Intelligent path planning method for robot to operate outside object

Abstract

The invention discloses an intelligent path planning method for a robot to work outside an object, and relates to the technical field of robot automation, and the intelligent path planning method comprises the steps of 1, primary machining path planning, 2, machining effect evaluation and 3, secondary machining scheme analysis. According to the method, the outer surface of the target object is subjected to region segmentation, the vertical visual angle depth image of each region is obtained, the optimal slice direction and the optimal slice width of each region are obtained through the neural network technology, and the slice accompanying path of each region is obtained through analysis according to the optimal slice direction and the optimal slice width; and analyzing to obtain a primary processing accompanying path of the outer surface of the target object, performing primary processing on the outer surface of the target object according to the primary processing accompanying path, analyzing a processing effect after the processing is completed, and analyzing a secondary processing scheme of the outer surface of the target object based on the processing effect to obtain a slicing effect with higher precision and better quality.

Description

Technical Field

[0001] This application relates to the field of robot automation technology, specifically to an intelligent path planning method for robots to perform external tasks. Background Technology

[0002] To enable robots to perform automated cleaning, spraying, polishing, and grinding on workpiece surfaces, the path and motion planning of the robotic arm on the workpiece surface is inevitably involved. Currently, the robotic arm's motion path can be planned by dividing the surface area of ​​the workpiece and combining it with specific processing requirements, but the processing accuracy still needs to be improved. Therefore, this application proposes an intelligent path planning method for robots to perform conformal operations on the outer surface of objects.

[0003] Existing technology, such as the invention patent application with publication number CN113012149A, discloses an intelligent cleaning robot path planning method and system, which includes: acquiring image data of feature parts and soiled areas of the object to be cleaned using a depth camera and performing image preprocessing; using a convolutional neural network to identify and segment the image data of the soiled areas of the object to be cleaned; and combining prior information about the shape and size of the object to be cleaned to plan the cleaning path for the cleaning robot. This invention can utilize image information to target and clean key parts and soiled areas, saving water resources, improving cleaning quality, and more effectively removing stains. It can flexibly equip high-pressure water guns or force-controlled floating brush heads to achieve different cleaning methods for robots in various scenarios. Multiple cleaning paths are designed for different types of equipment or parts to be cleaned, and each cleaning path is individually packaged as a cleaning path process package, improving the system's operability.

[0004] The above solutions have the following technical problems: 1. Current technologies mainly determine the cutting path of the robot on the target object through manual teaching or offline programming. For complex shapes and frequently changing cutting tasks, the programming workload is large and time-consuming, and it requires high professional skills and experience from the operators. In addition, the robot can usually only cut according to the preset program and path, and it is difficult to adjust the cutting strategy and path autonomously in real time according to the actual cutting situation. This results in insufficient intelligence of the current path planning method and the need to improve the processing accuracy.

[0005] 2. Current technology lacks analysis of the processing effect after processing the target object, and therefore lacks planning for secondary processing of the object's outer surface. The neglect of this aspect in current technology will lead to problems such as unstable processing quality and reduced processing accuracy, thus resulting in a lack of perfection and rationality in current technology. Summary of the Invention

[0006] The purpose of this application is to provide an intelligent path planning method for robots to perform external operations, which solves the problems existing in the background technology.

[0007] To solve the above technical problems, this application adopts the following technical solution: This application provides an intelligent path planning method for robots to perform external operations on objects, including: Step 1, primary processing path planning: analyzing the primary following processing path of the outer surface of the target object.

[0008] Step 2: Processing effect evaluation: Analyze the processing effect of the robot on the target object through a single processing path, and then determine whether the target object needs secondary processing.

[0009] Step 3: Secondary processing scheme analysis: Based on the analysis of the primary processing effect of the target object, a secondary processing scheme is obtained.

[0010] Preferably, the analysis of the one-time processing path of the outer surface of the target object is as follows: S1, perform region segmentation of the outer surface of the target object based on the normal vector of the outer surface of the target object, and obtain the vertical viewing angle depth image corresponding to each region.

[0011] S2. Using the vertical perspective depth image of each region as input, output the optimal slicing direction and optimal slicing width for each region through neural network technology.

