Welding track generation method and device, computer equipment and storage medium

Abstract

The invention relates to a welding track generation method and device, computer equipment and a storage medium. The method comprises the steps that a workpiece point cloud of a target workpiece is obtained, and the target workpiece comprises a plurality of rib plates and a bottom plate; rib plate point clouds corresponding to the multiple rib plates are extracted from the workpiece point clouds, multiple rib plate straight lines are extracted from the rib plate point clouds, and the rib plate straight lines represent projection positions of the rib plates on the bottom plate; according to the intersection of the rib plate straight lines and the multiple rib plate straight lines, a welding seam point cloud path is generated; and according to loop information in the welding seam point cloud path, a welding track is extracted from the welding seam point cloud. By adopting the method, the welding track generation accuracy of the target workpiece can be improved.

Description

Technical Field

[0001] This application relates to the field of weld trajectory recognition technology, and in particular to a welding trajectory generation method, apparatus, computer equipment, and storage medium. Background Technology

[0002] Currently, ship sub-assembly structures are generally composed of a bottom plate and stiffening ribs welded together. Weld path planning for these sub-assembly structures typically relies on point cloud-based weld extraction methods, which often depend on planar feature segmentation and geometric fitting, such as using the RANSAC algorithm to fit planes or lines to extract rib edges.

[0003] However, in complex welding scenarios, the weld locations of small assembled structural components are numerous, complex in type, and vary in spatial orientation. Traditional weld extraction methods are prone to weld path breaks or repetitions. Summary of the Invention

[0004] Therefore, it is necessary to provide a welding trajectory generation method, apparatus, computer equipment, computer-readable storage medium, and computer program product that can improve the accuracy of welding trajectory generation in response to the above-mentioned technical problems.

[0005] In a first aspect, this application provides a method for generating welding trajectories, including:

[0006] Obtain the point cloud of the target workpiece, which includes multiple stiffeners and a base plate;

[0007] Extract multiple stiffener point clouds corresponding to stiffeners from the workpiece point cloud, and extract multiple stiffener straight lines from the stiffener point cloud. The stiffener straight lines represent the projection position of the stiffener on the base plate.

[0008] Based on the intersection of the straight lines of each stiffener and multiple straight lines of stiffeners, generate the weld point cloud path;

[0009] Welding trajectory is extracted from weld point cloud based on loop information in weld point cloud path.

[0010] Secondly, this application also provides a welding trajectory generation device, comprising:

[0011] The point cloud acquisition module is used to acquire the point cloud of the target workpiece, wherein the target workpiece includes multiple stiffeners and a base plate;

[0012] The rib plate straight line extraction module is used to extract multiple rib plate point clouds corresponding to multiple rib plates from the workpiece point cloud, and extract multiple rib plate straight lines from the rib plate point cloud. The rib plate straight lines represent the projection position of the rib plate on the base plate.

[0013] The weld point cloud path generation module is used to generate weld point cloud paths based on the intersection between the straight lines of each stiffener and multiple straight lines of the stiffener.

[0014] The welding trajectory generation module is used to extract the welding trajectory from the weld point cloud based on the loop information in the weld point cloud path.

[0015] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in any of the above-described welding trajectory generation method embodiments.

[0016] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps in any of the above-described embodiments of the welding trajectory generation method.

[0017] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps in any of the above-described embodiments of the welding trajectory generation method.

[0018] The aforementioned welding trajectory generation method, apparatus, computer equipment, computer-readable storage medium, and computer program product first acquire the workpiece point cloud of the target workpiece, which can be a ship assembly workpiece including multiple stiffeners and a bottom plate. Second, multiple stiffener point clouds corresponding to the stiffeners are extracted from the workpiece point cloud, and multiple stiffener straight lines are extracted from the stiffener point cloud. These stiffener straight lines represent the projection position of the stiffeners on the bottom plate, thus facilitating the location of the weld between the stiffeners and the bottom plate. Then, to improve the continuity and completeness of weld trajectory recognition, a weld point cloud path is generated based on the intersection of the stiffener straight lines and the multiple stiffener straight lines. Further, by searching for loop information in the weld point cloud path, the welding trajectory is extracted from the weld point cloud. This improves the continuity and completeness of weld recognition, reduces the possibility of repeated recognition, and realizes automatic planning of the welding path. This is beneficial for improving the intelligence level and production efficiency of ship assembly welding, and reducing manual teaching costs and programming time. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is an application environment diagram of the welding trajectory generation method in one embodiment;

[0021] Figure 2 This is a flowchart illustrating a welding trajectory generation method in one embodiment;

[0022] Figure 3 This is a schematic diagram of the workpiece point cloud in one embodiment;

[0023] Figure 4 This is a flowchart illustrating the welding trajectory generation method in another embodiment;

[0024] Figure 5 This is a schematic diagram of the first weld point cloud circuit in one embodiment;

[0025] Figure 6 This is a flowchart illustrating the welding trajectory generation method in yet another embodiment;

[0026] Figure 7 This is a schematic diagram of a weld path segment of an L-shaped weld type in one embodiment;

[0027] Figure 8 This is a schematic diagram of a weld path segment of a concave weld type in one embodiment;

[0028] Figure 9 This is a schematic diagram of the straight lines of the stiffeners in one embodiment;

[0029] Figure 10 This is a structural block diagram of a welding trajectory generation device in one embodiment;

[0030] Figure 11 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0032] The welding trajectory generation method provided in this application embodiment can be applied to, for example... Figure 1In the application environment shown, the point cloud acquisition device 102 is communicatively connected to the control terminal 104. Specifically, the point cloud acquisition device 102 transmits the workpiece point cloud of the target workpiece to the control terminal 104. Then, the control terminal 104 acquires the workpiece point cloud of the target workpiece, which includes multiple stiffeners and a base plate. Next, it extracts the stiffener point clouds corresponding to the multiple stiffeners from the workpiece point cloud and extracts multiple stiffener straight lines from the stiffener point cloud. The stiffener straight lines represent the projection position of the stiffeners on the base plate. Then, based on the intersection between the stiffener straight lines and the multiple stiffener straight lines, it generates a weld seam point cloud path. Finally, based on the loop information in the weld seam point cloud path, it extracts the welding trajectory from the weld seam point cloud. The point cloud acquisition device 102 can be, but is not limited to, devices such as laser scanners and structured light scanners.

