A non-irregular thin-walled part machining path identification method based on projection and bidirectional ring search

Abstract

The application discloses a kind of non-heterogeneous thin-walled part machining path identification method based on projection and bidirectional search ring.The method includes first obtaining the stretching direction according to the shape characteristics of the non-heterogeneous thin-walled part based on model data;The normal of the topological surface in the non-heterogeneous thin-walled part and the topological surface perpendicular to the stretching direction are obtained as the projected surface;The center point of the non-heterogeneous thin-walled part is taken as the origin, and the stretching direction is taken as the normal to construct the projection plane, and each projected surface is projected onto the projection plane along the stretching direction to obtain the corresponding two-dimensional projection edge;According to the two-dimensional projection edge set, the outer wall surface set and the inner wall surface set are obtained respectively;According to the requirement, the outer wall surface set or the inner wall surface set is selected to obtain the machining path set and is processed.The application utilizes the structural characteristics of the non-heterogeneous thin-walled part, converts the three-dimensional feature recognition to the two-dimensional plane for processing, effectively reduces the operation complexity, maintains high recognition accuracy, and significantly reduces the dependence on manual intervention.

Description

Technical Field

[0001] This invention belongs to the field of industrial parts manufacturing, and specifically relates to a method for identifying the processing path of non-irregular thin-walled parts based on projection and bidirectional loop search. Background Technology

[0002] Laser cutting technology, as an advanced processing method, boasts significant advantages such as high processing speed, high precision, and good flexibility, and is widely used in the processing of thin-walled parts in fields such as medical devices, electronic products, automobile manufacturing, and aerospace. However, in actual production processes, due to the diversity of thin-walled part structures and the complexity of cutting shapes, traditional laser cutting systems struggle to automatically identify workpiece features, requiring manual selection of processing features, which severely restricts the improvement of processing efficiency.

[0003] Existing feature recognition methods mainly include rule-based methods, graph-based methods, volume decomposition-based methods, and deep learning-based methods. Rule-based and graph-based methods generate a connectivity attribute table or graph based on the part's topological information and match it with an existing feature library to generate a processing path. However, due to the limitations of the feature library, it is difficult to accurately identify complex non-standard features. Furthermore, matching is an NP-complete problem, requiring significant computational time. Volume decomposition-based methods decompose the model into the smallest identifiable sub-features according to a specific model, which is prone to problems such as non-convergence. Deep learning-based methods rely on the accuracy of the training point cloud data; if the data is noisy, smaller features are difficult to identify. All of the above methods are general processing methods for general parts and do not utilize the unique structural characteristics of non-irregular thin-walled parts for specialized recognition. Summary of the Invention

[0004] To address the problems existing in the background technology, this invention provides a method for identifying the processing path of non-irregular thin-walled parts based on projection and bidirectional loop search. This method solves the technical problem of lacking specialized identification based on the structural characteristics of non-irregular thin-walled parts in the current technology, and achieves accurate extraction of geometric feature information of parts, improving the accuracy and robustness of feature recognition and reducing manual operation.

[0005] The technical solution adopted in this invention is:

[0006] I. A method for identifying machining paths of non-irregularly shaped thin-walled parts based on projection and bidirectional loop search:

[0007] S1. Obtain model data of non-irregular thin-walled parts, and analyze the shape characteristics of non-irregular thin-walled parts based on the model data to obtain the stretching direction of non-irregular thin-walled parts.

[0008] S2. Traverse all topological surfaces of non-irregular thin-walled parts, and determine whether the normal of each topological surface is perpendicular to the stretching direction. If it is perpendicular, the topological surface is used as the projection surface, thus obtaining the set of projection surfaces.

[0009] S3. Construct a projection plane with the center point of the non-irregular thin-walled part as the origin and the stretching direction as the normal. Project each projection surface in the final set of projection surfaces onto the projection plane along the stretching direction to obtain the corresponding two-dimensional projection edge, thus obtaining the two-dimensional projection edge set.

[0010] S4. Obtain the outer wall surface set and inner wall surface set of the non-irregular thin-walled part based on the two-dimensional projection edge set.

[0011] S5. Select the outer wall surface set or the inner wall surface set according to the requirements to obtain the processing path set, and then perform processing according to the processing path set.

[0012] Step S1 specifically involves: acquiring model data of a non-irregular thin-walled part using the boundary representation method; generating a triangular mesh model by performing surface mesh generation on the non-irregular thin-walled part based on the model data; acquiring all non-repeating vertices in the triangular mesh model as point cloud data of the non-irregular thin-walled part; performing principal component analysis on the point cloud data to obtain multiple eigenvectors and the eigenvalues ​​corresponding to each eigenvector; and using the eigenvector corresponding to the largest eigenvalue as the stretching direction of the non-irregular thin-walled part.

