Additive method and system for guiding a robot based on target 3D vision laser positioning

Abstract

The application discloses an additive method and system for guiding a robot based on target 3D visual laser positioning, the method comprising: establishing a digital model of a workpiece; obtaining attitude information of the workpiece through a 3D sensor; scanning the target to obtain point cloud data; and then processing the point cloud data, taking the digital model as a source point cloud data set S, and taking the scanned target point cloud as a target point cloud data set T, performing registration, sending the registration result data to the robot, guiding the robot to perform additive welding, and repeating the process until the whole workpiece is welded. The application automatically positions an additive welding starting point through a 3D visual system, guides the robot to intelligently work, thereby improving additive welding efficiency, reducing additive welding cost, and being applicable to workpieces of different sizes and to the case that the workpiece is continuously displaced during production.

Description

Technical Field

[0001] This invention relates to the field of robotic welding additive manufacturing technology, and in particular to an additive manufacturing method and system for a robot guided by target 3D vision laser positioning. Background Technology

[0002] Currently, most welding additive manufacturing in my country still relies on traditional manual methods. Due to the large workload and high labor intensity, welding quality is easily compromised, and prolonged exposure to welding poses safety risks and potential physical harm to operators. With the rapid development of intelligent manufacturing and the widespread adoption of industrial robots, these robots are being rapidly applied in the field of welding additive manufacturing. The high quality, high efficiency, and stable welding achieved using industrial robots in fully automated welding have been widely recognized.

[0003] Current welding additive manufacturing technology requires technicians to design the additive manufacturing trajectory based on the surface shape of the workpiece. Then, offline programming of robot code generates the additive motion trajectory, and the robot executes the code to perform the additive manufacturing operation according to the trajectory. The additive height of each layer is set based on the operator's experience, making it impossible to control the quality of each layer during the additive manufacturing process. If problems occur during the additive manufacturing process, it will lead to waste of raw materials and an increase in the total cost of rework.

[0004] After several layers of additive manufacturing, if the production process requires repositioning of the workpiece, the additive trajectory needs to be redesigned. Offline programming of robot code generates the additive motion trajectory, and the robot executes the code to perform the additive operation. This process involves several workpiece repositionings throughout the production process, resulting in low automation, longer product cycle times, and compromised additive quality.

[0005] For example, Chinese patent document CN108145332A, published on June 12, 2018, discloses a robotic arc welding additive and subtractive forming apparatus and method. The apparatus involved includes a positioner, a work platform, a substrate, a wire feeder, a robot control cabinet, a robot, a welding torch and its welding power supply, and a laser cutting device and its power supply. A single robot is connected to both the welding additive equipment and the laser cutting equipment. The robot control cabinet communicates with both the welding and cutting power supplies, controlling their alternating operation. This method utilizes CAD technology to create a corresponding three-dimensional model of the product to be processed, obtaining processing control information codes. Then, it sequentially uses a robotic arc welding additive and laser cutting subtractive composite processing method, following a pre-set path, to process the product and obtain the required qualified parts.

[0006] It is evident that existing robotic additive manufacturing technologies require individual programming for workpieces of different shapes and sizes, resulting in low efficiency and increased labor and project management costs. Summary of the Invention

[0007] The main objective of this invention is to provide an additive manufacturing method and system based on target 3D vision laser positioning and robot guidance. This method aims to eliminate the need for manual pre-setting of additive manufacturing trajectories for different workpieces. After the workpiece is repositioned, the 3D vision system automatically locates the starting point of additive welding and guides the robot to operate, thereby improving additive welding efficiency and reducing additive welding costs.

[0008] To achieve the above objectives, the present invention provides an additive manufacturing method for a target-guided 3D vision laser positioning robot, the method comprising the following steps:

[0009] S1: Establish a digital model of the workpiece to be added with materials and extract the source point cloud dataset of the workpiece to be added with materials;

[0010] S2: Use a 3D vision system to obtain the orientation information of the workpiece to be additively manufactured;

[0011] S3: Using a 3D vision system, scan the target set on the workpiece to be added based on the welding signal to obtain a target point cloud dataset;

[0012] S4: Register the source point cloud dataset and the target point cloud dataset so that the data points in the source point cloud dataset correspond to those in the target point cloud dataset;

[0013] S5: Send the registration results of the source point cloud dataset and the target point cloud dataset to the robot, so that the robot moves to the welding position corresponding to the posture of the workpiece to be added based on the registration results, and performs additive welding.