[0033] In one exemplary embodiment, such as Figure 2 As shown, a welding trajectory generation method is provided, which can be applied to... Figure 1 Taking control terminal 104 as an example, the explanation includes the following steps (hereinafter referred to as S): S100 to S400. Wherein:

[0034] S100, acquire the point cloud of the target workpiece, wherein the target workpiece includes multiple stiffeners and a base plate.

[0035] In this embodiment, the welding of a sub-assembly workpiece during the sub-assembly stage is used as an application scenario. The target workpiece may include a base plate and multiple stiffening plates, and the weld of the target workpiece is the position where the stiffening plate meets the base plate. The base plate may be a steel plate forming the bottom of a ship's hull. The stiffening plates may be plates used for welding or fixing to the base plate, designed to enhance the rigidity and strength of the structure.

[0036] In specific implementation, a pre-configured system can acquire the workpiece point cloud of the target workpiece. This pre-configured system may include a welding robot, a ground-rail cantilever, a line scan camera, a weld seam tracking sensor, a welding machine, and a control terminal. The welding robot is suspended upside down on the ground-rail cantilever, and the line scan camera is mounted on the cantilever, maintaining a constant relative position with the robot's base coordinate system. During the actual point cloud acquisition process, the ground rail moves at a constant speed along a preset path, synchronously driving the line scan camera to scan the target workpiece, thus obtaining the workpiece point cloud. In other embodiments, after acquiring the workpiece point cloud, visual calibration parameters can be used to convert the workpiece point cloud to the robot's base coordinate system to facilitate subsequent weld seam trajectory planning. Figure 3 As shown, the point cloud of the target workpiece is displayed.

[0037] S200: Extract multiple stiffener point clouds corresponding to stiffeners from the workpiece point cloud, and extract multiple stiffener straight lines from the stiffener point cloud. The stiffener straight lines represent the projection position of the stiffener on the base plate.

[0038] In practice, the workpiece point cloud can be pre-fitted using a point cloud plane fitting algorithm. Based on the fitting results, the workpiece point cloud can be separated into the base plate point cloud and the rib plate point cloud. After obtaining the rib plate point cloud, it is projected onto the base plate plane, and then denoised using a filtering algorithm. The point cloud plane fitting algorithm includes, but is not limited to, least squares, principal component analysis, and RANSCA (random sample consensus). The filtering algorithm includes, but is not limited to, statistical filtering and pass-through filtering.

[0039] Subsequently, the filtered stiffener point cloud is fitted with straight lines using a point cloud straight line fitting algorithm to obtain multiple stiffener straight lines. These algorithms include, but are not limited to, least squares, RANSCA, Hough transform, region-growing-based point cloud fitting algorithms, and clustering-based point cloud fitting algorithms.

[0040] S300 generates weld point cloud paths based on the intersections between the straight lines of each stiffener and multiple straight lines of the stiffener.

[0041] The intersection between the straight lines of the stiffener plate represents the point of intersection between the straight lines of the stiffener plate.

[0042] In this embodiment, based on the welding requirements of the target workpiece (welding trajectory planning is required at both ends of the intersection between the stiffening plate and the base plate), for each stiffening plate straight line, the point cloud on the stiffening plate straight line can be used as a node, and two edges can be added between adjacent nodes. Then, for stiffening plate straight lines with intersections, duplicate point clouds and duplicate edges are removed to generate the weld point cloud path. Alternatively, for each stiffening plate straight line, the point cloud on the stiffening plate straight line can be used as a node, and directed edges can be added between adjacent nodes. Then, for stiffening plate straight lines with intersections, duplicate point clouds and duplicate edges are removed to generate the weld point cloud path.

[0043] S400 extracts the welding trajectory from the weld point cloud based on the loop information in the weld point cloud path.

[0044] The loop information includes closed loop paths in the weld point cloud path.

[0045] In practice, for a weld point cloud path with directed edges, first query the closed path in the weld point cloud path, and then query whether there is a path in the weld point cloud that starts from a node and returns to the node after passing through several paths. If a closed path is found, the point cloud in the closed path is converted to the coordinate system of the welding robot end effector to obtain the welding trajectory of the closed path.

[0046] Subsequently, the closed path can be removed from the weld point cloud path with undirected edges. An undirected graph loop search algorithm can then be used to find the loop path, and the point cloud within the loop path can be transformed into the coordinate system of the welding robot's end effector to obtain the welding trajectory of the loop path. The undirected graph loop search algorithm can include, but is not limited to, depth-first search, breadth-first search, disjoint-set data structure, and degree-based deletion methods.