[0013] The specific steps for obtaining the two-dimensional projection edge corresponding to each surface to be projected are as follows: First, obtain the parameter domain axis corresponding to the stretching direction in the surface parameter domain. Map all outer boundary edges on the current surface to be projected onto the two-dimensional parameter domain of the surface. For each outer boundary edge parallel to the stretching direction, extract the span interval in the direction of the non-parameter domain axis to obtain the span interval of all outer boundary edges in the direction of the non-parameter domain axis. Take the union of all span intervals to obtain the maximum span interval. After fixing any value on the parameter domain axis, generate a line segment as the isoparametric line in the two-dimensional parameter domain with the maximum span interval as the boundary. Obtain the isoparametric line in the three-dimensional space corresponding to the isoparametric line in the two-dimensional parameter domain. Project the isoparametric line in the three-dimensional space onto the projection plane to obtain the two-dimensional projection edge on the projection plane.

[0014] Step S4 specifically involves:

[0015] S41. In the set of two-dimensional projected edges, the first two-dimensional projected edge is taken as the edge to be processed. The edge to be processed is traversed and deduplicated with each subsequent two-dimensional projected edge. After deduplication, the set of two-dimensional projected edges is updated.

[0016] S42. In the updated set of two-dimensional projected edges, take the next two-dimensional projected edge of the edge to be processed as the new edge to be processed, and perform deduplication on the new edge to be processed and each subsequent two-dimensional projected edge of the new edge to be processed. After deduplication, update the set of two-dimensional projected edges.

[0017] S43. Repeat step S42 until the new edge to be processed is the last two-dimensional projection edge in the updated two-dimensional projection edge set. The final updated two-dimensional projection edge set is the deduplicated two-dimensional projection edge set.

[0018] S44. In the deduplicated set of two-dimensional projected edges, perform bidirectional loop search along the vector direction of one of the two-dimensional projected edges and the opposite direction of the vector direction to obtain several two-dimensional loops, thus obtaining a set of two-dimensional loops.

[0019] S45. Obtain the outer wall surface set and inner wall surface set of non-irregular thin-walled parts based on the two-dimensional ring set.

[0020] The deduplication process is as follows: if the current edge to be processed overlaps with the two-dimensional projection edge that needs to be deduplicated together, then the overlapping part is deleted from the two-dimensional projection edge that needs to be deduplicated together and the deduplication process ends; if the current edge to be processed does not overlap with the two-dimensional projection edge that needs to be deduplicated together, then no processing is performed and the deduplication process ends.

[0021] Step S44 specifically involves: selecting a two-dimensional projection edge as the initial connecting edge from the deduplicated set of two-dimensional projection edges, traversing and searching along the vector direction of the initial connecting edge to trace the connection relationship of adjacent two-dimensional projection edges, thereby constructing several closed minimum two-dimensional loops; then traversing and searching in the opposite direction of the vector direction of the initial connecting edge to trace the connection relationship of adjacent two-dimensional projection edges, thereby constructing a closed outer boundary two-dimensional loop, and summing all the minimum two-dimensional loops and the outer boundary two-dimensional loops to obtain a set of two-dimensional loops.

[0022] Step S45 specifically involves:

[0023] S451. Divide each two-dimensional ring in the set of two-dimensional rings into clockwise two-dimensional rings or counterclockwise two-dimensional rings.

[0024] S452. First, for the counterclockwise two-dimensional ring, the surface to be projected corresponding to each two-dimensional projection edge in the two-dimensional ring is taken as the outer wall surface, thus obtaining the set of outer wall surfaces.

[0025] S453. Then, traverse all clockwise two-dimensional rings: if any two-dimensional projection edge in the current two-dimensional ring has been identified as an outer wall, no processing is done, and the next clockwise two-dimensional ring is traversed; if no two-dimensional projection edge in the current two-dimensional ring has been identified as an outer wall, then each two-dimensional projection edge in the current two-dimensional ring is identified as an inner wall, thus obtaining the set of inner wall surfaces, and the traversal ends.

[0026] The processing path set is obtained through the following steps: Based on topological connections, all wall boundary topological edges are selected from the outer wall set or the inner wall set; all wall boundary topological edges are then aggregated to obtain the wall boundary topological edge set; all topological edges in the wall boundary topological edge set are divided into an in-plane topological edge set and an edge topological edge set; for the in-plane topological edge set, each in-plane topological edge is a loop processing path, thus the in-plane topological edge set becomes a loop processing path set; for the edge topological edge set, a directed edge search is performed on all edge topological edges within the set to obtain several directed connected lines, each of which is a directed processing path; all directed processing paths are aggregated to obtain a directed processing path set; the loop processing path set and the directed processing path set are then aggregated to obtain the complete processing path set.

[0027] The wall boundary topological edge is the topological edge used by a single wall; the in-plane topological edge is the topological edge located within the face of the wall; the edge topological edge is the topological edge located at the edge of the wall.

[0028] II. A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described method.

