Sand blasting robot path automatic planning method based on line laser scanning

Abstract

The invention discloses a sand blasting robot path automatic planning method based on line laser scanning, and relates to the field of material surface machining. The method comprises the following steps: acquiring surface point cloud data of a target workpiece uploaded by a line laser scanner; gridding processing is conducted on the surface point cloud data, a gridding model of the target workpiece is generated, the gridding model is composed of a plurality of gridding areas, and each gridding area comprises a plurality of curved surface points and stores curvature values; according to the thermodynamic parameters of the target workpiece, curvature sensitivity coefficients of the multiple grid areas are calculated; and generating a sand blasting path of the sand blasting robot based on the curvature sensitivity coefficients of the plurality of grid areas. By implementing the technical scheme provided by the invention, the problem of over-spraying or under-spraying in the prior art is solved.

Description

Technical Field

[0001] This application relates to the technical field of material surface processing, specifically to an automatic path planning method for a sandblasting robot based on line laser scanning. Background Technology

[0002] With the rapid improvement of my country's industrial manufacturing level, sandblasting technology is widely used in the surface treatment of metal workpieces, which is of great significance for ensuring workpiece performance and improving product quality; however, in actual industrial scenarios, especially in high-pressure sandblasting operations, complex problems are gradually emerging.

[0003] In traditional sandblasting operations, the planning of the sandblasting path is usually based on a static workpiece model. However, in actual industrial scenarios, the surface temperature of the workpiece (aluminum alloy, carbon steel, etc.) will rise instantly under the impact of the abrasive in high-pressure sandblasting, resulting in thermal deformation and changes in the actual contour of the workpiece. Existing technologies lack the ability to compensate for the path of dynamic deformation of the workpiece, causing the sandblasting trajectory to deviate from the target area, thus resulting in over-blasting or under-blasting. Summary of the Invention

[0004] To address the over- or under-spraying problems in existing technologies, this application provides an automatic path planning method for sandblasting robots based on line laser scanning.

[0005] This application provides an automatic path planning method for a sandblasting robot based on line laser scanning, the method comprising:

[0006] Acquire the surface point cloud data of the target workpiece uploaded by a line laser scanner;

[0007] The surface point cloud data is processed into a grid to generate a grid model of the target workpiece. The grid model consists of multiple grid regions, each of which contains multiple surface points and stores curvature values.

[0008] Based on the thermodynamic parameters of the target workpiece, the curvature sensitivity coefficients of multiple mesh regions are calculated;

[0009] The sandblasting path of the sandblasting robot is generated based on the curvature sensitivity coefficients of multiple grid regions.

[0010] Optionally, before acquiring the surface point cloud data of the target workpiece uploaded by the line laser scanner, the following steps are included:

[0011] Obtain the coarse sweep data of the target workpiece;

[0012] Based on the coarse scan data, a low-resolution surface model is constructed.

[0013] Identify multiple feature regions of the low-resolution surface, including edges, holes, and uneven surfaces;

[0014] Calculate the area corresponding to each of the multiple feature regions, and match the scanning speed of the multiple feature regions from the scanning speed reference table;

[0015] Based on the scanning speed of the multiple feature regions, and by scanning the multiple feature regions in ascending order of region area, detailed scan data of the multiple feature regions are obtained;

[0016] The coarse sweep data and the fine sweep data are used to construct the surface point cloud data of the target workpiece.

[0017] Optionally, the step of performing meshing processing on the surface point cloud data to generate a meshed model of the target workpiece specifically includes:

[0018] Calculate the centroid of the first grid region based on multiple surface points within the first grid region, where the first grid region is any one of the multiple grid regions;

[0019] Calculate the point distances between multiple surface points within the first grid region and the centroid of the first grid region;

[0020] Anomaly detection is performed on the point distances of multiple curved surface points within the first grid region to obtain the distances of multiple valid points within the first grid region;

[0021] Based on the distances between multiple effective points within the first grid region, the influence weights of multiple curved surfaces within the first grid region are determined.

