Automatic weld seam recognition method based on t-joint
By using CAD models and bounding box contact relationships to identify weld seams in T-joints of complex steel structure workpieces, the problem of time-consuming and labor-intensive weld seam identification in existing technologies has been solved, and the accuracy and stability of weld seam identification have been improved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- WUHAN HYPERION SOFTWARE CO LTD
- Filing Date
- 2024-01-09
- Publication Date
- 2026-07-21
AI Technical Summary
Existing visual guidance and recognition methods for welds are time-consuming and labor-intensive in complex steel structures, and cannot effectively identify welds in T-joints, affecting the efficiency and accuracy of robotic automatic welding.
Using a CAD model of a complex steel structure workpiece, the weld seam of a T-joint is identified by using a bounding box and defining contact relationships. By establishing an AABB bounding box, the contact relationship of the bounding box is calculated, and the surfaces related to the weld seam are identified and extracted, thus optimizing the weld seam identification algorithm.
It improves the accuracy and stability of weld seam identification, eliminates background interference and the influence of non-weld seam objects, and optimizes calculation speed and efficiency.
Smart Images

Figure CN117798559B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the fields of machine vision and weld seam recognition technology, and specifically to an automatic weld seam recognition method based on a T-joint. Background Technology
[0002] Current visual-guided weld seam recognition typically relies on robot teaching for initial positioning during the visual imaging stage. This is labor-intensive and time-consuming for welding multiple weld seams on complex steel structures, failing to leverage the advantages of automated robotic welding. Identifying and extracting T-joints within complex steel structures, and extracting the surfaces related to the weld seam, is crucial. These are the input data for calculating subsequent imaging and welding trajectory planning, and therefore essential elements that must be identified simultaneously.
[0003] This patent develops a weld seam recognition algorithm based on CAD models of complex steel structure workpieces. By using methods such as bounding boxes and defining contact relationships, the accuracy and stability of weld seam recognition are improved. Summary of the Invention
[0004] Purpose of the invention: In view of the problems pointed out in the background art, the present invention discloses an automatic weld identification method based on T-joints, which improves the accuracy and stability of weld identification by using bounding boxes and defining contact relationships.
[0005] Technical solution: This invention discloses an automatic weld identification method based on T-joints, comprising the following steps:
[0006] Step 1: Input the T-joint model data and process parameters;
[0007] Step 2: Traverse the model, build the corresponding AABB bounding box, and obtain the model S to be welded. i {S b S c}, S b S c For two welding models, F bq F bh F bz F by S b The four sides; F cq F ch F cz F cy S c The four sides; B q B h B z B y There are four weld seams, located on the outer contour of the weld surface;
[0008] Step 3: For model S iModel S b Perform calculations and traverse model S b For each face in the array, construct a bounding box, obtain information about the diagonal vertices, calculate the distances along the coordinate axes, and obtain the set of faces m whose maximum value differs from Long by a precision range. i And determine the contact relationship; the contact relationship is defined as follows:
[0009] 1) In the spatial relationship between two bounding boxes N1 and N2, if the coordinate values of the corresponding vertices of the two bounding boxes are subtracted along a certain axis, and the absolute value is less than the precision PC, then they are considered to be coincident along that axis; if the coordinate value of the vertex corresponding to N1 is within the precision range of the coordinate values of the two vertices corresponding to N2 along a certain coordinate axis, then they are considered to be contained along that two coordinate axis.
[0010] 2) If the two bounding boxes mentioned above have an inclusion relationship on all three axes, it is defined as relationship one in the contact relationship; if they have an overlap relationship on the x-axis and an inclusion relationship on the y and z axes, it is defined as relationship two in the contact relationship; if they have an overlap relationship on the y-axis and an inclusion relationship on the x and z axes, it is defined as relationship three in the contact relationship; if they have an overlap relationship on the z-axis and an inclusion relationship on the y and x axes, it is defined as relationship four in the contact relationship.