[0029] The beneficial effects of this invention are:

[0030] 1. This invention utilizes the structural characteristics of non-irregularly shaped thin-walled parts to transfer three-dimensional feature recognition to a two-dimensional plane for processing. While effectively reducing computational complexity, it maintains high recognition accuracy and significantly reduces reliance on manual intervention.

[0031] 2. Improved Stability: Compared to directly projecting model patches onto the projection plane, using isoparametric projection of the non-stretched direction of the model patches avoids the complex calculations of surface-to-plane projection and eliminates unnecessary projection of curves along the stretching direction. Furthermore, each patch has only one isoparametric line to be projected, which significantly reduces the number of curves requiring deduplication compared to directly projecting multiple non-stretched direction curves onto the patch, thus lowering the complexity of the 2D data.

[0032] 3. High reliability: It adopts bidirectional loop search processing and selects the maximum rotation angle to intuitively and accurately obtain the outer and inner wall surfaces of the model. Attached Figure Description

[0033] Figure 1 This is a flowchart of the method of the present invention.

[0034] Figure 2 This is a schematic diagram showing the normal of the stretching direction to the topological surface in Example 2.

[0035] Figure 3 This is a schematic diagram of isoparametric line extraction of the surface parameter domain in Example 2.

[0036] Figure 4 This is a schematic diagram of the projection result of the set of surfaces to be projected in Example 2.

[0037] Figure 5 This is a schematic diagram of the projection result of the outer wall surface assembly in Example 2.

[0038] Figure 6 This is a schematic diagram of the direction of the two-dimensional ring set in Example 2.

[0039] Figure 7 This is a schematic diagram of the two-dimensional projected edge and the tangent of the adjacent edge in Example 2.

[0040] Figure 8 This is a schematic diagram of the two-dimensional ring result corresponding to the inner and outer wall surfaces in Example 2.

[0041] Figure 9 This is a schematic diagram of the connecting pipe fittings in Example 3.

[0042] Figure 10 This is a schematic diagram of a two-dimensional ring corresponding to the model wall of the connecting pipe in Example 3.

[0043] Figure 11 This is a schematic diagram of the machining path for the outer wall surface of the connecting pipe in Example 3. Detailed Implementation

[0044] The present invention will now be described in more detail with reference to the accompanying drawings and embodiments. However, the present invention is not limited thereto. For those skilled in the art, several improvements and modifications can be made without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention. Contents not described in detail in this specification are prior art known to those skilled in the art.

[0045] Non-irregularly shaped thin-walled parts: These refer to thin-walled parts formed by cutting stretched bodies (such as cuboids, cylinders, etc.) to create irregular geometric shapes or complex features. Their characteristics include maintaining a relatively thin overall profile, with the cross-sectional shape potentially changing along the axial direction. The outer and inner walls of the part are not connected. "Thin-walled" means the wall thickness is relatively small, typically within a small proportional range compared to the overall size of the part.

[0046] Outer wall surface: refers to the outer surface area of ​​a non-irregularly shaped thin-walled part, the surface that is in direct contact with the external environment. This definition specifically refers to the outer curved surface of a non-irregularly shaped thin-walled part, excluding the end face areas at both ends.

[0047] Inner wall surface: refers to the inner surface area of ​​a non-irregularly shaped thin-walled part, which is parallel to and corresponds to the outer wall surface. It is the surface that constitutes the inner cavity of the non-irregularly shaped thin-walled part, and has similar geometric features to the outer wall surface but in the opposite direction.

[0048] Cutting surface: refers to the new combined surface produced during the cutting process, which connects the outer and inner wall surfaces of non-irregular thin-walled parts.

[0049] Processing path: refers to the movement trajectory of processing tools (such as laser cutting heads, CNC tools, etc.) according to a predefined spatial trajectory to achieve a specific processing purpose during the manufacturing process.

[0050] Wall-shared topological edge: refers to a topological edge that is shared by two topological faces, and both topological faces belong to the wall.

[0051] Wall boundary topological edge: refers to a topological edge that is used by a topological face and that topological face belongs to a wall, or is shared by two topological faces but only one of them belongs to a wall.

[0052] Example 1:

[0053] like Figure 1 As shown, the method for identifying the machining path of non-irregularly shaped thin-walled parts in this embodiment includes the following steps:

[0054] S1. Obtain model data of non-irregular thin-walled parts, and analyze the shape characteristics of non-irregular thin-walled parts based on the model data to obtain the stretching direction of non-irregular thin-walled parts.

[0055] Step S1 specifically involves: acquiring model data of a non-irregular thin-walled part using the boundary representation method for topological representation; generating a triangular mesh model by performing surface mesh generation on the non-irregular thin-walled part based on the model data; acquiring all non-repeating vertices in the triangular mesh model as point cloud data of the non-irregular thin-walled part; performing principal component analysis on the distribution of the point cloud data to obtain multiple eigenvectors and the eigenvalues ​​corresponding to each eigenvector; and using the eigenvector corresponding to the largest eigenvalue as the stretching direction of the non-irregular thin-walled part.