[0047] In the above welding trajectory generation method, firstly, the workpiece point cloud of the target workpiece is obtained. The target workpiece can be a ship assembly workpiece including multiple stiffeners and a bottom plate. Secondly, the stiffener point clouds corresponding to multiple stiffeners are extracted from the workpiece point cloud, and multiple stiffener straight lines are extracted from the stiffener point cloud. These stiffener straight lines represent the projection position of the stiffener on the bottom plate, which is beneficial for locating the weld position between the stiffener and the bottom plate. Then, to improve the continuity and completeness of weld trajectory recognition, a weld point cloud path is generated based on the intersection between each stiffener straight line and multiple stiffener straight lines. Furthermore, by searching for loop information in the weld point cloud path, the welding trajectory is extracted from the weld point cloud. This improves the continuity and completeness of weld recognition, reduces the possibility of repeated recognition, and realizes automatic planning of the welding path. This is beneficial for improving the intelligence level and production efficiency of ship assembly welding, and reducing manual teaching costs and programming time.

[0048] In an exemplary embodiment, a weld point cloud path, including S301 to S302, is generated based on the intersection between the straight lines of each stiffener and multiple straight lines of the stiffener. Wherein:

[0049] S301, determine the point cloud of intersections between the straight lines of each stiffener.

[0050] In practice, the line models of each pair of stiffening plates can be combined to solve for the point cloud of the intersection points between the two pairs of stiffening plates.

[0051] S302, for a target stiffening plate straight line with intersecting point clouds, a first weld point cloud path and a second weld point cloud path are generated based on the point cloud in the target stiffening plate straight line. The first weld point cloud path represents a directed weld path, and the second weld point cloud path represents an undirected weld path. The weld point cloud path includes the first weld point cloud path and the second weld point cloud path.

[0052] In practice, target stiffening lines with intersection point clouds are selected from multiple stiffening lines. Understandably, there are multiple target stiffening lines. Then, for each target stiffening line, the distance between each point cloud and the starting point cloud (including the starting point cloud, intersection point cloud, and ending point cloud) is determined, and the point clouds are sorted according to this distance. This process is repeated for each target stiffening line. Duplicate point clouds are then removed to obtain a set of weld point clouds. Next, using each point cloud in the weld point cloud set as a node, directed edges are added between adjacent nodes to obtain a directed graph containing the first weld point cloud path. Finally, using each point cloud in the weld point cloud set as a node, two edges are added between adjacent nodes to obtain an undirected graph containing the second weld point cloud path.

[0053] In this embodiment, by constructing directed and undirected graphs of the point cloud in the stiffener straight line, it is beneficial to determine the welding trajectory that meets the continuity requirement.

[0054] In one exemplary embodiment, the weld point cloud includes a first weld point cloud path and a second weld point cloud path, such as... Figure 4 As shown, based on the loop information in the weld point cloud path, the welding trajectory is extracted from the weld point cloud, including S401 to S404, where:

[0055] S401, find the first weld point cloud loop from the first weld point cloud path, remove the first weld point cloud loop from the second weld point cloud path, and obtain the target weld point cloud path.

[0056] In specific implementation, a depth-first search algorithm can be used to backtrack and find simple loops in the path of the first weld point cloud to obtain the first weld point cloud loop. A simple loop is characterized in a graph vertex sequence where, except for the first and last vertices being the same, the remaining vertices do not appear repeatedly. In other implementations, the simple loop search algorithm may also include a breadth-first search algorithm, etc.

[0057] For example, such as Figure 5 The diagram illustrates the first weld point cloud loop found from the first weld point cloud path. After obtaining the first weld point cloud loop, it is removed from the second weld point cloud path, i.e., the edges between nodes in the first weld point cloud loop are deleted from the second weld point cloud path to obtain the target weld point cloud path.

[0058] S402, determine the weld intersection type corresponding to the intersection point cloud in the target weld point cloud path.

[0059] The weld intersection type represents the type of weld at which the intersection is located. For example, a first weld intersection type, a second weld intersection type, and a third weld intersection type are preset. The first weld intersection type represents an L-type weld intersection type, the second weld intersection type represents a T-type weld intersection type, and the third weld intersection type represents an X-type weld intersection type.

[0060] In practical applications, corresponding discrimination conditions can be pre-set for different weld intersection types. For example, for an L-shaped weld intersection, the discrimination condition could be: the intersection point cloud is simultaneously an endpoint point cloud of both stiffening plate lines. For a T-shaped weld intersection, the discrimination condition could be: the intersection point cloud is an interior point (non-endpoint point cloud) of one stiffening plate line and an endpoint point cloud of the other stiffening plate line. For an X-shaped weld intersection, the discrimination condition could be: both intersection point clouds are interior points (non-endpoint point clouds) of both stiffening plate lines.

[0061] In practice, for a pair of straight stiffeners with intersecting point clouds, the endpoint point clouds of each straight stiffener in the pair can be determined. Then, based on the endpoint point clouds, the intersection point clouds, and the preset discrimination conditions, the weld intersection type of the intersection point clouds can be determined.

[0062] S403, based on the spatial position characteristics between node point clouds in the target weld point cloud path and the weld intersection type, find the second weld point cloud loop from the target weld point cloud path.

[0063] The spatial location features may include the angle between the line vector of the stiffening line to which the node point cloud belongs and the line vector of the stiffening line to which the adjacent node point cloud belongs.

[0064] In practical implementation, the Hierholzer algorithm can be used to find an Eulerian circuit from the target weld point cloud path to obtain the second weld point cloud circuit. During the algorithm backtracking process, when visiting adjacent node point clouds, if there are multiple traversable edges in the node point cloud, and each traversable edge corresponds to an adjacent node, the angle between the line vector of the stiffener line to which the node point cloud belongs and the line vector of the stiffener line to which the adjacent node point cloud belongs can be determined. The adjacent node corresponding to the smallest angle is then selected, and the edge between the node point cloud and that adjacent node is visited. After the backtracking is completed, the second weld point cloud circuit is obtained.