[0011] Step 4: Determine the connection model, let m i {m1, m2}, where m1 and m2 are S b Given two end faces, obtain m for each face. i and the corresponding model S b traverse S i China is different from S b Model S c Construct surface m respectively i and S c For bounding boxes, obtain information about their diagonal vertices and determine if two bounding boxes have any of the following contact relationships: then face m... i Model S b With model S c Connected and this surface m i For welding surface;
[0012] Step 5: Traverse face m i Find all edges in the array, determine the relationship between the edges and the surface, and locate the required weld L. i and the corresponding face F1 i ;
[0013] Step 6: Update face F1 i traversing face F1 i Update the rounded corner face to S c The corresponding side F in the middle cqand F ch ;
[0014] Step 7: Obtain face F2 i Traversing model S b For all faces, construct a bounding box using the coordinate information of two diagonal vertices, calculate the distance along the axis, and obtain the face with the maximum distance satisfying Long. Then determine the relationship between this face and the weld seam L. i If any one of the following contact relationships exists, then this surface is the weld L. i Corresponding face F2 i ;
[0015] Step 8: Based on face F1 i Kneading F2 i Calculate the welding torch posture during the welding process.
[0016] Furthermore, before obtaining the model to be welded in step 2, models that do not need to be welded need to be excluded, as follows:
[0017] Traverse each model and construct an AABB bounding box for it. Obtain the coordinate information of the two diagonal vertices of the bounding box, P1(x1, y1, z1) and P2(x2, y2, z2), and calculate the distance d in the three axis directions. i Determine the distance d i If it equals Long within the accuracy range, then it is the required welding model S. i {S b S c}, where Long is the length of the model; the AABB bounding box is constructed by obtaining the maximum and minimum values on the three coordinate axes based on the model data to be constructed, and the two diagonal vertices are defined as follows;
[0018] P min =[x min y min , z min ]
[0019] P max =[x max y max , z max ]
[0020] Distance d in the three axial directions i for;
[0021] d1=|x min -x max |
[0022] d2=|y min -y max |
[0023] d3=|zmin -z max |
[0024] Furthermore, the definition of the contact relationship in step 3 is as follows:
[0025] First, obtain the corresponding vertices P1 of the two bounding boxes. min and P1 max With P2 min and P2 max Perform the calculation:
[0026] P1 min =(x 1min y 1min , z 1min P1 max =(x 1max y 1max , z 1max )
[0027] P2 min =(x 2min y 2min , z 2min P2 max =(x 2max y 2max , z 2max )
[0028] 1) Relationship 1: It satisfies the condition by having an inclusion relationship on all three axes;
[0029] x 1min ≥x 2min -PC, x 1max ≤x 2max +PC
[0030] y 1min ≥y 2min -PC, y 1max ≤y 2max +PC
[0031] az 1min ≥z 2min -PC, z 1max ≤z 2max +PC
[0032] 2) Relationship 2: The x-axis coincides, and the y and z axes contain each other, which satisfies the condition;
[0033] {|x 1min -x 2min |≤PC,|x 1max -x 2min |≤PC or|x 1min -x 2max |≤PC,|x1max -x 2max |≤PC}
[0034] y 1min ≥y 2min -PC, y 1max ≤y 2max +PC, z 1min ≥z 2min -PC, z 1max ≤z 2max +PC
[0035] 3) Relationship 3: The y-axis coincides, and the x and z axes contain each other, which means the condition is met;
[0036] {|y 1min -y 2min |≤PC,|y 1max -y 2min |≤PC or|y 1min -y 2max |≤PC,|y 1max -y 2max |≤PC}
[0037] x 1min ≥x 2min -PC, x 1max ≤x 2max +PC, z 1min ≥z 2min -PC, z 1max ≤z 2max +PC
[0038] 4) Relationship 4: The z-axis coincides, and the y and x-axis contain each other, which means the conditions are met;
[0039] {|z 1min -z 2min |≤PC,|z 1max -z 2min |≤PC or|z 1min -z 2max |≤PC,|z 1max -z 2max |≤PC
[0040] x 1min ≥x 2min -PC, x 1max ≤x 2max +PC, y 1min ≥y 2min -PC, y 1max ≤y 2max +PC.