[0056] In practice, the discrete parameters during the surface mesh generation process are either user-defined or use the algorithm's default configuration, such as the default edge discrete offset accuracy being 0.1 and the default face discrete offset accuracy being 0.1. This embodiment uses the default configuration.

[0057] S2. Traverse all topological surfaces of non-irregular thin-walled parts, and determine whether the normal of each topological surface is perpendicular to the stretching direction. If the topological surface is perpendicular to the stretching direction, then the topological surface is taken as the surface to be projected, thus obtaining the set of surfaces to be projected.

[0058] S3. Construct a projection plane with the center point of the non-irregular thin-walled part as the origin and the stretching direction as the normal. Project each projection surface in the final set of projection surfaces onto the projection plane along the stretching direction to obtain the corresponding two-dimensional projection edge. Summarize all the two-dimensional projection edges to obtain the two-dimensional projection edge set.

[0059] The specific steps for obtaining the two-dimensional projection edge corresponding to each surface to be projected are as follows: First, obtain the parameter domain axis (v-axis or u-axis) corresponding to the stretching direction in the surface parameter domain. Map all outer boundary edges on the current surface to be projected onto the two-dimensional parameter domain of the surface. For each outer boundary edge parallel to the stretching direction, extract the span interval in the non-parameter domain axis direction to obtain the span interval of all outer boundary edges in the non-parameter domain axis direction. Take the union of all span intervals to obtain the maximum span interval. After fixing any value on the parameter domain axis (v0 is fixed), generate a line segment as the isoparametric line (u1, v0) -> (u4, v0) in the two-dimensional parameter domain with the maximum span interval as the boundary. Obtain the isoparametric line in the three-dimensional space corresponding to the isoparametric line in the two-dimensional parameter domain. Project the isoparametric line in the three-dimensional space onto the projection plane to obtain the two-dimensional projection edge on the projection plane.

[0060] If the stretching direction corresponds to the v-axis in the surface parameter domain, then the span interval is the interval along the u-axis.

[0061] S4. Obtain the outer wall surface set and inner wall surface set of the non-irregular thin-walled part based on the two-dimensional projection edge set; Step S4 specifically involves:

[0062] S41. In the set of two-dimensional projected edges, the first two-dimensional projected edge is taken as the edge to be processed. The edge to be processed is traversed and deduplicated with each subsequent two-dimensional projected edge. After deduplication, the set of two-dimensional projected edges is updated.

[0063] S42. In the updated set of two-dimensional projected edges, take the next two-dimensional projected edge of the edge to be processed as the new edge to be processed, and perform deduplication on the new edge to be processed and each subsequent two-dimensional projected edge of the new edge to be processed. After deduplication, update the set of two-dimensional projected edges.

[0064] S43. Repeat step S42 until the new edge to be processed is the last two-dimensional projection edge in the updated two-dimensional projection edge set. The final updated two-dimensional projection edge set is the deduplicated two-dimensional projection edge set.

[0065] The deduplication process is as follows: if the current edge to be processed overlaps with the two-dimensional projection edge that needs to be deduplicated together, then the overlapping part is deleted from the two-dimensional projection edge that needs to be deduplicated together and the deduplication process ends; if the current edge to be processed does not overlap with the two-dimensional projection edge that needs to be deduplicated together, then no processing is performed and the deduplication process ends.

[0066] Two-dimensional projected edges are vectors in the data dimension, so each one has a vector direction.

[0067] S44. In the deduplicated set of two-dimensional projected edges, perform bidirectional loop search along the vector direction of one of the two-dimensional projected edges and the opposite direction of the vector direction to obtain several two-dimensional loops, thus obtaining a set of two-dimensional loops.

[0068] Step S44 is as follows: In the deduplicated set of two-dimensional projected edges, select one two-dimensional projected edge as the initial connecting edge, and perform a traversal search along the vector direction of the initial connecting edge to trace the connection relationship of adjacent two-dimensional projected edges, thereby constructing several closed minimum two-dimensional loops; then perform a traversal search in the opposite direction of the vector direction of the initial connecting edge to trace the connection relationship of adjacent two-dimensional projected edges, thereby constructing a closed outer boundary two-dimensional loop, and summing up all the minimum two-dimensional loops and the outer boundary two-dimensional loops to obtain a set of two-dimensional loops.

[0069] In practice, the smallest two-dimensional ring is one that does not contain any smaller two-dimensional rings inside a clockwise closed two-dimensional ring.

[0070] Furthermore, obtaining all the smallest clockwise closed two-dimensional loops is carried out by the following steps:

[0071] Step P1, Initialization and Tracing of the First Loop:

[0072] D1. Use the current 2D projection edge as the initial connection edge and push it onto the first empty stack.

[0073] D2. Obtain the set of adjacent two-dimensional projected edges according to the tail endpoint of the vector direction of the initial connecting edge.

[0074] D3. Calculate the clockwise rotation angle of the tangent of the current two-dimensional projection edge at the tail end of the vector direction to the tangent of each adjacent two-dimensional projection edge.

