Automobile part modeling method based on feature recognition

Abstract

The invention discloses an automobile part modeling method based on feature recognition, and the method comprises the following steps: a, design analysis: analyzing the design requirements of parts, including the appearance, size, assembly requirements, stress conditions and the like, b, scanning the automobile parts through a three-dimensional scanner, recognizing the key features of the parts according to the design requirements, and carrying out the modeling of the automobile parts through the three-dimensional scanner. The key feature information of the part is presented in the form of point cloud data. According to the method, the deep learning algorithm is adopted, the geometric features of the parts are recognized by training the neural network model, especially for complex and irregular parts, so that the requirement for manual operation can be reduced, the feature recognition speed and accuracy can be improved, and multi-physical field simulation and optimization, such as multi-physical field simulation and optimization, are introduced in the feature recognition and modeling process. When the geometric features of the part are created, the mechanical property, thermal expansion, fatigue and other factors of the part can be considered at the same time, coupling analysis is conducted in the modeling stage, and the precision and reliability of part design are greatly improved.

Description

Technical Field

[0001] The invention relates to the technical field of automobile parts production, in particular to an automobile parts modeling method based on feature recognition. Background Art

[0002] The automobile parts modeling method is one of the key technologies in computer-aided design (CAD) and computer-aided engineering (CAE). It aims to create three-dimensional geometric models of automobile parts through various modeling techniques. These models are not only used for virtual design and simulation, but also for manufacturing, assembly, analysis and other links. According to the complexity and needs of modeling, common automobile parts modeling methods can be divided into the following types: feature-based modeling method, free-form surface-based modeling method, solid modeling and parameter-based modeling method, etc.; At present, the process of feature recognition still relies on the experience and manual operation of designers. Especially when dealing with complex or irregular shapes, the proportion of manual intervention is high, which cannot meet the needs of today's market. In order to improve efficiency and reduce human errors, the degree of automation of feature recognition can be further improved. Here, a method of automotive parts modeling based on feature recognition is introduced specifically. Summary of the invention

[0003] The purpose of the present invention is to provide an automobile part modeling method based on feature recognition, which has the advantage of high feature recognition efficiency and solves the problem that the existing automobile part modeling method relies on the experience and manual operation of designers, especially when dealing with complex or irregular shapes, the proportion of manual intervention is high.

[0004] To achieve the above object, the present invention provides the following technical solution: a method for modeling automobile parts based on feature recognition, comprising the following steps: Step a: Design analysis, analyzing the design requirements of parts, including shape, size, assembly requirements, stress conditions, etc.; Step b: Data collection: Scan the auto parts with a 3D scanner and identify the key features of the parts according to the design requirements. The key feature information of the parts is presented in the form of point cloud data, which contains the geometric shape information of the parts. Step c: Add geometric constraints: Add appropriate constraint relationships (such as parallel, vertical, concentric, etc.) to the geometric elements (such as line segments, arcs, etc.) in the sketch to ensure that the geometric constraint relationships of the sketch are clear; Step d: Data preprocessing, denoising the collected data, because the scanning process may be disturbed by environmental factors and generate noise points. After denoising, data simplification is performed. The point cloud data may be very large. The simplification operation can reduce the amount of data and improve the efficiency of subsequent processing without affecting the expression of the main features of the parts; Step e: Feature recognition, rule-based recognition: Identify features based on pre-defined geometric and topological rules, and use existing annotated part data to train the model, such as using a convolutional neural network (CNN). The trained model can classify and identify features in new part data; Step f: Feature parameter extraction. For the identified features, such as bosses, it is necessary to extract the dimensional parameters such as height, length, and width. If it is a slot feature, it is necessary to extract the depth and width of the slot. These parameters are used to accurately describe the geometric properties of the feature. Step g: Define parametric features: Define adjustable parameters for each feature of the part (such as holes, slots, bosses, etc.). You can use the parametric tools of the CAD software to create dimensional variables so that the part can be flexibly adjusted; Step h: Geometric modeling: Based on the extracted geometric features, the shape of the parts is described by constructing corresponding mathematical models (such as points, lines, surfaces, curved surfaces, etc.), and the automobile part model is constructed in the modeling software. According to the positional relationship and assembly relationship between the features, the various features are combined to finally generate a complete three-dimensional model of the automobile part; Step i: Correction and optimization: After modeling is completed, the part model needs to be checked and optimized to correct possible geometric inaccuracies or inaccuracies, such as correcting the hole diameter, uneven thickness, etc. Step j: Accuracy check: Check the accuracy of the model to confirm whether it meets the design requirements and whether it can be successfully implemented in the actual manufacturing process; Step k: Storage format selection: Save the completed part model in a standard format (such as STEP, IGES, STL, etc.) to facilitate subsequent use, processing or data exchange with other software.