[0012] S3. Slice the point cloud of each region based on the optimal slicing direction and the optimal slicing width to obtain the contour of the point cloud slice of each region. Generate the slice following path of each region based on the contour and normal vector of the point cloud slice of each region, and analyze the slice following path sequence of each region to obtain the first following processing path of the outer surface of the target object.

[0013] Preferably, the process of segmenting the outer surface of the target object based on the normal vector of the outer surface of the target object and obtaining the vertical viewing depth image corresponding to each region is as follows: A1. Obtain any point on the outer surface of the target object, and denote it as... Point, and get with Any number of points that are adjacent to each other and equidistant from each other are denoted as . The points adjacent to the point will be obtained Point and Connect any two adjacent points of a point to construct a... The planes containing the points are measured using a 3D laser scanner. Obtain the normal vectors of the planes containing the point. For each plane containing a point, the planes with the same normal vector are used to construct a region on the outer surface of the target object, and thus obtain each region on the outer surface of the target object.

[0014] A2. Use a depth camera to obtain the vertical viewing depth image of a region along the normal vector direction of a certain region on the outer surface of the target object, and then obtain the vertical viewing depth images corresponding to each region on the outer surface of the target object.

[0015] Preferably, the process of using the vertical viewing depth image of each region as input is as follows: The value of each pixel is obtained from the vertical viewing depth image of each region and recorded as... ,in The numbers corresponding to the vertical view depth images of each region. , It is any integer greater than 2. This refers to the number corresponding to each pixel in the vertical view depth image. , For any integer greater than 2, according to the calculation formula: The vertical view depth images of each region are standardized, among which... For the first The vertical perspective depth image of the region is the first The input value of each pixel. Indicates the first The minimum number of pixels in the vertical viewing depth image of each region. Indicates the first The maximum pixel value of the vertical viewing depth image of each region.

[0016] Preferably, the point cloud slicing of each region based on the optimal slicing direction and optimal slicing width to obtain the contour of each region's point cloud slice is as follows: B1. Establish a three-dimensional Cartesian coordinate system with any vertex of the target object as the origin, where The axis is horizontal. The axis is in the vertical direction. The axes are parallel to the spatial height, and the positive direction of each coordinate axis is determined using the right-hand rule. The preset length of the coordinate axis is set to represent one unit length, thereby obtaining the three-dimensional coordinate data of the point cloud of each region.

[0017] B2. The optimal slicing direction and optimal slicing width for each region of the outer surface of the target object are respectively denoted as... and Based on the 3D coordinate data of point clouds in each region, according to the plane The point cloud of each region is sliced, among which The numbers corresponding to each point in the point cloud. , It is any integer greater than 2. Indicates the first In the point cloud of each region, the first The cutting coordinates of each point are determined according to the optimal slicing direction of the target object's outer surface. According to the plane Construct slice trails for each region on the outer surface of the target object.

[0018] Preferably, the process of analyzing the slice path sequence of each region to obtain the first-order processing path of the outer surface of the target object is as follows: the optimal slice width of each region of the target object is compared with each other, and the region corresponding to the maximum slice width is obtained and recorded as the first slice region. Then, the optimal slice width of each adjacent region of the first slice region is compared with each other, and the region corresponding to the maximum slice width of each adjacent region of the first slice region is obtained and recorded as the second slice region. Based on this, the slice sequence of each region of the outer surface of the target object is obtained, and then the first-order processing path of the outer surface of the target object is constructed.

[0019] Preferably, the analysis robot analyzes the processing effect of the target object through a single processing path. The specific analysis process is as follows: Based on the processing effect of the target object's outer surface in one pass, the absolute value of the difference between the actual slice width and the optimal slice width for each region of the target object's outer surface is obtained. The absolute value of the difference between the actual slice width and the optimal slice width is compared with a preset slice width error threshold. When the absolute value of the difference between the actual slice width and the optimal slice width for a certain region is greater than the preset slice width error threshold, the robot's processing of the target object's outer surface in one pass is recorded as unqualified slice width; otherwise, the robot's processing of the target object's outer surface in one pass is recorded as qualified slice width.