[0075] D4. Select the adjacent 2D projection edge corresponding to the maximum turning angle as the next connecting edge and push it onto the current stack, and establish the connection relationship.

[0076] D5. Repeat steps D1-D4 until all the traced connecting edges are closed at both ends, forming a clockwise two-dimensional loop.

[0077] Step P2, Closure of the loop and backtracking of the stack:

[0078] D6. After the loop is closed, determine whether all adjacent edges of the top connecting edge of the stack are already in the stack:

[0079] If all adjacent edges of the top connecting edge are already in the stack, pop the top connecting edge from the stack and re-evaluate whether all adjacent edges of the new top connecting edge are already in the stack. If at least one adjacent edge of the top connecting edge is not in the stack, push the top connecting edge onto a new empty stack (as the starting point for the new two-dimensional ring tracing) and return to step one to start tracing the new two-dimensional ring.

[0080] Step P3: Stack backtracking and algorithm termination:

[0081] If the current stack is empty, backtrack to the top of the previous stack, use the edge connecting the top of the stack as the new starting point, and jump to step two to continue execution. When all stacks are empty, it means that all clockwise closed minimum two-dimensional cycles have been extracted, and the algorithm terminates.

[0082] Furthermore, after all the smallest clockwise closed two-dimensional loops are obtained, the loop search is performed in the opposite direction of the vector direction of the initial connecting edge to obtain the outer boundary two-dimensional loop that is closed counterclockwise.

[0083] S45. Obtain the outer wall surface set and inner wall surface set of non-irregular thin-walled parts based on the two-dimensional ring set.

[0084] S451. Divide each two-dimensional ring in the set of two-dimensional rings into clockwise two-dimensional rings or counterclockwise two-dimensional rings.

[0085] S452. First, for the counterclockwise two-dimensional ring, the surface to be projected corresponding to each two-dimensional projection edge in the two-dimensional ring is taken as the outer wall surface, thus obtaining the set of outer wall surfaces.

[0086] S453. Then, traverse all clockwise two-dimensional rings: if any two-dimensional projection edge in the current two-dimensional ring has been identified as an outer wall, no processing is done, and the next clockwise two-dimensional ring is traversed; if no two-dimensional projection edge in the current two-dimensional ring has been identified as an outer wall, then each two-dimensional projection edge in the current two-dimensional ring is identified as an inner wall, thus obtaining the set of inner wall surfaces, and the traversal ends.

[0087] Furthermore, in the set of surfaces to be projected, all surfaces to be projected that are determined to be either outer or inner wall surfaces are used as cutting surfaces, thus obtaining a set of cutting surfaces.

[0088] S5. Select the outer wall surface set or the inner wall surface set according to the requirements to obtain the processing path set, and then perform processing according to the processing path set.

[0089] The processing path set is obtained through the following steps: Based on topological connections, all wall boundary topological edges are selected from the outer or inner wall set. All wall boundary topological edges are then aggregated to form a wall boundary topological edge set. All topological edges in this set are further divided into an in-plane topological edge set and an edge topological edge set. For the in-plane topological edge set, each in-plane topological edge represents a loop processing path, thus the in-plane topological edge set becomes a loop processing path set. For the edge topological edge set, a directed edge search is performed on all edge topological edges within the set to obtain several directed connected lines. Each directed connected line represents a directed processing path. All directed processing paths are aggregated to form a directed processing path set. The loop processing path set and the directed processing path set are then combined to obtain the complete processing path set. Wall boundary topological edges are the topological edges used by a single wall; in-plane topological edges are the topological edges located within the wall's surface; edge topological edges are the topological edges located at the edges of the wall.

[0090] Example 2:

[0091] like Figure 2 As shown, this embodiment takes a square tube model (non-irregular thin-walled part) with two processing paths as an example.

[0092] Step 1: Identify the stretching direction of the square tube model.

[0093] The input square tube model is denoted as T, and the topological representation of the model uses the boundary representation method. Surface mesh generation is performed on the square tube model T to obtain a triangular mesh model M, where the discretization parameters use a default edge discretization offset precision of 0.1 and a default face discretization offset precision of 0.1. The mesh vertices in M ​​are the point cloud data corresponding to model T. Then, principal component analysis is used to analyze the distribution of the point cloud data, obtaining multiple eigenvectors and the eigenvalues ​​corresponding to each eigenvector. The eigenvector corresponding to the largest eigenvalue is used as the stretching direction V. D ,like Figure 2 As shown.

[0094] Step 2: Obtain the set of surfaces to be projected.

