A method for identifying machining features of broken surfaces based on attributed adjacency graph

Abstract

The present application is about a method for identifying processing features containing broken faces based on an attribute adjacency graph, and relates to the field of intelligent manufacturing. The method includes: obtaining a target processing entity model, which includes at least one processing feature; extracting node features from the target processing entity model to obtain an initial attribute adjacency graph corresponding to the target processing entity model; determining at least one broken face node in the attribute adjacency graph based on a broken face set judgment rule; screening the broken face nodes in the initial attribute adjacency graph to obtain an attribute adjacency graph corresponding to the target processing entity; and generating a processing feature recognition result corresponding to the target processing entity model based on the attribute adjacency graph. In the process of feature processing recognition, by identifying and screening broken faces, a more accurate recognition of the processing features in the target processing entity is achieved, thereby improving the accuracy of feature recognition on a three-dimensional model containing broken faces.

Description

Technical Field

[0001] The present application relates to the field of intelligent manufacturing, and in particular to a method for identifying processing features of fragmented surfaces based on an attribute adjacency graph. Background Art

[0002] Machining features refer to geometric entities on a part that can be formed using basic machining methods. They are the prerequisite for determining the machining process for a part. Typical examples include holes, cavities, and chamfers. With the advent of the information age and the rapid development of computer technology, technologies such as Computer Aided Design (CAD) and Computer Aided Manufacturing (CAM) have been widely used in the manufacturing industry. Due to the complexity of actual parts, selecting machining features in CAM software often requires the participation of process personnel. This process is subject to frequent human-computer interaction and is time-consuming. If the process of identifying machining features is automatically implemented through algorithms, it will have important practical significance for improving CNC machining operating systems, increasing a company's production and manufacturing efficiency, and reducing the total manufacturing cost of parts.

[0003] In related technologies, machining feature recognition methods are mainly divided into two routes based on the representation form of the model: boundary representation-based and volume decomposition-based. Compared with the removed parts of the blank, the volume decomposition-based method has a higher dimension of feature representation when analyzing the model, which increases the computational complexity. Therefore, the mainstream method is the boundary-based method, which analyzes and solves based on the boundary representation information of the model. Based on the type of convenient representation information, machining feature recognition methods can be further divided into graph-based (attribute adjacency graph), point cloud-based, grid-based, etc. In particular, the popularity of artificial intelligence has driven the application of computer vision technology in the field of machining feature recognition. The advantage of neural networks in extracting key features from a large amount of geometric information has promoted the advancement of technical methods such as point clouds and grids.

[0004] In related technologies, the model features of open source CAD part datasets such as MFInstSeg are relatively simple and the number of model faces is relatively small. In the process of feature processing and recognition containing broken faces, the neural network trained based on such datasets does not perform well on complex models. Summary of the Invention

[0005] This application relates to a method for identifying surface processing features containing broken pieces based on an attribute adjacency graph, which improves the recognition effect of surface processing features containing broken pieces. The method is applied to a computer device and includes:

[0006] Acquire a target processing entity model, wherein the target processing entity model includes at least one processing feature;

[0007] Extracting node features from the target processing entity model to obtain an initial attribute adjacency graph corresponding to the target processing entity model, wherein the initial attribute adjacency graph includes at least two feature nodes, and two adjacent feature nodes are connected by a feature connection line;

[0008] Determine at least one fragment node in the attribute adjacency graph based on a fragment set determination rule;

[0009] The broken face nodes in the initial attribute adjacency graph are screened out to obtain the attribute adjacency graph corresponding to the target processing entity;

[0010] The machining feature recognition results corresponding to the target machining entity model are generated based on the attribute adjacency graph.

[0011] In an optional embodiment, node feature extraction is performed on the target processing entity model to obtain an initial attribute adjacency graph corresponding to the target processing entity model, including:

[0012] Extracting node features in the target machining solid model, the node features including at least one of plane node features, cylindrical node features, conical node features, spherical node features, torus node features, spline surface node features, and other surface node features;

[0013] Generate feature nodes in the initial attribute adjacency graph based on node features;

[0014] Extract edge features between two adjacent feature nodes in the target processing entity model. The edge features include directed edge features with the same value and directed edge features with different values.

