Welding method based on weld joint recognition, computer equipment and readable storage medium

Abstract

The invention provides a welding method based on weld joint recognition, computer equipment and a readable storage medium. The method comprises the steps that point cloud data of a to-be-welded position area of a workpiece is obtained; wherein the point cloud data is obtained by scanning by taking a welding seam of the to-be-welded position area as a center; determining an effective point cloud surface of the weld joint according to the point cloud data; wherein the effective point cloud surface refers to a point cloud gathering area which is closest to the weld joint and has the largest area in the point cloud data; the effective point cloud surface is subjected to multiple times of segmentation with different granularities, and welding seam point cloud data in the point cloud data are found out; the welding track is generated according to the welding seam point cloud data, the robot is guided to conduct welding based on the welding track, the method is suitable for workpieces of different sizes and welding seams of different types, the corresponding accurate welding track can be given even under the complex conditions that pre-machining craters, defects and the like exist, and the precision and reliability of the welding process are improved.

Description

Technical Field

[0001] The present invention relates to the field of welding technology, and in particular to a welding method based on weld identification, a computer device and a readable storage medium. Background Art

[0002] Welding is a manufacturing process and technology that uses heat, high temperature or high pressure to join metals or other thermoplastic materials such as plastics. It is widely used in industry. For example, in the power and communication industries, steel pipe poles are used as brackets and support towers for power transmission or communication equipment, and the demand for welding is particularly huge.

[0003] Existing welding technology usually uses special welding machines and customized welding tools for welding. It is difficult to achieve compatibility with workpieces of different thicknesses or irregular welds, and lacks flexibility and adaptability. In addition, the pre-processing links before welding, such as cutting and grinding, often cause scars or unevenness on the workpiece surface due to insufficient precision, which brings considerable challenges to subsequent automated welding.

[0004] Therefore, the existing welding system is difficult to adapt to complex and changeable working conditions, and has problems of low accuracy, reliability and poor adaptability. Summary of the invention

[0005] In view of this, the purpose of the present invention is to provide a welding method, a computer device and a readable storage medium based on weld recognition, which are used to adapt to complex and changeable working conditions and improve welding accuracy and efficiency. In order to achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0006] In a first aspect, the present invention provides a welding method based on weld identification, the method comprising: obtaining point cloud data of a workpiece in a region where welding is to be performed; wherein the point cloud data is obtained by scanning with the weld in the region where welding is to be performed as the center; determining an effective point cloud surface of the weld based on the point cloud data; wherein the effective point cloud surface refers to a point cloud clustering region in the point cloud data that is closest to the weld and has the largest area; performing multiple segmentations of different granularities on the effective point cloud surface to find weld point cloud data in the point cloud data; generating a welding trajectory based on the weld point cloud data, and guiding a robot to perform welding based on the welding trajectory.

[0007] In a second aspect, the present invention provides a computer device comprising a processor and a memory, wherein the memory stores machine executable instructions that can be executed by the processor, and the processor can execute the machine executable instructions to implement the welding method based on weld identification as described in any of the aforementioned embodiments.

[0008] In a third aspect, the present invention provides a readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the welding method based on weld identification as described in any one of the aforementioned embodiments is implemented.

[0009] The welding method, computer device and readable storage medium based on weld identification provided by the embodiment of the present invention first obtain the point cloud data of the area of ​​the workpiece to be welded, ensure that the information obtained is highly correlated with the actual position of the weld, determine the effective point cloud surface of the weld according to the point cloud data, and lay the foundation for the subsequent accurate identification of the true position of the weld. The effective point cloud surface is segmented multiple times with different granularities, and then the weld point cloud data is extracted, which can more carefully analyze the shape and characteristics of the weld, thereby improving the accuracy of the welding trajectory, and generate a welding trajectory according to the extracted weld point cloud data. Finally, the robot is guided to perform welding operations based on the generated welding trajectory. The embodiment of the present invention is not only suitable for workpieces of different sizes and different types of welds, but also can perform visual trajectory correction and give corresponding welding trajectories even in complex situations such as pre-processed weld scars and defects. Finally, the accuracy and reliability of the welding process can be improved.

[0010] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0012] Figure 1 A system architecture diagram of a welding method based on weld seam recognition provided by an embodiment of the present invention;

[0013] Figure 2 A schematic flow chart of a welding method based on weld identification provided by an embodiment of the present invention;

[0014] Figure 3 A functional module diagram of a welding device based on weld seam recognition provided by an embodiment of the present invention;

[0015] Figure 4 A structural block diagram of a computer device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0016] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0017] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present invention.