[0014] Optionally, after step S3, the method further includes: preprocessing the target point cloud data to remove noise and discrete points; wherein the preprocessing includes: noise reduction processing, homogenization correction processing, and interpolation completion processing.

[0015] Optionally, after the homogenization correction process, the target point cloud data has a row of data formed by several independent points along the laser extension direction of the 3D vision system; wherein, the independent points include square points and cross mark points, the square points are data points that are not vibrated during the acquisition process, and the cross mark points are data points that are distorted due to vibration during the acquisition process.

[0016] Optionally, the interpolation completion specifically includes: after the target point cloud data is homogenized and corrected, for the interference points, blind collection points and reflective invalid points that appear, the collected target point cloud data is interpolated and completed using a method based on the domain information to continuously expand the boundary.

[0017] Optionally, step S4 specifically includes: performing ICP precise registration processing on the source point cloud dataset and the target point cloud dataset to obtain the registration result; and transforming the matrix... After the transformation, the source point cloud dataset corresponds to the data points in the target point cloud dataset;

[0018] The expression corresponding to the data points in the source point cloud dataset and the target point cloud dataset is specifically as follows:

[0019]

[0020] In the formula, It is a translation matrix. For rotation matrix, This represents the source point cloud data for point p. Target point cloud data representing point p; translation matrix With rotation matrix Make the objective function Take the minimum value, at this time and It is excellent.

[0021] Optionally, the step of performing ICP precise registration processing on the source point cloud dataset and the target point cloud dataset specifically involves: establishing a KD tree between the initially transformed source point cloud data and the target point cloud data to accelerate nearest neighbor search; and using a normal vector threshold... Remove erroneous point pairs and use singular value decomposition to calculate the transformation matrix. , so that the objective function Minimize the minimum value and iterate repeatedly to obtain the optimal registration result.

[0022] Optionally, in step S5, the method further includes: before and after single-layer additive manufacturing, the robot performs data scanning on the additive surface of the workpiece to be additively manufactured, compares the data before and after additive manufacturing, performs fixed-point sampling, performs least squares plane fitting on the sampling area, calculates the additive thickness, and determines whether the workpiece is qualified based on the thickness of the current layer of additive manufacturing; wherein, the least squares method is implemented using the method of solving overdetermined equations based on SVD singular value decomposition.

[0023] Optionally, least squares plane fitting is performed on the sampling region; where:

[0024] (1) The expression for the plane is as follows:

[0025]

[0026] In the formula, a, b, c, and d are planar parameters, and x, y, and z are the coordinates of points in the point cloud data;

[0027] (2) The constraints are:

[0028]

[0029] (3) The plane needs to satisfy:

[0030]

[0031] In the formula, di is any point p in the point cloud data. i (x i , y i , z i The distance d from the fitted plane. i =|ax i +by i +cz i +d|.

[0032] Optionally, in step S5, the method further includes:

[0033] The digital model established in step S1 is used to obtain the additive thickness and the number of additive layers required for each posture, and to determine whether the additive surface of the workpiece in the current posture has been welded.

[0034] If the welding of the additive surface in the current posture is not completed, adjust the height of the robot according to the thickness of the additive material and continue welding until the welding of the additive surface in the current posture is completed.

[0035] If the current orientation of the additive surface is welded, move the angle of the workpiece to be added and judge whether the overall welding has been completed; if not, return to step S2, obtain the current orientation information of the workpiece to be added, and continue to steps S3-S5 until the entire workpiece is welded.

[0036] Furthermore, to achieve the above objectives, the present invention also provides an additive manufacturing system based on a target-based 3D vision laser positioning and guiding robot, used to execute an additive manufacturing method based on a target-based 3D vision laser positioning and guiding robot. The system includes: a mounting frame, a robot, and a welding torch. The welding torch is mounted on the working end of the robot, and the working end of the robot drives the welding torch to move and adjust the welding position. The robot is mounted on the mounting frame, and the mounting frame adjusts the overall displacement of the robot. The working end of the robot is also equipped with a 3D sensor, which is used to scan the workpiece to be additively manufactured.