[0015] Generate feature connection lines in the initial attribute adjacency graph based on edge features;

[0016] Based on the feature connection lines and feature nodes, an initial attribute adjacency graph corresponding to the target processing entity model is generated.

[0017] In an optional embodiment, extracting edge features between two adjacent feature nodes in the target processing entity model includes:

[0018] Determine an edge curve type between two adjacent feature nodes in a target machining entity model, where the edge curve type includes at least one of a straight line type, a circular arc type, an elliptical arc type, a spiral line type, a spline curve type, and other curve types;

[0019] Based on the edge curve type and the surface shape assignment rule corresponding to the edge curve type, 1 generates edge features between two adjacent feature nodes.

[0020] In an optional embodiment, the method further includes:

[0021] Based on the edge features, the feature connection lines are assigned numbers.

[0022] In an optional embodiment, the facet set determination rules include face type restriction rules, adjacent edge restriction rules, and directed edge attribute restriction rules.

[0023] In an optional embodiment, determining at least one fragment node in the attribute adjacency graph based on the fragment set determination rule includes:

[0024] In response to at least two adjacent feature nodes meeting the fragment face set determination rule, the at least two adjacent feature nodes are determined to be fragment face nodes and belong to the same fragment face node set.

[0025] In an optional embodiment, the feature connection line further includes a broken surface feature connection line;

[0026] The method also includes:

[0027] In response to at least two adjacent feature nodes being broken surface nodes, a connection line connecting the two adjacent broken surface nodes is modified into a broken surface feature connection line.

[0028] In an optional embodiment, the broken face nodes in the initial attribute adjacency graph are screened out to obtain an attribute adjacency graph corresponding to the target processing entity, including:

[0029] Extract key broken surface nodes from the broken surface node set;

[0030] The key broken surface nodes are retained, and other broken surface nodes in the broken surface node set are screened out to obtain the attribute adjacency graph corresponding to the target processing entity.

[0031] In an optional embodiment, the method further includes:

[0032] Determine the association relationship between the broken face nodes in the broken face node set and other feature nodes in the initial attribute adjacency graph;

[0033] Adjust the topological relationship between the key fragment nodes and other feature nodes in the initial attribute adjacency graph based on the association relationship.

[0034] In an optional embodiment, generating a machining feature recognition result corresponding to the target machining entity model based on the attribute adjacency graph includes:

[0035] The processing feature recognition results in text form corresponding to the target processing entity model are generated based on the attribute adjacency graph.

[0036] The technical effects of each embodiment of the present application include at least:

[0037] During the process of identifying machining features for a target machining entity, after generating an initial attribute adjacency graph corresponding to the target machining entity, the initial attribute adjacency graph is used to identify and remove broken face nodes. After these nodes are removed, the machining feature recognition results corresponding to the target machining entity model are determined. During the feature machining recognition process, by identifying and removing broken faces, more accurate identification of machining features in the target machining entity is achieved, thereby improving the accuracy of feature recognition on 3D models containing broken faces. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0039] Figure 1 A flow chart of a method for identifying features of surface processing containing fragments based on an attribute adjacency graph provided by an exemplary embodiment of the present application is shown.

[0040] Figure 2 A flow chart of another method for identifying features of surface processing containing fragments based on an attribute adjacency graph provided by an exemplary embodiment of the present application is shown.

[0041] Figure 3 A schematic diagram of a target processing entity model provided by an exemplary embodiment of the present application is shown.

[0042] Figure 4 A schematic diagram of an initial attribute adjacency graph provided by an exemplary embodiment of the present application is shown. DETAILED DESCRIPTION

[0043] In order to make the objectives, technical solutions and advantages of this application clearer, the implementation methods of this application will be further described in detail below with reference to the accompanying drawings.