[0041] Furthermore, the required weld L found in step 5 iand the corresponding face F1 i When doing so, the non-welded edges need to be deleted, as follows:
[0042] 1) Calculate the curvature of the edges and obtain the edge B with zero curvature. i ;
[0043] 2) Update edge B i traverse edge B i For each edge, construct a bounding box to obtain vertex information, and then compare it with the face m. i The bounding box vertex information is calculated and judged. If two bounding boxes simultaneously satisfy any two of the contact relationships, namely, relationship two, three, and four, then the edge is retained; otherwise, it is deleted.
[0044] 3) Traverse the updated edge B as described above. i Then construct the bounding box and traverse model S again. c For each face in the array, construct a bounding box, and determine whether two bounding boxes are contained within each other. If they are, then edge B is selected. i For weld L i , corresponding to S c The face in the middle is face F1 i .
[0045] Furthermore, the non-welded edges are deleted, as follows:
[0046] 1) Traverse and judge all edges of the welding surface, calculate the curvature information of each edge, and obtain the edge B with zero curvature. i The parametric equations of a three-dimensional curve are generally in the form of: x = x(t), y = y(t), z = z(t), where t is a parameter, and the curvature formula is as follows:
[0047] K=[x′(t)^2+y′(t)^2+z′(t)^2]^(3 / 2) / [x″(t)y′(t)-x′(t)y″(t)]
[0048] Where x′, y′, z′ represent the first derivatives of x, y, z respectively, and x″, y″, z″ represent the second derivatives of x, y, z respectively;
[0049] 2) Update edge B i Since the weld joint is located on the outer contour of the weld surface, the inner edge of the weld surface is an incorrect edge and needs to be deleted, as detailed below:
[0050] Traverse edge B i For each edge, construct a bounding box to obtain vertex information. Compare this information with the vertex information of the bounding box of face m. If both bounding boxes satisfy any two of the contact relationships (relationships two, three, and four), retain the edge; otherwise, delete it. Internal edges have only one contact relationship, while external edges have two contact relationships. This update results in B.i {B q B h B z B y}, edge B q There are two types of contact relationships: the second and the third; edge B h There are two types of contact relationships: the second and the third; edge B z There are two types of contact relationships: the second and the fourth; edge B y There are two types of contact relationships: the second and the fourth.
[0051] Furthermore, F1 i For {F t1 F t2 F cy}, where F t1 F t2 For the two outer rounded corner faces, update face F1 in step 6. i hour:
[0052] First, traverse face F1. i Each time will be different from face F cy face F t1 F t2 Traverse each edge, construct a bounding box, use the information of the two diagonal vertices to make a judgment, calculate the distance in the three coordinate axes, and find the edge with the maximum distance greater than Long, denoted as la. i ;
[0053] For surface F t1 ,la i Given {l11, l12}, obtain edge la. i Center and surface F cy Edge l12, which has no contact relationship, is used to obtain model S. c The middle part has any one of the following contact relationships with this edge l12, and it is different from the surface F. t1 face F cq F1 i {F t1 F t2 F cy}Middle surface F t1 Updated to F cq ;
[0054] Similarly, for surface F t2 ,la i Given {l21, l22}, obtain edge la. i Center and surface F cy Edge l22 has no contact relationship, then obtain model S. c The middle part has any one of the following contact relationships with this edge l22, and it is different from the surface F. t2 face Fch F1 i {F t1 F t2 F cy}Middle surface F t2 Updated to F ch Thus far, F1 i For {F cq F ch F cy}
[0055] Further, in step 7, surface F2 is obtained. i At that time, the details are as follows:
[0056] Traversal model S b For all faces, construct a bounding box using the coordinate information of two diagonal vertices, calculate the distance along the axis, and find the face with the maximum distance greater than Long. Then determine the face that is connected to the weld seam L. i If an inclusion relationship exists, then this surface is weld L. i Corresponding face F2 i That is, F2 i For {F bq F bh F bz F by}
[0057] Beneficial effects:
[0058] This invention develops a weld seam recognition algorithm based on the CAD model of complex steel structure workpieces. It identifies and extracts T-joints in complex steel structure workpieces, extracting the surfaces related to the weld seam. Using bounding boxes and defined contact relationship methods, the accuracy and stability of weld seam recognition are improved.