[0020] Based on the processing effect of the target object's outer surface in one operation, the deviation angle between the actual slicing direction and the optimal slicing direction of the robot for each region of the target object's outer surface is obtained. The deviation angle between the actual slicing direction and the optimal slicing direction is compared with a preset threshold value for the deviation angle of the slicing direction. When the deviation angle between the actual slicing direction and the optimal slicing direction of a certain region is greater than the preset threshold value for the deviation angle of the slicing direction, the robot's processing of the object's outer surface in one operation is recorded as unqualified in slicing direction; otherwise, the robot's processing of the object's outer surface in one operation is recorded as qualified in slicing direction.

[0021] When the robot's slice width and slice direction are both qualified in a single processing of the target object's outer surface, the deviation angle values ​​of the slice width and slice direction in each region are obtained. The analysis yields the robot's single processing quality evaluation coefficient for the target object's outer surface. This coefficient is then compared with a preset threshold. If the robot's single processing quality evaluation coefficient is greater than or equal to the preset threshold, it indicates that the robot's single processing quality for the target object's outer surface is qualified; otherwise, it indicates that the robot's single processing quality for the target object's outer surface is unqualified.

[0022] Preferably, the process of obtaining the secondary processing scheme based on the analysis of the primary processing effect of the target object is as follows: If the slicing direction of the robot's primary processing of the outer surface of the target object is unqualified, the deviation angle value between the slicing direction of each region with unqualified slicing direction and the optimal slicing direction is obtained. Then, the deviation angle values ​​between the slicing direction of each region with unqualified slicing direction and the optimal slicing direction are sorted from largest to smallest. The secondary processing trajectory of the robot on each region of the outer surface of the target object is obtained according to the sorting result. Similarly, the secondary processing trajectory of the robot on each region of the outer surface of the target object is obtained when the processing quality of the robot's primary processing of the outer surface of the target object is unqualified.

[0023] If the quality of the first processing of the target object's outer surface by the robot is unqualified, the evaluation coefficient of the first processing quality of each region of the target object's outer surface by the robot is analyzed based on the slice width and slice direction deviation angle of each region of the target object's outer surface. The evaluation coefficients of the first processing quality of each region of the target object's outer surface by the robot are sorted from smallest to largest, and the secondary processing trajectory of each region of the target object's outer surface by the robot is obtained according to the sorting result.

[0024] Based on this, the secondary processing trajectory is obtained when the slice width, slice direction, and slice quality are unacceptable in the first processing of the outer surface of the target object by the robot, and the outer surface of the target object is processed in a secondary manner according to the secondary processing trajectory.

[0025] The beneficial effects of this application are as follows: 1. This application provides an intelligent path planning method for robots to perform external operations on objects. By segmenting the outer surface of the target object into regions and obtaining vertical perspective depth images of each region, the optimal slicing direction and optimal slicing width of each region are obtained through neural network technology. Based on the optimal slicing direction and optimal slicing width, the slicing following path of each region is analyzed, and then the primary processing following path of the outer surface of the target object is obtained. Based on this, the outer surface of the target object is processed once. After the processing is completed, the processing effect is analyzed, and the secondary processing scheme of the outer surface of the target object is analyzed based on the processing effect to obtain a slicing effect with higher accuracy and better quality.

[0026] 2. This application segments the outer surface of the target object into regions and obtains vertical perspective depth images of each region, laying the foundation for the analysis of the subsequent follow-up processing path of the target object's outer surface. Based on the vertical perspective depth images of each region, neural network technology is used to determine the optimal slicing direction and optimal slicing width for each region. This ensures that errors are minimized during the cutting process, improving the slicing accuracy of each region, resulting in a smoother and neater cut surface, thus improving slicing quality. It also effectively avoids problems such as burrs and deformation caused by improper slicing angles, thereby improving the processing and assembly accuracy of the target object. Simultaneously, it avoids the waste of raw materials caused by improper slicing angles and directions. Finally, based on the optimal slicing direction and optimal slicing width, the slicing follow-up path of each region is analyzed, and then the primary follow-up processing path of the target object's outer surface is obtained. This ensures that the robot slices the target object's outer surface according to the optimal path, reducing cutting time and improving production efficiency. This efficiency improvement is particularly significant when handling complex-shaped objects or large-scale production tasks, helping to reduce production costs and improve economic benefits.