[0044] Figure 1 A flow chart of a method for identifying surface processing features containing broken pieces based on an attribute adjacency graph provided by an exemplary embodiment of the present application is shown. The method is described using an example of a computer device. The method includes:

[0045] Step 101: Acquire a target processing entity model.

[0046] In the embodiment of the present application, the target processing entity model is a physical model corresponding to the part to be processed. Optionally, the target processing entity model can be a model obtained based on artificial intelligence technology, which corresponds to the part entity.

[0047] Step 102 : extract node features from the target processing entity model to obtain an initial attribute adjacency graph corresponding to the target processing entity model.

[0048] In an embodiment of the present application, the initial attribute adjacency graph is implemented in a computer device as a picture. The initial attribute adjacency graph includes at least two feature nodes, and two adjacent feature nodes are connected by a feature connection line.

[0049] Step 103: Determine at least one fragment node in the attribute adjacency graph based on the fragment set determination rule.

[0050] In the embodiment of the present application, the fragment face set determination rule is used to determine whether the initial attribute adjacency graph contains nodes obtained by extracting fragment face features. In the embodiment of the present application, the number of fragment face feature nodes is usually more than 2, which are used to characterize the fragment face.

[0051] Step 104 , filtering out the broken face nodes in the initial attribute adjacency graph to obtain an attribute adjacency graph corresponding to the target processing entity.

[0052] This process is the process of filtering out broken face nodes. In some embodiments of this application, broken face nodes are completely filtered out; in other embodiments of this application, broken face nodes are replaced with newly created feature nodes; in still other embodiments of this application, broken face nodes and their corresponding topological relationships are organized to form an attribute adjacency graph. This application does not limit the actual form of the attribute adjacency graph.

[0053] Step 105 : generating a machining feature recognition result corresponding to the target machining entity model based on the attribute adjacency graph.

[0054] In the embodiment of the present application, the processing feature recognition result is a result corresponding to the target processing entity, including the type, content and corresponding value of the processing feature in the target processing entity. In one example, the processing feature recognition result can be implemented in text form.

[0055] In summary, the method provided in the embodiments of the present application, during the process of identifying machining features for a target machining entity, generates an initial attribute adjacency graph corresponding to the target machining entity, identifies broken surface nodes in the initial attribute adjacency graph, and screens out the broken surface nodes. After screening out, the machining feature identification result corresponding to the target machining entity model is determined. During the process of feature machining identification, by identifying and screening out broken surfaces, more accurate identification of machining features in the target machining entity is achieved, thereby improving the accuracy of feature identification on three-dimensional models containing broken surfaces.

[0056] Figure 2 A flow chart of another method for identifying surface processing features containing broken pieces based on an attribute adjacency graph provided by an exemplary embodiment of the present application is shown. This method is described using a computer device as an example. The method includes:

[0057] Step 201: Acquire a target processing entity model.

[0058] This process corresponds to the process shown in step 101 and will not be described in detail here.

[0059] Step 202: extract node features in the target processing entity model.

[0060] It should be noted that, in the embodiment of the present application, before performing node feature extraction on the target processing entity model, the target processing entity model needs to be extracted as a wireframe model, that is, a model that only represents the lines of the target processing entity but not the color of the target processing entity. In one example, the target processing entity model 310 is in the form of Figure 3 shown.

[0061] In an embodiment of the present application, the node feature includes at least one of a plane node feature, a cylindrical node feature, a conical node feature, a spherical node feature, a torus node feature, a spline surface node feature, and other surface node features.

[0062] Step 203: Generate feature nodes in the initial attribute adjacency graph based on the node features.

[0063] It should be noted that the computer device will assign values ​​corresponding to different node features. In an example, the value of the node is defined as N=0~5, where 1~5 corresponds to plane / cylinder / cone / sphere / torus, and 0 represents other planes that do not belong to the above types.