[0059] This invention effectively eliminates background interference and interference from non-weld objects by using a bounding box treatment on the weld seam, thereby improving the accuracy of weld seam detection. It also optimizes computational speed and efficiency. The method used in this patent is based on physical principles and geometric models, and the computational methods employed are relatively simple, thus improving computational speed and efficiency. Attached Figure Description
[0060] Figure 1 The steel structure workpiece of the T-type joint in the embodiment of the present invention;
[0061] Figure 2 This is a schematic diagram illustrating the relevant symbol definitions for the T-type connector according to an embodiment of the present invention;
[0062] Figure 3 This is an overall flowchart of the present invention;
[0063] Figure 4This is a schematic diagram illustrating the relationship between the two bounding boxes in the three coordinate axes of an embodiment of the present invention.
[0064] Figure 5 The welding surface and model S of this invention c Schematic diagram of the enclosure;
[0065] Figure 6 This is a schematic diagram of a zero-curvature edge according to an embodiment of the present invention;
[0066] Figure 7 This is a schematic diagram of the surface and inner and outer edges surrounding the box of the present invention;
[0067] Figure 8 For the present invention, side B i Schematic diagram of the enclosure;
[0068] Figure 9 For weld B of the present invention z B y and face F cy Schematic diagram;
[0069] Figure 10 For weld B of the present invention q B h and face F t1 F t2 Schematic diagram;
[0070] Figure 11 For the present invention, surface F1 i Updated diagram;
[0071] Figure 12 For the present invention, surface F2 i Schematic diagram. Detailed Implementation
[0072] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.
[0073] This invention discloses an automatic weld identification method based on T-joints, see [link to relevant documentation]. Figure 1 , Figure 1 The figure shown is a T-type connector model according to an embodiment of the present invention.
[0074] The model's length, width, height, and thickness are (Long, Wide, High, Thick), respectively. b S c There are two welding models. The model surrounded by the red dashed line is the one that needs to be deleted. i {m1, m2}, where m1 and m2 are S b The two end faces of F. bq Fbh F bz F by S b The four sides, F cq F ch F cz F ty S c The four sides. B q B h B z B y These are four weld seams located on the outer contour of surface m2, and PC represents the data accuracy error.
[0075] For detailed operation and identification procedures, please refer to [link / reference]. Figure 3 It includes the following steps:
[0076] Step 1: Input the T-type connector model data and process parameters.
[0077] Step 2: Traverse the model, build the corresponding AABB bounding box, and obtain the model S to be welded. i {S b S c}, S b S c For two welding models, F bq F bh F bz F by S b The four sides; F cq F ch F cz F cy S c The four sides; B q B h B z B y There are four welds, located on the outer contour of the weld surface.
[0078] Before obtaining the model to be welded, the models that do not need to be welded need to be excluded, as follows:
[0079] Traverse each model and construct an AABB bounding box for it. Obtain the coordinate information of the two diagonal vertices of the bounding box, P1(x1, y1, z1) and P2(x2, y2, z2), and calculate the distance d in the three axis directions. i Determine the distance d i If it equals Long within the accuracy range, then it is the required welding model S. i {S b S c}, where Long is the length of the model; the AABB bounding box is constructed by obtaining the maximum and minimum values on the three coordinate axes based on the model data to be constructed, and the two diagonal vertices are defined as follows;
[0080] P min =[x min y min , z min ]
[0081] P max =[x max y max , z max ]
[0082] Distance d in the three axial directions i for;
[0083] d1=|x min -x max |
[0084] d2=|y min -y max |
[0085] d3=|z min -z max |
[0086] Step 3: For model S i Model S b Perform calculations and traverse model S b For each face in the array, construct a bounding box, obtain information about the diagonal vertices, calculate the distances along the coordinate axes, and obtain the set of faces m whose maximum value differs from Long by a precision range. i The contact relationship is determined by analyzing the relationship between two bounding boxes along the three coordinate axes. This reveals the relationships between the enclosed geometric elements. The contact relationship is defined as follows:
[0087] 1) In the spatial relationship between two bounding boxes N1 and N2, if the coordinate values of the corresponding vertices of the two bounding boxes are subtracted along a certain axis and the absolute value is less than the precision PC, then they are considered to be in a coincident relationship along that axis. If the coordinate value of the vertex corresponding to N1 is within the precision range of the coordinate values of the two vertices corresponding to N2 along a certain coordinate axis, then they are considered to be in an inclusive relationship along those two coordinate axes.
