Teaching-free laser shock enhancement method and system for 3D models and laser point clouds

Abstract

This invention belongs to the field of laser processing modification technology, and relates to a teaching-free laser shock strengthening method and system based on a 3D model and laser point cloud. The system includes a 3D modeling module, a point cloud acquisition module, a point cloud processing module, a path planning module, and a control module. It identifies the area to be strengthened using a 3D information model and plans the initial processing path. High-precision point clouds are acquired using a line laser sensor and monocular vision fusion method, and the deviation matrix of the workpiece to be processed is calculated through registration. Based on the deviation matrix, the initial processing path is transformed and corrected point by point, ultimately generating a robot execution program to control the laser processing device to complete the automated laser shock strengthening operation. This invention achieves fully automated operation from 3D information model to robot operation, effectively solving the problems of traditional methods such as reliance on manual teaching, low positioning accuracy, and difficulty in assessing accessibility.

Description

Technical Field

[0001] This invention relates to the field of laser processing and modification technology, and in particular to a teachless laser shock strengthening method and system based on a three-dimensional model and laser point cloud. Background Technology

[0002] Laser shock peening technology can introduce residual compressive stress on the surface of metal materials through shock waves induced by high-power lasers, significantly improving the fatigue life and stress corrosion resistance of metal components, and has wide applications in aerospace, energy equipment, and other fields. However, current laser shock peening operations mainly rely on manual teaching or traditional offline programming methods, which have the following technical drawbacks: Manual teaching leads to low efficiency, requiring several days or even weeks of on-site debugging time for complex curved surface components; traditional offline programming methods lack the ability to compensate for actual clamping deviations of the workpiece, resulting in a mismatch between the planned path and the actual path; there is a lack of effective means to assess the accessibility of complex structures, which can easily lead to interruption of the strengthening process due to interference between the laser head and the workpiece; when strengthening multiple areas, there is a lack of overall optimization of the processing sequence, resulting in long idle movement paths and the risk of collision.

[0003] Existing patent CN117327896A discloses a laser shock blasting shaping control method for blades. This method applies 3D models and point cloud technology to laser shock blasting, primarily addressing blade deformation control. It establishes a model by acquiring the 3D coordinate data of the impeller and combines it with a point cloud data extraction system to achieve blade strengthening. However, this technology does not address teach-free planning and accessibility assessment of multi-region strengthening paths, making it difficult to meet the demands for efficient and precise automated processing of complex components. Therefore, there is an urgent need for a teach-free laser shock blasting system capable of automatically generating robot processing programs from 3D models and possessing actual position perception and compensation capabilities. Summary of the Invention

[0004] To address the aforementioned technical challenges, this invention provides a teach-free laser shock enhancement method and system based on a 3D model and laser point cloud. As a teach-free laser shock enhancement system with actual position perception and compensation capabilities, it can automatically generate robot processing programs, thus solving the technical problems of existing technologies such as reliance on manual labor, inaccurate positioning, difficulty in assessing accessibility, and low efficiency. To achieve the above objectives, this invention utilizes the following technical solutions: A teach-free laser shock enhancement method based on a 3D model and laser point cloud includes the following steps: Step S1: Analyze the three-dimensional information model of the workpiece to be processed, automatically identify the areas to be strengthened according to preset rules, and automatically generate an initial processing path and offline program for each area to be strengthened. Step S2: Acquire surface data of the workpiece to be processed by fusion of line laser and monocular vision to obtain a high-precision point cloud dataset of the workpiece to be processed; Step S3: Register the high-precision point cloud dataset of the workpiece to be processed obtained in step S2 with the three-dimensional information model in step S1, and calculate the deviation matrix of the workpiece to be processed. Step S4: Perform coordinate transformation and point-by-point correction on the initial machining path according to the deviation matrix to generate an actual machining path that matches the actual workpiece; Step S5: Based on the actual processing path, generate a robot execution program to control the laser processing device to complete the automated impact strengthening operation.