[0064] Step 204 : extracting edge features between two adjacent feature nodes in the target processing entity model.

[0065] In the embodiment of the present application, the edge features include same-value directed edge features and different-value directed edge features.

[0066] In an embodiment of the present application, the edge between two feature nodes has edge type characteristics and edge phase characteristics. Among them, the computer device will first obtain the internal and external relationship between the edge and the corresponding surface, the concave and convex state of the edge, and the shape of the edge, and finally generate the edge characteristics. Optionally, the computer device defines the value E=XYZ of the directed edge. X=1~2, corresponding to the outer edge of surface i / the inner edge of surface i; Y=1~3, corresponding to concave edge / convex edge / smooth edge; Z=0~4, where 1~4 corresponds to straight line / circular arc / elliptical arc / spiral, and 0 represents other directed edges that do not belong to the above geometric types. Optionally, the edge curve type includes at least one of a straight line type, a circular arc type, an elliptical arc type, a spiral type, a spline curve type, and other curve types.

[0067] Step 205: Generate feature connection lines in the initial attribute adjacency graph based on the edge features.

[0068] It should be noted that in the embodiment of the present application, there is a process of assigning values ​​to feature connecting lines based on edge features.

[0069] Step 206 : generating an initial attribute adjacency graph corresponding to the target processing entity model based on the feature connection lines and the feature nodes.

[0070] In one embodiment of the present application, the form of the initial attribute adjacency graph is as follows: Figure 4 As shown. Figure 4 , including feature nodes 401 to 408, and there is a feature connection line with assigned values ​​between two adjacent feature nodes.

[0071] Step 207 : In response to at least two adjacent feature nodes meeting the fragment face set determination rule, the at least two adjacent feature nodes are determined to be fragment face nodes, and the nodes belong to the same fragment face node set.

[0072] In the embodiment of the present application, the rules for determining the fragmented face set include face type restriction rules, adjacent edge restriction rules, and directed edge attribute restriction rules. This corresponds to the fact that in an entity, the necessary and sufficient condition for judging that two faces belong to the same fragmented face set is that the two faces are of the same type and the transition edges are smooth. By decomposing the corresponding node features and edge features, we can obtain three conditions: the same face type, smooth adjacent edges, and the same attributes of the directed edges adjacent to the same face. When adjacent feature nodes meet at least three of the above conditions, it can be determined that the corresponding nodes all belong to fragmented face nodes, and the corresponding fragmented face nodes belong to the same fragmented face node set.

[0073] In an embodiment of the present application, in response to at least two adjacent feature nodes being fractal nodes, the connection line connecting the two adjacent fractal nodes is modified to a fractal feature connection line. In the initial feature adjacency graph, the connection line between the fractal nodes can be equivalently represented by a dotted line.

[0074] Step 208: extract key fragment nodes from the fragment node set.

[0075] This process is the key surface node extraction process based on step 207.

[0076] Step 209 : retain the key broken surface nodes, filter out other broken surface nodes in the broken surface node set, and obtain an attribute adjacency graph corresponding to the target processing entity.

[0077] It should be noted that in this process, it is necessary to determine the association relationship between the fragment nodes in the fragment node set and other feature nodes in the initial attribute adjacency graph; and based on the association relationship, adjust the topological relationship between the key fragment nodes and other feature nodes in the initial attribute adjacency graph, so as to delete some fragment nodes and retain at least one fragment node in a fragment node set to obtain the attribute adjacency graph corresponding to the target processing entity.

[0078] Step 210 : Generate a processing feature recognition result in text form corresponding to the target processing entity model based on the attribute adjacency graph.

[0079] In embodiments of the present application, the processing feature recognition results can be implemented in text form. In one example, a computer device stores a database corresponding to processing features in the prior art. Based on a comparison of the database's attribute adjacency graph, the computer device generates at least one processing feature corresponding to the target processing entity and generates visual information in text form to illustrate the result.