[0088] 2) If the two bounding boxes mentioned above have an inclusion relationship on all three axes, it is defined as relationship one in the contact relationship; if they have an overlap relationship on the x-axis and an inclusion relationship on the y and z axes, it is defined as relationship two in the contact relationship; if they have an overlap relationship on the y-axis and an inclusion relationship on the x and z axes, it is defined as relationship three in the contact relationship; if they have an overlap relationship on the z-axis and an inclusion relationship on the y and x axes, it is defined as relationship four in the contact relationship.
[0089] First, obtain the corresponding vertices P1 of the two bounding boxes. min and P1 max With P2 min and P2 max Perform the calculation:
[0090] P1 min =(x 1min y 1min , z 1min P1 max =(x 1max y 1max , z 1max )
[0091] P2 min =(x 2min y 2min , z 2min P2 max =(x 2max y 2max , z 2max )
[0092] 1. Relationship 1: The condition is satisfied if there is an inclusion relationship on all three axes, such as... Figure 4 (a).
[0093] x 1min ≥x 2min -PC, x 1max ≤x 2max +PC
[0094] y 1min ≥y 2min -PC, y 1max ≤y 2max +PC
[0095] az 1min ≥z 2min -PC, z 1max ≤z 2max +PC
[0096] 2. Relationship Two: The x-axis coincides, and the y and z axes contain each other, thus satisfying the condition, such as... Figure 4 (b)
[0097] {|x1min -x 2min |≤PC,|x 1max -x 2min |≤PC or|x 1min -x 2max |≤PC,|x 1max -x 2max |≤PC}
[0098] y 1min ≥y 2min -PC, y 1max ≤y 2max +PC, z 1min ≥z 2min -PC, z 1max ≤z 2max +PC
[0099] 3. Relationship Three: The y-axis coincides, and the x and z axes contain each other, thus satisfying the condition, such as... Figure 4 (c)
[0100] {|y 1min -y 2min |≤PC,|y 1max -y 2min |≤PC or|y 1min -y 2max |≤PC,|y 1max -y 2max |≤PC}
[0101] x 1min ≥x 2min -PC, x 1max ≤x 2max +PC, z 1min ≥z 2min -PC, z 1max ≤z 2max +PC
[0102] 4. Relationship Four: The z-axis coincides, and the y and x-axis contain each other, thus satisfying the condition; for example... Figure 4 (d)
[0103] {|z 1min -Z 2min |≤PC,|z 1max -z 2min |≤PC or|z 1min -z 2max |≤PC,|z 1max -z 2max |≤PC
[0104] x 1min ≥x 2min -PC, x 1max≤x 2max +PC, y 1min ≥y 2min -PC, y 1max ≤y 2max +PC
[0105] Step 4: Determine the connection model, let m i {m1, m2}, where m1 and m2 are S b Given two end faces, obtain m for each face. i and the corresponding model S b traverse S i China is different from S b Model S c Construct surface m respectively i and S c For bounding boxes, obtain information about their diagonal vertices and determine if two bounding boxes have any of the following contact relationships: then face m... i Model S b With model S c Connected and this surface m i This is the welding surface.
[0106] The welding in this paper is done at both ends of the model, therefore S needs to be welded. b The model's faces are evaluated, and the faces at both ends are selected. If this face matches another model S... c If they are joined, this surface is the welding surface m, as detailed below:
[0107] 1) First, the above welding models S are respectively... b The process involves iterating through the face, constructing a bounding box for each face, and calculating the distances along the three coordinate axes based on the coordinates of two vertices. Then, it's determined whether the maximum value falls within the precision range of Long, thus obtaining the face set m. i {m1, m2}.
