CN112488176B - Processing characteristic recognition method based on triangular mesh and neural network - Google Patents
Processing characteristic recognition method based on triangular mesh and neural network Download PDFInfo
- Publication number
- CN112488176B CN112488176B CN202011351826.4A CN202011351826A CN112488176B CN 112488176 B CN112488176 B CN 112488176B CN 202011351826 A CN202011351826 A CN 202011351826A CN 112488176 B CN112488176 B CN 112488176B
- Authority
- CN
- China
- Prior art keywords
- processing
- data
- neural network
- triangle
- triangular mesh
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000012545 processing Methods 0.000 title claims abstract description 136
- 238000000034 method Methods 0.000 title claims abstract description 57
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 41
- 238000012549 training Methods 0.000 claims abstract description 27
- 238000013461 design Methods 0.000 claims abstract description 12
- 239000013598 vector Substances 0.000 claims description 26
- 238000012360 testing method Methods 0.000 claims description 8
- 230000003746 surface roughness Effects 0.000 claims description 6
- 238000013075 data extraction Methods 0.000 claims description 3
- 238000003672 processing method Methods 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 2
- 238000004519 manufacturing process Methods 0.000 abstract description 8
- 230000000630 rising effect Effects 0.000 abstract description 2
- 229920006934 PMI Polymers 0.000 description 14
- 238000011161 development Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000000354 decomposition reaction Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000011960 computer-aided design Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/04—Manufacturing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Business, Economics & Management (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- Manufacturing & Machinery (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Numerical Control (AREA)
- Image Analysis (AREA)
Abstract
The application discloses a processing characteristic identification method based on a triangular mesh and a neural network, which comprises the following steps: importing a part design model, and obtaining processing characteristics through PMI positioning; performing triangular mesh division on the processing characteristics, and deriving triangular mesh data; extracting and processing the triangle mesh data; creating a custom processing feature data set; putting the custom processing characteristic data set into a neural network for training to obtain an optimal neural network training model; and inputting the processed data into an optimal neural network training model for processing feature recognition. The application supports the user-defined processing characteristics, the user can pertinently create the characteristic data set, and the problems of the increase of the processing characteristic types of parts, the complex and changeable shape and the like caused by the rising of the production modes of small batches and multiple varieties are solved, thereby improving the characteristic recognition efficiency and the accuracy.
Description
Technical Field
The application belongs to the field of mechanical design and manufacture and automation thereof, and relates to a processing characteristic identification method based on a triangular mesh and a neural network.
Background
With the development of manufacturing industry, the development mode of enterprises is changed from mass production to multi-variety and small-scale production, the types of processing characteristics of parts are increased, the shapes of the parts are complex and changeable, and new characteristics are required to be continuously added into the system. And the types of the processing characteristics of products in different fields are often different, however, the identification of various processing characteristics in multiple fields is difficult to realize in the prior art.
The processing feature recognition technology is a key support for intelligent design and manufacture. Feature identification is an essential fundamental building block in all computer aided design and manufacturing systems that analyze and make decisions by means of features. There are many current methods for identifying processing features. Among them, more classical methods can be generalized to graph-based methods, volume decomposition-based methods, rule-based methods, and trace-based methods.
The method based on the graph is difficult to effectively identify the intersecting characteristics, because the method based on the volume decomposition is only suitable for parts with certain shape requirements, the operation amount is large, and characteristic interpretation combination explosion can be caused to complex parts; the rule-based method is non-unique in definition of the characteristic rules, has no completeness, needs to be matched in a large amount, and is low in efficiency; whereas trace-based methods feature trace generation and continuation algorithms depend on specific feature types, it is difficult to add new feature types.
In recent years, the great success of neural networks in computer vision and pattern recognition has verified the powerful recognition and classification capabilities, and analogy to the neural network method has great development potential in the field of processing feature recognition. The basic idea of the neural network method is to implement a feature recognition task by inputting sample features into the network and repeatedly training the network according to the desired output result by means of the learning ability of the neural network. Therefore, what sample feature data is input to the neural network, i.e. how to pre-process and pre-encode the features, makes it easy to understand and infer the features a first problem to be solved.