[0037] The technical solution of the present invention has the following beneficial effects:

[0038] (1) The entire process of implementing this invention can be fully automated, which not only reduces costs but also improves production efficiency;

[0039] (2) By introducing an additive welding 3D vision system, this invention not only improves the welding additive yield, but also improves the level of digital production in the factory. Moreover, the entire 3D vision system is based on the industrial Internet platform, which can monitor the robot status, welding status and quality in real time, operate remotely and diagnose remotely.

[0040] (3) The present invention has product adaptability, can be compatible with different products of various specifications, realizes programming-free, greatly reduces the intensity of manual labor, and realizes automated, intelligent and digital additive welding;

[0041] (4) The present invention establishes a digital model of the additive workpiece to be welded, and registers it with the scanned target point cloud data to achieve automatic positioning. Using the target method, no matter how the additive workpiece is displaced, as long as the target is within the field of view of the sensor, the starting point of welding can be automatically located. Attached Figure Description

[0042] Figure 1 This is a flowchart illustrating an additive manufacturing method for a target-based 3D vision laser positioning and guidance robot according to an embodiment of the present invention.

[0043] Figure 2 This is a schematic diagram after the homogenization correction algorithm is applied in an embodiment of the present invention.

[0044] Figure 3 This is a schematic diagram illustrating the process of building a KD tree to accelerate nearest neighbor search in an embodiment of the present invention.

[0045] Figure 4 This is a schematic diagram illustrating the registration process of partial data collected by a line laser sensor in an embodiment of the present invention.

[0046] Figure 5 This is a schematic diagram of an additive manufacturing system for a target-based 3D vision laser positioning and guiding robot performing additive manufacturing operations, as described in an embodiment of the present invention.

[0047] Reference numerals: 1-Mounting frame; 2-Lifting platform; 3-Guide rail; 4-Robot; 5-Welding torch; 6-3D sensor and protective cover; 7-Workpiece to be added material; 8-Target; 9-Blade; 10-Blade with completed welding and additive manufacturing; 11-Angle adjustment shaft.

[0048] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0049] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0050] Currently, in the relevant technical fields, existing robotic additive manufacturing technologies require programming each workpiece of different shapes and sizes individually, which is inefficient and increases labor and project management costs.

[0051] To address this issue, various embodiments of the additive manufacturing method and system based on target 3D vision laser positioning and guidance robot of the present invention are proposed. The additive manufacturing method and system based on target 3D vision laser positioning and guidance robot provided by the present invention automatically locates the starting point of additive welding through a 3D vision system, guiding the robot to operate intelligently, thereby improving additive welding efficiency, reducing additive welding costs, and being applicable to workpieces of different sizes and situations where the workpieces are constantly shifting during the production process.

[0052] Example 1:

[0053] This invention provides an additive manufacturing method for a target-guided 3D vision laser positioning robot, referring to... Figure 1 , Figure 1 This is a flowchart illustrating an embodiment of the additive manufacturing method for a target-based 3D vision laser positioning and guidance robot according to the present invention.

[0054] In this embodiment, the additive manufacturing method for a target-based 3D vision laser positioning and guidance robot includes the following steps:

[0055] S1. Establish a digital model of the workpiece to be added;

[0056] The S2.3D vision system acquires the posture information of the workpiece to be additively manufactured;

[0057] The S3.3D vision system scans the target based on the welding signal to obtain target point cloud data;

[0058] The S4 3D vision system preprocesses the obtained point cloud data to remove noise and discrete points.

[0059] The preprocessing performed includes, in sequence:

[0060] S41. Perform mean filtering and noise reduction on the point cloud data to filter out interference points.

[0061] Specifically, by calculating the first difference (i.e., taking the derivative) of all data points, the rate of change between adjacent data points is obtained, and data points with a rate of change significantly higher than the mean rate of change are filtered out.