[0018] It should be noted that relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprises" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0019] Taking into account that the existing robot automatic welding methods are difficult to adapt to complex and changeable working conditions and have problems such as low reliability and poor adaptability, an embodiment of the present invention provides a welding method based on weld recognition, which is suitable for workpieces with scars or uneven surfaces during the pre-processing process. It can provide corresponding welding trajectories based on visual trajectories, thereby guiding the robot to achieve high-quality automated welding of the workpiece.

[0020] The preferred embodiments of the present invention are described below in conjunction with the drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention. In addition, the embodiments of the present invention and the features in the embodiments may be combined with each other if there is no conflict.

[0021] See also Figure 1 , Figure 1 A system architecture diagram of a welding method based on weld identification provided by an embodiment of the present invention, such as Figure 1 As shown, the system architecture may include: a camera 101 , a computer device 102 and a robot 103 .

[0022] The camera 101 is used to obtain a spatial point cloud dataset of a position area to be welded of the workpiece, and send the spatial point cloud dataset to the computer device 102 .

[0023] In one embodiment of the present invention, the camera 101 may be placed in, but is not limited to: connected to a mechanical arm of the robot 103; or installed above a workbench for placing workpieces.

[0024] In an optional embodiment, the camera 101 may be, but is not limited to, a line laser camera.

[0025] The computer device 102 has data processing and control capabilities, and is used to receive the spatial point cloud data set and execute the welding method based on weld seam recognition provided by the present invention, and guide the robot 103 to perform automated welding.

[0026] The robot 103 is used to weld the to-be-welded position area of ​​the workpiece according to the weld trajectory sent by the computer device 102 under the control of the computer device 102 .

[0027] Combined with the above Figure 1 See the system architecture shown in Figure 2 , Figure 2 The schematic flow chart of the welding method based on weld identification provided by the embodiment of the present invention, the execution subject of the method may be a computer device 102, and the workpiece to which the method can be applied may include but is not limited to steel pipes. The method mainly includes steps S201 to S204, which are described as follows:

[0028] S201: Obtaining point cloud data of the area of ​​the workpiece to be welded; wherein the point cloud data is obtained by scanning with the weld seam of the area of ​​the workpiece to be welded as the center;

[0029] S202: determining an effective point cloud surface of the weld according to the point cloud data; wherein the effective point cloud surface refers to a point cloud gathering area in the point cloud data that is closest to the weld and has the largest area;

[0030] S203: segmenting the effective point cloud surface multiple times with different granularities to find out the weld point cloud data in the point cloud data;

[0031] S204: Generate a welding trajectory according to the weld point cloud data, and guide the robot to perform welding based on the welding trajectory.

[0032] The above-mentioned welding method based on weld identification provided by the embodiment of the present invention, first, the computer device obtains the point cloud data of the workpiece to be welded position area, this data is obtained by scanning with the weld in the weld position area as the center, ensuring that the information obtained is highly correlated with the actual position of the weld, next, the computer device determines the effective point cloud surface of the weld according to the point cloud data, which lays the foundation for the subsequent accurate identification of the true position of the weld, and then the computer device performs multiple segmentations of different granularities on the effective point cloud surface, and then extracts the weld point cloud data. Through this segmentation, the device can analyze the shape and characteristics of the weld more carefully, thereby improving the accuracy of the welding trajectory, generating a welding trajectory according to the extracted weld point cloud data, and finally guiding the robot to perform welding operations based on the generated welding trajectory. The embodiment of the present invention relies on the scanned and reconstructed point cloud data around the weld to be welded, and can accurately identify the weld. This method is not only suitable for workpieces of different sizes and different types of welds, but also can perform visual trajectory correction and give corresponding welding trajectories even in complex situations such as pre-processing weld scars and defects. Finally, the implementation of the embodiment of the present invention enables the robot to perform high-quality automated welding on the workpiece, improving the accuracy and reliability of the welding process.

[0033] Next, the embodiment of the present invention will introduce the above steps S202 to S204 in detail with reference to relevant examples.

[0034] In step S201, Figure 1 As shown, the embodiment of the present invention can obtain the original point cloud data of the position area to be welded through the camera 101. For example, as an example, the robot 103 can be equipped with a line laser camera, which is aimed at the position area to be welded for scanning, so as to obtain the original point cloud data of the position area to be welded, which is recorded as pointcloud_org. Pointcloud_org is composed of a series of points and can be expressed as pointcloud_org = {p1, p2, p3, ..., p N}, where p i represents the i-th point in the original point cloud data, and N represents the total number of points in the point cloud data.