[0080] In summary, the method provided in the embodiments of the present application, during the process of identifying machining features for a target machining entity, generates an initial attribute adjacency graph corresponding to the target machining entity, identifies broken surface nodes in the initial attribute adjacency graph, and screens out the broken surface nodes. After screening out, the machining feature identification result corresponding to the target machining entity model is determined. During the process of feature machining identification, by identifying and screening out broken surfaces, more accurate identification of machining features in the target machining entity is achieved, thereby improving the accuracy of feature identification on three-dimensional models containing broken surfaces.

[0081] The above are only optional embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application.

Claims

1. A method for identifying surface processing features containing debris based on attribute adjacency graph, characterized in that: The method is applied to a computer device, and the method includes: Acquire a target processing entity model, wherein the target processing entity model includes at least one processing feature; Extracting node features from the target processing entity model to obtain an initial attribute adjacency graph corresponding to the target processing entity model, wherein the initial attribute adjacency graph includes at least two feature nodes, and two adjacent feature nodes are connected by a feature connection line; Determining at least one fragmented face node in the attribute adjacency graph based on a fragmented face set determination rule, wherein the fragmented face set determination rule includes a face type restriction rule, an adjacent edge restriction rule, and a directed edge attribute restriction rule; Screening out the broken face nodes in the initial attribute adjacency graph to obtain an attribute adjacency graph corresponding to the target processing entity; generating a machining feature recognition result corresponding to the target machining entity model based on the attribute adjacency graph; The determining of at least one fragment node in the attribute adjacency graph based on the fragment set determination rule includes: In response to at least two adjacent feature nodes meeting the fragment face set determination rule, determining the at least two adjacent feature nodes as the fragment face nodes and belonging to the same fragment face node set; The step of extracting node features from the target processing entity model to obtain an initial attribute adjacency graph corresponding to the target processing entity model includes: Extracting node features in the target machining entity model, wherein the node features include at least one of plane node features, cylindrical node features, conical node features, spherical node features, torus node features, and spline surface node features; generating a feature node in the initial attribute adjacency graph based on the node feature; Extracting edge features between two adjacent feature nodes in the target processing entity model, wherein the edge features include directed edge features with the same value and directed edge features with different values; generating feature connecting lines in the initial attribute adjacency graph based on the edge features; generating an initial attribute adjacency graph corresponding to the target processing entity model based on the feature connection lines and the feature nodes; Wherein, the feature connection lines also include broken surface feature connection lines; The method further comprises: In response to at least two adjacent feature nodes being the broken surface nodes, modifying a connection line connecting the two adjacent broken surface nodes into a broken surface feature connection line; The step of removing the broken face nodes in the initial attribute adjacency graph to obtain the attribute adjacency graph corresponding to the target processing entity includes: Extracting key broken surface nodes from the broken surface node set; The key broken surface nodes are retained, and other broken surface nodes in the broken surface node set are screened out to obtain the attribute adjacency graph corresponding to the target processing entity; The method further comprises: Determining association relationships between the fragment nodes in the fragment node set and other feature nodes in the initial attribute adjacency graph; The topological relationship between the key fragment node and other feature nodes in the initial attribute adjacency graph is adjusted based on the association relationship.

2. The method according to claim 1, characterized in that The step of extracting edge features between two adjacent feature nodes in the target processing entity model includes: Determining an edge curve type between two adjacent feature nodes in the target machining entity model, where the edge curve type includes at least one of a straight line type, a circular arc type, an elliptical arc type, a spiral line type, and a spline curve type; Based on the edge curve type and the in-plane shape assignment rule corresponding to the edge curve type, the edge feature between the two adjacent feature nodes is generated.

3. The method according to claim 2, characterized in that The method further comprises: Based on the edge features, the feature connecting lines are assigned numbers.

4. The method according to claim 1, wherein Generating a machining feature recognition result corresponding to the target machining entity model based on the attribute adjacency graph includes: A processing feature recognition result in text form corresponding to the target processing entity model is generated based on the attribute adjacency graph.

Images