[0095] Traverse the topological surfaces of the square tube model and obtain the normal V of each surface. N When the stretching direction V D With normal V N When perpendicular, that is, when abs(V) is satisfied. D ∙V N ) <TOL ZEROAdd the topological surface to the set of surfaces to be projected, F. PS TOL ZERO For the tolerance of externally input values, this embodiment uses 1.0e-7. After traversal, the set of surfaces to be projected includes the model's inner and outer walls and some cut surfaces. For example... Figure 2 As shown in the figure. V N1 V is the normal to a topological surface in the set of outer wall surfaces of the square tube model. N2 V is the normal to a topological surface in the set of inner wall surfaces of the square tube model. N3 and V N4 These are the normals of a topological surface at each of the two cut points of the square tube model.

[0096] Step 3: Project the set of surfaces to be projected.

[0097] With tension direction V D Using the normal direction and the center point of the bounding box of the square tube model as the origin, construct the projection plane. Iterate through the set F of the planes to be projected. PS The surface parameter domain corresponding to the surface to be projected is defined as the UV domain, and the range of the UV domain is: U: [umin,umax], V: [vmin,vmax]. The stretching direction V is determined. D Is the parameter domain axis corresponding to the UV domain of the current projection surface the U-axis or the V-axis? Assuming it corresponds to the U-axis, then fix the U-axis coordinates as the midpoint umid of the U-axis interval. Extract a v-direction isoparametric line based on the V-axis interval [vmin, vmax] of the current UV domain, such as... Figure 3 As shown. The extracted isoparametric lines in the two-dimensional parameter domain are projected along the stretching direction VD onto the projection plane to obtain the three-dimensional projection result. The corresponding two-dimensional projection results (two-dimensional projection edges on the projection plane) in the three-dimensional projection result are retained. Figure 4 Display the projection results of the set of surfaces to be projected from the square tube model. Figure 5 Displays the projection set of the outer wall surface set in the set of surfaces to be projected.

[0098] Step 4: Obtain the outer wall surface set and inner wall surface set of the non-irregular thin-walled part based on the two-dimensional projection edge set.

[0099] 1. In the set of 2D projected edges obtained in step 3, the first 2D projected edge is taken as the edge to be processed. This edge is then traversed and deduplicated with each subsequent 2D projected edge. After deduplication, the set of 2D projected edges is updated. 2. In the updated set of 2D projected edges, the next 2D projected edge of the edge to be processed is taken as the new edge to be processed. This new edge is then deduplicated with each subsequent 2D projected edge. After deduplication, the set of 2D projected edges is updated. 3. Step 2 is repeated until the new edge to be processed is the last 2D projected edge in the updated set of 2D projected edges. The final updated set of 2D projected edges is the deduplicated set of 2D projected edges.

[0100] Next, in the deduplicated set of 2D projected edges, one 2D projected edge is selected as the starting edge. A traversal search is performed along the vector direction of the starting edge to trace the connection relationships of adjacent 2D projected edges, thus constructing several closed minimal 2D loops. Then, a traversal search is performed in the opposite direction of the vector direction of the starting edge to trace the connection relationships of adjacent 2D projected edges, thereby constructing a closed outer boundary 2D loop. All minimal 2D loops and outer boundary 2D loops are combined to obtain a set of 2D loops, such as... Figure 6 As shown.

[0101] The process of obtaining a single two-dimensional ring is as follows: Traverse the set of deduplicated two-dimensional projected edges, take the current two-dimensional projected edge E as the connecting edge, obtain the endpoint P of E along the current traversal vector direction, and calculate the tangential inverse direction V of E at P. P Find the set of other unvisited two-dimensional projected edges sharing a common endpoint P, denoted as the adjacent edge set E. AS Traverse the set of adjacent edges E AS Calculate the tangential reverse V P The current adjacent edge is tangent to V at point P. AP For each clockwise turn, select the adjacent edge corresponding to the largest turn as the next connecting edge, and gradually add new connecting edges until all current connecting edges form a closed two-dimensional loop.

[0102] by Figure 7 For example, taking the current 2D projection edge as the starting edge E2, obtain the set of adjacent 2D projection edges (E1 and E3) according to the tail endpoint P of the vector direction of the starting edge E2, and calculate the tangential direction V of the current 2D projection edge E2 at the tail endpoint P of the vector direction. p Tangential (V) to each adjacent two-dimensional projection edge (E1 and E3) A1P and V A3P Along the clockwise rotation, select the adjacent two-dimensional projection edge V corresponding to the maximum rotation angle. A3P This will serve as the next connecting edge. After the bidirectional cycle search is completed, multiple ordered connected two-dimensional cycles can be obtained.

[0103] Finally, each two-dimensional ring in the set is divided into clockwise or counterclockwise two-dimensional rings. For the counterclockwise two-dimensional rings, the projection surface corresponding to each two-dimensional projection edge is treated as an outer wall surface, thus obtaining a set of outer wall surfaces. Then, all clockwise two-dimensional rings are traversed: if any projection surface corresponding to a two-dimensional projection edge in the current two-dimensional ring has already been determined as an outer wall surface, no further processing is done, and the next clockwise two-dimensional ring is traversed; if no projection surface corresponding to any two-dimensional projection edge in the two-dimensional ring has been determined as an outer wall surface, then each projection surface corresponding to a two-dimensional projection edge in the two-dimensional ring is treated as an inner wall surface, thus obtaining a set of inner wall surfaces, and the traversal ends. Figure 8 As shown, the two two-dimensional rings obtained in the end correspond to the outer wall surface set and the inner wall surface set, respectively.