[0005] Preferably, in the step a, in the actual design process, it is first necessary to obtain the geometric information of the part, which may come from traditional two-dimensional engineering drawings, scanned point cloud data, or existing three-dimensional CAD models.

[0006] Preferably, in step b, the key features of the part include supporting ribs, holes, threads, grooves, chamfers, fillets, etc. of the structural member. In addition to the 3D scanner, the basic outline of the part (such as a rectangle, circle, etc.) can also be drawn using a sketch tool in the CAD software. The sketch should contain all necessary dimensions and constraints.

[0007] Preferably, the feature recognition method in step e is, for example, if a set of point cloud data forms a closed cylindrical shape and the interior thereof is empty, it can be determined as a hole feature.

[0008] Preferably, the denoising process in step d includes but is not limited to using a filtering algorithm to remove abnormal points that deviate from the main data distribution; the denoising process keeps the data clear and accurately reflects the true characteristics of the object surface and shape.

[0009] Preferably, in step h, for complex features, it may be necessary to first perform shape decomposition and feature fusion, use feature-driven modeling methods to gradually build the three-dimensional shape of the part, and set constraints between each feature according to the part design requirements, such as size, position, symmetry, etc., to ensure that the feature combination meets the actual design requirements.

[0010] Preferably, after checking and optimizing the part model in step i, finite element analysis is performed using CAE software to simulate the performance of the part under different working conditions, such as stress, deformation, heat conduction, etc.

[0011] Preferably, in step k, after the completed part model is saved in a standard format, virtual simulation is performed: the function and performance of the part are verified through the simulation module in the CAD system, and the simulation analysis includes stress analysis, thermal analysis, etc. to confirm the reliability of the part in actual use.

[0012] Compared with the prior art, the present invention has the following beneficial effects: The present invention adopts a deep learning algorithm to identify the geometric features of parts by training a neural network model, especially for complex and irregular parts. This can reduce the need for manual operation and improve the speed and accuracy of feature recognition. It introduces simulation and optimization of multiple physical fields in the feature recognition and modeling process. For example, when creating the geometric features of a part, its mechanical properties, thermal expansion, fatigue and other factors can be considered at the same time, and coupling analysis can be performed in the modeling stage, which greatly improves the accuracy and reliability of part design. DETAILED DESCRIPTION