[0035] It is easy to understand that the collected original point cloud data pointcloud_org often contains noise points, which may interfere with the subsequent weld recognition results. In order to ensure the accuracy and efficiency of the welding process, after obtaining the original point cloud data pointcloud_org of the welding location area, the computer equipment can first pre-process the original data to purify the data and remove noise, laying a solid foundation for subsequent accurate weld recognition and welding.

[0036] In an optional implementation, the computer device may perform statistical outlier filtering on the original point cloud data pointcloud_org to eliminate the interference of noise points. After the noise points are eliminated, the computer device may further downsample the data to reduce the amount of data at a certain density. This process can reduce the time and space complexity of the recognition process as much as possible while ensuring the recognition accuracy. By selecting an appropriate sampling rate, an effective weld trajectory that meets specific accuracy requirements can be obtained more efficiently. For ease of distinction, the embodiment of the present invention marks the point cloud data after filtering and downsampling as pointcloud_pre_process.

[0037] Furthermore, considering that the amount of point cloud data is still huge despite the preprocessing, in order to quickly and accurately achieve the point cloud segmentation of the area to be welded, the computer device can also choose to further reduce the density of the preprocessed point cloud data pointcloud_pre_process, improve the processing speed, and ensure the accuracy and efficiency of the segmentation process, so that the welding trajectory is generated more quickly.

[0038] In an optional implementation manner, the processing method for reducing the density of point cloud data pointcloud_pre_process adopted in the embodiment of the present invention is described as follows:

[0039] First, the three-dimensional space where the pre-processed point cloud data is located is divided into a number of small cubic grids. For example, as an example, the computer device can divide the three-dimensional space where the point cloud data pointcloud_pre_process is located into a plurality of spatial cubic grids in the three directions of X, Y and Z with a step size of 0.3 mm. The volume of each cubic grid is 0.3 mm×0.3 mm×0.3 mm.

[0040] Then, each spatial cube grid is checked for point cloud data (i.e., points). If a spatial cube grid has at least one point, then the grid is considered to have a point cloud.

[0041] Finally, in the space cube grid where the point cloud data exists, a point is randomly selected from all the points in the grid as the representative of the grid. In this way, the points with the same number as the grid are obtained to form a new point cloud data set, so that the point cloud data with reduced density can be obtained, which is recorded as pointcloud_corse.

[0042] Based on the above processing steps, it can be understood that the "point cloud data" referred to in step S201 in the embodiment of the present invention refers to the point cloud data after preprocessing and density reduction processing, that is, pointcloud_corse. For the convenience of distinction and description, the embodiment of the present invention refers to the points in the point cloud data pointcloud_corse mentioned later as "grid points". It should be emphasized that this description is only for the convenience of explanation and distinction, and does not constitute any limitation on its function or use.

[0043] Next, based on the grid points obtained in step S201, step S202 in the embodiment of the present invention is introduced as follows:

[0044] S202: Determine the effective point cloud surface of the weld according to the point cloud data.

[0045] In the embodiment of the present invention, the so-called effective point cloud surface refers to the point cloud gathering area with the closest distance to the weld and the largest area identified in the point cloud data pointcloud_corse. The computer device can accurately segment the point cloud data pointcloud_corse using the curvature information of each grid point, thereby determining the effective point cloud surface of the weld. Therefore, a key step in the embodiment of the present invention is to first determine the curvature of each grid point, which lays the foundation for the subsequent weld identification and accurate planning of the welding trajectory.

[0046] In one embodiment of the present invention, the processing method of determining the curvature of each grid point by the computer device may include steps a1 to a4, which are described as follows:

[0047] Step a1: Get the normal vector corresponding to the target point.

[0048] It can be understood that the target point is each grid point in the point cloud data pointcloud_corse, which means that each grid point needs to determine its curvature through steps a1 to a4. The following will take a single grid point as an example for explanation.

[0049] In the embodiment of the present invention, the normal vector of the target point is obtained by performing feature analysis on the coordinates of its neighborhood points. In other words, before the computer device determines the curvature information of each grid point, it is first necessary to estimate the normal vector information of the target point, which is recorded as pointnormal. This provides the necessary basis for the subsequent curvature calculation and the determination of the effective point cloud surface of the weld.

[0050] As an optional implementation, the computer device may perform feature analysis on the coordinates of the neighborhood points corresponding to each grid point, and then use the minimum feature vector as the normal vector of each point.