[0108] 2) For the above face set m i Further determine the condition of each face in S, and traverse S. i China is different from S b Model S c Construct surfaces m1 and S respectively. c The bounding boxes are also evaluated based on their respective vertex information. Clearly, the two bounding boxes do not touch, therefore there is no contact relationship, and face m1 is not a welding face. Similarly, the bounding box for face m2 is constructed and evaluated. If a third type of contact relationship exists, then this face is a welding face m. Model S b With model S c Connected.
[0109] Step 5: Traverse face m i Find all edges in the array, determine the relationship between the edges and the surface, and locate the required weld L.i and the corresponding face F1 i .
[0110] Among the weld surfaces m identified above, there are some non-welded edges. Based on the contact relationship between the edges and the bounding boxes of the surfaces, the non-welded edges are determined and deleted, as follows:
[0111] 1) Traverse and judge all edges of the welding surface m, calculate the curvature information of each edge, and obtain the edge B with zero curvature. i .like Figure 6 As shown, the red edge is edge B. i .
[0112] The parametric equations of a three-dimensional curve are generally in the form of: x = x(t), y = y(t), z = z(t), where t is a parameter. The curvature formula is as follows:
[0113] K=[x′(t)^2+y′(t)^2+z′(t)^2]^(3 / 2) / [x″(t)y′(t)-x′(t)y″(t)]
[0114] Where x′, y′, z′ represent the first derivatives of x, y, z respectively, and x″, y″, z″ represent the second derivatives of x, y, z respectively.
[0115] 2) Update edge B i Since the weld joint is located on the outer contour of the weld surface, the edge inside the weld surface m is an incorrect edge and needs to be deleted, as detailed below:
[0116] Traverse edge B i For each edge, construct a bounding box to obtain vertex information. Compare this information with the vertex information of the bounding box of face m. If both bounding boxes satisfy any two of the contact relationships (relationships two, three, and four), retain the edge; otherwise, delete it. Internal edges have only one contact relationship, while external edges have two contact relationships. This update results in B. i {B q B h B z B y}
[0117] See Figure 7 and Figure 8 Side B q There are two types of contact relationships, the second and the third, with edge B. h There are two types of contact relationships: the second and the third. Edge B z There are two types of contact relationships, the second and the fourth, with edge B. y There are two types of contact relationships: the second and the fourth.
[0118] 3) Determine weld L iand the corresponding face F1 i
[0119] Traverse the updated edge B i Construct a bounding box for each edge in the model S, and iterate through the model S. c For each face in the array, construct a bounding box. Based on the contact relationship, determine if two bounding boxes belong to either type. If they do, then the edge is considered a weld L. i This face is face F1. i Thus far, F1 i For {F t1 F t2 F cy}, where F t1 F t2 These are two rounded outer corner surfaces. See details. Figure 9 ,and Figure 10 .
[0120] Step 6: Update face F1 i traversing face F1 i Update the rounded corner face to S c The corresponding side F in the middle cq and F ch .
[0121] Update F1 i First, traverse the aforementioned face F1. i Each time will be different from face F cy face F t1 F t2 Traverse each edge, construct a bounding box, use the information of the two diagonal vertices to make a judgment, calculate the distance in the three coordinate axes, and find the edge with the maximum distance greater than Long, denoted as la. i .
[0122] For surface F t1 ,la i Given {l11, l12}, obtain edge la. i Center and surface F cy Edge l12, which has no contact relationship, is used to obtain model S. c The middle part has any one of the following contact relationships with this edge l12, and it is different from the surface F. t1 face F cq F1 i {F t1 F t2 F cy}Middle surface F t1 Updated to F cq Similarly, for surface F... t2 ,la i Given {l21, l22}, obtain edge la. i Center and surface Fcy Edge l22 has no contact relationship, then obtain model S. c The middle part has any one of the following contact relationships with this edge l22, and it is different from the surface F. t2 face F ch F1 i {F t1 F t2 F cy}Middle surface F t2 Updated to F ch Thus far, F1 i For {F cq F ch F cy See details. Figure 11 .