[0065] S404 generates welding trajectories based on the first weld point cloud circuit and the second weld point cloud circuit, respectively.

[0066] In practice, the first weld point cloud loop and the second weld point cloud loop can be converted into the coordinate system of the actuator according to the posture of the actuator at the end of the welding robot (such as the posture of the welding gun) to obtain the welding trajectory.

[0067] In this embodiment, by analyzing the point cloud in the stiffener straight line into directed and undirected graphs, and by using loop finding and elimination, weld paths with closed simple loops and loops containing multiple different weld types are filtered out, thereby improving the continuity and integrity of the welding trajectory and thus improving welding accuracy.

[0068] In one exemplary embodiment, the spatial location feature includes an directional angle. For example... Figure 6 As shown, based on the weld intersection type and the spatial location characteristics between the node point clouds, the second weld point cloud loop is found from the target weld point cloud path, including S431 to S433, where:

[0069] Using the preset node point cloud in the target weld point cloud path as the current node point cloud, iteratively execute the following S431 to S433 until all edges of the node point cloud in the target weld point cloud path are in the visited state, and obtain the second weld point cloud loop.

[0070] S431, if there is an unvisited edge in the current node point cloud, determine the adjacent node point cloud as the current node point cloud and update the edge to the visited state.

[0071] S432, when the current node point cloud is an intersection point cloud and the weld intersection type is the target weld intersection type, from the target adjacent node point cloud corresponding to the minimum direction angle among multiple adjacent node point clouds, update the edge between the current node point cloud and the target adjacent node point cloud to the visited state, and determine the target adjacent node point cloud as the current node point cloud.

[0072] S433: If there are no unvisited edges in the current node's point cloud, perform node point cloud rollback.

[0073] In this embodiment, based on the improved Hierholzer algorithm, an Eulerian circuit is found from the target weld point cloud path to obtain the second weld point cloud circuit. An Eulerian circuit is defined as follows: if a path in the graph includes each edge exactly once, then the path is an Eulerian path; if a circuit is an Eulerian path, then the circuit is an Eulerian circuit.

[0074] In practice, one can pre-set a node point cloud in the target weld point cloud path as the current node point cloud (representing the starting point), use a stack to record the currently visited path, and then backtrack to execute the following steps:

[0075] For S431, based on the visited paths in the stack, determine whether the edge between the current node point cloud and the adjacent node point cloud is an unvisited edge. If the current node point cloud has only one unvisited edge, then the adjacent node corresponding to the unvisited edge is determined as the current node point cloud, the unvisited edge is pushed onto the stack, and the edge is updated to the visited state.

[0076] For S432, the target weld intersection type can include a second weld intersection type (representing a T-type weld intersection) and a third weld intersection type (representing an X-type weld intersection). Regarding the direction angle, let the current node point cloud be b, the previous node point cloud be a, the adjacent node point cloud be c, the direction vector between the previous node point cloud and the current node point cloud be (a, b), and the direction vector between the current node point cloud and the adjacent node point cloud be (b, c). Then, the direction angle represents the angle between the direction vectors (a, b) and (b, c).

[0077] If the current node point cloud does not meet the conditions in S431, it indicates that the current node point cloud has multiple unvisited adjacent node point clouds and corresponding edges. Further, it is determined whether the current node point cloud is an intersection point cloud. If it is, the directional angles between the current node point cloud and multiple adjacent nodes are determined, with the method for determining the directional angles referring to the above implementation method. Subsequently, the target adjacent node point cloud corresponding to the minimum directional angle is selected from the multiple adjacent node point clouds. The edge between the current node point cloud and the target adjacent node point cloud is pushed onto the stack and updated to a visited state. The target adjacent node point cloud is then updated to the current node point cloud. This helps improve the smoothness of the welding trajectory.

[0078] For S433, if the current node point cloud does not have any unvisited edges, then the current node point cloud is added to the second weld point cloud loop and rolled back.

[0079] Thus, by backtracking the improved Hilholzer algorithm, the point cloud sequence corresponding to the Eulerian circuit in the target weld point cloud path was found. Then, the order of the point cloud sequence was reversed to obtain the second weld point cloud circuit.

[0080] In this embodiment, an improved Hilholzer algorithm is used to find Eulerian circuits in the target weld point cloud path. On the one hand, this helps to improve the continuity and integrity of the weld path, reduce repetition, and adapt to the traversal of complex weld networks. On the other hand, the edge traversal mechanism based on the minimum directional angle is beneficial to the smoothness of the welding trajectory.

[0081] In an exemplary embodiment, generating the welding trajectory based on the first weld point cloud loop and the second weld point cloud loop includes steps S441 to S443, wherein:

[0082] S441, Based on the weld direction characteristics corresponding to the first weld point cloud loop, generate a first welding coordinate system, transform the node point cloud in the first weld point cloud loop to the first welding coordinate system, and obtain the first welding trajectory.

[0083] Among them, the weld direction feature represents the weld advancing direction of the first weld point cloud loop. Specifically, it can be the starting point cloud in the first weld point cloud loop.

[0084] In specific implementation, the direction vector between the starting point cloud and its adjacent point clouds in the first weld point cloud loop is determined to obtain the weld advancing direction. This weld advancing direction is then defined as the X-axis of the first welding coordinate system. Subsequently, the centroid point cloud of the first weld point cloud loop is determined, and the unit vector between the starting point cloud and the centroid point cloud is defined as the Y-axis of the first welding coordinate system. Then, the X-axis and Y-axis are cross-multiplied to obtain the Z-axis, and this calculated Z-axis is compared with the Z-axis of the world coordinate system. If the calculated Z-axis is opposite in direction to the Z-axis in the world coordinate system, the order of the node point clouds in the first weld point cloud loop is reversed. Finally, the node point clouds in the first weld point cloud loop are transformed into the first welding coordinate system to obtain the first welding trajectory.