At present, when the neural network is used for identifying the processing characteristics, the common characteristic preprocessing and encoding method mainly comprises the following steps: attribute-based adjacency graph coding, surface-based adjacency matrix coding, face-value vector coding, voxel coding methods, and the like. Based on attribute adjacency graph coding, surface adjacency matrix coding and face value vector coding, the methods have limitations and can only identify specific processing characteristics; the voxelization method is a characteristic preprocessing and encoding method of a three-dimensional CNN (convolutional neural network), and the patent 'an intelligent recognition and search method for three-dimensional processing characteristics' (patent application number: 201811594293.5) divides a three-dimensional model into voxel grids, performs binary assignment on voxels, further obtains input data of the CNN, but is greatly influenced by the resolution of the voxels, boundary characteristic information is lost if the resolution is low, and calculated amount rises exponentially if the resolution is high.
Therefore, a new solution is needed to solve the above problems.
Disclosure of Invention
The application aims to: in order to overcome the defects in the prior art, the processing feature recognition method based on the triangular mesh and the neural network is provided, the method solves the problems that the existing feature recognition technology is low in recognition efficiency, does not support user-defined features, cannot effectively recognize incomplete features and intersecting features and the like, solves the problems that feature data are difficult to input into the neural network and the like, and improves the feature recognition efficiency and accuracy.
The technical scheme is as follows: in order to achieve the above object, the present application provides a processing feature recognition method based on triangular mesh and neural network, comprising the following steps:
s1: importing a part design model, and obtaining processing characteristics through PMI positioning;
s2: performing triangular mesh division on the processing characteristics, and deriving triangular mesh data;
s3: extracting and processing the triangle mesh data;
s4: creating a custom processing feature data set;
s5: putting the custom processing characteristic data set into a neural network for training to obtain an optimal neural network training model;
s6: and (3) inputting the data processed in the step (S3) into an optimal neural network training model for processing feature recognition.
Further, the part design model in the step S1 is an MBD design model.
Further, in the step S1, the method for obtaining the processing characteristics through PMI positioning includes: firstly, acquiring a PMI label specific to processing characteristics; then obtain the face set f= { F associated with each PMI 1 ,f 2 ,...,f K -wherein k is the number of each PMI associated face; finally eachThe PMI associated surface set is the surface set which forms the corresponding processing characteristic.
Further, in the step S1, the PMI specific to the processing feature is denoted by a surface roughness PMI label, which is expressed as: p= { P 1 ,p 2 ,...,p t And (2) t is the number of the processing features to be identified on the imported model.
Further, the step S2 specifically includes: converting the processing characteristic association surface set F into a sheet body set S= { S 1 ,s 2 ,...,s K And performing triangular mesh division on the sheet body set, and deriving triangular mesh data.
Further, the triangular mesh data derived in the step S2 includes a normal vector of each triangle and coordinates of three vertex angles.
Further, the data extraction and processing method of the triangle mesh data in the step S3 is as follows:
a1: extracting normal vectors of each triangle and coordinates of three vertex angles in the triangle mesh data;
a2: calculating the center coordinates of each triangle;
a3: calculating the angle vector of each triangle (three vectors formed by the center and three vertex angles);
a4: carrying out normalization processing on the coordinate data;
a5: numbering each triangle after processing to obtain three adjacent surface numbers (less than three are complemented by own numbers) of each triangle.
Further, the creating method of the custom processing feature data set in step S4 is as follows:
b1: custom processing feature set n= { N 1 ,n 2 ,...,n S -wherein s is the number of custom machined features;
b2: for the ith processing feature type n i E N, creating such processing features m= { M of different sizes 1 ,m 2 ,...,m l -wherein i is the number created for each type of processing feature, the specific value of which is variable for each feature, steps B3 to B7 are performed;
B3：for the jth processing feature m j E, M, carrying out random rotation, and executing the steps B4 to B6;
b4: extracting processing characteristics;
b5: performing triangular mesh division on the processing characteristics, and deriving triangular mesh data;
b6: extracting and processing the triangle mesh data;
b7: the processed data is divided into a training set and a testing set.
Further, the data of the custom processing feature data set put into the training of the neural network in the step S5 includes the center coordinates, normal vectors, angle vectors, and adjacent triangle numbers of each triangle.