[0062] S42. Since the robotic arm carrying the data acquisition equipment will inevitably vibrate during the data acquisition process, the acquired data will be uneven and produce jitter and distortion. Therefore, the point cloud data after noise reduction needs to be straightened.

[0063] After performing uniformity correction on the noise-reduced point cloud data, along the laser extension direction of the 3D vision system, each independent point, including square points and crosshair mark points, forms a neat row of data; the square points are 3D points acquired without vibration during the acquisition process, while the crosshair mark points are distorted data points obtained during actual acquisition due to vibration. Figure 2 As shown, this is a side view of each row of data collected along the laser extension direction. The arrows indicate the effect after processing by the homogenization correction algorithm.

[0064] S43. Due to the inevitable occurrence of data interference points, blind acquisition points, and invalid reflection points during the acquisition of 3D point cloud data, after the point cloud data is aligned and rectified, a method based on neighborhood information to continuously expand the boundary is used to perform interpolation completion processing on the acquired 3D point cloud data.

[0065] S5. After completing the above point cloud data processing operations, using the digital model from step S1 as the source point cloud dataset S and the scanned target point cloud data as the target point cloud dataset T, perform improved ICP (Iterative Closest Point) precise registration processing to obtain the registration result, and then transform the matrix... After transformation, the data points in the source point cloud dataset correspond to those in the target point cloud dataset. The corresponding formula is:

[0066]

[0067] in: It is a translation matrix. Let sp be the rotation matrix, and tp be the target point cloud data of point p; and tp be the translation matrix. With rotation matrix Make the objective function Take the minimum value, at this time and It is excellent;

[0068] The improved ICP precise registration establishes a KD tree between the initially transformed source point cloud data and the target point cloud data to accelerate nearest neighbor search;

[0069] Thresholding by normal vector Remove erroneous point pairs and use singular value decomposition to calculate the transformation matrix. , so that the objective function The minimum value is obtained through iterative iteration to achieve the optimal registration result. Its working principle is as follows: Figure 3 As shown.

[0070] Reference Figure 4 , Figure 4 It is the process of registering a portion of the data collected by the line laser sensor of the 3D vision system.

[0071] Figure 4 In the middle, the right side is the target point cloud data map to be registered (dark target point cloud data on the left side will be generated during registration, which is the same as the target point cloud data on the right side, so that the dark target point cloud data on the left side will gradually be registered to the white point cloud data on the left side), and white is the source point cloud data; the target point cloud data needs to be registered to the source point cloud. During the registration process, each iteration can find that the dark target point cloud data map on the left side moves closer to the source point cloud data.

[0072] S6. Send the registration result data obtained in step S5 to the robot, guide the robot to move to the welding position corresponding to the posture of the workpiece to be added, and perform additive welding.

[0073] S7. Before and after single-layer additive manufacturing, the robot will scan the additive surface of the workpiece to be additively manufactured, compare the data before and after additive manufacturing, perform fixed-point sampling, and perform least squares plane fitting on the sampling area to calculate the additive thickness. The workpiece is qualified by judging the thickness of the current layer of additive manufacturing. The least squares method is implemented by solving overdetermined equations based on SVD singular value decomposition.

[0074] For the sampled n point cloud data, let the fitted plane equation be:

[0075]

[0076] Where a, b, c, and d are planar parameters, and x, y, and z are the coordinates of points in the point cloud data;

[0077] The constraints are:

[0078]

[0079] We can obtain the plane parameters a, b, c, and d. To obtain the optimal fitted plane, we need to minimize the sum of the squares of the distances from the k nearest neighbors to that plane, which satisfies:

[0080]

[0081] In the formula, d i It is any point p in the point cloud data i(x i , y i , z i The distance d from this plane i =|ax i +by i +cz i +d|; To make e→min, SVD matrix decomposition can be used; the derivation process is as follows:

[0082] The average coordinates of all points are ,but:

[0083]

[0084] Subtracting equation (4) from equation (1) yields:

[0085]

[0086] Assumption matrix:

[0087]

[0088] Column matrix:

[0089]

[0090] Equation (5) is equivalent to:

[0091]

