Search

CN-122024218-A - Method for recognizing a weld, computer device, readable storage medium and program product

CN122024218ACN 122024218 ACN122024218 ACN 122024218ACN-122024218-A

Abstract

The application relates to a method for recognizing a vertical welding seam, a computer device, a computer readable storage medium and a computer program product. The method comprises the steps of obtaining target simplex piece point clouds of a target workpiece, extracting rib plate point clouds corresponding to the rib plates from the target simplex piece point clouds, respectively projecting the rib plate point clouds to a plane where the base plate is located, downsampling the projected rib plate point clouds to obtain a plurality of weld point clouds, and detecting spatial position relations between an intersecting position and the first and second linear weld point clouds when the first and second linear weld point clouds are identified from the weld point clouds and intersect, and generating a vertical identification result according to the spatial position relations. By adopting the method, the accuracy of the vertical weld joint identification can be improved.

Inventors

  • SUN JUNDI
  • ZHAO LEI
  • DENG WENPING

Assignees

  • 湖南视比特机器人有限公司

Dates

Publication Date
20260512
Application Date
20260119

Claims (13)

  1. 1. A method of seam weld identification, the method comprising: acquiring a target single-workpiece point cloud of a target workpiece, wherein the target workpiece comprises a plurality of rib plates and a bottom plate; extracting rib plate point clouds corresponding to the rib plates from the target single-workpiece point clouds; projecting each rib plate point cloud to a plane where the bottom plate is located, and performing centroid downsampling on each projected rib plate point cloud to obtain a plurality of weld point clouds; Detecting a spatial position relationship between an intersecting position and the first and second linear weld point clouds in a case where the first and second linear weld point clouds are identified from among the weld point clouds and the first and second linear weld point clouds intersect; and generating a vertical weld recognition result according to the spatial position relation.
  2. 2. The method of claim 1, wherein the acquiring a target single-workpiece point cloud of the target workpiece comprises: acquiring multiple workpiece scanning point clouds and platform template point clouds of the platform, and dividing single workpiece scanning point clouds corresponding to each workpiece from the multiple workpiece scanning point clouds; According to the first straight-through filtering range corresponding to each workpiece, carrying out straight-through filtering on the platform template point cloud to obtain initial platform point clouds corresponding to each workpiece; Fitting the respective corresponding platform planes of the workpieces according to the respective initial platform point clouds, and projecting the respective initial platform point clouds to the respective corresponding platform planes to obtain respective corresponding target platform point clouds of the workpieces; according to the second direct filtering range corresponding to each workpiece, direct filtering is carried out on the single workpiece scanning point cloud corresponding to each workpiece, and the single workpiece point cloud corresponding to each workpiece is obtained; taking any workpiece as a target workpiece, and taking a single workpiece point cloud corresponding to the target workpiece as a target single workpiece point cloud; The maximum value of a first filtering sub-range perpendicular to the platform direction in the first straight-through filtering range is the sum of the maximum value of the corresponding single-workpiece scanning point cloud in the direction perpendicular to the platform direction and a preset first increment value, and the maximum value of a second filtering sub-range perpendicular to the platform direction in the second straight-through filtering range is the sum of the minimum value of the target platform point cloud in the direction perpendicular to the platform direction and a preset second increment value.
  3. 3. The method of claim 2, wherein the dividing the single workpiece scanning point cloud for each of the workpieces from the multiple workpiece scanning point cloud comprises Downsampling the multi-workpiece scanning point cloud to obtain a multi-workpiece downsampling point cloud; Removing platform background points matched with the platform template point cloud from the multi-workpiece downsampling point cloud to obtain a multi-workpiece initial point cloud; projecting the multi-workpiece initial point cloud to a preset reference level, and carrying out clustering segmentation on the projected multi-workpiece initial point cloud to obtain single-workpiece projection point clouds corresponding to the workpieces; Determining a third straight-through filtering range corresponding to each workpiece according to each single-workpiece projection point cloud, wherein the third straight-through filtering range does not restrict the direction perpendicular to the platform; and respectively carrying out direct filtering on the multi-workpiece scanning point clouds according to the third direct filtering ranges to obtain single-workpiece scanning point clouds corresponding to the workpieces.
  4. 4. The method according to claim 1, wherein in a case where a first straight-line weld point cloud and a second straight-line weld point cloud are identified from among the weld point clouds, and the first straight-line weld point cloud intersects the second straight-line weld point cloud, before detecting a spatial positional relationship between an intersection position and the first straight-line weld point cloud and the second straight-line weld point cloud, the method further comprises: detecting an included angle between any two straight-line weld point clouds under the condition that a plurality of straight-line weld point clouds are identified from the weld point clouds; under the condition that an included angle between a first linear welding seam point cloud and a second linear welding seam point cloud is in a preset intersection angle range, determining that the first linear welding seam point cloud is intersected with the second linear welding seam point cloud, wherein the first linear welding seam point cloud is one of the linear welding seam point clouds, and the second linear welding seam point cloud is one of the linear welding seam point clouds.
  5. 5. The method of claim 4, wherein for each weld point cloud, in the event that a plurality of linear weld point clouds are identified from each of the weld point clouds, prior to detecting the included angle between two of the linear weld point clouds, the method further comprises: traversing points in the weld point cloud, and detecting the number of endpoints in the weld point cloud; When the number of the endpoints in the weld point cloud is zero, determining that the weld point cloud is a closed curve weld point cloud; when the number of the endpoints in the weld point cloud is two, constructing endpoint connecting line segments between the two endpoints, and detecting the interval distance between each weld point and the endpoint connecting line segments; and determining the weld point cloud to be a straight weld point cloud under the condition that the average value of the interval distances is smaller than a preset distance threshold value.
  6. 6. The method of claim 5, wherein the detecting the number of endpoints in the weld point cloud comprises: Determining at least one candidate endpoint from the weld point cloud according to a direction information set of each point in the weld point cloud pointing to other points; performing radius clustering on each candidate endpoint to obtain a plurality of clustering clusters; And determining the geometric center points of the clusters as end points, and counting the number of the end points.