[0051] It can be understood that the above-mentioned feature analysis process refers to: calculating the covariance matrix of the coordinates of the neighborhood points. Then the covariance matrix is ​​decomposed into eigenvalues ​​and eigenvectors to obtain three eigenvalues ​​and corresponding three eigenvectors. In an embodiment of the present invention, the eigenvector corresponding to the minimum eigenvalue is selected from the three eigenvectors. This eigenvector points to the vertical direction of the point cloud surface and can be used as the normal vector of the current point. The other two eigenvectors will be used in the process of determining the curvature.

[0052] In the embodiment of the present invention, "neighborhood points" refer to all points within a preset radius with the target point as the center. The preset radius range can be flexibly set according to actual needs, and the specific value is not limited. For example, assuming that the radius is set to 3mm, then with a certain grid point as the center, find all points that are less than 3mm away from the grid point. These points constitute the neighborhood points corresponding to the grid point, recorded as {p_n_0, p_n_1,..., p_n_k}, which contains a total of k+1 neighborhood points.

[0053] The computer device can perform eigenvalue and eigenvector decomposition on the coordinates of these neighborhood points to obtain three eigenvalues ​​and three eigenvectors. The eigenvector corresponding to the minimum eigenvalue is taken as the normal vector of the grid point. The above implementation method is repeated for each grid point to obtain the normal vector of each grid point in the point cloud data pointcloud_corse.

[0054] Step a2: Project all neighborhood points of the target point into the planes determined by the lines where the two largest eigenvectors obtained in the feature analysis are located, and find two target projection points that meet the preset conditions in each plane.

[0055] In the embodiment of the present invention, the preset conditions for finding the target projection point are designed as follows: the distance from the straight line is within a preset range, and the distance from the target point is the farthest and located on both sides of the target point.

[0056] To explain the above process more clearly, assume that the two eigenvectors are v1 and v2, and the current grid point p0 is used as the reference point. The eigenvector v1 and the point p0 together define a straight line l1. Similarly, the eigenvector v2 and the point p0 define another straight line l2. The two straight lines are perpendicular to each other. The embodiment of the present invention projects all neighborhood points onto the plane formed by the straight lines l1 and l2, and identifies the projection points that meet specific conditions.

[0057] For example, taking the plane where the line l1 is located as an example, if the preset distance threshold is 0.2mm, the computer device will filter out those projection points whose distance from the line l1 is less than 0.2mm. Among these projection points, the computer device further searches for the points farthest from the grid point p0, and these points must be located on both sides of p0 to ensure that they can be used as target projection points. In this way, the computer device can identify the target projection points that meet the conditions in the plane of the line l1.

[0058] The same method is also applicable to the plane where the line l2 is located. The computer device can find the target projection points that meet the above conditions in the plane. Through this process, the computer device can eventually obtain four target projection points, which will be used to determine the curvature of point p0, thereby providing key data for subsequent analysis and calculation.

[0059] Step a3: In each plane, based on the two target projection points and the radius of the circle determined by the target point, determine the curvature of the line where the feature vector lies.

[0060] Step a4: Calculate the curvature of the target point based on the curvature of the straight line where the two eigenvectors are located.

[0061] Continuing with the above example, for the plane where the line l1 is located, the two target projection points and p0 are on the same circle, so the radius of the circle can be determined to be r1, and the curvature in the direction of the line l1 is 1 / r1. Similarly, the curvature in the direction of l2 is 1 / r2. Then the curvature of p0 is calculated as: (1 / r1+1 / r2) / 2.0.

[0062] Continuing with the above example, in the plane where the line l1 is located, it can be found that the two determined target projection points and p0 are located on the circumference of the same circle. Based on this geometric relationship, the computer device can calculate the radius of this circle, recorded as r1. Therefore, the curvature in the direction of the line l1 can be defined as the reciprocal of its radius, that is, 1 / r1.

[0063] The same method is applied to the plane where the line l2 lies. Through the same geometric analysis, the computer equipment can determine the curvature in the direction of l2, which is recorded as 1 / r2.

[0064] Finally, in order to obtain the curvature of point p0, the embodiment of the present invention adopts the method of averaging the curvatures in the l1 and l2 directions. The specific calculation method is: (1 / r1+1 / r2) / 2.0. In this way, the local curvature change of each grid point can be accurately determined, which provides a basis for subsequent point cloud segmentation.

[0065] Through the above implementation, the computer device can determine the curvature of each grid point in the point cloud data pointcloud_corse. Next, the point cloud can be segmented based on the curvature, and then the effective point cloud surface of the weld can be determined based on the segmentation result.