[0104] Step 5: Obtain the set of processing paths.

[0105] The outer wall set is selected to identify the processing paths of the walls. The outer wall set is traversed, and the set of topological edges of the wall boundaries is obtained based on the topological connections of the walls. All topological edges in the set of topological edges of the wall boundaries are divided into an in-plane topological edge set and an edge topological edge set. For the in-plane topological edge set, each in-plane topological edge is a loop processing path, thus the in-plane topological edge set is a set of loop processing paths. For the edge topological edge set, a directed edge search is performed on all edge topological edges within the set to obtain several directed connected lines, each of which is a directed processing path. All directed processing paths are summarized to obtain a set of directed processing paths. The set of loop processing paths and the set of directed processing paths are summarized to obtain a set of all processing paths.

[0106] Example 3:

[0107] Taking a connecting pipe fitting with multiple processing paths as an example, such as Figure 9 As shown. The processing path of this pipe fitting model has two possible locations: one is located on a single surface, and the other is located on multiple surfaces. Following the above technical solution, using the default discrete parameter configuration with a numerical tolerance of 1.0e-7, the stretching direction is obtained by performing principal component analysis on the point cloud data corresponding to the pipe fitting model. Then, the selected projection surfaces are projected onto the corresponding projection planes along the stretching direction. A bidirectional loop search is used to obtain the pipe fitting model wall surface. The resulting two-dimensional loop corresponding to the model wall surface is shown in the figure. Figure 10 As shown, taking the identification of the outer wall surface as an example, the final result of the model processing path identification can be obtained, such as... Figure 11 The green area shown represents a total of 6 processing paths.

[0108] This invention transforms the 3D feature recognition problem into a 2D projective contour recognition problem by introducing projection analysis, significantly reducing computational complexity and effectively handling the complex geometric features of non-irregularly shaped pipes. Furthermore, this invention utilizes a bidirectional loop-finding strategy to accurately identify the outer and inner walls of the model, and obtains the processing path using the model's topological connection information.

[0109] The above embodiments are merely preferred embodiments provided to fully illustrate the present invention, and the scope of protection of the present invention is not limited thereto. Equivalent substitutions or modifications made by those skilled in the art based on the present invention are all within the scope of protection of the present invention. The scope of protection of the present invention is defined by the claims.

Claims

1. A method for identifying the machining path of non-irregular thin-walled parts based on projection and bidirectional loop search, characterized in that, Includes the following steps: S1. Obtain model data of non-irregular thin-walled parts, and analyze the shape characteristics of non-irregular thin-walled parts based on the model data to obtain the stretching direction of non-irregular thin-walled parts; S2. Traverse all topological surfaces of non-irregular thin-walled parts, and determine whether the normal of each topological surface is perpendicular to the stretching direction. If it is perpendicular, the topological surface is used as the projection surface, thus obtaining the set of projection surfaces. S3. Construct a projection plane with the center point of the non-irregular thin-walled part as the origin and the stretching direction as the normal. Project each projection surface in the final set of projection surfaces onto the projection plane along the stretching direction to obtain the corresponding two-dimensional projection edge, thus obtaining the two-dimensional projection edge set. The specific steps for obtaining the two-dimensional projection edge corresponding to each surface to be projected are as follows: First, obtain the parameter domain axis corresponding to the stretching direction in the surface parameter domain. Map all outer boundary edges on the current surface to be projected onto the two-dimensional parameter domain of the surface. For each outer boundary edge parallel to the stretching direction, extract the span interval in the non-parameter domain axis direction to obtain the span interval of all outer boundary edges in the non-parameter domain axis direction. Take the union of all span intervals to obtain the maximum span interval. After fixing any value on the parameter domain axis, generate a line segment as the isoparametric line in the two-dimensional parameter domain with the maximum span interval as the boundary. Obtain the isoparametric line in the three-dimensional space corresponding to the isoparametric line in the two-dimensional parameter domain. Project the isoparametric line in the three-dimensional space onto the projection plane to obtain the two-dimensional projection edge on the projection plane. S4. Obtain the outer wall surface set and inner wall surface set of the non-irregular thin-walled part based on the two-dimensional projection edge set; S5. Select the outer wall surface set or the inner wall surface set according to the requirements to obtain the processing path set, and then perform processing according to the processing path set.

2. The method for identifying the machining path of non-irregular thin-walled parts based on projection and bidirectional loop search according to claim 1, characterized in that, Step S1 specifically involves: Obtain model data of non-irregular thin-walled parts using the boundary representation method. Based on the model data, perform surface mesh generation on the non-irregular thin-walled parts to obtain a triangular mesh model. Obtain all non-repeating vertices in the triangular mesh model as point cloud data of the non-irregular thin-walled parts. Perform principal component analysis on the point cloud data to obtain multiple eigenvectors and the eigenvalues ​​corresponding to each eigenvector. Use the eigenvector corresponding to the largest eigenvalue as the stretching direction of the non-irregular thin-walled parts.