[0027] 3. This application analyzes the effect of a robot's primary processing of the target object's outer surface, and analyzes the secondary processing scheme based on the obtained primary processing effect. This greatly ensures the slicing accuracy of the target object's outer surface, improves the processing quality of the target object, and thus obtains a smoother and more accurate cut surface. This ensures the perfection and rationality of the robot's intelligent path planning method for external object operations. Attached Figure Description

[0028] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 This is a flowchart illustrating the steps involved in implementing the method described in this application. Detailed Implementation

[0030] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0031] Reference Figure 1As shown, this application provides an intelligent path planning method for robots to perform external operations on objects, including the following steps: Step 1, primary processing path planning: analyzing the primary following processing path of the outer surface of the target object.

[0032] In a specific example, the analysis of the processing path of the outer surface of the target object is as follows: S1, the outer surface of the target object is segmented into regions based on the normal vector of the outer surface of the target object, and the vertical viewing depth image corresponding to each region is obtained.

[0033] S2. Using the vertical perspective depth image of each region as input, output the optimal slicing direction and optimal slicing width for each region through neural network technology.

[0034] S3. Slice the point cloud of each region based on the optimal slicing direction and the optimal slicing width to obtain the contour of the point cloud slice of each region. Generate the slice following path of each region based on the contour and normal vector of the point cloud slice of each region, and analyze the slice following path sequence of each region to obtain the first following processing path of the outer surface of the target object.

[0035] It should be noted that the neural network technology uses the vertical perspective depth image of the outer surface of the target object as input neurons, and outputs two output neurons through the convolutional neural network model, namely the optimal slice direction and the optimal slice width. The training method of the convolutional neural network model is an existing technology, so it will not be described in detail.

[0036] In a specific example, the process of segmenting the outer surface of the target object based on the normal vector of the target object's outer surface and obtaining the vertical viewing depth image corresponding to each region is as follows: A1. Obtain any point on the outer surface of the target object, and denote it as... Point, and get with Any number of points that are adjacent to each other and equidistant from each other are denoted as . The points adjacent to the point will be obtained Point and Connect any two adjacent points of a point to construct a... The planes containing the points are measured using a 3D laser scanner. Obtain the normal vectors of the planes containing the point. For each plane containing a point, the planes with the same normal vector are used to construct a region on the outer surface of the target object, and thus obtain each region on the outer surface of the target object.

[0037] A2. Use a depth camera to obtain the vertical viewing depth image of a region along the normal vector direction of a certain region on the outer surface of the target object, and then obtain the vertical viewing depth images corresponding to each region on the outer surface of the target object.

[0038] It should be noted that, The planes containing the points coincide with the outer surface of the target object.

[0039] In a specific example, the process of taking the vertical viewing depth image of each region as input is as follows: A1. Obtain the value of each pixel from the vertical viewing depth image of each region and record it as... ,in The numbers corresponding to the vertical view depth images of each region. , It is any integer greater than 2. This refers to the number corresponding to each pixel in the vertical view depth image. , For any integer greater than 2, according to the calculation formula: The vertical view depth images of each region are standardized, among which... For the first The vertical perspective depth image of the region is the first The input value of each pixel. Indicates the first The minimum number of pixels in the vertical viewing depth image of each region. Indicates the first The maximum pixel value of the vertical viewing depth image of each region.

[0040] It should be noted that the input values ​​of each pixel in the vertical view depth image of each region are within the range Inside.

[0041] In a specific example, the point cloud slicing of each region based on the optimal slicing direction and optimal slicing width to obtain the contour of each region's point cloud slice is as follows: B1. Establish a three-dimensional Cartesian coordinate system with any vertex of the target object as the origin, where... The axis is horizontal. The axis is in the vertical direction. The axes are parallel to the spatial height, and the positive direction of each coordinate axis is determined using the right-hand rule. The preset length of the coordinate axis is set to represent one unit length, thereby obtaining the three-dimensional coordinate data of the point cloud of each region.