The design principle of the application is as follows:
the beneficial effects are that: compared with the prior art, the application has the following advantages:
1. according to the application, the PMI is used for rapidly positioning and acquiring the processing characteristics for the first time, so that the efficiency and accuracy of the whole characteristic identification process are improved;
2. the method supports user-defined processing characteristics, a user can pertinently create a characteristic data set, and the problems of increased processing characteristic types, complex and changeable shapes and the like of parts caused by the rising of production modes of small batches and multiple varieties are solved, so that the characteristic recognition efficiency and the accuracy are improved;
3. the method uses the triangular mesh as a data source, can effectively ensure the integrity of the characteristic information, and simultaneously uses the neural network as an algorithm basis, thereby ensuring the calculation efficiency and improving the characteristic identification capability;
4. the neural network is applied to feature recognition, has strong self-learning and generalization capability, can recognize similar features, and enhances the applicability of feature recognition.
Drawings
FIG. 1 is a general flow chart of the present application;
FIG. 2 is a flow chart of a method for data extraction and processing in the present application;
FIG. 3 is a flow chart of a method of processing feature dataset creation in accordance with the present application;
FIG. 4 is a flow chart of a method for extracting and processing data when creating a data set according to the present application;
FIG. 5 is a schematic view of an adjacent triangle;
FIG. 6 is a diagram showing a display of a type of processing feature;
fig. 7 is a processing feature identification display diagram.
Detailed Description
The present application is further illustrated in the accompanying drawings and detailed description which are to be understood as being merely illustrative of the application and not limiting of its scope, and various modifications of the application, which are equivalent to those skilled in the art upon reading the application, will fall within the scope of the application as defined in the appended claims.
The application provides a processing characteristic recognition method based on a triangular mesh and a neural network, which comprises two parts of processing characteristic recognition and neural network training. Before processing feature recognition, training of the neural network is first performed. The data set used for training the neural network can be automatically created by a user according to the requirements of the user after the processing characteristic type is determined. The processing characteristic identification process mainly comprises the following steps: acquiring processing characteristics; performing triangular mesh division on the processing characteristics, and deriving triangular mesh data; extracting and processing the triangle mesh data; and (5) putting the processed data into an optimal neural network training model for processing feature recognition.
For a better illustration of the process according to the application, as shown in fig. 1, the specific steps of the process according to the application are as follows:
step one: importing a part MBD design model to obtain processing characteristics;
and (3) importing the part MBD design model into three-dimensional CAD software NX, and obtaining all PMI labels of the model by utilizing an NX secondary development technology. Since the surface roughness belongs to the PMI attribute specific to the machined feature, all machined features can be located by screening out all surface roughness PMIs. The surface set associated with the surface roughness PMI is the surface set f= { F constituting the corresponding processing feature 1 ,f 2 ,...,f K And (2) k is the number of each PMI associated surface.
Step two: performing triangular mesh division on the processing characteristics, and deriving triangular mesh data;
because three-dimensional CAD software can only perform triangular mesh division on a body, the processing characteristic association surface set F is firstly converted into a sheet body set S= { S 1 ,s 2 ,...,s K And then, carrying out triangular mesh division on the sheet set to derive stl-format files. The stl format file in this embodiment contains normal vectors for each triangle of the triangular meshAnd coordinates P of three vertex angles 1 、P 2 、P 3 。
Step three: extracting and processing the triangle mesh data;
as shown in fig. 2, a specific extraction and processing method is as follows:
(31) Extracting normal vector of each triangle in triangle mesh data from derived stl fileCoordinates P of three vertex angles 1 、P 2 、P 3 ；
(32) Calculate the center coordinates C of each triangle o ；
(33) Calculating the angle vector of each triangle (three vectors formed by the center and three vertex angles);
(34) Normalizing the coordinate data;
first calculate the distance between the vertex coordinates P (x, y, z) of all triangles and the originAnd (3) separating:comparing to obtain the maximum distance d max Then multiply each triangle vertex coordinates by +.>
(35) Each triangle is numbered and three adjacent face numbers (less than three complements with their own numbers) for each triangle are calculated.