  7. 7. The method according to any one of claims 1 to 6, wherein generating a standing seam recognition result from the spatial positional relationship comprises: When the intersection position is coincident with any end point of the first linear weld point cloud and the intersection position is coincident with any end point of the second linear weld point cloud, a first search vector is built by taking the intersection position as a starting point and taking the other end point, which is not coincident with the intersection position, of the first linear weld point cloud as an end point; Determining a first height according to the first linear weld point cloud and the first search vector, and determining a second height according to the second linear weld point cloud and the second search vector; determining that the number of the vertical welds is one under the condition that the first height and the second height are both larger than a preset height threshold value; taking the intersection position as a track starting point of the vertical welding seam, and taking the smaller one of the first height and the second height as the height of the vertical welding seam; and generating a vertical welding seam recognition result according to the track starting point, the height of the vertical welding seam, the first search vector and the second search vector.
  8. 8. The method according to any one of claims 1 to 6, wherein generating a standing seam recognition result from the spatial positional relationship comprises: When the intersection position is coincident with any end point of the first linear weld point cloud and the intersection position is not coincident with any end point of the second linear weld point cloud, constructing a third search vector by taking the intersection position as a starting point and the other end point which is not coincident with the intersection position in the first linear weld point cloud as an end point, and constructing a fourth search vector by taking the intersection position as a starting point and any end point which is not coincident with the intersection position in the second linear weld point cloud as an end point; Determining a third height according to the first linear weld point cloud and the third search vector, determining a fourth height according to the second linear weld point cloud and the fourth search vector, and determining a fifth height according to the second linear weld point cloud and the fifth search vector; Determining that the number of the vertical welds is two under the condition that the third height, the fourth height and the fifth height are all larger than a preset height threshold value; Taking the intersection position as a track starting point of the vertical welding seam, taking the smaller of the third height and the fourth height as the height of the first vertical welding seam, and taking the smaller of the third height and the fifth height as the height of the second vertical welding seam; generating a first vertical welding seam identifier result according to the track starting point, the height of the first vertical welding seam, the third search vector and the fourth search vector, generating a second vertical welding seam identifier result according to the track starting point, the height of the second vertical welding seam, the fourth search vector and the fifth search vector, and taking the first vertical welding seam identifier result and the second vertical welding seam identifier result as vertical welding seam identifier results.
  9. 9. The method according to any one of claims 1 to 6, wherein generating a standing seam recognition result from the spatial positional relationship comprises: When the intersection position is not coincident with any end point of the first linear weld point cloud and the intersection position is not coincident with any end point of the second linear weld point cloud, constructing a sixth search vector by taking the intersection position as a starting point and any end point of the first linear weld point cloud which is not coincident with the intersection position as an end point, constructing a seventh search vector by taking the intersection position as a starting point and another end point of the first linear weld point cloud which is not coincident with the intersection position as an end point, constructing an eighth search vector by taking the intersection position as a starting point and any end point of the second linear weld point cloud which is not coincident with the intersection position as an end point, and constructing a ninth search vector by taking the intersection position as a starting point and another end point of the second linear weld point cloud which is not coincident with the intersection position as an end point; Determining a sixth height according to the first straight-line weld point cloud and the sixth search vector, determining a seventh height according to the first straight-line weld point cloud and the seventh search vector, determining an eighth height according to the second straight-line weld point cloud and the eighth search vector, and determining a ninth height according to the second straight-line weld point cloud and the ninth search vector; Determining that the number of the vertical welds is four when the sixth height, the seventh height, the eighth height and the ninth height are all greater than a preset height threshold; Taking the intersection position as a track starting point of the vertical welding seam, taking the smaller of the sixth height and the seventh height as a height of a third vertical welding seam, taking the smaller of the seventh height and the eighth height as a height of a fourth vertical welding seam, taking the smaller of the eighth height and the ninth height as a height of a fifth vertical welding seam, and taking the smaller of the ninth height and the sixth height as a height of a sixth vertical welding seam; Generating a third vertical weld identifier result according to the track starting point, the height of a third vertical weld, the sixth search vector and the seventh search vector, generating a fourth vertical weld identifier result according to the track starting point, the height of a fourth vertical weld, the seventh search vector and the eighth search vector, generating a fifth vertical weld identifier result according to the track starting point, the height of a fifth vertical weld, the eighth search vector and the ninth search vector, generating a sixth vertical weld identifier result according to the track starting point, the height of a fifth vertical weld, the ninth search vector and the sixth search vector, and taking the third vertical weld identifier result, the fourth vertical weld identifier result, the fifth vertical weld identifier result and the sixth vertical weld identifier result as vertical weld identifier results.
  10. 10. A vertical weld identification apparatus, the apparatus comprising: the acquisition module is used for acquiring a target single-workpiece point cloud of a target workpiece, wherein the target workpiece comprises a plurality of rib plates and a bottom plate; the extraction module is used for extracting rib plate point clouds corresponding to the rib plates from the target single-workpiece point clouds; the projection module is used for respectively projecting each rib plate point cloud to the plane of the bottom plate, and carrying out centroid downsampling on each projected rib plate point cloud to obtain a plurality of weld point clouds; The detection module is used for detecting the spatial position relationship between the intersecting position and the first and second linear weld point clouds under the condition that the first and second linear weld point clouds are identified from the weld point clouds and the first and second linear weld point clouds intersect; and the generating module is used for generating a vertical weld joint recognition result according to the spatial position relation.
  11. 11. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any one of claims 1 to 9 when the computer program is executed.
  12. 12. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 9.
  13. 13. A computer program product comprising a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any one of claims 1 to 9.

