without the need for large-scale point cloud data storage, reducing the memory requirement. By using simple mathematical operations such as gradient calculation and normal vector calculation, it avoids deep learning and complex data fusion, has a low computational volume, and can run on ordinary industrial computers or embedded devices.

[0045] (3) Optimize the laser processing efficiency through adaptive grid merging The prior art with the publication number CN115761137A mainly focuses on surface reconstruction and does not use normal vectors for optimizing processing efficiency. The prior art with the publication number CN104408772A only realizes grid projection-assisted 3D reconstruction and also does not optimize processing efficiency. The present invention proposes normal vector deviation sorting and iterative calculation, performs adaptive grid merging on regions with small local curvature changes, and generates processing paths through grid line sampling. On the one hand, it realizes the integration from measurement to processing planning, reducing complex data processing steps. On the other hand, by reducing the frequent attitude adjustment of the processing head, it reduces the complexity of the processing path and improves the processing efficiency.

[0046] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A laser processing planning method for convex surface workpieces based on grid distortion, characterized in that, Including: Place the convex surface workpiece under the laser grid projector and project a laser grid onto the surface of the convex surface workpiece; Take the beam emission direction of the laser grid projector as the Z-axis direction of the space coordinate system. Four cameras are respectively arranged in the positive and negative directions of the X-axis and the positive and negative directions of the Y-axis, and the optical axes of the four cameras are all perpendicular to the Z-axis direction. Four distorted grid images are captured by the four cameras respectively; Extract the grid node information on the workpiece surface from the captured distorted grid images, record the lattice point coordinates of each grid, calculate the local gradient of the surface, and calculate the normal vector of the surface by the gradient cross product method; By setting the normal vector deviation threshold δ, sort and iteratively calculate the deviation of the adjacent grid normal vectors, automatically identify the areas with relatively small local curvature changes, and perform grid merging to generate block processing areas; Sample along the grid lines in the surface to generate laser processing path points, and generate a laser processing path in combination with the average normal vector after segmentation; 2. The method for laser processing planning of convex surface workpieces based on grid distortion according to claim 1, wherein, Denoise and perform lens distortion compensation on the captured distorted grid images, and extract the grid node information on the workpiece surface through an edge detection algorithm; 3. The laser processing planning method for convex surface workpieces based on grid distortion according to claim 2, characterized in that, Local Gradient of the Surface It represents the height change rate of the surface along the X and Y directions at the grid points of the mesh, and the calculation formula is as follows: ； ； ； in, Represents the surface at point The slope along the X direction, Represents the surface at point The slope along the Y direction, The coordinates are , is the height at which the laser intersects the surface.

4. The laser processing planning method for convex surface workpieces based on grid distortion according to claim 3, characterized in that The calculation formula of the normal vector N of the surface is as follows: ； ； ； ； ； Then perform normalization to obtain 。 5. The laser processing planning method for convex surface workpieces based on grid distortion according to claim 1, characterized in that Sort the normal vector deviations between all adjacent grids, and iteratively judge the adjacent grid with the smallest current normal vector deviation by setting the normal vector deviation threshold δ. If it does not exceed δ, merge the grids to generate block processing areas, and store the lattice point coordinates and normal vectors of all points in each block processing area respectively; 6. The laser processing planning method for convex surface workpieces based on grid distortion according to claim 1, characterized in that For the generated laser processing path, between blocks, smooth the change of the processing posture through the B-spline curve fitting algorithm; 7. The laser processing planning method for convex surface workpieces based on grid distortion according to claim 1, characterized in that, Adjust the height of the laser grid projector until the laser grid emitted by the laser grid projector completely covers the surface of the convex surface workpiece; 8. The laser processing planning method for convex surface workpieces based on grid distortion according to claim 1, characterized in that, The two cameras located in the positive and negative directions of the X-axis are on the same central axis, the two cameras located in the positive and negative directions of the Y-axis are on the same central axis, and the heights of the four cameras are adjustable and kept consistent; 9. A laser processing planning device for convex surface workpieces based on grid distortion, characterized in that, Used to implement the laser processing planning method for convex surface workpieces described in any one of claims 1 to 8, including a laser grid projector, a cross, columns, a work platform and four cameras. A telescopic component that can move along the Z-axis direction is installed in the middle position of the cross. The driving end of the telescopic component is installed with a laser grid projector. The work platform is located directly below the cross. Four vertical columns are respectively arranged on the top of the work platform. The tops of the four columns are respectively connected to the four ends of the cross in a butting manner. One camera is installed on each column; 10. The laser processing planning device for convex surface workpieces based on grid distortion according to claim 9, characterized in that, The installation height of the camera on the column is adjustable. Four feet are arranged at the bottom of the work platform, and the height of the feet is adjustable. A leveling component is arranged between the column and the cross end bracket.

Images