[0123] Step 7: Obtain face F2 i Traversing model S b For all faces, construct a bounding box using the coordinate information of two diagonal vertices, calculate the distance along the axis, and obtain the face with the maximum distance satisfying Long. Then determine the relationship between this face and the weld seam L. i If any one of the following contact relationships exists, then this surface is the weld L. i Corresponding face F2 i That is, F2 i For {F bq F bh F bz F by See} Figure 12 .
[0124] Step 8: Based on face F1 i Kneading F2 i It calculates the welding torch posture during the welding process, including information such as the walking direction and the welding torch normal.
[0125] The above embodiments are only for illustrating the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. They should not be construed as limiting the scope of protection of the present invention. All equivalent transformations or modifications made in accordance with the spirit and essence of the present invention should be covered within the scope of protection of the present invention.
Claims
1. A method for automatic weld identification based on T-joints, characterized in that, Includes the following steps: Step 1: Input the T-joint model data and process parameters; Step 2: Traverse the model, build the corresponding AABB bounding box, and obtain the model to be welded. , For two welding models, They are respectively The four sides; They are respectively The four sides; There are four weld seams, located on the outer contour of the weld surface; Step 3: Model Medium model Perform calculations, model Length is Traversal Model For each face in the model, create a bounding box, obtain the information of the diagonal vertices, calculate the distance between the absolute values of the coordinates of corresponding vertices of two bounding boxes along a certain axis of the coordinate system, and obtain the maximum value and the model length. surface set within the phase difference accuracy range And determine the contact relationship; the contact relationship is defined as follows: 1) Place the two enclosure boxes and In spatial relationships, the coordinates of corresponding vertices of two bounding boxes along a certain axis are subtracted, and the absolute value is taken. If it is less than the precision... If they are coincident along this axis, then they are considered to be in a coincident relationship along this axis; if they are in a certain coordinate axis direction... The corresponding vertex coordinates are in If the precision range of the corresponding two vertex coordinate values is between those values, then the two coordinate axes are considered to be inclusive. 2) Regarding the two enclosure boxes mentioned above and A relationship that has an inclusion relationship on all three axes is defined as relationship one in the contact relationship; The axes have a coincident relationship. , If there is an inclusion relationship on the axis, it is defined as relationship two in the contact relationship; for The axes have a coincident relationship. , An inclusion relationship on an axis is defined as relationship three in a contact relationship; for The axes have a coincident relationship. , If there is an inclusion relationship on the axis, it is defined as relationship four in the contact relationship; Step 4: Determine the connection model, assuming... , for Two end faces, obtain each face and the corresponding model traversal China is different model Construct surfaces respectively and For bounding boxes, obtain information about the diagonal vertices and determine if two bounding boxes have any of the following contact relationships: model With model Connected and this surface For welding surface; Step 5: Traverse the face Find all edges in the array, determine the relationship between the edges and the surfaces, and locate the required weld. and the corresponding surface ; Step 6: Update the face traversing the face Update the rounded corner surfaces to The corresponding side and ; Step 7: Obtain the face Traversal Model For all faces, construct a bounding box using the coordinate information of two diagonal vertices, calculate the distance along the axis, and obtain the maximum distance that satisfies the model length. The surface is used to determine the relationship between this surface and the weld. If any form of contact relationship exists, then this surface is a weld. Corresponding face ; Step 8: According to the surface Kneading noodles Calculate the welding torch posture during the welding process.