The numbering of step 35 is described here: if two triangles are adjacent, then the two triangles share one edge, i.e., have two common vertex angles. Firstly, numbering each triangle; then, for each surface, judging whether the two surfaces have two same vertex angles at the same time by traversing all other surfaces to determine whether the surfaces are adjacent surfaces; finally, the adjacent face number of each face is recorded.
As shown in FIG. 5, the triangle with the number a in the present embodiment is adjacent to the triangle with the numbers b, c, d, denoted as L a ＝{b,c,d}
Step four: creating a custom processing feature data set;
as shown in fig. 3, the specific creation method is as follows:
(41) Custom processing feature type set n= { N 1 ,n 2 ,...,n S -wherein s is the number of custom machined features; the user generalizes the required feature type set N= { N according to specific requirements 1 ,n 2 ,...,n S }。
(42) For the ith processing feature type n i E N, creating such processing features m= { M of different sizes 1 ,m 2 ,...,m l -wherein i is the number created for each type of processing feature, the specific value of which is variable for each feature, steps (43) to (47) are performed;
for each type of processing feature in the processing feature type set, a user needs to prepare a plurality of processing features of the type, a three-dimensional CAD software can be used for creating a model of the processing features, and the processing features can be obtained from the existing model. The number of specific processing features is not a fixed requirement. Steps (43) to (47) are performed.
(43) For the jth processing feature m j E M, performing random rotation, and executing the steps (44) to (46); for a certain processing feature of a certain class, a random rotation operation is performed on the processing feature in order to enhance the data set.
In the embodiment, the random rotation operation uses the origin O of coordinates of the three-dimensional model where the processing feature is located as the origin P of the rotation axis; randomly generating a unit vectorForms a rotation axis with the origin P>Randomly generating rotation angle theta, theta epsilon [0,2 pi ]]The method comprises the steps of carrying out a first treatment on the surface of the The mould is +.>And rotating theta. Steps (44) to (46) are performed.
(44) Extracting processing characteristics;
in order to facilitate the extraction of the processing features, the method of acquiring the processing features in the first step is utilized to firstly associate the surfaces contained in the processing features with a surface roughness, and then a secondary development technology is utilized to acquire the processing feature surface set.
(45) Performing triangular mesh division on the processing characteristics, and deriving triangular mesh data;
firstly, converting the processing characteristic surface set into a sheet set, and then carrying out triangular mesh division on the sheet set to derive stl-format files. The stl format file in this embodiment contains normal vectors for each triangle of the triangular meshAnd coordinates P of three vertex angles 1 、P 2 、P 3 。
(46) Extracting and processing the triangle mesh data;
as shown in fig. 4, the specific method is as follows;
1) Extracting normal vector of each triangle in triangle mesh data from derived stl fileCoordinates P of three vertex angles 1 、P 2 、P 3 ；
2) Calculate the center coordinates C of each triangle o ；
3) Calculating the angle vector of each triangle (three vectors formed by the center and three vertex angles);
4) Normalizing the coordinate data; first, the distance between the vertex coordinates P (x, y, z) of all triangles and the origin is calculated:comparing to obtain the maximum distance d max Then multiply each triangle vertex angle coordinate
5) Numbering each triangle, and calculating three adjacent surface numbers (less than three are complemented by own numbers) of each triangle; if two triangles are adjacent, then the two triangles share one edge, i.e., have two common vertex angles. Firstly, numbering each triangle; then, for each surface, judging whether the two surfaces have two same vertex angles at the same time by traversing all other surfaces to determine whether the surfaces are adjacent surfaces; finally, the adjacent face number of each face is recorded.
6) The data are stored in a folder named by the processing characteristic type, the processed data are stored in the folder, and the processed data mainly refer to center coordinates, normal vectors, angle vectors and adjacent triangle numbers of triangles and are stored in a npz file format. The naming of the feature type is to use the name of the folder as a label at the time of training.
(47) And dividing the processed processing characteristic data of each type into a training set and a testing set. In this embodiment, the processed data are divided into a training set and a test set according to a ratio of 7:3 for training test of the neural network.
Step five: and putting the training set and the testing set into a neural network for training to obtain an optimal neural network training model.