[0022] The curvature value of the first grid region is adjusted by using multiple surface influence weights of the first grid region, and the adjusted curvature value is stored in the first grid region.

[0023] Optionally, the step of performing anomaly detection on the point distances of multiple surface points within the first grid region to obtain multiple valid point distances within the first grid region specifically includes:

[0024] Calculate the local density deviation of multiple surface points within the first grid region;

[0025] Based on a preset local density deviation threshold, multiple candidate abnormal surface points are selected from multiple surface points within the first grid region.

[0026] Calculate the overall surface deviation between the normal vectors of multiple candidate abnormal surface points and the normal vectors of the first grid region, and calculate the local surface deviation between the curvature of multiple candidate abnormal surface points and the mean curvature of their respective neighborhoods.

[0027] Based on the preset overall deviation threshold and local deviation threshold of the surface, the multiple candidate abnormal surface points are divided into multiple key surface points and multiple noisy surface points.

[0028] The point distances corresponding to the key surface points are determined as the effective point distances within the first grid area.

[0029] Optionally, determining the point distance corresponding to the key surface point as the effective point distance within the first grid region further includes:

[0030] Surface reconstruction is performed on the neighborhood surfaces corresponding to each of the multiple key surface points to obtain the reconstructed neighborhood surfaces corresponding to each of the multiple key surface points;

[0031] Extract the reconstructed key surface points from the reconstructed neighborhood surface corresponding to each of the multiple key surface points;

[0032] The multiple key surface points are weighted and fused with their respective corresponding reconstructed key surface points to obtain the fused surface points corresponding to the multiple key surface points;

[0033] The effective point distance is defined as the distance between the fused surface point corresponding to the multiple key surface points and the centroid point of the first grid region.

[0034] Optionally, calculating the curvature sensitivity coefficients of multiple mesh regions based on the thermodynamic parameters of the target workpiece specifically includes:

[0035] Obtain the sandblasting process parameters of the target workpiece and calculate the heat input of the second grid region, wherein the second grid region is any one of the multiple grid regions;

[0036] Based on the thermodynamic parameters of the target workpiece, the temperature rise of the second grid region is solved using the transient heat conduction equation;

[0037] The thermal deformation of the second grid region is calculated based on the curvature value and thermal expansion coefficient of the second grid region.

[0038] The curvature sensitivity coefficient of the second grid region is determined based on the amount of thermal deformation in the second grid region.

[0039] Optionally, generating the sandblasting path of the sandblasting robot based on the curvature sensitivity coefficients of multiple grid regions specifically involves:

[0040] a. Select the maximum curvature sensitivity coefficient from the curvature sensitivity coefficients of the multiple grid regions;

[0041] b. Use the grid area corresponding to the maximum curvature sensitivity coefficient as the starting point of the sandblasting path;

[0042] c. Remove the grid region corresponding to the maximum curvature sensitivity coefficient from the plurality of grid regions;

[0043] d. Repeat steps a to c until the sandblasting path covers multiple of the grid areas, and output the sandblasting path.

[0044] Optionally, generating the sandblasting path of the sandblasting robot based on the curvature sensitivity coefficients of multiple grid regions further includes:

[0045] Obtain the standard sandblasting speed of the target workpiece;

[0046] The sandblasting speed of the third grid region is calculated based on the standard sandblasting speed of the target workpiece, the curvature sensitivity coefficient of the third grid region, and the curvature change rate of the third grid region.

[0047] Optionally, the step of calculating the sandblasting speed of the third grid region based on the standard sandblasting speed of the target workpiece, the curvature sensitivity coefficient of the third grid region, and the curvature change rate of the third grid region specifically involves:

[0048]

[0049] Where v is the sandblasting speed, Where k is the standard sandblasting speed, and k is the rate of change of curvature. This is the curvature sensitivity coefficient.

[0050] In summary, one or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0051] This application performs meshing of the point cloud data of the target workpiece surface, storing curvature values ​​in mesh regions and matching them with the effective coverage area of ​​the nozzle. Then, based on the thermodynamic parameters of the target workpiece, it calculates the curvature sensitivity coefficient of each mesh region. This coefficient is related to the thermodynamic characteristics of the workpiece and the constraints of the sandblasting process, and can reflect the degree of influence of thermodynamic factors on the sandblasting effect in different curvature regions. It should be noted that the larger the curvature sensitivity coefficient, the more significantly the mesh region is affected by thermodynamic parameters, and the sandblasting path needs to be adjusted accordingly. For example, edge regions with high curvature sensitivity are more prone to sandblasting deviations due to thermal deformation. Low-curvature sensitive planar areas are relatively stable. Under complex working conditions, if only conventional path planning is followed, over-blasting or under-blasting may occur due to differences in thermodynamic response in different areas. However, by utilizing the curvature sensitivity coefficient, the sandblasting path can be adapted to the characteristics of each area, with precise compensation for highly sensitive areas and reasonable allocation of process resources for low-sensitive areas. Therefore, generating sandblasting paths based on the curvature sensitivity coefficient can improve sandblasting uniformity and accuracy, reduce quality problems caused by workpiece deformation and poor process adaptability, provide reliable algorithm support for efficient and high-quality sandblasting operations, and adapt to the sandblasting needs of diverse and complex workpieces. Attached Figure Description

[0052] Figure 1 This is a flowchart illustrating an automatic path planning method for a sandblasting robot based on line laser scanning, provided in an embodiment of this application. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0054] For sandblasting of metal workpieces, traditional methods mostly rely on static workpiece models to determine the trajectory. However, during high-pressure sandblasting, the abrasive impacts the workpiece surface at high speed (typically 80-120 m / s), converting a large amount of kinetic energy into heat energy. This causes the local surface temperature of the workpiece to rise instantaneously (the temperature rise can reach 30-80℃). Due to the thermal expansion characteristics of metal materials such as aluminum alloys and carbon steel, the actual contour will undergo dynamic thermal deformation, resulting in a deviation from the theoretical state of the initial static workpiece model.

[0055] Currently, while the industry utilizes environmental monitoring to assist in the planning and control of sandblasting paths, it lacks effective mechanisms to address the dynamic deformation caused by high-pressure sandblasting. During actual sandblasting, because the workpiece deviates from the theoretical model, the sandblasting trajectory also deviates from the target area, leading to frequent over-blasting and under-blasting. This not only affects the surface roughness and smoothness of the workpiece but also shortens nozzle life due to ineffective sandblasting, increasing production costs and maintenance difficulty, thus hindering the application of sandblasting technology in high-precision manufacturing scenarios.

[0056] To address the aforementioned problems, this application provides an automatic path planning method for sandblasting robots based on line laser scanning, such as... Figure 1 As shown, the method includes steps S101 to S104, which are as follows:

[0057] S101. Acquire the surface point cloud data of the target workpiece uploaded by the line laser scanner.

[0058] In the above steps, most of the target workpieces that need to be sandblasted are non-standard workpieces with irregular surface contours. At this time, when the line laser scanner collects the surface point cloud data of the target workpiece, point cloud loss often occurs due to the excessive scanning speed, which requires multiple rescans and reduces modeling efficiency. Therefore, before acquiring the surface point cloud data of the target workpiece, this application first obtains the coarse scan data of the target workpiece. The coarse scan data is the workpiece surface data obtained by a line laser scanner scanning according to a preset scanning path. Then, a low-resolution surface model is constructed based on the coarse scan data. Next, multiple feature regions in the low-resolution surface model are identified. Feature regions are regions with curvature greater than or equal to the curvature threshold (e.g., edges, holes, uneven surfaces, etc.). Then, the area of ​​multiple feature regions is calculated. Finally, based on the area of ​​multiple feature regions, the scanning speed of multiple region areas is matched from the scanning speed reference table. It should be noted that for high curvature regions, the smaller the area, the more drastic the contour change, and the more likely the point cloud is to be missing, and the lower the scanning speed. Finally, based on the scanning speed of multiple feature regions, multiple feature regions are scanned in ascending order of area to obtain the fine scan data of multiple feature regions. Finally, the coarse scan data and the fine scan data are combined to construct the surface point cloud data of the target workpiece, thereby ensuring modeling efficiency while improving modeling accuracy.

[0059] S102. Perform meshing processing on the surface point cloud data to generate a meshed model of the target workpiece. The meshed model consists of multiple mesh regions, and each mesh region stores curvature values.

[0060] In the above steps, the surface point cloud data is meshed according to the effective coverage area of ​​the nozzle of the sandblasting robot, resulting in multiple mesh regions of equal area. The area of ​​each mesh region is equal to the effective coverage area of ​​the nozzle, thus matching the sandblasting scenario of the nozzle. Then, the curvature value of each mesh region is calculated and stored in each mesh region. Specifically, the curvature value of the mesh region can be calculated using a surface curvature calculation method based on the neighborhood normal vector, as follows:

[0061]

[0062] Where k is the surface curvature, and n is the normal vector of the surface centroid. Let p be the normal vector of the i-th point on the surface, and let p be the coordinates of the centroid of the surface. Let S be the coordinates of the i-th surface point in the surface, and S be the total number of surface points in the surface.

[0063] In the above formula, the contribution of each surface point to the overall curvature of the surface is evaluated by judging the difference between the normal vector of the average position of the surface center (centroid point) and the normal vector of each surface point.

[0064] In one possible implementation, since the curvature influence region of the surface is closer to the centroid region, when the distances from each surface point to the centroid are different, the surface point closer to the centroid can better reflect the overall curvature of the surface. Therefore, it is also necessary to consider the influence of the distances between each surface point and the centroid on the overall curvature of the surface. Specifically, this application first calculates the point distances between each surface point and the centroid, then performs anomaly judgment on the point distances of each surface point, removes abnormal surface points, and then converts the remaining surface points into surface influence weights, specifically: ,in Let i be the surface influence weight of the i-th surface point on the surface. Let be the distance from the i-th point on the surface to the centroid. A larger distance results in a smaller surface influence weight. Then, the surface influence weights of multiple surface points are normalized. Finally, based on the surface influence weights of each surface point, the contribution of each surface point to the overall curvature of the surface is adjusted. Specifically:

[0065]

[0066] in, Let be the surface influence weight of the i-th surface point on the surface.

[0067] In one possible implementation, during the process of anomaly detection and removal of abnormal surface points based on the distances between various surface points, some surface points may have low confidence levels due to factors such as optical reflection and occlusion. However, if these surface points contribute significantly to the curvature of the surface, they cannot be directly removed. In this case, to ensure the accuracy of the surface curvature assessment, this application first employs a local outlier factor algorithm to calculate the local density deviation of each surface point. Then, based on the local density deviation threshold, multiple candidate abnormal surface points are selected, i.e., surface points whose local density deviation is greater than the local density deviation threshold. If a point is identified as a candidate anomalous surface point, then the overall deviation of the surface normal vector from the mean of the surface normal vectors of each candidate anomalous surface point is calculated. Additionally, the local deviation of the curvature of each candidate anomalous surface point from the mean curvature of its corresponding neighborhood is also calculated. If the overall deviation of the candidate anomalous surface point is less than the overall deviation threshold, and the local deviation is greater than the local deviation threshold, then the candidate anomalous surface point contributes significantly to the overall curvature of the surface but contributes little to the local curvature. In this case, it can be identified as a critical surface point. Otherwise, it is identified as a noisy surface point, which can be removed.

[0068] After selecting multiple key surface points, a local reconstruction method is used to reconstruct the neighborhood surfaces of the key surface points. Specifically, with the key surface point as the center, multiple confidence neighborhoods are divided, where the outermost layer has the lowest confidence and the innermost layer has the highest confidence. Then, based on the confidence corresponding to each neighborhood, the least squares method is used to fit the best surface layer by layer from the outermost to the innermost layer, resulting in the final reconstructed neighborhood surface. For example, if the key surface point has three confidence neighborhoods, with the first layer being the innermost and the third layer being the outermost, the second and third layers of surfaces are fitted first. Then, the fitted surfaces of the second and third layers are fitted with the first layer to obtain the reconstructed neighborhood surface. Next, the point information of the key surface points in the reconstructed neighborhood surface is identified. The point information includes the normal vector, curvature, and point coordinates. Then, the point information of the original key surface points and the point information of the new key surface points are weighted and fused to obtain the fused surface points. Finally, the original key surface points are replaced with the fused surface points.

[0069] S103. Calculate the curvature sensitivity coefficients of multiple grid regions based on the thermodynamic parameters of the target workpiece.

[0070] In the above steps, the curvature sensitivity coefficient can be understood as the magnitude of curvature change in the grid region under the influence of thermal effects. The larger the curvature sensitivity coefficient, the greater the influence of thermal effects on the workpiece material in the grid region, the greater the magnitude of curvature change, and the greater the material deformation. For the target workpiece, the processing material is generally consistent, and the thermodynamic parameters of the target workpiece in each grid region are also consistent, including thermal conductivity, specific heat capacity, coefficient of thermal expansion, and density. However, due to the different curvatures of each grid region, the stress distribution is different. In grid regions with high curvature, the stress is more concentrated, and the curvature sensitivity coefficient is larger. Therefore, this application calculates the curvature sensitivity coefficients of multiple grid regions of the target workpiece to determine the deformation magnitude of each grid region during sandblasting, providing a basis for the rational planning of the subsequent sandblasting path. Specifically:

[0071] First, based on the sandblasting process parameters, the heat input of a single grid region is calculated using a thermodynamic model, i.e.:

[0072]

[0073] Where Q is the heat input, denoted as the heat conversion coefficient, P as the nozzle air pressure, v as the sand particle velocity, t as the residence time, and A as the area of ​​a single grid region.

[0074] In the above formula, the heat input of a single grid region mainly comes from the conversion of the kinetic energy of sand particles when they impact the workpiece surface. The conversion efficiency is represented by the heat conversion coefficient. The sand particle velocity multiplied by the residence time can be understood as the impact flow rate of the sand particles within the residence time t. The nozzle air pressure can be understood as the force of the sand particles when they impact the workpiece. Finally, multiplying by the impact area gives the heat input of a single grid region.

[0075] Then, based on the thermodynamic parameters of the target workpiece, the temperature rise in the grid region is calculated using the transient heat conduction equation, specifically using the following calculation method:

[0076]

[0077] in, , denoted as density, specific heat capacity, and thermal conductivity of the target workpiece, respectively; T is the temperature of the target workpiece; t is the residence time; and Q is the unit heat flux density of the grid region.

[0078] In the above formula, the left side of the equation represents the change in the internal energy of the material per unit time, the first term on the right side is the heat conduction term, and the second term is the heat source term. This equation describes the temperature rise of the grid region caused by the amount of heat input from sand particles to the grid region.

[0079] Then, based on the curvature and thermal expansion coefficient of the mesh region, the thermal deformation of the mesh region is calculated, specifically:

[0080]

[0081] in, For the thermal deformation of the grid region, The coefficient of thermal expansion is... The curvature value of the grid region. This represents the temperature change within the grid region.

[0082] Finally, the thermal deformation of multiple grid regions is normalized to obtain the curvature sensitivity coefficient of each grid region.

[0083] S104. Based on the curvature sensitivity coefficients of multiple grid regions, generate the sandblasting path of the sandblasting robot.

[0084] In the above steps, when sandblasting a certain grid area, due to the conductivity of heat, neighboring grid areas will also absorb some heat, resulting in an actual temperature rise higher than the ideal temperature rise. If the curvature sensitivity coefficient of the neighboring grid areas is large, their deformation will increase accordingly. Therefore, when planning the sandblasting path of the sandblasting robot, this application first identifies the curvature sensitivity coefficients of multiple grid areas, then selects the grid area with the largest curvature sensitivity coefficient from among them, and uses the grid area corresponding to this curvature sensitivity coefficient as the starting point of the sandblasting path. Then, it iterates through multiple curvature sensitivity coefficients of the neighboring grid areas of the initial grid area, and selects the grid area corresponding to the largest curvature sensitivity coefficient from among the multiple curvature sensitivity coefficients of the neighboring grid areas of the initial grid area as the second sandblasting area of ​​the sandblasting path, and so on, until the sandblasting path completely covers all grid areas. This method prioritizes sandblasting grid areas with high curvature sensitivity coefficients, thereby reducing the thermal deformation impact caused by sandblasting of neighboring grid areas, thus improving sandblasting uniformity and accuracy, and reducing quality problems caused by workpiece deformation and poor process adaptability.

[0085] In one possible implementation, for high-curvature grid areas, due to their uneven surface structure, under-blasting is very likely to occur. In order to accurately compensate for this part of the grid area, this application dynamically adjusts the sandblasting speed of each area according to the curvature change rate of each grid area, specifically as follows:

[0086]

[0087] Where v is the sandblasting speed, Where k is the standard sandblasting speed, and k is the rate of change of curvature. This is the curvature sensitivity coefficient.

[0088] In the above formula, if the rate of change of curvature is large, it means that the surface shape of the grid area has changed significantly. This can easily lead to deviation of the sandblasting trajectory. Therefore, it is necessary to slow down the sandblasting speed to ensure the sandblasting precision. The curvature sensitivity coefficient can be understood as a constraint factor. For grid areas with low curvature sensitivity coefficient and high rate of change of curvature, it means that the deformation of the grid area is not easily affected. In this case, a longer spraying time may lead to over-spraying. Therefore, the sandblasting speed is constrained and adjusted according to the curvature sensitivity coefficient of the grid area to further reduce the possibility of under-spraying or over-spraying.

[0089] The above description is merely an exemplary embodiment of this disclosure and should not be construed as limiting the scope of this disclosure. Any equivalent changes and modifications made in accordance with the teachings of this disclosure shall still fall within the scope of this disclosure. Other embodiments of this disclosure will be readily apparent to those skilled in the art upon consideration of the specification and the disclosure of practical truths.

[0090] This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not described in this disclosure. The specification and embodiments are to be considered exemplary only, and the scope and spirit of this disclosure are defined by the claims.

Claims

1. A linear laser scanning-based automatic planning method for a sandblasting robot path, characterized in that, The method comprises: acquiring surface point cloud data of a target workpiece uploaded by a line laser scanner; performing meshing processing on the surface point cloud data to generate a meshing model of the target workpiece, the meshing model being composed of a plurality of meshing areas, each meshing area containing a plurality of curved surface points and storing a curvature value; calculating curvature sensitivity coefficients of the plurality of meshing areas according to thermodynamic parameters of the target workpiece; generating a sandblasting path of a sandblasting robot based on the curvature sensitivity coefficients of the plurality of meshing areas.

2. The method of claim 1, wherein, Before the acquiring of the surface point cloud data of the target workpiece uploaded by the line laser scanner, the method comprises: acquiring coarse scanning data of the target workpiece; constructing a low-resolution surface model according to the coarse scanning data; identifying a plurality of feature areas of the low-resolution surface, the plurality of feature areas including edges, holes and concave-convex surfaces; calculating respective area sizes of the plurality of feature areas and matching scanning speeds of the plurality of feature areas from a scanning speed reference table; scanning the plurality of feature areas according to the scanning speeds of the plurality of feature areas and in the order of the area sizes from small to large to obtain fine scanning data of the plurality of feature areas; constructing the coarse scanning data and the fine scanning data into the surface point cloud data of the target workpiece.

3. The method of claim 1, wherein, The meshing processing on the surface point cloud data to generate the meshing model of the target workpiece specifically comprises: calculating a centroid point of a first meshing area according to a plurality of curved surface points in the first meshing area, the first meshing area being any one of the plurality of meshing areas; calculating point distances between the plurality of curved surface points in the first meshing area and the centroid point of the first meshing area; performing abnormality judgment on the point distances of the plurality of curved surface points in the first meshing area to obtain a plurality of effective point distances in the first meshing area; determining a plurality of curved surface influence weights of the first meshing area based on the plurality of effective point distances in the first meshing area; adjusting the curvature value of the first meshing area by using the plurality of curved surface influence weights of the first meshing area, and storing the adjusted curvature value in the first meshing area.

4. The method of claim 3, wherein, The abnormality judgment on the point distances of the plurality of curved surface points in the first meshing area to obtain the plurality of effective point distances in the first meshing area specifically comprises: calculating local density deviation degrees of the plurality of curved surface points in the first meshing area; screening a plurality of candidate abnormal curved surface points from the plurality of curved surface points in the first meshing area according to a preset local density deviation degree threshold; calculating a curved surface overall deviation degree of normal vectors of the plurality of candidate abnormal curved surface points and a normal vector of the first meshing area, and calculating a local curved surface deviation degree of curvatures of the plurality of candidate abnormal curved surface points and average curvatures of respective corresponding neighborhoods; dividing the plurality of candidate abnormal curved surface points into a plurality of key curved surface points and a plurality of noise curved surface points according to a preset curved surface overall deviation degree threshold and a curved surface local deviation degree threshold; determining the point distances of the key curved surface points as the effective point distances in the first meshing area.

5. The method of claim 4, wherein, The effective point distance corresponding to the key curved surface point is determined as an effective point distance in the first grid region, and the method further includes: Carrying out surface reconstruction on the respective corresponding neighborhood curved surface of each of the plurality of key curved surface points to obtain a respective corresponding reconstructed neighborhood curved surface of each of the plurality of key curved surface points; Extracting a reconstructed key curved surface point in the respective corresponding reconstructed neighborhood curved surface of each of the plurality of key curved surface points; Weighted fusion of the plurality of key curved surface points and the respective corresponding reconstructed key curved surface points to obtain a plurality of fusion curved surface points corresponding to the plurality of key curved surface points; The point distance from the plurality of fusion curved surface points corresponding to the plurality of key curved surface points to the centroid point of the first grid region is taken as the effective point distance.

6. The method of claim 1, wherein, The curvature sensitivity coefficient of each of the plurality of grid regions is calculated according to the thermodynamic parameters of the target workpiece, and specifically includes: Obtaining the sandblasting process parameters of the target workpiece, calculating the heat input of a second grid region, and the second grid region is any one of the plurality of grid regions; Solving the temperature rise of the second grid region by using a transient heat conduction equation according to the thermodynamic parameters of the target workpiece; Based on the curvature value and the thermal expansion coefficient of the second grid region, the thermal deformation amount of the second grid region is calculated; Based on the thermal deformation amount of the second grid region, the curvature sensitivity coefficient of the second grid region is determined.

7. The method of claim 1, wherein, The sandblasting path of the sandblasting robot is generated based on the curvature sensitivity coefficients of the plurality of grid regions, and specifically includes: a. Selecting a maximum curvature sensitivity coefficient from the curvature sensitivity coefficients of the plurality of grid regions; b. Taking the grid region corresponding to the maximum curvature sensitivity coefficient as the starting point of the sandblasting path; c. Excluding the grid region corresponding to the maximum curvature sensitivity coefficient from the plurality of grid regions; d. Repeating steps a to c until the sandblasting path covers the plurality of grid regions, and outputting the sandblasting path.

8. The method of claim 1, wherein, The sandblasting path of the sandblasting robot is generated based on the curvature sensitivity coefficients of the plurality of grid regions, and specifically further includes: Obtaining the standard sandblasting speed of the target workpiece; According to the standard sandblasting speed of the target workpiece, the curvature sensitivity coefficient of a third grid region, and the curvature change rate of the third grid region, the sandblasting speed of the third grid region is calculated.

9. The method of claim 8, wherein, The sandblasting speed of the third grid region is calculated according to the standard sandblasting speed of the target workpiece, the curvature sensitivity coefficient of the third grid region, and the curvature change rate of the third grid region, and specifically includes: where v is the sandblasting speed, is the standard sandblasting speed, k is the curvature change rate, is the curvature sensitivity coefficient.

Images