[0066] In one embodiment of the present invention, the process of segmenting the point cloud data pointcloud_corse based on curvature mainly includes steps c1 to c2, which are described as follows:

[0067] Step c1: Obtain the curvature of the neighborhood points corresponding to each point.

[0068] Step c2: If the curvature difference between each point and its corresponding neighboring points is less than a preset value, each point and its corresponding neighboring points are divided into the same point cloud data block until each point is located in a point cloud data block.

[0069] In an embodiment of the present invention, the "preset value" can be adjusted according to specific application scenarios and requirements, and its specific value is not rigidly stipulated here. Take a specific example to illustrate, assuming that the preset value is set to 1.2. Based on this preset value, if it is found that the curvature difference between the neighborhood point and the current point is less than 1.2, then the two points can be regarded as belonging to the same area. Through such an iterative comparison and classification process, point cloud blocks with similar curvature can eventually be identified. Each point cloud set with similar curvature is classified into a cluster, thereby realizing grid point cloud segmentation based on curvature characteristics.

[0070] It can be understood that the above curvature-based segmentation method provides a coarse-grained point cloud data processing method. Since the point cloud data is scanned with the weld as the center, the limited support surface of the weld should be the two largest clusters in the segmentation result. Based on this understanding, on the basis of coarse segmentation, valid point cloud surfaces can be further screened out. The specific steps are as follows: Among the multiple different point cloud blocks obtained after curvature-based point cloud segmentation, identify the two point cloud blocks with the largest area. These two point cloud blocks with the largest area will be regarded as the two valid point cloud surfaces of the weld, recorded as corse_part1 and corse_part2 respectively.

[0071] Through the above implementation, the embodiment of the present invention can ensure that the key supporting surface of the weld can be accurately located, and provide a solid foundation for the subsequent accurate identification of the weld.

[0072] Next, the embodiment of the present invention can more accurately find the weld point cloud data in the point cloud data by segmenting the effective point cloud surface multiple times with different granularities, which provides a basis for subsequently identifying the accurate weld trajectory based on the weld point cloud data, see step S203.

[0073] In one embodiment of the present invention, step S203 may include steps d1 to d3, which are described as follows:

[0074] Step d1: Obtain a coarsely segmented point cloud surface corresponding to the valid point cloud surface.

[0075] In an embodiment of the present invention, the coarse segmentation point cloud surface is composed of the effective point cloud surface and the points within the effective point cloud surface and other points allocated from the preprocessed original point cloud data. Specifically, the present invention first takes the effective point cloud surfaces corse_part1 and corse_part2 as the reference, and then performs point allocation based on the preprocessed original point cloud data pointcloud_pre_process. Through this process, two coarse segmentation point cloud surfaces can be obtained, which are recorded as dense_part1 and dense_part2 respectively. These two coarse segmentation point cloud surfaces not only contain the precise data of the effective point cloud surface, but also integrate other point cloud information related thereto, thereby providing more comprehensive and detailed point cloud data support for the welding process.

[0076] In the process of allocating point cloud data, the computer device may perform the following processing on each point in the preprocessed point cloud data set pointcloud_pre_process: calculate the distance from the point to all points in the valid point cloud surface corse_part1, and identify the minimum value of these distances, recorded as dist1. Similarly, calculate the distance from the point to all points in the valid point cloud surface corse_part2, and find the minimum distance, recorded as dist2. Based on these calculation results, the decision process will determine the attribution of each point: if dist1 is less than dist2, it means that the point is closer to corse_part1, so it is allocated to the coarse segmentation point cloud surface dense_part1 corresponding to corse_part1; conversely, if dist2 is smaller, the point is allocated to the coarse segmentation point cloud surface dense_part2 corresponding to corse_part2.

[0077] Through the above iterative comparison and allocation process, the computer device can allocate all point cloud data in pointcloud_pre_process to the corresponding valid point cloud surface. Finally, dense_part1 corresponding to corse_part1 and dense_part2 corresponding to corse_part2 can be obtained. These two roughly segmented point cloud surfaces provide more accurate point cloud data support for subsequent welding work.

[0078] Step d2: According to the preset search strategy, find the coarse-grained point cloud cluster containing the weld from the coarsely segmented point cloud surface.

[0079] In the embodiment of the present invention, the preset search strategy is: based on the set distance threshold, find the points adjacent to each other in two coarsely segmented point cloud surfaces, and merge them into a point cloud cluster, recorded as corner_cloud. In other words, the computer device can obtain the distance from each point in one of the coarsely segmented point cloud surfaces to all points in the other coarsely segmented point cloud surface; find all points in one of the coarsely segmented point cloud surfaces whose minimum distance to all points is less than the set threshold, and form a point cloud cluster.