[0092] Ideally, all points lie on the plane, and equation (6) holds true; in reality, some points lie outside the plane. The purpose of fitting is to minimize the sum of the distances from the plane to all points, so the objective function is:

[0093]

[0094] The constraints are:

[0095]

[0096] If matrix A can be decomposed using singular value decomposition:

[0097]

[0098] but:

[0099]

[0100] in: Let be a column matrix, and:

[0101]

[0102] Since the diagonal elements of D are singular values, assuming the last diagonal element is the smallest singular value, then it is true if and only if:

[0103]

[0104] When the time is right, equation (10) can reach its minimum value, that is, equation (7) holds true. At this time:

[0105]

[0106] The optimal solution of objective function (7) under constraint (8) is:

[0107]

[0108] Therefore, the minimum value of e is the minimum eigenvalue of matrix A, and the corresponding eigenvectors are the plane parameters a, b, and c. d can be obtained using the centroid.

[0109] S8. Obtain the additive thickness (i.e. the number of additive layers required) for each posture using the digital model established in step S1, and determine whether the additive surface of the workpiece in the current posture has been welded.

[0110] If the welding of the additive surface in the current posture is not completed, adjust the height of the robot according to the additive thickness and continue welding until the welding of the additive surface in the current posture is completed.

[0111] If the current orientation of the additive surface is completed, move the angle of the workpiece to be added and judge whether the overall welding has been completed; if not, return to step S2, obtain the current orientation information of the workpiece to be added, and continue to steps S3-S8 until the entire workpiece is welded.

[0112] Based on the above method, an additive manufacturing system that operates according to this method includes a fixed mounting frame, a robot, and a welding torch. The welding torch is mounted on the working end of the robot, and the working end of the robot drives the welding torch to move and adjust the welding position. The robot is mounted on the mounting frame, and the mounting frame adjusts the overall displacement of the robot. The working end of the robot is also equipped with a 3D sensor and a matching protective cover. The protective cover is opened and closed in a controlled manner, and the 3D sensor scans the workpiece to be additively manufactured.

[0113] This embodiment provides an additive manufacturing method for a target-based 3D vision laser positioning and guidance robot. This method can effectively perform automated additive manufacturing through steps such as point cloud data preprocessing, point cloud uniformity correction, point cloud interpolation completion, point cloud positioning and registration, and geometric feature fitting.

[0114] Example 2:

[0115] This invention provides an additive manufacturing system for a target-guided 3D vision laser positioning robot, referring to... Figure 5 , Figure 1 This is a schematic diagram of the structure of an additive manufacturing system for a target-based 3D vision laser positioning and guidance robot according to an embodiment of the present invention.

[0116] In this embodiment, the additive manufacturing system based on target 3D vision laser positioning and guiding robot includes a fixed mounting frame 1, a robot 4, and a welding torch 5.

[0117] The welding torch 5 is installed on the working end of the robot 4. The working end of the robot 4 drives the welding torch 5 to move and adjust the welding position. The robot 4 is assembled on the mounting frame 1 through the lifting platform 2 and the guide rail 3 to adjust the overall displacement of the robot 4. The working end of the robot 4 is also equipped with a 3D sensor and a protective cover 6, which is opened and closed in a controlled manner.

[0118] After the protective cover is opened, the workpiece to be added is scanned by a 3D sensor. After scanning, the additive manufacturing process can be carried out according to the additive manufacturing method in the above embodiment. Then, by adjusting the angle and position of the workpiece to be added, multiple additive manufacturing operations can be automatically continued.

[0119] In a preferred embodiment, for a workpiece 7 to be additively manufactured, the workpiece includes a plurality of additive blades 9 to be welded, and there are two targets 8 on the additive blades 9 to be welded. The additive manufacturing system is used to perform additive manufacturing operations on the workpiece according to the additive manufacturing method. The specific operation is as follows:

[0120] First, import the digital model of the workpiece 7 to be added, and obtain the current attitude information of the additive blade 9 to be welded on the workpiece 7 through a 3D sensor.

[0121] Subsequently, based on the welding signal, the 3D sensor scans the targets 8 on both sides of the workpiece to be welded additive blade 9 to obtain point cloud data of the targets 8.