[0042] B2. The optimal slicing direction and optimal slicing width for each region of the outer surface of the target object are respectively denoted as... and Based on the 3D coordinate data of point clouds in each region, according to the plane The point cloud of each region is sliced, among which The numbers corresponding to each point in the point cloud. , It is any integer greater than 2. Indicates the first In the point cloud of each region, the first The cutting coordinates of each point are determined according to the optimal slicing direction of the target object's outer surface. According to the plane Construct slice trails for each region on the outer surface of the target object.

[0043] In a specific example, the process of analyzing the slice path sequence of each region to obtain the first-order processing path of the outer surface of the target object is as follows: the optimal slice width of each region of the target object is compared with each other, and the region corresponding to the maximum slice width is obtained and recorded as the first slice region. Then, the optimal slice width of each adjacent region of the first slice region is compared with each other, and the region corresponding to the maximum slice width of each adjacent region of the first slice region is obtained and recorded as the second slice region. Based on this, the slice sequence of each region of the outer surface of the target object is obtained, and then the first-order processing path of the outer surface of the target object is constructed.

[0044] Step 2: Processing effect evaluation: Analyze the processing effect of the robot on the target object through a single processing path, and then determine whether the target object needs secondary processing.

[0045] In a specific example, the analysis robot analyzes the processing effect of the target object through a single processing path. The specific analysis process is as follows: Based on the processing effect of the target object's outer surface in one pass, the absolute value of the difference between the actual slice width and the optimal slice width for each region of the target object's outer surface is obtained. The absolute value of the difference between the actual slice width and the optimal slice width is compared with a preset slice width error threshold. When the absolute value of the difference between the actual slice width and the optimal slice width for a certain region is greater than the preset slice width error threshold, the robot's processing of the target object's outer surface in one pass is recorded as unqualified slice width; otherwise, the robot's processing of the target object's outer surface in one pass is recorded as qualified slice width.

[0046] Based on the processing effect of the target object's outer surface in one operation, the deviation angle between the actual slicing direction and the optimal slicing direction of the robot for each region of the target object's outer surface is obtained. The deviation angle between the actual slicing direction and the optimal slicing direction is compared with a preset threshold value for the deviation angle of the slicing direction. When the deviation angle between the actual slicing direction and the optimal slicing direction of a certain region is greater than the preset threshold value for the deviation angle of the slicing direction, the robot's processing of the object's outer surface in one operation is recorded as unqualified in slicing direction; otherwise, the robot's processing of the object's outer surface in one operation is recorded as qualified in slicing direction.

[0047] When the robot's slice width and slice direction are both qualified in a single processing of the target object's outer surface, the deviation angle values ​​of the slice width and slice direction in each region are obtained. The analysis yields the robot's single processing quality evaluation coefficient for the target object's outer surface. This coefficient is then compared with a preset threshold. If the robot's single processing quality evaluation coefficient is greater than or equal to the preset threshold, it indicates that the robot's single processing quality for the target object's outer surface is qualified; otherwise, it indicates that the robot's single processing quality for the target object's outer surface is unqualified.

[0048] In a specific example, the analysis yields a quality assessment coefficient for the robot's first-pass machining of the target object's outer surface. The specific analysis process is as follows: the slice width and slice direction deviation angle values ​​for each region are denoted as... and According to the calculation formula: The analysis yielded the evaluation coefficient for the robot's single-process machining of the target object's outer surface. ,in and These represent the weighting factors corresponding to the slice width and the slice direction deviation angle during a single processing of the target object's outer surface by the robot.

[0049] It should be noted that, , , .

[0050] It should be noted that the weighting factors corresponding to the slice width error value and the slice direction deviation angle value during the first processing of the target object's outer surface by the robot are obtained through factor analysis. First, the spatial path attenuation information of the slice width error value and slice direction deviation angle value during the first processing of the target object's outer surface by each farmer's robot is condensed. Then, the variance explained rate after rotation is obtained, and the weights are obtained by dividing the accumulated variance explained rate.