[0013] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0014] A method for modeling automobile parts based on feature recognition comprises the following steps: Step a: Design analysis, analyzing the design requirements of parts, including shape, size, assembly requirements, stress conditions, etc.; Step b: Data collection: Scan the auto parts with a 3D scanner and identify the key features of the parts according to the design requirements. The key feature information of the parts is presented in the form of point cloud data, which contains the geometric shape information of the parts. Step c: Add geometric constraints: Add appropriate constraint relationships (such as parallel, vertical, concentric, etc.) to the geometric elements (such as line segments, arcs, etc.) in the sketch to ensure that the geometric constraint relationships of the sketch are clear; Step d: Data preprocessing, denoising the collected data, because the scanning process may be disturbed by environmental factors and generate noise points. After denoising, data simplification is performed. The point cloud data may be very large. The simplification operation can reduce the amount of data and improve the efficiency of subsequent processing without affecting the expression of the main features of the parts; Step e: Feature recognition, rule-based recognition: Identify features based on pre-defined geometric and topological rules, and use existing annotated part data to train the model, such as using a convolutional neural network (CNN). The trained model can classify and identify features in new part data; Step f: Feature parameter extraction. For the identified features, such as bosses, it is necessary to extract the dimensional parameters such as height, length, and width. If it is a slot feature, it is necessary to extract the depth and width of the slot. These parameters are used to accurately describe the geometric properties of the feature. Step g: Define parametric features: Define adjustable parameters for each feature of the part (such as holes, slots, bosses, etc.). You can use the parametric tools of the CAD software to create dimensional variables so that the part can be flexibly adjusted; Step h: Geometric modeling: Based on the extracted geometric features, the shape of the parts is described by constructing corresponding mathematical models (such as points, lines, surfaces, curved surfaces, etc.), and the automobile part model is constructed in the modeling software. According to the positional relationship and assembly relationship between the features, the various features are combined to finally generate a complete three-dimensional model of the automobile part; Step i: Correction and optimization: After modeling is completed, the part model needs to be checked and optimized to correct possible geometric inaccuracies or inaccuracies, such as correcting the hole diameter, uneven thickness, etc. Step j: Accuracy check: Check the accuracy of the model to confirm whether it meets the design requirements and whether it can be successfully implemented in the actual manufacturing process; Step k: Storage format selection: Save the completed part model in a standard format (such as STEP, IGES, STL, etc.) to facilitate subsequent use, processing or data exchange with other software.

[0015] Example 1 A method for modeling automobile parts based on feature recognition comprises the following steps: Step a: Design analysis, analyzing the design requirements of parts, including shape, size, assembly requirements, stress conditions, etc.; Step b: Data collection: Scan the auto parts with a 3D scanner and identify the key features of the parts according to the design requirements. The key feature information of the parts is presented in the form of point cloud data, which contains the geometric shape information of the parts. Step c: Add geometric constraints: Add appropriate constraint relationships (such as parallel, vertical, concentric, etc.) to the geometric elements (such as line segments, arcs, etc.) in the sketch to ensure that the geometric constraint relationships of the sketch are clear; Step d: Data preprocessing, denoising the collected data, because the scanning process may be disturbed by environmental factors and generate noise points. After denoising, data simplification is performed. The point cloud data may be very large. The simplification operation can reduce the amount of data and improve the efficiency of subsequent processing without affecting the expression of the main features of the parts; Step e: Feature recognition, rule-based recognition: Identify features based on pre-defined geometric and topological rules, and use existing annotated part data to train the model, such as using a convolutional neural network (CNN). The trained model can classify and identify features in new part data; Step f: Feature parameter extraction. For the identified features, such as bosses, it is necessary to extract the dimensional parameters such as height, length, and width. If it is a slot feature, it is necessary to extract the depth and width of the slot. These parameters are used to accurately describe the geometric properties of the feature. Step g: Define parametric features: Define adjustable parameters for each feature of the part (such as holes, slots, bosses, etc.). You can use the parametric tools of the CAD software to create dimensional variables so that the part can be flexibly adjusted; Step h: Geometric modeling: Based on the extracted geometric features, the shape of the parts is described by constructing corresponding mathematical models (such as points, lines, surfaces, curved surfaces, etc.), and the automobile part model is constructed in the modeling software. According to the positional relationship and assembly relationship between the features, the various features are combined to finally generate a complete three-dimensional model of the automobile part; Step i: Correction and optimization: After modeling is completed, the part model needs to be checked and optimized to correct possible geometric inaccuracies or inaccuracies, such as correcting the hole diameter, uneven thickness, etc. Step j: Accuracy check: Check the accuracy of the model to confirm whether it meets the design requirements and whether it can be successfully implemented in the actual manufacturing process; Step k: Storage format selection: Save the completed part model in a standard format (such as STEP, IGES, STL, etc.) to facilitate subsequent use, processing or data exchange with other software.

[0016] Example 2 In Example 1, the following steps are added: Step a: In the actual design process, you first need to obtain the geometric information of the part, which may come from traditional two-dimensional engineering drawings, scanned point cloud data, or existing three-dimensional CAD models.