[0122] Then, the point cloud data is processed and registered and synthesized with the digital model of the workpiece 7 to be added to obtain the corresponding spatial position and attitude transformation matrix data; the robot is guided to move using the spatial position and attitude transformation matrix data to perform additive welding.

[0123] After welding a certain number of layers, a certain angle of the workpiece 7 to be added is changed. For example, in this embodiment, the angle adjustment shaft 11 is used to change the angle of the workpiece 7 to be added. The robot drives the 3D sensor to continue scanning the double-sided target 8. According to the additive manufacturing method, the starting point of the robot to be welded is obtained. In this way, the additive manufacturing operation of a single blade is completed. The blade that has been welded is the blade 10 that has been welded and added.

[0124] In this embodiment, each additive blade 9 to be welded has two targets. Using a single target would amplify the welding error on the side farther from the target, potentially leading to defective products after additive manufacturing. By scanning targets on both sides, the welding accuracy is at least twice as high.

[0125] For workpieces requiring angular changes, i.e., after adding a certain thickness, the angle needs to be changed to continue, and this action is repeated N times, then the method and system of this invention can be used to perform additive manufacturing on such workpieces.

[0126] This embodiment provides an additive manufacturing system based on target 3D vision laser positioning and guidance robot. The additive manufacturing process according to this embodiment can effectively reduce costs and increase efficiency. One system can replace 3-4 workers and realize the automation of the entire process. At the same time, based on the industrial Internet platform, the robot status, welding status and quality can be monitored in real time, remotely operated and remotely diagnosed, which is suitable for the widespread use of industrial additive manufacturing processes.

[0127] Other embodiments or specific implementations of the additive manufacturing system for target 3D vision laser positioning and guidance robots of the present invention can be referred to the above-described method embodiments, and will not be repeated here.

[0128] Furthermore, embodiments of the present invention also propose a storage medium storing an additive manufacturing method program for a target-based 3D vision laser positioning and guiding robot. When executed by a processor, the additive manufacturing method program for a target-based 3D vision laser positioning and guiding robot implements the steps of the additive manufacturing method for a target-based 3D vision laser positioning and guiding robot as described above. Therefore, it will not be repeated here. Additionally, the beneficial effects of using the same method will not be repeated. For technical details not disclosed in the computer-readable storage medium embodiments of this application, please refer to the description of the method embodiments of this application. As an example, program instructions may be deployed to execute on a single computing device, or on multiple computing devices located in one location, or on multiple computing devices distributed across multiple locations and interconnected via a communication network.

[0129] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0130] It should also be noted that the system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0131] Through the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus necessary general-purpose hardware, and of course, it can also be implemented by special hardware including application-specific integrated circuits, special CPUs, special memory, special components, etc. Generally, any function performed by a computer program can be easily implemented by corresponding hardware, and the specific hardware structure used to implement the same function can also be diverse, such as analog circuits, digital circuits, or special-purpose circuits. However, for the present invention, software program implementation is more often the preferred implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a readable storage medium, such as a computer floppy disk, USB flash drive, portable hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

Claims

1. An additive manufacturing method for a target-guided 3D vision laser positioning robot, characterized in that, The method includes the following steps: S1: Establish a digital model of the workpiece to be added with materials and extract the source point cloud dataset of the workpiece to be added with materials; S2: Use a 3D vision system to obtain the orientation information of the workpiece to be additively manufactured; S3: Using a 3D vision system, scan the target set on the workpiece to be added based on the welding signal to obtain a target point cloud dataset; S4: Register the source point cloud dataset and the target point cloud dataset so that the data points in the source point cloud dataset correspond to those in the target point cloud dataset; S5: Send the registration results of the source point cloud dataset and the target point cloud dataset to the robot, so that the robot moves to the welding position corresponding to the posture of the workpiece to be added based on the registration results, and performs additive welding; In step S5, the method further includes: before and after single-layer additive manufacturing, the robot scans the additive surface of the workpiece to be additively manufactured, compares the data before and after additive manufacturing, performs fixed-point sampling, performs least squares plane fitting on the sampling area, calculates the additive thickness, and determines whether the workpiece is qualified based on the thickness of the current layer of additive manufacturing; wherein, the least squares method is implemented using the method of solving overdetermined equations based on SVD singular value decomposition. In step S5, the method further includes: The digital model established in step S1 is used to obtain the additive thickness and the number of additive layers required for each posture, and to determine whether the additive surface of the workpiece in the current posture has been welded. If the welding of the additive surface in the current posture is not completed, adjust the height of the robot according to the thickness of the additive material and continue welding until the welding of the additive surface in the current posture is completed. If the current orientation of the additive surface is welded, move the angle of the workpiece to be added and judge whether the overall welding has been completed; if not, return to step S2, obtain the current orientation information of the workpiece to be added, and continue to steps S3-S5 until the entire workpiece is welded.