[0085] S442, based on the spatial position characteristics between the node point clouds in the second weld point cloud loop, the target weld point cloud path is divided into weld path segments of multiple weld types.

[0086] Among them, spatial location features may include collinearity between node point clouds and the angle between the path segment vectors to which the node point clouds belong.

[0087] In specific implementation, the weld type of the weld path segment to which the node point cloud belongs can be determined based on the spatial position characteristics between the node point clouds in the second weld point cloud loop, and the second weld point cloud loop can be divided into multiple weld path segments according to the weld type. Specifically:

[0088] (1) Determine the node type of each node point cloud in the second weld point cloud loop. The node type includes endpoint and weld intersection types. Starting from the first endpoint point cloud in the second weld point cloud loop, create a point cloud sequence to represent the segmented weld path segment and add the starting point to the point cloud sequence.

[0089] (2) Traverse and visit the next node point cloud. At the same time, determine whether the next node point cloud is collinear with the last point cloud in the point cloud sequence (located on the same stiffener line). If it is determined to be collinear and the point cloud sequence does not contain the next node point cloud, then add the next node point cloud to the point cloud sequence.

[0090] (3) If the next node point cloud is collinear with the last point cloud in the point cloud sequence and the point cloud contains the next node point cloud, then according to the point cloud in the point cloud sequence, the weld path segment is segmented from the second weld point cloud loop, and the weld type of the weld path segment is determined to be a straight weld type. Then, taking the last point in the point cloud sequence as the new starting point, a new point cloud sequence is created, and the process returns to step (2).

[0091] (4) Traverse the next node point cloud. If the next node point cloud is not collinear with the last point cloud in the point cloud sequence, add the next node point cloud to the point cloud sequence. At the same time, determine whether the next node point cloud is an endpoint. If it is determined to be an endpoint, segment the weld path segment from the second weld point cloud loop according to the point clouds in the point cloud sequence. Then, for the point clouds in the point cloud sequence, determine the vector from the starting point to the second point cloud and the vector from the second-to-last point to the last point cloud. Calculate the angle between these two vectors and determine the weld type of the weld path segment based on the angle between the vectors.

[0092] If the included angle is greater than a preset first included angle threshold (e.g., 150°), the weld type of the weld path segment is determined to be a concave weld type. Otherwise, if the included angle is less than a preset second included angle threshold (e.g., 20°), the weld type of the weld path segment is determined to be an unclosed polygonal weld. Otherwise, the weld type of the weld path segment is determined to be an L-shaped weld. Figure 7 As shown, weld path segments for multiple L-shaped weld types are illustrated. Figure 8 As shown, weld path segments for multiple concave weld types are illustrated.

[0093] S443, for each weld path segment, generate a second welding coordinate system corresponding to the weld path segment based on the weld direction characteristics corresponding to the weld path segment, transform the node point cloud in each weld path segment to the second welding coordinate system, and obtain the second welding trajectory, which includes the first welding trajectory and the second welding trajectory.

[0094] In specific implementation, a second welding coordinate system is constructed for each weld path segment. Here, it can be the implementation method described in the above embodiment, which generates the first welding coordinate system based on the weld direction characteristics corresponding to the first weld point cloud loop. Specifically, for each weld path segment, the vector between the starting point cloud and the adjacent point cloud of the weld path segment is determined as the X-axis of the second weld coordinate system corresponding to that weld path segment. The centroid point cloud of the weld path segment is determined. Then, the unit vector pointing from the starting point cloud to the centroid point cloud is determined as the Y-axis of the second welding coordinate system. The cross product of the X-axis and Y-axis is used to obtain the Z-axis of the second welding coordinate system. The Z-axis is compared with the Z-axis in the world coordinate system. If the directions are opposite, the order of the point clouds is reversed.

[0095] Traverse each weld path segment to obtain the second welding coordinate system corresponding to each weld path segment, and transform the point cloud in the weld path segment to the corresponding second welding coordinate system. Sort the transformed weld path segments according to the segmentation order to obtain the second welding trajectory.

[0096] In this embodiment, the weld morphology is determined by the spatial position characteristics between the node point clouds in the second weld point cloud loop, thereby realizing intelligent classification and segmentation of the path. This is beneficial for improving the accuracy of welding path planning by combining path optimization algorithms.

[0097] In an exemplary embodiment, multiple rib lines are extracted from the rib point cloud, including S201 to S202, wherein:

[0098] S201, perform linear fitting on the rib point cloud to obtain multiple rib line segments.

[0099] Among them, the stiffener line segment includes multiple stiffener point clouds with the same stiffener straight line vector.

[0100] In practice, a preset point cloud line fitting algorithm can be used to perform line fitting on the rib point cloud to obtain multiple fitted rib line vectors. Interior points of the rib lines can then be extracted from the rib point cloud to obtain multiple rib line segments. For example, the RANSAC algorithm can be used to perform line fitting on the rib point cloud to extract interior points until the number of remaining rib point clouds is less than a preset point threshold, thus obtaining multiple rib line segments.

[0101] S202, based on the spatial position characteristics between pairs of stiffening plate segments, perform straight line fitting on the point cloud of multiple stiffening plate segments to obtain multiple stiffening plate straight lines.

[0102] Among them, spatial location features may include, but are not limited to, the angle between the direction vectors of the two stiffener line segments, the distance between the intersection points of their extensions, etc.