2. The automatic weld identification method based on T-joints according to claim 1, characterized in that, Before obtaining the model to be welded in step 2, models that do not need to be welded need to be excluded, as follows: Iterate through each model and construct an AABB bounding box for it, obtaining the coordinate system information of the two diagonal vertices of the bounding box. and Calculate the distances along the three axes. Determine distance If it is equal to within the precision range This is the required welding model. { }, where the AABB bounding box is constructed by obtaining the maximum and minimum values on the three coordinate axes based on the model data to be constructed, and the two diagonal vertices are defined as follows; ; ; Distance in the three axial directions for; ; ; 。 3. The automatic weld identification method based on T-joints according to claim 1, characterized in that, The definition of the contact relationship in step 3 is as follows: First, obtain the corresponding vertices of the two bounding boxes. and and and Perform the calculation: , ; ; 1) Relationship 1: It satisfies the condition by having an inclusion relationship on all three axes; , ; , ; , ; 2) Relationship Two: Coincident axis relationship , The axis containment relationship satisfies the condition; , | or ,| ; , , , ; 3) Relationship Three: Coincident axis relationship , The axis containment relationship satisfies the condition; , | or | ,| ; , , , ; 4) Relationship Four: Coincident axis relationship , The axis containment relationship satisfies the condition; , | or ,| ; , , , 。 4. The automatic weld identification method based on T-joints according to claim 1, characterized in that, In step 5, the required weld seam is located. and the corresponding surface When doing so, the non-welded edges need to be deleted, as follows: 1) Calculate the curvature of the edges and obtain the edges with zero curvature. ; 2) Update edges traverse edges For each edge, construct a bounding box to obtain vertex information, and then compare it with the face. The bounding box vertex information is calculated and judged. If two bounding boxes simultaneously satisfy any two of the contact relationships, namely, relationship two, three, and four, then the edge is retained; otherwise, it is deleted. 3) Traverse the updated edges as described above. Then construct the bounding box and traverse the model again. For each face in the array, construct a bounding box, determine if two bounding boxes are in an inclusive relationship, and if so, add edges. For weld ,correspond The face in the middle is a face .
5. The automatic weld identification method based on T-joints according to claim 4, characterized in that, Delete the non-welded edges as follows: 1) Traverse all edges of the welding surface, calculate the curvature information of each edge, and obtain the edges with zero curvature. The parametric equations of three-dimensional space curves are generally in the form of: in These are parameters, and the curvature formula is as follows: ; in, They represent The first derivative, They represent The second derivative; 2) Update edges Since the weld joint is located on the outer contour of the weld surface, the inner edge of the weld surface is an incorrect edge and needs to be deleted, as detailed below: Traverse edges For each edge, construct a bounding box to obtain vertex information, and then compare it with the face. The bounding box vertex information is calculated and judged. If two bounding boxes simultaneously satisfy any two of the contact relationships (relationships two, three, and four), the edge is retained; otherwise, it is deleted. Internal edges have only one contact relationship, while external edges have two contact relationships. This update completes the process. , ,side There are two types of contact relationships: the second and the third; edge There are two types of contact relationships: the second and the third; edge There are two types of contact relationships: the second and the fourth; edge There are two types of contact relationships: the second and the fourth.
6. The automatic weld identification method based on T-joints according to claim 4, characterized in that, for ,in For the two outer rounded corner faces, update the faces in step 6. hour: First, traverse the face. Each time will be different from the previous one. face Traverse each edge, construct a bounding box, use the information of the two diagonal vertices to make a judgment, calculate the distance in the three coordinate axis directions, and obtain the maximum distance that is greater than the model length. The side note is ; For the face , for , obtain edge Middle and surface no contact relationship Finally, obtain the model. Middle and this side Having any type of contact relationship and different from a surface face ,Will middle noodle Updated to ; Similarly, for surfaces , for , obtain edge Middle and surface no contact relationship Then obtain the model Middle and this side Having any type of contact relationship and different from a surface face ,Will middle noodle Updated to Thus far for .
7. The automatic weld identification method based on T-joints according to claim 4, characterized in that, In step 7, the surface is obtained. At that time, the specific details are as follows: Traversal Model For all faces, construct a bounding box using the coordinate information of two diagonal vertices, calculate the distance along the axis, and obtain the maximum distance greater than the model length. The surface is used to determine the relationship between this surface and the weld. If an inclusion relationship exists, then this surface is a weld. Corresponding face ,Right now for .