Step six: and D, inputting the data processed in the third step into the optimal neural network training model obtained in the fifth step for processing feature recognition.
In order to verify the effect of the above method, in this embodiment, the test is performed by using the above method, specifically taking a marine diesel engine key part as an example, 7 kinds of processing features are defined in total, and as shown in fig. 6, the processing features are respectively as follows: plane, through hole, circular arc groove, straight flute, flat bottom blind hole, taper bottom blind hole, shoulder hole. 2-5 processing features are selected for each type of feature, each processing feature rotates 10 times around 15 rotation axes which are randomly generated, each rotation is at a random angle, and then the processing features are derived into triangular mesh data and are subjected to data processing. Thus, 2678 files were generated in total, and divided into training and test sets according to a 7:3 ratio.
And finally, 32 processing characteristics are selected on a marine diesel engine frame, a cylinder cover and a connecting rod for verification, as shown in fig. 7, the 32 processing characteristics are completely identified correctly, and the accuracy reaches 100% and is obviously higher than that of the existing identification method.
Claims (6)
1. A processing characteristic recognition method based on triangular meshes and a neural network is characterized by comprising the following steps of: the method comprises the following steps:
s1: importing a part design model, and obtaining processing characteristics through PMI positioning;
s2: performing triangular mesh division on the processing characteristics, and deriving triangular mesh data;
s3: extracting and processing the triangle mesh data;
s4: creating a custom processing feature data set;
s5: putting the custom processing characteristic data set into a neural network for training to obtain an optimal neural network training model;
s6: inputting the data processed in the step S3 into an optimal neural network training model for processing feature recognition;
the triangular mesh data derived in the step S2 comprise normal vectors of each triangle and coordinates of three vertex angles;
the data extraction and processing method of the triangular mesh data in the step S3 comprises the following steps:
a1: extracting normal vectors of each triangle and coordinates of three vertex angles in the triangle mesh data;
a2: calculating the center coordinates of each triangle;
a3: calculating the angle vector of each triangle, wherein the angle vector is three vectors formed by the center and three vertex angles;
a4: carrying out normalization processing on the coordinate data;
a5: numbering each triangle after processing to obtain three adjacent surface numbers of each triangle;
the data of the custom processing feature data set put into the training of the neural network in the step S5 includes the center coordinates, normal vectors, angle vectors, and adjacent triangle numbers of each triangle.
2. The processing feature recognition method based on the triangular mesh and the neural network according to claim 1, wherein the processing feature recognition method is characterized by comprising the following steps of: the part design model in the step S1 is an MBD design model.
3. The processing feature recognition method based on the triangular mesh and the neural network according to claim 1, wherein the processing feature recognition method is characterized by comprising the following steps of: the method for acquiring the processing characteristics through PMI positioning in the step S1 comprises the following steps: firstly, obtaining a PMI mark special for processing characteristicsInjecting; then obtain the face set f= { F associated with each PMI 1 ,f 2 ,...,f K -wherein k is the number of each PMI associated face; and finally, each PMI associated surface set is a surface set forming corresponding processing characteristics.
4. A method for identifying processing features based on triangular meshes and neural networks according to claim 3, wherein: in the step S1, the PMI label specific to the processing feature is a surface roughness PMI label, which is expressed as: p= { P 1 ,p 2 ,...,p t And (2) t is the number of the processing features to be identified on the imported model.
5. A method for identifying processing features based on triangular meshes and neural networks according to claim 3, wherein: the step S2 specifically comprises the following steps: converting the processing characteristic association surface set F into a sheet body set S= { S 1 ,s 2 ,...,s K And performing triangular mesh division on the sheet body set, and deriving triangular mesh data.