[0103] In specific implementation, for any two stiffening plate segments, the angle between the direction vectors of these two stiffening plate segments can be determined. If the angle is less than a preset angle threshold, then these two stiffening plate segments are determined to be a pair of stiffening plate segments that need to be fitted with a straight line. In other implementations, after filtering out stiffening plate segment pairs with an angle less than a preset angle threshold, the distance between the intersection point of the extension lines of one stiffening plate segment and the endpoint of the other stiffening plate segment can be determined. If the intersection point distance is less than a preset distance threshold, then the pair of stiffening plate segments is determined to be a pair of stiffening plate segments that need to be fitted with a straight line.

[0104] Then, the point clouds of the stiffener plate segments that need to be fitted with straight lines can be merged into a set, and straight line fitting can be performed on the point clouds in the set to obtain multiple stiffener plate straight lines. These stiffener plate straight lines represent the weld between the stiffener plate and the base plate.

[0105] In this embodiment, the stiffener point cloud is fitted with a straight line again based on the spatial position characteristics between the stiffener segments to obtain a stiffener straight line that represents the weld between the stiffener and the base plate, thereby improving the continuity and completeness of weld identification.

[0106] In an exemplary embodiment, the spatial position features include included angle and minimum distance. Based on the spatial position features between pairs of stiffening plate segments, straight line fitting is performed on the point cloud of multiple stiffening plate segments to obtain multiple stiffening plate straight lines, including S220 to S221, wherein:

[0107] S220: Select target stiffener line segment pairs from multiple stiffener line segments, where the included angle is less than a preset included angle threshold and the minimum distance is less than a preset distance threshold.

[0108] Here, the included angle represents the cosine of the angle between the straight-line direction vectors of the two stiffener segments. The minimum distance represents the minimum distance between the point clouds in the two stiffener segments.

[0109] In practice, the process first iterates through combinations of multiple stiffening segments, with each combination containing two stiffening segments. For each combination, assuming it contains stiffening segment a and stiffening segment b, the angle between the line direction vectors of stiffening segment a and b is determined. This angle is compared to a preset angle threshold. If the angle is less than the preset threshold, the distance between the endpoint point clouds of stiffening segment a and the point clouds of each stiffening segment in stiffening segment b is further determined, and the minimum distance is selected. Then, the minimum distance is compared to a preset distance threshold. If the minimum distance is less than the preset minimum distance threshold, it indicates that stiffening segment a and stiffening segment b meet the conditions for segment merging, and they are identified as the target stiffening segment pair.

[0110] S221, perform straight line fitting on the point cloud of the target stiffener line segment to obtain the stiffener line.

[0111] In practice, following the steps above, after selecting multiple target stiffener line segment pairs, an adjacency graph is constructed based on these pairs. Specifically, taking the stiffener line segments in the target stiffener line segment pairs as nodes, edges are added between the nodes corresponding to the stiffener line segments in the target stiffener line segment pairs where the included angle is less than a preset included angle threshold and the minimum distance is less than a preset distance threshold, thus obtaining the adjacency graph.

[0112] Subsequently, a node in the adjacency graph can be determined as the starting node. Starting from the starting node, all adjacent nodes are visited layer by layer until all reachable nodes are visited, thus obtaining the access path. The stiffener segments corresponding to the nodes in the access path represent the stiffener segments that need to be merged into a single stiffener line.

[0113] Next, the stiffener point clouds of the stiffener line segments corresponding to the nodes in the access path are merged into a single point cloud set. The boundary of this point cloud set is identified to obtain the boundary point cloud. Then, a straight line is fitted to the boundary point cloud to obtain the stiffener straight line, which represents the weld between the stiffener and the base plate. The point cloud boundary identification methods include, but are not limited to, methods based on neighborhood analysis, such as determining the boundary by calculating the rate of change of the normal vector of points in the neighborhood; methods based on density analysis, using point cloud density differences to identify boundary regions; and methods based on curvature estimation, calculating the curvature change of the point cloud surface to identify boundary points with higher curvature. The straight line fitting method refers to the straight line fitting method in the above embodiment, and will not be repeated here. For example, as shown... Figure 9 As shown, this illustrates the fitted straight lines of multiple stiffeners.

[0114] In this embodiment, a dual threshold judgment based on angle and distance is introduced to construct an adjacency graph of stiffener segments. Then, the adjacency graph traversal algorithm is used to visit mergeable segments layer by layer to filter out weld segment groups belonging to the same stiffener, thereby improving the accuracy of weld identification.

[0115] To provide a clearer explanation of the welding trajectory generation method provided in this application, a specific embodiment is described below, which includes the following steps:

[0116] S1, Obtain the point cloud of the target workpiece, wherein the target workpiece includes multiple stiffeners and a base plate.

[0117] S2, extract the stiffener point cloud corresponding to multiple stiffeners from the workpiece point cloud, perform linear fitting on the stiffener point cloud to obtain multiple stiffener line segments, select target stiffener line segment pairs from the multiple stiffener line segments with included angle less than a preset included angle threshold and minimum distance less than a preset distance threshold, perform linear fitting on the point cloud of the target stiffener line segment pairs to obtain stiffener straight lines, where stiffener straight lines represent the projection position of the stiffener on the base plate.

[0118] S3, determine the point cloud of intersection points between the straight lines of each stiffener plate. For the target stiffener plate straight line with intersection point clouds, generate the first weld point cloud path and the second weld point cloud path respectively based on the point cloud in the target stiffener plate straight line. The first weld point cloud path represents the directed weld path, and the second weld point cloud path represents the undirected weld path.

[0119] S4, find the first weld point cloud loop from the first weld point cloud path, remove the first weld point cloud loop from the second weld point cloud path, and obtain the target weld point cloud path.

[0120] S5, determine the weld intersection type corresponding to the intersection point cloud in the target weld point cloud path.