[0080] In an embodiment of the present invention, a search strategy based on a preset distance threshold is proposed: based on the set distance threshold, points adjacent to each other on two roughly segmented point cloud surfaces are found and merged into a point cloud cluster, recorded as corner_cloud. This processing method aims to identify and merge points adjacent to each other between two roughly segmented point cloud surfaces. Specifically:

[0081] The computer device first calculates the distance from each point in one coarsely segmented point cloud surface to all points in another coarsely segmented point cloud surface. Then, from these calculations, the points whose distance to any point in the other coarsely segmented point cloud surface is less than a preset threshold are selected. These points are considered to be adjacent to each other because their distance to the other supporting surface is within a specific range, and are therefore merged into a point cloud cluster.

[0082] For ease of understanding, assume that two point clouds are recorded as point cloud A and point cloud B. For each point in point cloud A, calculate its distance {d_A i _B j}, i = 1, 2, ..., n; j = 1, 2, ..., m. For each point in A, that is, when i is fixed, find {d_A i _B j}Minimum distance, denoted as d_A i _B j _min, if this value is less than the given threshold of 20mm, then point i in point cloud A is considered to be very close to point B and is a neighboring point. Repeat this operation for all points in point cloud A to find all points in point cloud A that are close to point cloud B. Similarly, all points in point cloud B that are close to point cloud A can be found using the same method as above.

[0083] To illustrate this process more intuitively, assume that there are two point clouds, denoted as point cloud A and point cloud B. The computer device will calculate the distance from each point in point cloud A to each point in point cloud B. These distance sets are represented as {d_Ai_Bj}, where i represents the index of the point in point cloud A, from 1 to n; j represents the index of the point in point cloud B, from 1 to m.

[0084] For each point in point cloud A (i.e., fixed i), the computer device will find the minimum distance to all points in point cloud B, and this minimum value is recorded as d_Ai_Bj_min. If d_Ai_Bj_min is less than the threshold we set, such as 20mm, then it is considered that the distance between point i in point cloud A and point cloud B is close enough to be considered as a neighboring point. The computer device will perform this operation on all points in point cloud A to identify all points that are adjacent to point cloud B.

[0085] The same method is applied to point cloud B. The computer device will perform the same calculation and screening process on each point in point cloud B to find all points that are adjacent to point cloud A.

[0086] Through this strategy, the computer equipment can identify points that are close to each other between two roughly segmented point cloud surfaces and combine them into a unified point cloud cluster. This process enhances the understanding of weld characteristics and ensures accurate identification of welds during the welding process.

[0087] Step d3: Segment the coarse-grained point cloud clusters, and find the weld point cloud again according to the preset search strategy based on the segmentation results.

[0088] After successfully identifying the coarse-grained point cloud cluster containing the weld, the computer device will enter the next processing flow, which is to further subdivide the two coarse-grained point cloud clusters. In this refinement process, the computer device still divides each coarse-grained point cloud cluster into two finer clusters based on the curvature of each point in the point cloud cluster. This curvature-based segmentation method is the same as the steps described above, so it will not be repeated.

[0089] In the embodiment of the present invention, the two segmentation results obtained by performing the segmentation again in step d3 are recorded as precise_part1 and precise_part2, and then the neighboring point search is performed again according to the above search strategy, and finally the weld point cloud is obtained, which is recorded as welding_cloud_org. Based on the weld point cloud, the accurate weld trajectory can be obtained, please refer to step S204.

[0090] In step S204, in order to ensure the accuracy of the weld trajectory, the weld point cloud welding_cloud_org found in step S203 may be denoised first.

[0091] As an optional implementation, the computer device can find neighboring points based on each point in the welding point cloud data welding_cloud_org with a given distance as a threshold, fit these neighboring points into a straight line, and then project each point in the welding point cloud data welding_cloud_org onto the straight line to obtain the denoised welding point cloud welding_cloud_denoise.

[0092] Furthermore, the computer device can sample the denoised weld point cloud data in an orderly manner and generate a weld trajectory based on the sampling points. In other words, the computer device can set a threshold value, sample the denoised weld point cloud in an orderly manner, obtain an ordered sampled weld point cloud ordered_cloud, add a fixed welding posture to the weld point cloud sequence obtained by the orderly sampling, generate a welding trajectory pose sequence, and send the welding trajectory pose sequence to the robot to guide the robot to perform welding.