2. The additive manufacturing method for a target-based 3D vision laser positioning and guidance robot as described in claim 1, characterized in that, After step S3, the method further includes: preprocessing the target point cloud data to remove noise and discrete points; wherein the preprocessing includes: noise reduction processing, homogenization correction processing and interpolation completion processing.

3. The additive manufacturing method for a target-based 3D vision laser positioning and guidance robot as described in claim 2, characterized in that, After the alignment correction process, the target point cloud data has a row of data formed by several independent points along the laser extension direction of the 3D vision system; wherein, the independent points include square points and cross mark points, the square points are data points that are not vibrated during the acquisition process, and the cross mark points are data points that are distorted due to vibration during the acquisition process.

4. The additive manufacturing method for a target-based 3D vision laser positioning and guidance robot as described in claim 2, characterized in that, The interpolation completion specifically includes: after the target point cloud data is homogenized and corrected, for the interference points, blind collection points and reflective invalid points that appear, the method of continuously expanding the boundary based on the domain information is used to perform interpolation completion processing on the collected target point cloud data.

5. The additive manufacturing method for a target-based 3D vision laser positioning and guidance robot as described in claim 1, characterized in that, Step S4 specifically includes: performing ICP precise registration processing on the source point cloud dataset and the target point cloud dataset to obtain the registration result; and transforming the matrix... After the transformation, the source point cloud dataset corresponds to the data points in the target point cloud dataset; The expression corresponding to the data points in the source point cloud dataset and the target point cloud dataset is specifically as follows: In the formula, It is a translation matrix. For rotation matrix, This represents the source point cloud data for point p. Target point cloud data representing point p; translation matrix With rotation matrix Make the objective function Take the minimum value, at this time and It is excellent.

6. The additive manufacturing method for a target-based 3D vision laser positioning and guidance robot as described in claim 5, characterized in that, The ICP precise registration process for the source point cloud dataset and the target point cloud dataset specifically involves: establishing a KD tree between the initially transformed source point cloud data and the target point cloud data to accelerate nearest neighbor search; and using a normal vector threshold... Remove erroneous point pairs and use singular value decomposition to calculate the transformation matrix. , so that the objective function Minimize the minimum value and iterate repeatedly to obtain the optimal registration result.

7. The additive manufacturing method for a target-based 3D vision laser positioning and guidance robot as described in claim 1, characterized in that, The sampling region is fitted using the least squares plane fitting method; where: (1) The expression for the plane is as follows: In the formula, a, b, c, and d are planar parameters, and x, y, and z are the coordinates of points in the point cloud data; (2) The constraints are: (3) The plane needs to satisfy: In the formula, di is any point p in the point cloud data. i (x i , y i , z i The distance d from the fitted plane. i =|ax i +by i +cz i +d|.

8. An additive manufacturing system for a target-guided 3D vision laser positioning robot, characterized in that, For performing the additive manufacturing method of a target-based 3D vision laser positioning and guiding robot as described in any one of claims 1-7, the system includes: a mounting frame, a robot, and a welding torch. The welding torch is mounted on the working end of the robot, and the working end of the robot drives the welding torch to move and adjust the welding position. The robot is mounted on the mounting frame, and the mounting frame adjusts the overall displacement of the robot. The working end of the robot is also equipped with a 3D sensor, which is used to scan the workpiece to be additively manufactured.

Images