[0121] S6, taking the preset node point cloud in the target weld point cloud path as the current node point cloud, iteratively execute S7 to S9 until all edges of the node point cloud in the target weld point cloud path are in the visited state, and obtain the second weld point cloud loop.

[0122] S7. If there is an unvisited edge in the current node point cloud, determine the adjacent node point cloud as the current node point cloud and update the edge to the visited state.

[0123] S8. When the current node point cloud is an intersection point cloud and the weld intersection type is the target weld intersection type, from the target adjacent node point cloud corresponding to the minimum direction angle among multiple adjacent node point clouds, update the edge between the current node point cloud and the target adjacent node point cloud to the visited state, and determine the target adjacent node point cloud as the current node point cloud.

[0124] S9. If there are no unvisited edges in the current node's point cloud, perform node point cloud rollback processing.

[0125] S10. Based on the weld direction characteristics corresponding to the first weld point cloud loop, generate a first welding coordinate system, and transform the node point cloud in the first weld point cloud loop to the first welding coordinate system to obtain the first welding trajectory.

[0126] S11, based on the spatial position characteristics between the node point clouds in the second weld point cloud loop, the target weld point cloud path is divided into weld path segments of multiple weld types.

[0127] S12, for each weld path segment, generate the second welding coordinate system corresponding to the weld path segment based on the weld direction characteristics of the weld path segment, and transform the node point cloud in each weld path segment to the second welding coordinate system to obtain the second welding trajectory.

[0128] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0129] In one exemplary embodiment, such as Figure 10As shown, a welding trajectory generation device 600 is provided, including: a point cloud acquisition module 610, a stiffener straight line extraction module 620, a weld point cloud path generation module 630, and a welding trajectory generation module 640, wherein:

[0130] The point cloud acquisition module 610 is used to acquire the point cloud of the target workpiece, wherein the target workpiece includes multiple stiffeners and a base plate.

[0131] The rib plate straight line extraction module 620 is used to extract multiple rib plate point clouds corresponding to multiple rib plates from the workpiece point cloud, and extract multiple rib plate straight lines from the rib plate point cloud, wherein the rib plate straight lines represent the projection position of the rib plate on the base plate.

[0132] The weld point cloud path generation module 630 is used to generate weld point cloud paths based on the intersection between the straight lines of each stiffener and multiple straight lines of the stiffener.

[0133] The welding trajectory generation module 640 is used to extract the welding trajectory from the weld point cloud based on the loop information in the weld point cloud path.

[0134] In an exemplary embodiment, the weld point cloud path generation module 630 is further configured to determine the intersection point cloud between each stiffening plate straight line; for a target stiffening plate straight line with intersection point clouds, a first weld point cloud path and a second weld point cloud path are generated according to the point cloud in the target stiffening plate straight line, the first weld point cloud path representing a directed weld path and the second weld point cloud path representing an undirected weld path; the weld point cloud path includes the first weld point cloud path and the second weld point cloud path.

[0135] In an exemplary embodiment, the welding trajectory generation module 640 is further configured to: find a first weld point cloud loop from the first weld point cloud path; remove the first weld point cloud loop from the second weld point cloud path to obtain a target weld point cloud path; determine the weld intersection type corresponding to the intersection point cloud in the target weld point cloud path; find a second weld point cloud loop from the target weld point cloud path based on the spatial position characteristics between node point clouds and the weld intersection type; and generate a welding trajectory based on the first weld point cloud loop and the second weld point cloud loop, respectively.

[0136] In an exemplary embodiment, the welding trajectory generation module 640 is further configured to iteratively execute the following steps using a preset node point cloud in the target weld point cloud path as the current node point cloud, until all edges of the node point clouds in the target weld point cloud path are in a visited state, thereby obtaining a second weld point cloud loop: if there is an unvisited edge in the current node point cloud, the adjacent node point cloud is determined as the current node point cloud, and the edge is updated to a visited state; if the current node point cloud is an intersection point cloud, and the weld intersection type is the target weld intersection type, the edge between the current node point cloud and the target adjacent node point cloud is updated to a visited state from the target adjacent node point cloud corresponding to the minimum direction angle among multiple adjacent node point clouds, and the target adjacent node point cloud is determined as the current node point cloud; if there is no unvisited edge in the current node point cloud, node point cloud rollback processing is performed.

[0137] In an exemplary embodiment, the welding trajectory generation module 640 is further configured to generate a first welding coordinate system based on the weld direction features corresponding to the first weld point cloud loop, and convert the node point clouds in the first weld point cloud loop to the first welding coordinate system to obtain a first welding trajectory; based on the spatial position features between the node point clouds in the second weld point cloud loop, divide the target weld point cloud path into weld path segments of multiple weld types; for each weld path segment, based on the weld direction features corresponding to the weld path segment, generate a second welding coordinate system corresponding to the weld path segment, and convert the node point clouds in each weld path segment to the second welding coordinate system to obtain a second welding trajectory; the welding trajectory includes the first welding trajectory and the second welding trajectory.

[0138] In an exemplary embodiment, the rib line extraction module 620 is further configured to perform line fitting on the rib point cloud to obtain multiple rib line segments; and to perform line fitting on the point cloud of multiple rib line segments based on the spatial position characteristics between pairs of rib line segments to obtain multiple rib lines.

[0139] In an exemplary embodiment, the rib line extraction module 620 is further configured to filter out target rib line segment pairs from multiple rib line segments, wherein the included angle is less than a preset included angle threshold and the minimum distance is less than a preset distance threshold; and to perform line fitting on the point cloud of the target rib line segment pairs to obtain rib lines.