[0005] Preferably, the area to be strengthened is a stress concentration area, which includes corners, edges, and mating surfaces of the workpiece to be processed; The initial processing path includes the starting point, ending point, midpoint of the path, and normal vector information of the region to be enhanced; The offline program includes the robot's motion trajectory, the laser head's posture, and the laser shock strengthening process parameters.

[0006] Preferably, the preset rules in step S1 include: When the rate of change of principal curvature in a certain region exceeds a set threshold and the angle between the neighboring normal vectors is greater than a preset angle, it is identified as a corner region; when performing adjacent surface angle analysis on edge features, if the angle is less than a certain angle, it is identified as an edge region; for mating surfaces, identification is based on the consistency of their normal vectors and the assembly relationship of adjacent components.

[0007] Preferably, step S2 includes: Step S21: Use a monocular camera to acquire images of the workpiece to be processed from multiple perspectives, and reconstruct the initial point cloud dataset of the workpiece to be processed through feature matching and triangulation. Step S22: Use a line laser sensor to scan the surface of the workpiece to be processed along a preset scanning path to obtain line laser point cloud data of the workpiece to be processed with absolute dimensional information. Step S23: The initial point cloud dataset and the line laser point cloud data are fused together. The absolute scale of the line laser point cloud data is used to correct the fuzzy scale of the initial point cloud dataset to obtain a high-precision point cloud dataset of the workpiece to be processed.

[0008] Preferably, the registration includes coarse registration and fine registration; the fine registration is performed after the coarse registration. The coarse registration is implemented as follows: global features are extracted from the high-precision point cloud dataset and the 3D information model, and then feature matching is used to quickly match the two sets of features in the high-precision point cloud dataset and the 3D information model to obtain an initial deviation matrix; the global features include edge contours, corner points, and reference holes; The implementation method of fine registration is as follows: the three-dimensional information model is transformed to the actual coordinate system of the workpiece to be processed using the initial deviation matrix, and the rough position of the area to be strengthened in the actual space is determined. The guide line laser sensor collects a high-density point cloud dataset of the area to be strengthened. The high-density point cloud dataset is iteratively registered with the corresponding area of ​​the three-dimensional information model after transformation using the initial deviation matrix using the nearest point ICP, and the deviation matrix is ​​optimized to obtain the deviation matrix.

[0009] Preferably, the method further includes the following steps before step S4: Step S3': For each initial processing path, accessibility is assessed, unreachable paths are identified based on the spatial open angle algorithm, and the laser head processing posture is calculated for reachable paths; The steps for implementing the accessibility assessment include: Step S31': For each initial processing path, extract the center point and normal vector of the current region to be strengthened; Step S32': Construct a spatial open angle calculation domain with the center point as the center of the sphere and the working distance of the laser head as the radius; Step S33': Simulate the laser head model within the open angle calculation domain of the space, and detect collisions angle by angle. Calculate the proportion of the non-collision angle range to the total angle range, and use it as the open angle. If the open angle is less than the preset threshold, it is marked as an unreachable path; otherwise, it is a reachable path, and the center direction of the open angle is selected as the optimal processing posture.

[0010] Preferably, step S3' further includes: A multi-objective optimization model with overlap uniformity constraints is established to optimize the processing sequence of multiple reachable paths. The optimal processing sequence is solved by using a genetic algorithm or particle swarm optimization algorithm, and the optimal processing sequence and the corresponding empty movement path are output.

[0011] Preferably, the multi-objective optimization model is:

[0012] In the formula, This is the total empty movement path length. To account for unevenness in the overlapping area. This is a penalty item for collision risk. The weighting coefficient for the total idle path length. This is a weighting coefficient for the unevenness of the overlapping area. Let be the weighting coefficient of the collision risk penalty term, and It is between [0,1].

[0013] A teach-free laser shock enhancement system based on a 3D model and laser point cloud, comprising: The 3D analysis module is used to analyze the 3D information model of the workpiece to be processed, automatically identify one or more areas to be strengthened, complete the initial path planning, and generate the initial processing path. The point cloud acquisition module, including a line laser sensor and a monocular camera, is used to acquire surface data of the workpiece to be processed and obtain a high-precision point cloud dataset. The point cloud processing module is connected to the point cloud acquisition module and the three-dimensional analysis module respectively. It is used to perform point cloud registration between the high-precision point cloud dataset of the workpiece to be processed and the three-dimensional information model, and calculate the deviation matrix of the workpiece to be processed. The path planning module, connected to the point cloud processing module, is used to perform coordinate transformation and point-by-point correction on the initial processing path according to the deviation matrix, and generate an actual processing path that matches the actual workpiece. The control module, connected to the path planning module, is used to generate a robot execution program based on the actual processing path and control the laser processing device to complete automated laser shock strengthening operations.

[0014] Preferably, the path planning module further includes a reachability analysis unit, a job sequencing unit, and an idle movement planning unit; The reachability analysis unit is used to evaluate the reachability of the initial processing path, identify unreachable paths, and calculate the laser head processing posture for reachable paths; The job sorting unit is used to execute the spatial open angle algorithm and detect and calculate the intersection point with the workpiece model; The air travel planning unit is used to optimize the processing sequence of multiple reachable paths using a multi-objective optimization model.

[0015] Preferably, the point cloud processing module further includes a point cloud registration unit and a solution unit; the point cloud registration unit is used to perform registration operations and sequentially perform coarse registration and fine registration; the solution unit is used to solve the deviation matrix of the workpiece to be processed.

[0016] Preferably, the control module further includes a signal acquisition unit, a feature extraction unit, a parameter calculation and decision-making unit, and an instruction generation unit; The signal acquisition unit is used to acquire workpiece processing signals and start the robot to begin processing operations. The feature extraction unit is used to extract processing features and send the features back to the parameter calculation and decision unit; The parameter calculation and decision-making unit is used to call up the processing parameters in the historical processing database based on the processing characteristics, and to decide which parameters to use for processing; The instruction generation unit is used to generate and output the robot execution program required for laser shock reinforcement.

[0017] The present invention has the following advantages over the prior art: 1. This invention obtains the initial processing path from the three-dimensional information model to generate the robot program without human intervention, achieving full-process teaching-free automation, which can significantly shorten the production preparation cycle and improve the processing efficiency of complex components. 2. This invention uses a line laser sensor and monocular vision fusion method to reconstruct a high-precision point cloud model, uses point cloud fusion to achieve high-precision positioning, and combines a coarse-fine two-level registration strategy to effectively compensate for workpiece clamping deviation and ensure accurate positioning of the area to be strengthened. 3. This invention achieves intelligent assessment of the accessibility of complex structures based on the spatial open angle algorithm, which can automatically identify unreachable paths and calculate safe processing postures, effectively avoiding the risk of laser head collision and achieving full coverage reinforcement of complex structures; 4. This invention adopts an industrial-grade offline program generation strategy, which can automatically generate standard robot execution files using templates. It does not require real-time communication, avoids data loss and equipment jitter issues, and has good engineering compatibility. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the execution process of the method of the present invention; Figure 2 This is a schematic diagram of the overall structure of the system of the present invention; Figure 3 A schematic diagram illustrating the technical principle of obtaining a high-precision point cloud model in step S2; Figure 4 This is a schematic diagram of the internal structure of the control module. Detailed Implementation

[0019] The present invention will now be further described with reference to the accompanying drawings and specific embodiments: To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of the embodiments 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.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0021] Example 1: like Figure 1 and Figure 3 As shown, this embodiment provides a teach-free laser shock strengthening method based on a 3D model and laser point cloud. This method achieves full automation of the laser shock strengthening process by constructing a digital closed loop from virtual design model to on-site physical processing, eliminating the need for manual intervention in the teaching stage. Specifically, it includes the following steps: Step S1: Analyze the 3D information model of the workpiece to be processed, automatically identify the areas to be strengthened according to preset rules, and automatically generate an initial processing path and offline program for each area to be strengthened. The areas to be strengthened in this step are also stress concentration areas, which include corners, edges, mating surfaces, etc. of the workpiece to be processed. The initial processing path includes the start point, end point, midpoint of the path, and normal vector information of the current area to be strengthened. The offline program includes the robot motion trajectory, laser head posture, and laser shock strengthening process parameters.

[0022] Specifically, the 3D information model of the workpiece to be processed is imported from an external source, such as directly importing a 3D information model obtained from CAD. The analysis process involves discretizing the 3D information model to extract the principal curvature and Gaussian curvature of each facet. After analysis, preset rules are executed to identify the areas to be strengthened, i.e., stress concentration areas. Specifically, when the rate of change of principal curvature in a certain area exceeds a set threshold and the angle between the neighboring normal vectors is greater than a preset angle, it is identified as a corner area. When analyzing the angle between adjacent surfaces of edge features, if the angle is less than a certain angle, such as less than 120°, it is identified as an edge area. For mating surfaces, identification is based on the consistency of their normal vectors and the assembly relationship of adjacent parts. After automatically identifying all stress concentration areas that meet the conditions, they are merged to generate the areas to be strengthened.

[0023] Step S2: Acquire surface data of the workpiece to be processed by fusing line laser and monocular vision to obtain a high-precision point cloud dataset of the workpiece; the specific implementation method of this step is as follows: Step S21: Use a monocular camera to acquire images of the workpiece to be processed from multiple perspectives, and reconstruct the initial point cloud dataset of the workpiece to be processed through feature matching and triangulation. The operation method for reconstructing the initial point cloud dataset in this step is as follows: a monocular camera is used to acquire two-dimensional images of the workpiece to be processed from multiple preset viewpoints, and ORB or SIFT feature points are extracted from each two-dimensional image. The correspondence between two-dimensional images is established by feature point matching. Then, the three-dimensional spatial coordinates of the feature points are calculated using epipolar geometric constraints and triangulation. The three-dimensional spatial coordinates of all preset viewpoints are then fused and downsampled to generate the initial point cloud dataset.

[0024] Step S22: Use a line laser sensor to scan the surface of the workpiece to be processed along a preset scanning path to obtain a line laser point cloud dataset of the workpiece to be processed with absolute dimensional information. Step S23: Fusion of the initial point cloud dataset and the line laser point cloud dataset, using the absolute scale of the line laser point cloud dataset to correct the fuzzy scale of the initial point cloud dataset, to obtain a high-precision point cloud dataset of the workpiece to be processed.

[0025] Step S3: Register the high-precision point cloud dataset of the workpiece to be processed obtained in Step S2 with the 3D information model constructed in Step S1, and calculate the deviation matrix of the workpiece to be processed. Specifically, receive the high-precision point cloud dataset of the workpiece to be processed from Step S2, and register the high-precision point cloud dataset of the workpiece to be processed with the 3D information model of the workpiece to be processed obtained in Step S1. The registration includes coarse registration and fine registration. Coarse registration must be performed first, followed by fine registration, and then the deviation matrix of the workpiece to be processed is calculated. It should be noted that to obtain the deviation matrix, the high-precision point cloud dataset and the 3D information model must first be unified to the same coordinate system. First, the initial deviation matrix is ​​obtained through coarse registration, and then the accurate deviation matrix is ​​obtained through fine registration for local optimization. The calculated rotation matrix R and translation vector T between the actual point cloud and the 3D information model are combined to form a 4×4 deviation matrix M, expressed as:

[0026] Since the high-precision point cloud dataset can reflect the actual coordinate position of the workpiece to be processed, while the three-dimensional information model can reflect the theoretical coordinate position of the workpiece to be processed, the coarse registration is implemented by first extracting global features such as edge contours, corner points and reference holes from the high-precision point cloud dataset and the three-dimensional information model. Then, feature matching is used to quickly match the two sets of features in the high-precision point cloud dataset and the three-dimensional information model to estimate the deviation relationship between the actual coordinate position and the theoretical coordinate position of the workpiece to be processed. Finally, the initial deviation matrix is ​​calculated to provide the initial deviation value for the subsequent fine registration.

[0027] The fine-fit criterion involves transforming the 3D information model to the actual coordinate system of the workpiece after coarse registration using the initial deviation matrix, thereby determining the approximate location of the area to be strengthened in actual space. Based on this, the line laser sensor is guided to collect a high-density point cloud dataset of the area again. Then, the high-density point cloud dataset of the area is iteratively registered with the corresponding area of ​​the 3D information model transformed using the initial deviation matrix using the nearest point ICP. In other words, the precise deviation matrix is ​​obtained by optimizing the initial deviation matrix.

[0028] Fine registration essentially involves narrowing the search range back to the vicinity of the region to be enhanced after coarse registration. This is achieved by collecting a local high-density point cloud dataset from the vicinity of the region to be enhanced and performing iterative nearest-point registration with the fine features of the corresponding region in the transformed 3D information model, thereby obtaining an accurate deviation matrix.

[0029] Step S4: Perform coordinate transformation and point-by-point correction on the initial machining path according to the deviation matrix to generate an actual machining path that matches the actual workpiece. The specific operation is as follows: According to the calculated deviation matrix, perform coordinate transformation and point-by-point correction on all the initial machining path points in the initially generated initial machining path to generate an actual machining path that perfectly matches the workpiece to be processed.

[0030] Specifically, the coordinate transformation and point-by-point correction in this step are implemented as follows: Let the initial processing path point set be P={p1,p2,…,p n}, where n is a non-zero natural number, and p i =(x i ,y i ,z i ,n i ), x i ,y i ,z i n represents the coordinates of the initial processing path point. i It is the normal vector. Let i be the deviation matrix, i is [1, n]; and the transformed actual path points satisfy:

[0031] Wherein, the transformed actual normal vector After transformation by the rotation matrix R, normalization is required. All initial machining path points are corrected sequentially to generate the actual machining path point set. By drawing the final processing path point set Obtain the final processing path.

[0032] Step S5: Based on the final processing path, generate a robot execution program to control the laser processing device to complete the automated impact strengthening operation. The robot execution program is automatically generated according to a preset motion instruction template, and the instruction types of the robot execution program include linear interpolation and joint movements, etc.

[0033] Furthermore, step S4 is preceded by: Step S3': Achieve reachability assessment for each initial processing path, identify unreachable paths based on the spatial open angle algorithm, and calculate the laser head processing posture for reachable paths; The implementation steps of accessibility assessment include: Step S31': For each initial processing path, extract the center point and normal vector of the current region to be strengthened; Step S32': Construct a spatial open angle calculation domain with the center point as the center of the sphere and the working distance of the laser head as the radius; Step S33': Simulate the laser head model within the open angle calculation domain of the space, and detect whether there is a collision with the workpiece or fixture at each angle. Calculate the proportion of the non-collision angle range to the total angle range, and take it as the open angle. If the open angle is less than the preset threshold, it is marked as an unreachable path; otherwise, it is a reachable path, and the center direction of the open angle is selected as the optimal processing posture.

[0034] The specific content of the spatial open angle algorithm is as follows: taking the center point of the region to be strengthened as the origin, and then using the vector... The principal axis direction; in spherical coordinates, with , Discrete sampling is used as the incident direction; for each incident direction From the origin along The laser emits a ray in a certain direction and detects the intersection with the workpiece model. If there is no intersection or the intersection is outside the safe distance of the laser head, the direction is deemed feasible. The ratio of the total solid angle of the feasible direction to the total solid angle of the hemisphere is the open angle of space.

[0035] Step S3' also includes: establishing a multi-objective optimization model with overlap uniformity constraints to optimize the processing sequence of multiple reachable paths. The multi-objective optimization model here is:

[0036] in, This is the total empty movement path length. To account for unevenness in the overlapping area. This is a penalty item for collision risk. The weighting coefficient for the total idle path length. This is a weighting coefficient for the unevenness of the overlapping area. Let be the weighting coefficient of the collision risk penalty term, and It is between [0,1].

[0037] Then, a genetic algorithm or particle swarm optimization algorithm is used to solve for the optimal order, and the optimal processing order and the corresponding empty movement path are output.

[0038] Example 2: like Figure 2 and Figure 4 As shown, a teach-free laser shock enhancement system based on a 3D model and laser point cloud includes: The 3D analysis module is used to analyze the 3D information model of the workpiece to be processed, automatically identify one or more areas to be strengthened, complete the initial path planning, and generate the initial processing path. The point cloud acquisition module, including a line laser sensor and a monocular camera, is used to acquire surface data of the workpiece to be processed and obtain a high-precision point cloud dataset. The point cloud processing module is connected to the point cloud acquisition module and the three-dimensional analysis module respectively. It is used to perform point cloud registration between the high-precision point cloud dataset of the workpiece to be processed and the three-dimensional information model, and calculate the deviation matrix of the workpiece to be processed. The path planning module, connected to the point cloud processing module, is used to perform coordinate transformation and point-by-point correction on the initial processing path according to the deviation matrix, and generate an actual processing path that matches the actual workpiece. The control module, connected to the path planning module, is used to generate a robot execution program based on the actual processing path and control the laser processing device to complete automated laser shock strengthening operations.

[0039] The control module also includes a signal acquisition unit, a feature extraction unit, a parameter calculation and decision unit, and an instruction generation unit. The signal acquisition unit is used to acquire workpiece processing signals and start the robot to start the processing operation. The feature extraction unit is used to extract processing features and send the features back to the parameter calculation and decision unit. The parameter calculation and decision unit is used to call the processing parameters in the historical processing database based on the processing features and decide which parameters to use for processing. The instruction generation unit is used to generate and output the robot execution program required for laser shock strengthening.

[0040] The point cloud processing module also includes a point cloud registration unit and a solution unit; the point cloud registration unit is used to perform registration operations and to perform coarse registration and fine registration in sequence; the solution unit is used to solve the deviation matrix of the workpiece to be processed.

[0041] The path planning module also includes a reachability analysis unit, a job sequencing unit, and an idle movement planning unit. The reachability analysis unit is used to evaluate the reachability of the initial processing path, identify unreachable paths, and calculate the laser head processing posture for reachable paths. The job sequencing unit is used to execute the spatial open angle algorithm and detect and calculate the intersection point with the workpiece model. The idle movement planning unit is used to optimize the processing sequence of multiple reachable paths using a multi-objective optimization model.

[0042] The aforementioned units can interact at high speed via an internal data bus or shared memory area. The point cloud processing module's calculation unit transmits the deviation matrix to the path planning module; each unit in the path planning module processes the data sequentially, generating the final path data which is then transmitted to the control module; the units in the control module work collaboratively to ultimately generate the robot's execution program. This modular and unit-based design not only clarifies the boundaries of each functional component, facilitating system maintenance and upgrades, but also improves the real-time performance and reliability of the system through the decoupling of hardware logic.

[0043] The user imports the generated code file into the robot controller and starts the program. The robot will automatically travel along the corrected, precise path to each area to be strengthened. The laser processing device completes the laser shock strengthening operation of all processing points according to the set process parameters. The entire processing process requires no manual intervention, and the system monitors the operating status in real time to ensure processing safety. Through the progressive process of "initial path planning based on the 3D model - precise workpiece positioning based on point cloud - coordinate transformation and path correction - final program generation and execution," this invention achieves high-precision, fully automated closed-loop control from the 3D model to the robot operation.

[0044] This invention automatically identifies the areas to be strengthened and completes initial path planning by constructing a 3D model of the workpiece, generating an initial processing path and offline program. After the robot is in place, it acquires a high-precision point cloud of the workpiece through the fusion of line laser and monocular vision, registers the actual point cloud with the 3D model, and calculates the actual deviation of the workpiece. Based on the actual deviation, it performs coordinate transformation and correction on the initial processing path, generates the final robot execution program, and controls the laser processing device to complete the automated impact strengthening operation. This invention realizes full-process automation from 3D information model to robot operation, effectively solving the problems of traditional methods such as reliance on manual teaching, low positioning accuracy, and difficulty in assessing the accessibility of complex structures.

[0045] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A teach-free laser shock enhancement method based on a 3D model and laser point cloud, characterized in that, Includes the following steps: Step S1: Analyze the three-dimensional information model of the workpiece to be processed, automatically identify the areas to be strengthened according to preset rules, and automatically generate an initial processing path and offline program for each area to be strengthened. Step S2: Acquire surface data of the workpiece to be processed by fusion of line laser and monocular vision to obtain a high-precision point cloud dataset of the workpiece to be processed; Step S3: Register the high-precision point cloud dataset of the workpiece to be processed obtained in step S2 with the three-dimensional information model in step S1, and calculate the deviation matrix of the workpiece to be processed. Step S4: Perform coordinate transformation and point-by-point correction on the initial machining path according to the deviation matrix to generate an actual machining path that matches the actual workpiece; Step S5: Based on the actual processing path, generate a robot execution program to control the laser processing device to complete the automated impact strengthening operation.

2. The method for teaching-free laser shock enhancement based on a three-dimensional model and laser point cloud as described in claim 1, characterized in that, The area to be strengthened is a stress concentration point, which includes the corners, edges, and mating surfaces of the workpiece to be processed. The initial processing path includes the starting point, ending point, midpoint of the path, and normal vector information of the region to be enhanced; The offline program includes the robot's motion trajectory, the laser head's posture, and the laser shock strengthening process parameters.

3. The method for teaching-free laser shock enhancement based on a three-dimensional model and laser point cloud as described in claim 1, characterized in that, The preset rules in step S1 include: When the rate of change of principal curvature in a certain region exceeds a set threshold and the angle between the neighboring normal vectors is greater than a preset angle, it is identified as a corner region; when performing adjacent surface angle analysis on edge features, if the angle is less than a certain angle, it is identified as an edge region; for mating surfaces, identification is based on the consistency of their normal vectors and the assembly relationship of adjacent components.

4. The method for teaching-free laser shock enhancement based on a three-dimensional model and laser point cloud as described in claim 1, characterized in that, Step S2 includes: Step S21: Use a monocular camera to acquire images of the workpiece to be processed from multiple perspectives, and reconstruct the initial point cloud dataset of the workpiece to be processed through feature matching and triangulation. Step S22: Use a line laser sensor to scan the surface of the workpiece to be processed along a preset scanning path to obtain line laser point cloud data of the workpiece to be processed with absolute dimensional information. Step S23: The initial point cloud dataset and the line laser point cloud data are fused together. The absolute scale of the line laser point cloud data is used to correct the fuzzy scale of the initial point cloud dataset to obtain a high-precision point cloud dataset of the workpiece to be processed.

5. The method for teaching-free laser shock enhancement based on a three-dimensional model and laser point cloud as described in claim 1, characterized in that, The registration includes coarse registration and fine registration; the fine registration is performed after the coarse registration. The coarse registration is implemented as follows: global features are extracted from the high-precision point cloud dataset and the 3D information model, and then feature matching is used to quickly match the two sets of features in the high-precision point cloud dataset and the 3D information model to obtain an initial deviation matrix; the global features include edge contours, corner points, and reference holes; The implementation method of fine registration is as follows: the three-dimensional information model is transformed to the actual coordinate system of the workpiece to be processed using the initial deviation matrix, and the rough position of the area to be strengthened in the actual space is determined. The guide line laser sensor collects a high-density point cloud dataset of the area to be strengthened. The high-density point cloud dataset is iteratively registered with the corresponding area of ​​the three-dimensional information model after transformation using the initial deviation matrix using the nearest point ICP, and the deviation matrix is ​​optimized to obtain the deviation matrix.