Description

Method for recognizing a weld, computer device, readable storage medium and program product Technical Field The present application relates to the field of welding technology, and in particular, to a method for recognizing a vertical weld, a computer device, a computer readable storage medium, and a computer program product. Background With the continuous evolution of the technology of automation and intellectualization of ship manufacturing, automatic identification and path planning of welding seams of small assembling structural members of ships have become a core link for improving welding production efficiency and process quality. In the conventional technology, weld joint recognition is generally performed by adopting a planar or linear feature fitting method based on an obtained three-dimensional point cloud of a workpiece, for example, geometric elements representing rib plates or weld joint edges are segmented from the point cloud by using algorithms such as RANSAC (Random Sample Consensus, random sampling agreement) and the like, so that the weld joint position is determined. However, when facing small assembly actual scenes of coexistence of multiple workpieces and complicated structure, the recognition result of the traditional method is often presented as a series of isolated two-dimensional plane line segments, and it is difficult to effectively construct a three-dimensional space topological relation between welding seams, especially when recognizing a vertical welding seam formed by intersection of a rib plate and a bottom plate, the situation of missing recognition, false recognition or incomplete path information and the like often occurs, so that the recognition accuracy of the vertical welding seam is obviously insufficient. Disclosure of Invention In view of the foregoing, it is desirable to provide a method, a computer device, a computer-readable storage medium, and a computer program product for recognizing a weld that can improve accuracy of weld recognition. In a first aspect, the present application provides a method for identifying a vertical weld, comprising: acquiring a target single-workpiece point cloud of a target workpiece, wherein the target workpiece comprises a plurality of rib plates and a bottom plate; extracting rib plate point clouds corresponding to the rib plates from the target single-workpiece point clouds; projecting each rib plate point cloud to the plane of the base plate respectively, and sampling the mass center of each projected rib plate point cloud downwards to obtain a plurality of weld point clouds; Under the condition that a first linear welding seam point cloud and a second linear welding seam point cloud are identified from each welding seam point cloud and the first linear welding seam point cloud is intersected with the second linear welding seam point cloud, detecting the spatial position relation between the intersection position and the first linear welding seam point cloud and the second linear welding seam point cloud; And generating a vertical weld recognition result according to the spatial position relation. In a second aspect, the present application also provides a device for identifying a vertical weld, including: the acquisition module is used for acquiring a target single-workpiece point cloud of a target workpiece, wherein the target workpiece comprises a plurality of rib plates and a bottom plate; the extraction module is used for extracting rib plate point clouds corresponding to the rib plates from the target single-workpiece point clouds; the projection module is used for respectively projecting each rib plate point cloud to the plane of the bottom plate, and carrying out centroid downsampling on each projected rib plate point cloud to obtain a plurality of weld point clouds; The detection module is used for detecting the spatial position relationship between the intersecting position and the first and second linear weld point clouds under the condition that the first and second linear weld point clouds are identified from the weld point clouds and the first and second linear weld point clouds intersect; and the generating module is used for generating a vertical weld joint recognition result according to the spatial position relation. In a third aspect, the present application also provides a computer device comprising a memory and a processor, the memory storing a computer program, the processor implementing the following steps when executing the computer program: acquiring a target single-workpiece point cloud of a target workpiece, wherein the target workpiece comprises a plurality of rib plates and a bottom plate; extracting rib plate point clouds corresponding to the rib plates from the target single-workpiece point clouds; projecting each rib plate point cloud to the plane of the base plate respectively, and sampling the mass center of each projected rib plate point cloud downwards to obtain a plurality of weld point clouds;