[0140] Each module in the aforementioned welding trajectory generation device 600 can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.

[0141] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 8 As shown, this computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network. When executed by the processor, the computer program implements a welding trajectory generation method.

[0142] Those skilled in the art will understand that Figure 8 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0143] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in any of the above embodiments of the welding trajectory generation method.

[0144] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the steps in any of the above embodiments of the welding trajectory generation method.

[0145] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in any of the above embodiments of the welding trajectory generation method.

[0146] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0147] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0148] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0149] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for generating welding trajectories, characterized in that, The method includes: Obtain the point cloud of the target workpiece, wherein the target workpiece includes multiple stiffeners and a base plate; Extract the rib point cloud corresponding to the multiple ribs from the workpiece point cloud, and extract multiple rib straight lines from the rib point cloud, wherein the rib straight lines represent the projection position of the rib on the base plate; Based on the intersection between the straight lines of each stiffener and the multiple straight lines of the stiffener, a weld point cloud path is generated; The welding trajectory is extracted from the weld point cloud based on the loop information in the weld point cloud path.

2. The method according to claim 1, characterized in that, The step of generating a weld point cloud path based on the intersection between the straight lines of each stiffener and the multiple straight lines of the stiffener includes: Determine the point cloud of intersections between the straight lines of each of the aforementioned stiffeners; For a target stiffening plate straight line with intersecting point clouds, a first weld point cloud path and a second weld point cloud path are generated based on the point cloud in the target stiffening plate straight line. The first weld point cloud path represents a directed weld path, and the second weld point cloud path represents an undirected weld path. The weld seam point cloud path includes the first weld seam point cloud path and the second weld seam point cloud path.

3. The method according to claim 1, characterized in that, The weld point cloud includes a first weld point cloud path and a second weld point cloud path; the step of extracting the welding trajectory from the weld point cloud based on the loop information in the weld point cloud path includes: Find the first weld point cloud loop from the first weld point cloud path, and remove the first weld point cloud loop from the second weld point cloud path to obtain the target weld point cloud path. Determine the weld intersection type corresponding to the intersection point cloud in the target weld point cloud path; Based on the spatial position characteristics between node point clouds in the target weld point cloud path and the weld intersection type, a second weld point cloud loop is found from the target weld point cloud path. Welding trajectories are generated based on the first weld point cloud circuit and the second weld point cloud circuit, respectively.

4. The method according to claim 3, characterized in that, The spatial location features include the included angle; the step of finding the second weld point cloud loop from the target weld point cloud path based on the weld intersection type and the spatial location features between the node point clouds includes: Using a preset node point cloud in the target weld point cloud path as the current node point cloud, iteratively execute the following steps until all edges of the node point cloud in the target weld point cloud path are visited, thus obtaining a second weld point cloud loop: If there is an unvisited edge in the current node point cloud, the adjacent node point cloud is determined as the current node point cloud, and the edge is updated to a visited state. When the current node point cloud is an intersection point cloud and the weld intersection type is the target weld intersection type, from the target adjacent node point cloud corresponding to the minimum direction angle among multiple adjacent node point clouds, update the edge between the current node point cloud and the target adjacent node point cloud to the visited state, and determine the target adjacent node point cloud as the current node point cloud. If there are no unvisited edges in the current node point cloud, perform node point cloud rollback processing.

5. The method according to claim 3, characterized in that, The step of generating welding trajectories based on the first weld point cloud circuit and the second weld point cloud circuit respectively includes: Based on the weld direction features corresponding to the first weld point cloud loop, a first welding coordinate system is generated, and the node point cloud in the first weld point cloud loop is transformed to the first welding coordinate system to obtain the first welding trajectory. Based on the spatial position characteristics between the node point clouds in the second weld point cloud loop, the target weld point cloud path is divided into weld path segments of multiple weld types. For each weld path segment, a second welding coordinate system is generated based on the weld direction characteristics corresponding to the weld path segment. The node point cloud in each weld path segment is transformed into the second welding coordinate system to obtain the second welding trajectory. The welding trajectory includes the first welding trajectory and the second welding trajectory.

6. The method according to claim 1, characterized in that, The extraction of multiple straight lines from the rib point cloud includes: Linear fitting is performed on the point cloud of the stiffener to obtain multiple stiffener line segments; Based on the spatial positional characteristics between pairs of stiffening plate segments, straight lines are fitted to the point clouds of multiple stiffening plate segments to obtain multiple stiffening plate straight lines.

7. The method according to claim 6, characterized in that, The spatial location features include the included angle and the minimum distance; The process involves fitting straight lines to the point clouds of multiple stiffening plate segments based on the spatial positional characteristics between pairs of stiffening plate segments, resulting in multiple stiffening plate straight lines, including: Select target rib plate line segment pairs from the plurality of rib plate line segments, where the included angle is less than a preset included angle threshold and the minimum distance is less than a preset distance threshold; The point cloud of the target stiffener line segment is fitted with a straight line to obtain the stiffener line.

8. A welding trajectory generation device, characterized in that, The device includes: The point cloud acquisition module is used to acquire the point cloud of the target workpiece, wherein the target workpiece includes multiple stiffeners and a base plate; The rib plate straight line extraction module is used to extract the rib plate point cloud corresponding to the multiple rib plates from the workpiece point cloud, and extract multiple rib plate straight lines from the rib plate point cloud, wherein the rib plate straight lines represent the projection position of the rib plate on the base plate. The weld point cloud path generation module is used to generate weld point cloud paths based on the intersection between the straight lines of each stiffener and the multiple straight lines of the stiffener. The welding trajectory generation module is used to extract the welding trajectory from the weld point cloud based on the loop information in the weld point cloud path.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images