6. The method for teaching-free laser shock enhancement based on a three-dimensional model and laser point cloud as described in claim 1, characterized in that, The procedure preceding step S4 also includes: Step S3': For each initial processing path, accessibility is assessed, unreachable paths are identified based on the spatial open angle algorithm, and the laser head processing posture is calculated for reachable paths; The steps for implementing the accessibility assessment include: Step S31': For each initial processing path, extract the center point and normal vector of the current region to be strengthened; Step S32': Construct a spatial open angle calculation domain with the center point as the center of the sphere and the working distance of the laser head as the radius; Step S33': Simulate the laser head model within the open angle calculation domain of the space, and detect collisions angle by angle. Calculate the proportion of the non-collision angle range to the total angle range, and use it as the open angle. If the open angle is less than the preset threshold, it is marked as an unreachable path; otherwise, it is a reachable path, and the center direction of the open angle is selected as the optimal processing posture.

7. The method for teaching-free laser shock enhancement based on a three-dimensional model and laser point cloud as described in claim 6, characterized in that, The step S3' further includes: A multi-objective optimization model with overlap uniformity constraints is established to optimize the processing sequence of multiple reachable paths. The optimal processing sequence is solved by using a genetic algorithm or particle swarm optimization algorithm, and the optimal processing sequence and the corresponding empty movement path are output.

8. The method for teaching-free laser shock enhancement based on a three-dimensional model and laser point cloud as described in claim 7, characterized in that, The multi-objective optimization model is as follows: In the formula, This is the total empty movement path length. To account for unevenness in the overlapping area. This is a penalty item for collision risk. The weighting coefficient for the total idle path length. This is a weighting coefficient for the unevenness of the overlapping area. This is the weighting coefficient for the collision risk penalty item, and It is between [0,1].

9. A teach-free laser shock enhancement system based on a three-dimensional model and laser point cloud, characterized in that, include: The 3D analysis module is used to analyze the 3D information model of the workpiece to be processed, automatically identify one or more areas to be strengthened, complete the initial path planning, and generate the initial processing path. The point cloud acquisition module, including a line laser sensor and a monocular camera, is used to acquire surface data of the workpiece to be processed and obtain a high-precision point cloud dataset. The point cloud processing module is connected to the point cloud acquisition module and the three-dimensional analysis module respectively. It is used to perform point cloud registration between the high-precision point cloud dataset of the workpiece to be processed and the three-dimensional information model, and calculate the deviation matrix of the workpiece to be processed. The path planning module, connected to the point cloud processing module, is used to perform coordinate transformation and point-by-point correction on the initial processing path according to the deviation matrix, and generate an actual processing path that matches the actual workpiece. The control module, connected to the path planning module, is used to generate a robot execution program based on the actual processing path and control the laser processing device to complete automated laser shock strengthening operations.

10. The teach-free laser shock enhancement system based on a three-dimensional model and laser point cloud according to claim 9, characterized in that, The path planning module also includes an accessibility analysis unit, a job sequencing unit, and an idle movement planning unit; The reachability analysis unit is used to evaluate the reachability of the initial processing path, identify unreachable paths, and calculate the laser head processing posture for reachable paths; The job sorting unit is used to execute the spatial open angle algorithm and detect and calculate the intersection point with the workpiece model; The air travel planning unit is used to optimize the processing sequence of multiple reachable paths using a multi-objective optimization model.

11. The teach-free laser shock enhancement system based on a three-dimensional model and laser point cloud as described in claim 9, characterized in that, The point cloud processing module further includes a point cloud registration unit and a solution unit; the point cloud registration unit is used to perform registration operations and to perform coarse registration and fine registration in sequence; the solution unit is used to solve the deviation matrix of the workpiece to be processed.

12. The teach-free laser shock enhancement system based on a three-dimensional model and laser point cloud as described in claim 9, characterized in that, The control module also includes a signal acquisition unit, a feature extraction unit, a parameter calculation and decision-making unit, and an instruction generation unit; The signal acquisition unit is used to acquire workpiece processing signals and start the robot to begin processing operations. The feature extraction unit is used to extract processing features and send the features back to the parameter calculation and decision unit; The parameter calculation and decision-making unit is used to call up the processing parameters in the historical processing database based on the processing characteristics, and to decide which parameters to use for processing; The instruction generation unit is used to generate and output the robot execution program required for laser shock reinforcement.

Images