[0051] It should be noted that factor analysis is a well-known technique. It is a multivariate statistical analysis method that starts by studying the internal dependencies of variables and reduces some variables with complex relationships to a few comprehensive factors. Information condensation is expressed as the calculation of the median. The variance explained rate is the amount of information extracted by the factors. Variance explained rate = eigenvalues ​​ / total number of analysis terms. The rotated variance explained rate is expressed as the variance explained by the factors after maximum variance rotation.

[0052] In a specific example, the process of determining whether a target object needs secondary processing is as follows: when the slice width of the first processing of the outer surface of the target object by the robot is not up to standard, it is determined that the target object needs secondary processing.

[0053] If the slicing direction of the robot's first processing of the outer surface of the target object is not up to standard, it is determined that the target object needs to be processed a second time.

[0054] When the quality of the first processing of the outer surface of the target object by the robot is not up to standard, it is determined that the target object needs to be processed a second time.

[0055] Step 3: Secondary processing scheme analysis: Based on the analysis of the primary processing effect of the target object, a secondary processing scheme is obtained.

[0056] In a specific example, the process of obtaining a secondary processing scheme based on the analysis of the primary processing effect of the target object is as follows: If the slice width of the primary processing of the outer surface of the target object by the robot is unqualified, the absolute value of the difference between the slice width of each region with unqualified slice width and the optimal slice width is obtained. Then, the absolute values ​​of the difference between the slice width of each region with unqualified slice width and the optimal slice width are sorted from largest to smallest. The secondary processing trajectory of each region of the outer surface of the target object by the robot is obtained according to the sorting result.

[0057] If the slicing direction of the robot's primary processing of the target object's outer surface is unqualified, the deviation angle value between the slicing direction of each region with unqualified slicing direction and the optimal slicing direction is obtained. Then, the deviation angle values ​​between the slicing direction of each region with unqualified slicing direction and the optimal slicing direction are sorted from largest to smallest. Based on the sorting result, the secondary processing trajectory of the robot for each region of the target object's outer surface is obtained. Similarly, the secondary processing trajectory of the robot for each region of the target object's outer surface when the quality of the primary processing of the target object's outer surface is unqualified is obtained.

[0058] Based on this, the secondary processing trajectory is obtained when the slice width, slice direction, and slice quality are unacceptable in the first processing of the outer surface of the target object by the robot, and the outer surface of the target object is processed in a secondary manner according to the secondary processing trajectory.

[0059] It should be noted that the evaluation coefficient of the first processing quality of each region is calculated in the same way as the evaluation coefficient of the first processing of the target object's outer surface by the robot, so it will not be repeated here.

[0060] This application provides an intelligent path planning method for robots to perform external operations on objects. The method involves segmenting the outer surface of the target object into regions and acquiring vertical perspective depth images of each region. Then, neural network technology is used to obtain the optimal slicing direction and width for each region. Based on the optimal slicing direction and width, the slicing path for each region is analyzed, leading to the first-stage processing path for the target object's outer surface. The target object's outer surface is then processed once. After processing, the processing effect is analyzed, and a secondary processing scheme for the target object's outer surface is analyzed based on the processing effect to obtain slicing results with higher accuracy and better quality.

[0061] The above content is merely an example and illustration of the concept of this application. Those skilled in the art can make various modifications or additions to the specific embodiments described or use similar methods to replace them, as long as they do not deviate from the inventive concept or exceed the scope defined in this application, they should all fall within the protection scope of this application.

Claims

1. An intelligent path planning method for robots performing tasks outside objects, characterized in that, include: Step 1: Single-step machining path planning: Analyze the single-step machining path of the outer surface of the target object; Step 2, Processing Effect Evaluation: Analyze the processing effect of the robot on the target object through a single following processing path, and then determine whether the target object needs secondary processing; Step 3: Secondary processing scheme analysis: Based on the analysis of the primary processing effect of the target object, a secondary processing scheme is obtained.