6. The processing feature recognition method based on the triangular mesh and the neural network according to claim 1, wherein the processing feature recognition method is characterized by comprising the following steps of: the creating method of the custom processing feature data set in the step S4 is as follows:
b1: custom processing feature set n= { N 1 ,n 2 ,...,n S -wherein s is the number of custom machined features;
b2: for the ith processing feature type n i E N, creating such processing features m= { M of different sizes 1 ,m 2 ,...,m l -wherein i is the number created for each class of processing features;
b3: for the jth processing feature m j E, M, carrying out random rotation;
b4: extracting processing characteristics;
b5: performing triangular mesh division on the processing characteristics, and deriving triangular mesh data;
b6: extracting and processing the triangle mesh data;
b7: the processed data is divided into a training set and a testing set.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011351826.4A CN112488176B (en) | 2020-11-26 | 2020-11-26 | Processing characteristic recognition method based on triangular mesh and neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011351826.4A CN112488176B (en) | 2020-11-26 | 2020-11-26 | Processing characteristic recognition method based on triangular mesh and neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112488176A CN112488176A (en) | 2021-03-12 |
| CN112488176B true CN112488176B (en) | 2023-11-21 |
Family
ID=74935373
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011351826.4A Active CN112488176B (en) | 2020-11-26 | 2020-11-26 | Processing characteristic recognition method based on triangular mesh and neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112488176B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110570503A (en) * | 2019-09-03 | 2019-12-13 | 浙江大学 | Method for acquiring normal vector, geometry and material of three-dimensional object based on neural network |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190164055A1 (en) * | 2017-11-29 | 2019-05-30 | Microsoft Technology Licensing, Llc | Training neural networks to detect similar three-dimensional objects using fuzzy identification |
-
2020
- 2020-11-26 CN CN202011351826.4A patent/CN112488176B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110570503A (en) * | 2019-09-03 | 2019-12-13 | 浙江大学 | Method for acquiring normal vector, geometry and material of three-dimensional object based on neural network |
Non-Patent Citations (2)
| Title |
|---|
| 基于RBF神经网络的三维网格实体模型重构的研究;鄢腊梅, 袁友伟, 周锋;机械科学与技术(10);全文 * |
| 面向三维模型分割的边界感知点云神经网络;关柏良;周凡;林淑金;罗笑南;;计算机辅助设计与图形学学报(01);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112488176A (en) | 2021-03-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Shi et al. | Manufacturability analysis for additive manufacturing using a novel feature recognition technique | |
| CN109117560B (en) | Three-dimensional process design method and platform for machining parts of automotive typical machine based on MBD | |
| Jayaraman et al. | Uv-net: Learning from boundary representations | |
| Shi et al. | Manufacturing feature recognition with a 2D convolutional neural network | |
| CN108073682A (en) | Based on parameter view functional query database | |
| Peddireddy et al. | Deep learning based approach for identifying conventional machining processes from CAD data | |
| CN107679333A (en) | A kind of method of two-dimensional topology optimum results geometry reconstruction | |
| Ning et al. | Part machining feature recognition based on a deep learning method | |
| Khorolska et al. | Application of a convolutional neural network with a module of elementary graphic primitive classifiers in the problems of recognition of drawing documentation and transformation of 2D to 3D models | |
| Ip et al. | A 3D object classifier for discriminating manufacturing processes | |
| Fu et al. | Improved dexel representation: A 3-d cnn geometry descriptor for manufacturing cad | |
| Anwer et al. | Toward a classification of partitioning operations for standardization of geometrical product specifications and verification | |
| Dommaraju et al. | Identifying topological prototypes using deep point cloud autoencoder networks | |
| CN112488176B (en) | Processing characteristic recognition method based on triangular mesh and neural network | |
| CN111914480B (en) | Processing feature intelligent recognition method based on point cloud semantic segmentation | |
| Luo et al. | An ELM-embedded deep learning based intelligent recognition system for computer numeric control machine tools | |
| Wei et al. | Manufacturing data-driven process adaptive design method | |
| Jia et al. | Machining feature recognition method based on improved mesh neural network | |
| Meltzer et al. | UVStyle-Net: Unsupervised few-shot learning of 3D style similarity measure for B-reps | |
| Lim et al. | Machining feature recognition using descriptors with range constraints for mechanical 3D models | |
| Lee et al. | BRepGAT: Graph neural network to segment machining feature faces in a B-rep model | |
| CN110046335B (en) | Method for rapidly generating appearance detection report | |
| CN109658489B (en) | Three-dimensional grid data processing method and system based on neural network | |
| Roj et al. | Classification of CAD-Models Based on Graph Structures and Machine Learning | |
| Yu et al. | Clustering-enhanced pointcnn for point cloud classification learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |