A mechanical arm automatic polishing method based on 3D model edge recognition

Abstract

This invention discloses an automated deburring method for robotic arms based on 3D model edge recognition in the field of intelligent manufacturing technology. The method includes: acquiring and loading a 3D model of the part to be deburred; determining the edge region to be deburred; performing edge feature quantization calculations based on the 3D model data to generate a suitable deburring trajectory and corresponding posture matrix; verifying the accuracy of the deburring trajectory, optimizing parameters and regenerating the trajectory if it fails; and sending the qualified trajectory and posture to the robotic arm controller to drive the automated deburring operation. This invention eliminates the need for industrial vision equipment for on-site image acquisition and edge recognition, avoiding interference from external factors such as lighting and surface contaminants, significantly improving the stability and reliability of edge deburring while reducing system deployment costs. It can adaptively adapt to edge features of different shapes and sharpness, is easy to operate, and is suitable for automated deburring of various mechanical parts.

Description

Technical Field

[0001] This invention relates to the field of intelligent manufacturing technology, and in particular to an automatic grinding method for robotic arms based on 3D model edge recognition. Background Technology

[0002] In machinery manufacturing, casting, sheet metal processing, and 3D product production, deburring and chamfering of workpiece edges are crucial processes to ensure assembly accuracy, surface quality, and operational safety. With the increasing level of industrial automation, more and more manufacturing companies are adopting robotic arm-based automated grinding solutions to replace traditional manual grinding, thereby improving processing efficiency and quality stability.

[0003] Existing automated grinding solutions mostly employ 3D vision scanning, point cloud reconstruction, or full-model automatic recognition to achieve edge detection and trajectory planning. For example, industrial cameras are used to collect images of the workpiece surface, and image recognition algorithms are used to extract the edges to be ground; 3D scanning is used to obtain the workpiece point cloud model, and edge features are automatically identified and grinding paths are generated; 3D vision is used to perform a full-size scan of the workpiece, and the area to be ground is automatically matched with a pre-stored standard model.

[0004] However, existing automated polishing solutions are not mature enough for practical applications and still have the following shortcomings: 1. High system cost: It requires high-precision industrial cameras, 3D scanning equipment, image processing units and other hardware. The deployment cost of the complete system is high, which is difficult for small and medium-sized manufacturing enterprises to afford. 2. Insufficient recognition stability: Relying on fully automatic visual recognition algorithms, it is easily affected by external factors such as ambient lighting conditions, workpiece clamping deviation, surface texture, reflection, and stains. This can easily lead to problems such as missed edge recognition, incorrect recognition, and distortion of contour extraction, resulting in deviation of the grinding trajectory and affecting the processing quality. 3. Weak adaptability: Existing solutions generate grinding trajectories and postures with fixed parameters, which cannot be adaptively adjusted according to the geometric features of the edge (such as the sharpness of the edges and changes in curvature). When faced with edges of different shapes, problems such as over-grinding, edge chipping, or grinding residue are likely to occur. The process parameters need to be manually adjusted for each piece, resulting in poor adaptability. 4. High operational complexity: Operators need to have professional skills such as visual calibration, point cloud processing, and algorithm parameter tuning. The operation threshold is high, which is not conducive to its promotion and application in traditional manufacturing enterprises. Summary of the Invention

[0005] To address the aforementioned shortcomings in existing technologies, this invention provides an automated edge polishing method using a robotic arm based on 3D model edge recognition. This method solves the problems of high system cost, poor stability, and weak adaptability caused by reliance on visual recognition in existing technologies. It achieves low-cost, high-stability, and highly adaptable automated edge polishing, eliminating the need for complex vision equipment and professional operating skills. Ordinary operators can quickly complete the polishing process configuration.

[0006] This application provides an automated grinding method for robotic arms based on 3D model edge recognition, comprising the following steps: S1: Obtain the 3D model of the part to be polished and load it into the industrial computer; S2: Receive the selection command input by the user through the interactive interface to determine the edge area to be polished in the 3D model; S3: Calculate the edge features of the area to be polished based on 3D model data, and generate the polishing trajectory and the corresponding polishing posture matrix; S4: Perform accuracy verification on the generated grinding trajectory. If the verification fails, adjust the trajectory calculation parameters and regenerate the grinding trajectory until the preset accuracy requirements are met. S5: Send the verified grinding trajectory and corresponding grinding posture matrix to the robotic arm controller, driving the robotic arm to drive the end-effector grinding tool to complete the grinding operation of the edge to be ground.

[0007] The beneficial effects of the above embodiments are as follows: Based on a standard 3D model of the workpiece to be processed, the present invention determines the edge to be polished through interactive selection, and adaptively generates the optimal polishing trajectory and posture by combining quantitative calculation of edge geometric features. After multi-dimensional accuracy verification, it drives a robotic arm to complete automated polishing. It eliminates the need for on-site image acquisition and edge recognition using industrial vision equipment, avoiding interference from external factors such as lighting and surface contaminants, significantly improving the stability and reliability of edge polishing while reducing system deployment costs. It can adaptively adapt to edge features of different shapes and sharpness, is easy to operate, and is particularly suitable for automated deburring of various mechanical parts.

[0008] Based on the above embodiments, this application can be further improved as follows: In one embodiment of this application, step S3, which involves calculating the edge features of the area to be polished, specifically includes: Obtain the basic parameter information of the 3D model's edge contour and adjacent surfaces. Calculate the edge angle, curvature, and radius of curvature for the edge to be ground, and determine the edge feature trajectory type. The calculation formula is as follows: Edge angle : ; Curvature of the edge and corresponding radius of curvature : ; ; In the formula These are the normal vectors of the two sides of the edge to be polished; and It refers to the first and second order guiding quantities of the space curve of the edge to be polished; Determine the type of edge feature trajectory: ; In the formula, These are the optimal trajectory types corresponding to the edge features; TE1: Straight-line trajectory along the edge, TE2: Segmented straight-line trajectory along the edge, TE3: Conformal curve trajectory, TE4: Variable curvature conformal trajectory; L is the edge length. A preset threshold is set for the length of straight edge segments. A preset threshold is set for the radius of curvature of the curved edge.

[0009] Technical effect: This claim realizes the quantitative classification of the geometric features of the edge to be polished. By accurately calculating parameters such as edge angle, curvature, and radius of curvature, the edges of different shapes are divided into four types of trajectory, which provides a quantitative basis for the adaptive matching of subsequent trajectory parameters and ensures that the edges of different shapes can obtain the optimal polishing trajectory scheme.

[0010] In one embodiment of this application, generating the polishing trajectory in step S3 specifically includes: Based on the trajectory type TE corresponding to the edge to be polished, determine the initial feature adaptation coefficient kf and the polishing accuracy coefficient kp; Calculate the feature adaptation step size S and the accuracy adaptation overlap rate η by combining the grinding process parameters: ; ; ; ; In the formula, Based on step size, To adjust the feed rate of the grinding head, To set the grinding head speed, It is the feed rate per revolution of the grinding head; Material coefficient; To adapt the step size to the feature, The base overlap ratio is given, and D is the diameter of the grinding head. Based on the overlap rate; Along the edge to be polished, discrete points are sampled according to the feature adaptation step size S to generate an ordered set of coordinate points for the polishing trajectory: ; In the formula, These are the coordinates of the sampling point; This is the starting point for grinding the edge to be polished. It is the sampling point number; To adapt the step size to the features already solved in the previous section; Let be the unit tangent vector of the edge to be polished.

[0011] Technical effect: This claim realizes differentiated generation of grinding trajectories by adaptively matching the calculation step size and overlap rate parameters based on edge trajectory type: a larger step size is used for straight edges to improve grinding efficiency, and a smaller step size is used for complex edges with high curvature to ensure grinding accuracy. At the same time, the overlap rate parameter controls the trajectory coverage to avoid missed grinding or over-grinding, thus balancing grinding efficiency and grinding quality.

[0012] In one embodiment of this application, generating the polishing posture matrix in step S3 specifically includes: Calculate the normal vector of the edge to be ground: ; Calculate the grinding posture rotation matrix: Calculate the pitch compensation angle: In the formula, It is the preset maximum process avoidance angle; Solve for the corrected actual pitch angle: In the formula, It is the reference pitch angle; Construct a local coordinate system basis for the edge: ; In the formula It is the X-axis basis of the local coordinate system; It is the Y-axis basis of the local coordinate system; It is the Z-axis basis of the local coordinate system; Construct the end-position rotation matrix of the coupling edge angle: ; In the formula , , It is the transpose of the three-axis basis vectors of the coordinate system; It is the actual pitch angle. The cosine and sine function values.

[0013] Technical effect: This claim introduces the edge angle as a constraint parameter for attitude adjustment, realizing adaptive dynamic adjustment of grinding attitude: for sharp edges, the avoidance angle is automatically increased to avoid edge chipping and over-grinding; for gentle edges, the avoidance angle is automatically decreased to improve grinding fit. This solves the problem that the traditional fixed attitude grinding scheme cannot adapt to different edge characteristics. There is no need to manually adjust the process parameters edge by edge, which greatly improves the adaptability of grinding and processing quality.

[0014] In one embodiment of this application, the precision verification in S4 specifically includes: Verify trajectory fit: In the formula, The coordinates of the feature reference points are extracted by discrete sampling along the contour of the edge to be polished in the 3D model. The coordinates of the trajectory path points corresponding to the generated polishing trajectory; m is the number of sampling points; Verify trajectory coverage: In the formula, The actual area covered by the trajectory. This represents the actual area of ​​the feature to be polished. Verify the actual overlap rate of the trajectory: ; like > or < 0 or < , , If 0 is the preset value, the verification is considered to have failed.

[0015] Technical effect: This claim establishes a multi-dimensional trajectory accuracy verification mechanism, which comprehensively verifies the generated grinding trajectory from three dimensions: trajectory position deviation, area coverage, and path overlap rate. For trajectories that do not meet the accuracy requirements, the parameters are automatically iterated and optimized to ensure that the final output grinding trajectory fully meets the processing accuracy requirements and avoids grinding quality problems caused by unqualified trajectories.

[0016] In one embodiment of this application, the 3D model in S1 includes the contour, size, and surface topology data of the area to be polished; the 3D model file format includes STL and STP formats.

[0017] In one embodiment of this application, the automatic grinding method of the robotic arm is applied to an automatic grinding system of the robotic arm. The grinding system includes an industrial control computer, a robotic arm and its controller, and an end-effector grinding tool. The industrial control computer is communicatively connected to the robotic arm controller, and the end-effector grinding tool is fixedly installed at the end of the robotic arm. The industrial control computer has a memory, a processor, an interaction unit, and a communication unit. The memory is used to store 3D model files. The processor is used to read the 3D model, perform edge capture, calculate edge features, automatically generate grinding trajectories, and complete accuracy verification. The interaction unit is used by the operator to perform model viewing and edge selection operations. The communication unit sends trajectory instructions to the robotic arm controller to control the robotic arm to drive the end-effector grinding tool to complete edge grinding. Attached Figure Description

[0018] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0019] Figure 1 This is a flowchart illustrating the steps of an automated grinding method using a robotic arm based on 3D model edge recognition, as described in an embodiment of this application. Detailed Implementation

[0020] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. After reading the present invention, any modifications of the present invention in various equivalent forms by those skilled in the art will fall within the scope defined by the appended claims.

[0021] Example: This invention discloses an automatic grinding method for robotic arms based on 3D model edge recognition, applied to an automatic grinding system for robotic arms. The grinding system includes an industrial control computer, a robotic arm and its controller, and an end-effector grinding tool. The industrial control computer is communicatively connected to the robotic arm controller, and the end-effector grinding tool is fixedly installed at the end of the robotic arm. The industrial control computer internally includes a memory, a processor, an interaction unit, and a communication unit. The memory is used to store 3D model files. The processor is used to read the 3D model, perform edge capture, highlight selected edges, calculate edge features, automatically generate grinding trajectories, and complete accuracy verification. The interaction unit is used by the operator to perform model viewing and edge selection operations. The communication unit sends trajectory commands to the robotic arm controller, controlling the robotic arm to drive the end-effector grinding tool to complete edge grinding.

[0022] like Figure 1 As shown, an automated grinding method for robotic arms based on 3D model edge recognition includes the following steps: S1: Obtain the 3D model of the part to be polished and load it into the industrial computer; specifically: The operator copies the 3D model to the industrial computer's memory via USB interface or network transmission; the processor reads and loads the 3D model. The device to be polished should be basically consistent with the imported 3D model in terms of size and shape, with no obvious deviation. The surface of the device should not have large-area damage or severe deformation, ensuring that the edges and surface contours of the actual device correspond to the 3D model. The 3D model includes, but is not limited to, STL and STP formats. The 3D model must completely contain the core data such as the contour, size, and surface topology of the area to be polished to ensure the accuracy of subsequent trajectory planning and attitude adjustment, and avoid insufficient polishing accuracy due to excessive deviation between the actual device and the model.

[0023] S2: Receives selection commands from the user via the interactive interface to define the edge areas to be polished in the 3D model; specifically: The operator determines the edge to be polished in the 3D model through the software interaction unit; after receiving the selection command, the industrial control computer performs topological continuity detection on the selected edge, and after confirming that the edge is a continuous single edge, the selected edge is highlighted.

[0024] S3: Based on the 3D model data, edge feature calculation is performed on the area to be polished, generating the polishing trajectory and polishing posture matrix; specifically: S3.1: Obtain the basic parameter information of the 3D model's edge contour and adjacent surfaces. For the edges to be ground, calculate geometric features such as edge angle, curvature, and radius of curvature, and automatically determine the edge feature trajectory type; specifically including: Calculate the included angle of the edges: (1); Calculate the curvature of the edge and its corresponding radius of curvature: (2); (3); The included angle of the edges in the formula The sharpness is determined and used for adaptive adjustment of the subsequent grinding posture, serving as a bias constraint for the posture matrix. The normal vectors of the two sides of the edge to be polished are obtained by the processor by parsing the topological relationship of the 3D model and directly extracting the geometric information of the adjacent faces on both sides of the edge. For curvature, and These are the first and second order guiding quantities of the spatial curve of the edge to be polished at this location. They are calculated by the processor after parsing the contour and surface geometry information of the 3D model and performing curve fitting on the selected edge. The radius of curvature of the curved edge; Determine the type of edge feature trajectory: (4); It is the optimal trajectory type corresponding to the edge features; k is the edge curvature, L is the edge length, Let be the radius of curvature of the curved edge. A preset threshold is set for the length of straight edge segments. A preset threshold is set for the radius of curvature of the curved edge; S3.2: Adaptive generation of trajectory parameters, calculated using the following formula: (1) Calculate the basic step size: (5); (2) Calculate the feature fitting step size: (6); (3) Calculate the basic overlap rate: (7); (4) Calculation accuracy adapts to overlap rate: (8); In the formula, Based on step size, This represents the feed rate of the grinding head (mm / min). The grinding head rotation speed (r / min) is used. This is the feed rate per revolution of the grinding head, in mm / r; The material coefficient is 0.8 for steel, 1.2 for aluminum, and 1.5 for plastic. To adapt the step size to the feature, For feature fitting coefficients; The base overlap ratio is given, and D is the diameter of the grinding head (mm). This refers to the grinding precision coefficient; depending on different... Give Different initial preset values, such as : =0.7, =0.8; : =0.6, =0.7; : =0.5, =0.5; : =0.3, =0.2; The accuracy-adapted overlap rate is used to control the degree of overlap in the grinding trajectory, which determines the uniformity of edge grinding and the burr removal effect. A higher overlap rate results in more complete grinding trajectory coverage, more uniform grinding, and no missed grinding marks; a lower overlap rate results in larger gaps in the trajectory coverage, making it easier for grinding blind spots and burr residue to appear. Therefore, the required accuracy-adapted overlap rate... As a benchmark threshold, it is used to compare and determine the overlap rate with the actual trajectory in subsequent trajectory accuracy verification.

[0025] S3.3: Calculate and generate the grinding trajectory for the edge to be ground: (1) Determine the starting point and unit tangent vector of the edge to be ground; (2) Used to generate discrete trajectory points along the edge to be ground at fixed step lengths to form the robot arm's walking path; calculate the three-dimensional coordinates of each trajectory point in sequence to generate an ordered sequence of trajectory points; the edge point sampling formula is as follows: (9); In the formula: These are the coordinates of the sampling point; This is the starting point for grinding the edge to be polished. It is the sampling point number; To adapt the step size to the features already solved in the previous section; The unit tangent vector of the edge to be polished; the geometric parameters such as edge angle, curvature, radius of curvature, and edge length are solved in the early stage. The ultimate goal is to match the corresponding feature adaptation coefficient and accuracy coefficient, determine the reasonable step size for sampling and marking points on the original edge, and complete the sampling of discrete edge points on the edge to be polished; using the edge line to be polished provided by the 3D model as the reference path, the edge line is used to make discrete points at equal intervals along the original edge line at a set step size. The resulting ordered set of three-dimensional coordinate points constitutes the motion trajectory of the robotic arm for polishing. S3.4: Calculate the rotation matrix of the robotic arm's grinding posture: (1) Calculate the normal vector of the edge to be ground: (10); It is the normal vector of the edge to be polished. These are the normal vectors of the two sides of the edge to be polished; (2) Calculate the grinding posture rotation matrix: This solution introduces the included angle of the edge, based on the traditional solution. As a constraint, adaptive adjustment is achieved through complete quantization calculations, with the following specific steps: ① Calculate the pitch compensation angle: (11); In the formula It is the pitch angle compensation value of the grinding tool; this compensation value is then superimposed on the reference pitch angle to complete the attitude angle adaptive correction, and is substituted into the construction of the robot arm attitude rotation matrix; This is the preset maximum process avoidance angle, which is a fixed process constant of the system. The value range is 5°-15° and can be adjusted according to actual conditions. It is the included angle of the edges, which has been calculated above; where the included angle of the edges The smaller the edge and the sharper the corner, the more the system automatically increases the size of the grinding tool to avoid the sharp corner from being over-grinded and chipping the edge; the included angle of the edge The larger the size and the smoother the edge, the more automatically the avoidance angle is reduced, so that the grinding head fits the edge evenly for grinding. ② Solve for the corrected actual pitch angle: (12); In the formula It is the actual pitch attitude angle after correction based on the edge angle feature; It is the reference pitch angle under no attitude bias conditions, which is a preset fixed process parameter of the system, usually set to 10°~15°; This is the pitch angle compensation value, which has been calculated above; ③ Construct a local coordinate system basis for the edges: (13); In the formula It is the X-axis basis of the local coordinate system, and its orientation is consistent with that of the edge. It is the unit tangent vector of the edge to be polished, which is extracted from the discrete points of the edge contour of the 3D model. It is the Y-axis basis of the local coordinate system, and the normal vector of the edge to be ground. Same direction; It is the normal vector of the edge to be ground, which is calculated by equation (10); It is the basis of the local coordinate system Z-axis, determined by the tangent vector. With normal vector The vector cross product is obtained and is pairwise orthogonal to the X and Y axes. ④ Construct the end-position rotation matrix for the included angle of the coupled edges: (14); In the formula , , It is the transpose of the three-axis basis vectors of the coordinate system; It is the actual pitch attitude angle The cosine and sine function values, angle It can be obtained from equation (12); In the formula This is the rotation matrix for the end effector posture of the robotic arm.

[0026] S4: Perform accuracy verification on the generated grinding trajectory. If the verification fails, adjust the trajectory calculation parameters and regenerate the grinding trajectory until the preset accuracy requirements are met. The formula for trajectory accuracy verification is as follows: (1) Verify the trajectory fit: (15); (2) Verify trajectory coverage: (16); (3) Verify the actual overlap rate of the trajectory: (17); For trajectory fit, where The coordinates of the feature reference points are extracted by discrete sampling along the contour of the edge to be polished in the 3D model. The coordinates of the trajectory path points discretely extracted at the same sampling interval on the polishing trajectory generated by the algorithm above. The two sets of sampling points have the same interval and correspond one-to-one with each other; n is the number of sampling points. For trajectory coverage, the preset acceptable threshold is 95%. The actual area covered by the trajectory. This represents the actual area of ​​the feature to be polished. This represents the actual overlap rate of the trajectories. For the diameter of the grinding head, To adapt the step size to the features solved in the previous section; The default setting is 0.05mm; 0 is preset to 95%; like > or < 0 or < Then return to step S3.2 and re-optimize using equations (5~8): such as... Reduce the value by 10%, then regenerate the grinding trajectory with the new coefficients, and simultaneously recalculate and match the robotic arm's grinding posture matrix to complete the integrated verification of trajectory and posture. Iterate repeatedly until the accuracy threshold is met.

[0027] S5: Send the verified grinding trajectory and corresponding posture matrix to the robotic arm controller, driving the robotic arm to drive the end-effector grinding tool to complete the grinding operation of the edge to be ground.

[0028] The communication unit sends the grinding trajectory and grinding posture rotation matrix to the robotic arm controller; the robotic arm controller uses the B-spline interpolation algorithm to smooth the trajectory points and generate motion control commands for each joint; it controls the robotic arm to drive the end-effector grinding head to move according to the trajectory sequence, and at the same time adjusts the orientation of the grinding head in real time according to the posture matrix of each point to keep the contact angle between the grinding head and the edge stable; after all trajectories are executed, the robotic arm returns to the origin, and the grinding operation ends.

[0029] The technical solutions described in the embodiments of this application have at least the following technical effects or advantages: 1. The operator copies the 3D model to the industrial computer via USB interface or network transmission. The operator selects the edge to be polished, and the system highlights the selected edge and automatically completes the polishing trajectory planning. Finally, it drives the robotic arm and end-effector polishing tool to perform automated deburring and polishing of the model edge. This invention can achieve precise and controllable edge polishing, is easy to operate and highly adaptable, and is suitable for various automated deburring processing scenarios for parts edges.

[0030] 2. Existing technologies generally employ industrial camera vision combined with image algorithms to automatically identify workpiece edges and areas to be polished. However, these technologies are highly dependent on external factors such as ambient lighting, workpiece placement, surface reflection, and contaminants, which can lead to issues like missed or incorrect edge identification and distorted contour extraction. This method, on the other hand, eliminates the need for industrial vision equipment for outdoor image acquisition and edge recognition. Instead, it directly uses an imported standard 3D model, allowing the operator to manually select the designated edge area to be polished via a software interaction unit. The processor then highlights and locks the selected edge to complete the delineation of the processing area. This method is unaffected by ambient lighting, workpiece placement deviations, surface reflection, or contaminants, ensuring stable and reliable recognition. It can accurately select any straight edge, curved edge, rounded corner, or curved surface polishing area as needed, making it highly targeted and applicable to a wider range of applications.

[0031] 3. Because the traditional robotic arm's posture rotation matrix relies solely on the tangent vector of the edge to be ground. Normal vector Traditional fixed-position construction only provides fixed-posture positioning capabilities and cannot dynamically adapt to changes in edge sharpness or contour steepness. Consequently, the traditional approach maintains a fixed posture and tilt angle throughout the entire process, requiring the same grinding posture for edges of varying sharpness. This leads to over-grinding and chipping at sharp edges, while insufficient fit and residual burrs at smoother edges necessitate manual adjustment of the process offset angle for each edge, resulting in poor versatility and weak adaptability. This new approach introduces edge angle adjustments. As an attitude offset constraint, the included angle of the edge As a reference value for attitude adjustment, it eliminates the need for manual setting of process tilt angle offset parameters for each edge, enabling adaptive grinding under different edge morphologies.

[0032] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. An automated grinding method for robotic arms based on 3D model edge recognition, characterized in that, Includes the following steps: S1: Obtain the 3D model of the part to be polished and load it into the industrial computer; S2: Receive the selection command input by the user through the interactive interface to determine the edge area to be polished in the 3D model; S3: Calculate the edge features of the area to be polished based on 3D model data, and generate the polishing trajectory and the corresponding polishing posture matrix; S4: Perform accuracy verification on the generated grinding trajectory. If the verification fails, adjust the trajectory calculation parameters and regenerate the grinding trajectory. S5: Send the verified grinding trajectory and corresponding grinding posture matrix to the robotic arm controller to drive the robotic arm to complete the grinding operation.

2. The automatic grinding method using a robotic arm according to claim 1, characterized in that: The edge feature calculation of the area to be polished in step S3 specifically includes: Obtain the basic parameter information of the 3D model's edge contour and adjacent surfaces. Calculate the edge angle, curvature, and radius of curvature for the edge to be ground, and determine the edge feature trajectory type. The calculation formula is as follows: Edge angle : ； Curvature of the edge and corresponding radius of curvature : ； ； In the formula These are the normal vectors of the two sides of the edge to be polished; and It refers to the first and second order guiding quantities of the space curve of the edge to be polished; Determine the type of edge feature trajectory: ； In the formula, These are the optimal trajectory types corresponding to the edge features; TE1: Straight-line trajectory along the edge, TE2: Segmented straight-line trajectory along the edge, TE3: Conformal curve trajectory, TE4: Variable curvature conformal trajectory; L is the edge length. A preset threshold is set for the length of straight edge segments. A preset threshold is set for the radius of curvature of the curved edge.

3. The automatic grinding method using a robotic arm according to claim 2, characterized in that: The specific steps involved in generating the polishing trajectory in S3 are as follows: Based on the trajectory type TE corresponding to the edge to be polished, determine the initial feature adaptation coefficient kf and the polishing accuracy coefficient kp; Calculate the feature adaptation step size S and the accuracy adaptation overlap rate η by combining the grinding process parameters: ； ； ； ； In the formula, Based on step size, To adjust the feed rate of the grinding head, To set the grinding head speed, It is the feed rate per revolution of the grinding head; Material coefficient; To adapt the step size to the feature, The base overlap ratio is given, and D is the diameter of the grinding head. Based on the overlap rate; Along the edge to be polished, discrete points are sampled according to the feature adaptation step size S to generate an ordered set of coordinate points for the polishing trajectory: ； In the formula, These are the coordinates of the sampling point; This is the starting point for grinding the edge to be polished. It is the sampling point number; To adapt the step size to the features already solved in the previous section; Let be the unit tangent vector of the edge to be polished.

4. The automatic grinding method for robotic arms according to claim 3, characterized in that: The specific steps involved in generating the polishing posture matrix in S3 are as follows: Calculate the normal vector of the edge to be ground: ； Calculate the grinding posture rotation matrix: Calculate the pitch compensation angle: In the formula, It is the preset maximum process avoidance angle; Solve for the corrected actual pitch angle: In the formula, It is the reference pitch angle; Construct a local coordinate system basis for the edge: ； In the formula It is the X-axis basis of the local coordinate system; It is the Y-axis basis of the local coordinate system; It is the Z-axis basis of the local coordinate system; Construct the end-position rotation matrix of the coupling edge angle: ； In the formula , , It is the transpose of the three-axis basis vectors of the coordinate system; It is the actual pitch angle. The cosine and sine function values.

5. The automatic grinding method for robotic arms according to claim 4, characterized in that: The precision verification in S4 specifically includes: Verify trajectory fit: In the formula, The coordinates of the feature reference points are extracted by discrete sampling along the contour of the edge to be polished in the 3D model. The coordinates of the trajectory path points corresponding to the generated polishing trajectory; m is the number of sampling points; Verify trajectory coverage: In the formula, The actual area covered by the trajectory. This represents the actual area of ​​the feature to be polished. Verify the actual overlap rate of the trajectory: ; like > or < 0 or < , , If 0 is the preset value, the verification is considered to have failed.

6. The automatic grinding method using a robotic arm according to claim 1, characterized in that: The 3D model in S1 includes the outline, dimensions, and surface topology data of the area to be polished; the 3D model file format includes STL and STP formats.

7. The automatic grinding method using a robotic arm according to claim 1, characterized in that: This invention relates to an automated grinding system for robotic arms. The grinding system includes an industrial control computer, a robotic arm and its controller, and an end-effector grinding tool. The industrial control computer is communicatively connected to the robotic arm controller, and the end-effector grinding tool is fixedly installed at the end of the robotic arm. The industrial control computer contains a memory, a processor, an interaction unit, and a communication unit. The memory is used to store 3D model files. The processor is used to read the 3D model, perform edge capture, calculate edge features, automatically generate grinding trajectories, and complete accuracy verification. The interaction unit is used by the operator to perform model viewing and edge selection operations. The communication unit sends trajectory commands to the robotic arm controller, controlling the robotic arm to drive the end-effector grinding tool to complete edge grinding.

Images