2. The intelligent path planning method for robots performing tasks outside objects according to claim 1, characterized in that, The analysis process for a single processing path along the outer surface of the target object is as follows: S1. Based on the normal vector of the outer surface of the target object, perform region segmentation on the outer surface of the target object and obtain the vertical viewing depth image corresponding to each region; S2. Using the vertical perspective depth image of each region as input, output the optimal slicing direction and optimal slicing width of each region through neural network technology; S3. Slice the point cloud of each region based on the optimal slicing direction and the optimal slicing width to obtain the contour of the point cloud slice of each region. Generate the slice following path of each region based on the contour and normal vector of the point cloud slice of each region, and analyze the slice following path sequence of each region to obtain the first following processing path of the outer surface of the target object.

3. The intelligent path planning method for robots performing tasks outside objects according to claim 2, characterized in that, The process of segmenting the outer surface of the target object based on the normal vector of the outer surface and obtaining the vertical viewing depth image corresponding to each region is as follows: A1. Obtain any point on the outer surface of the target object, and denote it as... Point, and get with Any number of points that are adjacent to each other and equidistant from each other are denoted as . The points adjacent to the point will be obtained Point and Connect any two adjacent points of a point to construct a... The planes containing the points are measured using a 3D laser scanner. Obtain the normal vectors of the planes containing the point. For each plane containing a point, the planes with the same normal vector are used to construct a region on the outer surface of the target object, and thus obtain each region on the outer surface of the target object. A2. Use a depth camera to obtain the vertical viewing depth image of a region along the normal vector direction of a certain region on the outer surface of the target object, and then obtain the vertical viewing depth images corresponding to each region on the outer surface of the target object.

4. The intelligent path planning method for robots performing tasks outside objects according to claim 2, characterized in that, The process of taking the vertical perspective depth image of each region as input is as follows: The values ​​of each pixel are obtained from the vertical perspective depth images of each region and recorded as follows: ,in The numbers corresponding to the vertical view depth images of each region. , It is any integer greater than 2. This refers to the number corresponding to each pixel in the vertical view depth image. , For any integer greater than 2, according to the calculation formula: The vertical view depth images of each region are standardized, among which... For the first The vertical perspective depth image of the region is the first The input value of each pixel. Indicates the first The minimum number of pixels in the vertical viewing depth image of each region. Indicates the first The maximum pixel value of the vertical viewing depth image of each region.

5. The intelligent path planning method for robots performing tasks outside objects according to claim 2, characterized in that, The point cloud is sliced ​​based on the optimal slicing direction and optimal slicing width to obtain the contour of each region's point cloud slice. The specific process is as follows: B1. Establish a three-dimensional rectangular coordinate system with any vertex of the target object as the origin, where The axis is horizontal. The axis is in the vertical direction. The axes are parallel to the spatial height, and the positive direction of each coordinate axis is determined by the right-hand rule. The preset length of the coordinate axis is set to represent a unit length, thereby obtaining the three-dimensional coordinate data of the point cloud of each region. B2. The optimal slicing direction and optimal slicing width for each region of the outer surface of the target object are respectively denoted as... and Based on the 3D coordinate data of point clouds in each region, according to the plane The point cloud of each region is sliced, among which The numbers corresponding to each point in the point cloud. , It is any integer greater than 2. Indicates the first In the point cloud of each region, the first The cutting coordinates of each point are determined according to the optimal slicing direction of the target object's outer surface. According to the plane Construct slice trails for each region on the outer surface of the target object.

6. The intelligent path planning method for robots performing tasks outside objects according to claim 5, characterized in that, The analysis of the slice following path sequence in each region yields the first following processing path for the outer surface of the target object. The specific process is as follows: The optimal slice widths of each region of the target object are compared with each other, and the region corresponding to the maximum slice width is obtained and recorded as the first slice region. Then, the optimal slice widths of each adjacent region of the first slice region are compared with each other, and the region corresponding to the maximum slice width of each adjacent region of the first slice region is obtained and recorded as the second slice region. Based on this, the slice order of each region of the outer surface of the target object is obtained, and then a follow-up processing path of the outer surface of the target object is constructed.