3. The method for identifying the machining path of non-irregular thin-walled parts based on projection and bidirectional loop search according to claim 1, characterized in that, Step S4 specifically involves: S41. In the set of two-dimensional projected edges, the first two-dimensional projected edge is taken as the edge to be processed. The edge to be processed is traversed and each subsequent two-dimensional projected edge is deduplicated. After deduplication, the set of two-dimensional projected edges is updated. S42. In the updated set of two-dimensional projected edges, take the next two-dimensional projected edge of the edge to be processed as the new edge to be processed, and perform deduplication processing on the new edge to be processed and each subsequent two-dimensional projected edge. After deduplication processing, update the set of two-dimensional projected edges. S43. Repeat step S42 until the new edge to be processed is the last two-dimensional projection edge in the updated two-dimensional projection edge set. The final updated two-dimensional projection edge set is the deduplicated two-dimensional projection edge set. S44. In the deduplicated set of two-dimensional projected edges, perform bidirectional loop search along the vector direction of one of the two-dimensional projected edges and the opposite direction of the vector direction to obtain several two-dimensional loops, thus obtaining a set of two-dimensional loops. S45. Obtain the outer wall surface set and inner wall surface set of non-irregular thin-walled parts based on the two-dimensional ring set.

4. The method for identifying the machining path of non-irregular thin-walled parts based on projection and bidirectional loop search according to claim 3, characterized in that, The deduplication process specifically involves: If the current edge to be processed overlaps with the two-dimensional projection edge that needs to be deduplicated together, then the overlapping part is deleted from the two-dimensional projection edge that needs to be deduplicated together and the deduplication process ends; if the current edge to be processed does not overlap with the two-dimensional projection edge that needs to be deduplicated together, then no processing is done and the deduplication process ends.

5. The method for identifying the machining path of non-irregular thin-walled parts based on projection and bidirectional loop search according to claim 3, characterized in that, Step S44 specifically involves: In the deduplicated set of 2D projected edges, one 2D projected edge is selected as the initial connecting edge. The search is performed along the vector direction of the initial connecting edge to trace the connection relationship of adjacent 2D projected edges, thereby constructing several closed minimum 2D loops. Then, the search is performed in the opposite direction of the vector direction of the initial connecting edge to trace the connection relationship of adjacent 2D projected edges, thereby constructing a closed outer boundary 2D loop. All minimum 2D loops and outer boundary 2D loops are summarized to obtain the set of 2D loops.

6. The method for identifying the machining path of non-irregular thin-walled parts based on projection and bidirectional loop search according to claim 3, characterized in that, Step S45 specifically involves: S451. Divide each two-dimensional ring in the set of two-dimensional rings into clockwise two-dimensional rings or counterclockwise two-dimensional rings; S452. First, for the counterclockwise two-dimensional ring, the surface to be projected corresponding to each two-dimensional projection edge in the two-dimensional ring is taken as the outer wall surface, thus obtaining the set of outer wall surfaces. S453. Then, traverse all clockwise two-dimensional rings: if any two-dimensional projection edge in the current two-dimensional ring has been identified as an outer wall surface, then do not process it, and traverse the next clockwise two-dimensional ring. If no two-dimensional projection edge in the current two-dimensional ring corresponds to a projection surface that is determined to be an outer wall surface, then each two-dimensional projection edge in the current two-dimensional ring corresponds to a projection surface that is determined to be an inner wall surface, thus obtaining the set of inner wall surfaces and ending the traversal.

7. The method for identifying the machining path of non-irregular thin-walled parts based on projection and bidirectional loop search according to claim 1, characterized in that, The set of processing paths is obtained by performing the following steps: Based on the topological connectivity, all wall boundary topological edges are selected from the outer wall set or the inner wall set. All wall boundary topological edges are then aggregated to form a wall boundary topological edge set. All topological edges in this set are further divided into an in-plane topological edge set and an edge topological edge set. For the in-plane topological edge set, each in-plane topological edge represents a loop processing path, thus the in-plane topological edge set becomes a loop processing path set. For the edge topological edge set, a directed edge search is performed on all edge topological edges within the set to obtain several directed connected lines, each of which represents a directed processing path. All directed processing paths are aggregated to form a directed processing path set. The loop processing path set and the directed processing path set are then combined to form the complete processing path set.

8. The method for identifying the machining path of non-irregular thin-walled parts based on projection and bidirectional loop search according to claim 7, characterized in that: The wall boundary topological edge is the topological edge used by a single wall; the in-plane topological edge is the topological edge located within the face of the wall; the edge topological edge is the topological edge located at the edge of the wall.

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 8.

Images