[0093] In summary, the welding method based on weld recognition provided by the embodiment of the present invention has the following advantages: the embodiment of the present invention can accurately identify the weld by scanning and reconstructing the point cloud data around the weld to be welded. This method is not only applicable to workpieces of different sizes and different types of welds, but also can perform visual trajectory correction and give the corresponding welding trajectory even in complex situations such as pre-processing weld scars and defects. Finally, the implementation of the embodiment of the present invention enables the robot to perform high-quality automated welding on the workpiece, significantly improving the efficiency and reliability of the welding process.

[0094] In order to perform the corresponding steps in the above embodiments and various possible methods, a method for implementing a welding device based on weld seam recognition is given below. Figure 3 , Figure 3 A functional module diagram of a welding device based on weld seam recognition provided by an embodiment of the present invention. The welding device 30 based on weld seam recognition includes:

[0095] The acquisition module 301 is used to obtain point cloud data of the area of ​​the workpiece to be welded; wherein the point cloud data is obtained by scanning with the weld seam of the area of ​​the workpiece to be welded as the center;

[0096] The identification module 302 is used to determine the effective point cloud surface of the weld according to the point cloud data; wherein the effective point cloud surface refers to the point cloud gathering area in the point cloud data that is closest to the weld and has the largest area;

[0097] The identification module 302 is also used to segment the effective point cloud surface multiple times with different granularities to find the weld point cloud data in the point cloud data;

[0098] The recognition module 302 is used to generate a welding trajectory according to the weld point cloud data;

[0099] The control module 303 is used to guide the robot to perform welding based on the welding trajectory.

[0100] It is understandable that the acquisition module 301, the identification module 302 and the control module 303 can be executed in coordination. Figure 2 Each step in the process is performed to achieve the corresponding technical effect.

[0101] It should be noted that the welding device based on weld seam recognition provided in this embodiment has the same technical features as the welding method based on weld seam recognition provided in the above embodiment, so it can also solve the same technical problems and achieve the same technical effects. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working process of the device described above can refer to the corresponding process in the above method embodiment, and will not be repeated here.

[0102] Optionally, the above modules can be stored in the form of software or firmware. Figure 4 The memory shown in FIG. 100 or the operating system (OS) of the computer device 102 may be fixed therein and may be Figure 4 Meanwhile, the data and program codes required for executing the above modules can be stored in the memory.

[0103] See also Figure 4 , Figure 4 The computer device 102 includes a memory 1021, a processor 1022, and a communication interface 1023. The memory 1021, the processor 1022, and the communication interface 1023 are electrically connected to each other directly or indirectly to achieve data transmission or interaction. For example, these components can be electrically connected to each other through one or more communication buses or signal lines.

[0104] Optionally, the bus can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0105] In an embodiment of the present invention, the processor 1022 may be a general-purpose processor, a digital signal processor, an application-specific integrated circuit, a field programmable gate array or other programmable logic device, a discrete gate or transistor logic device, or a discrete hardware component, and may implement or execute the methods, steps, and logic block diagrams disclosed in the embodiment of the present invention. A general-purpose processor may be a microprocessor or any conventional processor, etc. The steps of the method disclosed in the embodiment of the present invention may be directly embodied as being executed by a hardware processor, or may be executed by a combination of hardware and software modules in the processor. The software module may be located in the memory 1021, and the processor 1022 reads the program instructions in the memory 1021 and completes the steps of the above method in combination with its hardware.

[0106] In the embodiment of the present invention, the memory 1021 may be a non-volatile memory, such as a hard disk (HDD) or a solid-state drive (SSD), etc., or a volatile memory (Volatile Memory), such as RAM. The memory may also be any other medium that can be used to carry or store the desired program executable code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory in the embodiment of the present invention may also be a circuit or any other device that can realize a storage function, for storing instructions and / or data.

[0107] The memory 1021 can be used to store software programs and modules, such as the instructions / modules of the welding device 30 based on weld recognition provided in the embodiment of the present invention, which can be stored in the memory 1021 in the form of software or firmware (Firmware) or solidified in the operating system (OS) of the computer device 102. The processor 1022 executes the software programs and modules stored in the memory 1021 to perform various functional applications and data processing. The communication interface 1023 can be used to communicate signaling or data with other node devices.

[0108] Understandably, Figure 4 The structure shown is for illustration only. The computer device 102 may also include Figure 4 More or fewer components as shown, or with Figure 4 Different configurations shown. Figure 4 The components shown may be implemented in hardware, software or a combination thereof.