7. The intelligent path planning method for robots performing tasks outside objects according to claim 6, characterized in that, The analysis process of the robot's processing effect on the target object through a single following processing path is as follows: Based on the processing effect of the target object's outer surface in one operation, the absolute value of the difference between the actual slice width and the optimal slice width of each region of the target object's outer surface by the robot is obtained. The absolute value of the difference between the actual slice width and the optimal slice width is compared with a preset slice width error threshold. When the absolute value of the difference between the actual slice width and the optimal slice width of a certain region is greater than the preset slice width error threshold, the robot's processing of the target object's outer surface in one operation is recorded as unqualified slice width; otherwise, the robot's processing of the target object's outer surface in one operation is recorded as qualified slice width. Based on the processing effect of the outer surface of the target object, the deviation angle between the actual slicing direction and the optimal slicing direction of the robot for each region of the outer surface of the target object is obtained. The deviation angle between the actual slicing direction and the optimal slicing direction is compared with a preset threshold value. When the deviation angle between the actual slicing direction and the optimal slicing direction of a certain region is greater than the preset threshold value, the first processing of the outer surface of the object by the robot is recorded as unqualified in slicing direction; otherwise, the first processing of the outer surface of the object by the robot is recorded as qualified in slicing direction. When the robot's slice width and slice direction are both qualified in a single processing of the target object's outer surface, the deviation angle values ​​of the slice width and slice direction in each region are obtained. The analysis yields the robot's single processing quality evaluation coefficient for the target object's outer surface. This coefficient is then compared with a preset threshold. If the robot's single processing quality evaluation coefficient is greater than or equal to the preset threshold, it indicates that the robot's single processing quality for the target object's outer surface is qualified; otherwise, it indicates that the robot's single processing quality for the target object's outer surface is unqualified.

8. The intelligent path planning method for robots performing tasks outside objects according to claim 7, characterized in that, The analysis yields a quality assessment coefficient for the robot's first-pass machining of the target object's outer surface. The specific analysis process is as follows: The slice width and slice direction deviation angle of each region are respectively denoted as... and According to the calculation formula: The analysis yielded the evaluation coefficient for the robot's single-process machining of the target object's outer surface. ,in and These represent the weighting factors corresponding to the slice width and the slice direction deviation angle during a single processing of the target object's outer surface by the robot.

9. The intelligent path planning method for robots performing tasks outside objects according to claim 8, characterized in that, The specific process for determining whether the target object requires secondary processing is as follows: When the width of the slice processed by the robot on the outer surface of the target object in one operation is not up to standard, it is determined that the target object needs to be processed a second time. When the slicing direction of the robot's first processing of the outer surface of the target object is not up to standard, it is determined that the target object needs to be processed a second time. When the quality of the first processing of the outer surface of the target object by the robot is not up to standard, it is determined that the target object needs to be processed a second time.

10. The intelligent path planning method for robots performing tasks outside objects according to claim 9, characterized in that, The secondary processing scheme is obtained based on the analysis of the primary processing effect of the target object. The specific process is as follows: If the slicing direction of the robot's first processing of the outer surface of the target object is unqualified, the deviation angle value between the slicing direction of each region with unqualified slicing direction and the optimal slicing direction is obtained. Then, the deviation angle values ​​between the slicing direction of each region with unqualified slicing direction and the optimal slicing direction are sorted from largest to smallest. According to the sorting result, the secondary processing trajectory of the robot for each region of the outer surface of the target object is obtained. Similarly, the secondary processing trajectory of the robot for each region of the outer surface of the target object when the processing quality of the robot's first processing of the outer surface of the target object is unqualified is obtained. If the quality of the first processing of the outer surface of the target object by the robot is unqualified, the quality evaluation coefficient of the first processing of each region of the outer surface of the target object is analyzed based on the slice width and slice direction deviation angle of each region of the outer surface of the target object. The quality evaluation coefficient of the first processing of each region of the outer surface of the target object by the robot is sorted from small to large, and the secondary processing trajectory of each region of the outer surface of the target object by the robot is obtained according to the sorting result. Based on this, the secondary processing trajectory is obtained when the slice width, slice direction, and slice quality are unacceptable in the first processing of the outer surface of the target object by the robot, and the outer surface of the target object is processed in a secondary manner according to the secondary processing trajectory.

Images