[0109] Based on the above embodiments, the present invention also provides a readable storage medium, in which a computer program is stored. When the computer program is executed by a computer, the computer executes the welding method based on weld recognition provided in the above embodiments to achieve the purpose of controlling the robot to perform welding tasks.

[0110] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by instructions. These instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0111] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0112] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0113] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A welding method based on weld identification, characterized in that: The method comprises: Obtaining point cloud data of the area of ​​the workpiece to be welded; wherein the point cloud data is obtained by scanning with the weld of the area of ​​the workpiece to be welded as the center; Determine the effective point cloud surface of the weld according to the point cloud data; wherein the effective point cloud surface refers to the point cloud gathering area in the point cloud data that is closest to the weld and has the largest area; Segment the effective point cloud surface multiple times with different granularities to find out the weld point cloud data in the point cloud data; A welding trajectory is generated according to the weld point cloud data, and a robot is guided to perform welding based on the welding trajectory.

2. The welding method based on weld identification according to claim 1, characterized in that: Determining the effective point cloud surface of the weld according to the point cloud data includes: Obtaining the curvature of each point in the point cloud data; Segmenting the point cloud data according to the curvature of each point to obtain a plurality of point cloud data blocks; The two point cloud data blocks with the largest areas are taken as the effective point cloud surfaces.

3. The welding method based on weld identification according to claim 2, characterized in that: The point cloud data is segmented according to the curvature of each point to obtain a plurality of point cloud data blocks, including: Obtaining the curvature of the neighborhood points corresponding to each point; wherein the neighborhood points are points within a preset radius with each point as the center; If the curvature difference between each point and its corresponding neighboring points is less than a preset value, each point and its corresponding neighboring points are divided into the same point cloud data block until each point is located in a point cloud data block.

4. The welding method based on weld identification according to claim 2, characterized in that: Determining the effective point cloud surface of the weld according to the point cloud data also includes: Obtaining a normal vector corresponding to a target point; wherein the target point is each point in the point cloud data; the normal vector is obtained by performing feature analysis on the coordinates of the neighborhood points corresponding to the target point; Projecting all neighborhood points of the target point into the planes determined by the straight lines where the two maximum eigenvectors obtained in the feature analysis are located, and finding two target projection points that meet preset conditions in each plane; The preset condition is: the distance from the straight line is within a preset range, and the distance from the target point is the farthest and located on both sides of the target point; In each plane, based on the two target projection points and the radius of the circle determined by the target point, determine the curvature of the straight line where the feature vector is located; The curvature of the target point is calculated according to the curvature of the straight line where the two eigenvectors are located.

5. The welding method based on weld identification according to claim 2, characterized in that: Determining the effective point cloud surface of the weld according to the point cloud data also includes: The coordinates of the neighborhood points corresponding to each point are subjected to feature analysis, and then the minimum feature vector is used as the normal vector of each point.

6. The welding method based on weld identification according to claim 1, characterized in that: The effective point cloud surface is segmented multiple times with different granularities to find the weld point cloud data in the point cloud data, including: Obtaining a coarsely segmented point cloud surface corresponding to the effective point cloud surface; wherein the coarsely segmented point cloud surface is composed of the effective point cloud surface and other points assigned to the effective point cloud surface from the pre-processed original point cloud data; According to a preset search strategy, finding a coarse-grained point cloud cluster containing the weld from the coarsely segmented point cloud surface; The coarse-grained point cloud cluster is segmented, and based on the segmentation result, the weld point cloud is found again according to the preset search strategy.

7. The welding method based on weld identification according to claim 6, characterized in that: Finding a coarse-grained point cloud cluster containing the weld from the coarsely segmented point cloud surface includes: Obtain the distance between each point in one of the roughly segmented point cloud surfaces and all points in another roughly segmented point cloud surface; All points within the coarsely segmented point cloud surface whose minimum distance to all points is less than a set threshold are found to form the coarse-grained point cloud cluster.

8. The welding method based on weld identification according to claim 1, characterized in that: Generating a welding trajectory according to the weld point cloud data includes: Performing denoising processing on the weld point cloud data; The weld point cloud data after orderly sampling and denoising is generated based on the sampling points.

9. A computer device, characterized in that: It comprises a processor and a memory, wherein the memory stores machine executable instructions that can be executed by the processor, and the processor can execute the machine executable instructions to implement the welding method based on weld identification as described in any one of claims 1-8.

10. A readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the welding method based on weld seam identification as described in any one of claims 1 to 8 is implemented.

Images