[0017] A method for modeling automobile parts based on feature recognition comprises the following steps: Step a: Design analysis, analyzing the design requirements of parts, including shape, size, assembly requirements, stress conditions, etc.; Step b: Data collection: Scan the auto parts with a 3D scanner and identify the key features of the parts according to the design requirements. The key feature information of the parts is presented in the form of point cloud data, which contains the geometric shape information of the parts. Step c: Add geometric constraints: Add appropriate constraint relationships (such as parallel, vertical, concentric, etc.) to the geometric elements (such as line segments, arcs, etc.) in the sketch to ensure that the geometric constraint relationships of the sketch are clear; Step d: Data preprocessing, denoising the collected data, because the scanning process may be disturbed by environmental factors and generate noise points. After denoising, data simplification is performed. The point cloud data may be very large. The simplification operation can reduce the amount of data and improve the efficiency of subsequent processing without affecting the expression of the main features of the parts; Step e: Feature recognition, rule-based recognition: Identify features based on pre-defined geometric and topological rules, and use existing annotated part data to train the model, such as using a convolutional neural network (CNN). The trained model can classify and identify features in new part data; Step f: Feature parameter extraction. For the identified features, such as bosses, it is necessary to extract the dimensional parameters such as height, length, and width. If it is a slot feature, it is necessary to extract the depth and width of the slot. These parameters are used to accurately describe the geometric properties of the feature. Step g: Define parametric features: Define adjustable parameters for each feature of the part (such as holes, slots, bosses, etc.). You can use the parametric tools of the CAD software to create dimensional variables so that the part can be flexibly adjusted; Step h: Geometric modeling: Based on the extracted geometric features, the shape of the parts is described by constructing corresponding mathematical models (such as points, lines, surfaces, curved surfaces, etc.), and the automobile part model is constructed in the modeling software. According to the positional relationship and assembly relationship between the features, the various features are combined to finally generate a complete three-dimensional model of the automobile part; Step i: Correction and optimization: After modeling is completed, the part model needs to be checked and optimized to correct possible geometric inaccuracies or inaccuracies, such as correcting the hole diameter, uneven thickness, etc. Step j: Accuracy check: Check the accuracy of the model to confirm whether it meets the design requirements and whether it can be successfully implemented in the actual manufacturing process; Step k: Storage format selection: Save the completed part model in a standard format (such as STEP, IGES, STL, etc.) to facilitate subsequent use, processing or data exchange with other software.

[0018] Example 3 In Example 2, the following steps are added: In step b, the key features of the parts include the supporting ribs, holes, threads, grooves, chamfers, fillets, etc. of the structural parts. In addition to the 3D scanner, you can also use the sketch tool in the CAD software to draw the basic outline of the part (such as rectangle, circle, etc.). The sketch should contain all necessary dimensions and constraints.

[0019] A method for modeling automobile parts based on feature recognition comprises the following steps: Step a: Design analysis, analyzing the design requirements of parts, including shape, size, assembly requirements, stress conditions, etc.; Step b: Data collection: Scan the auto parts with a 3D scanner and identify the key features of the parts according to the design requirements. The key feature information of the parts is presented in the form of point cloud data, which contains the geometric shape information of the parts. Step c: Add geometric constraints: Add appropriate constraint relationships (such as parallel, vertical, concentric, etc.) to the geometric elements (such as line segments, arcs, etc.) in the sketch to ensure that the geometric constraint relationships of the sketch are clear; Step d: Data preprocessing, denoising the collected data, because the scanning process may be disturbed by environmental factors and generate noise points. After denoising, data simplification is performed. The point cloud data may be very large. The simplification operation can reduce the amount of data and improve the efficiency of subsequent processing without affecting the expression of the main features of the parts; Step e: Feature recognition, rule-based recognition: Identify features based on pre-defined geometric and topological rules, and use existing annotated part data to train the model, such as using a convolutional neural network (CNN). The trained model can classify and identify features in new part data; Step f: Feature parameter extraction. For the identified features, such as bosses, it is necessary to extract the dimensional parameters such as height, length, and width. If it is a slot feature, it is necessary to extract the depth and width of the slot. These parameters are used to accurately describe the geometric properties of the feature. Step g: Define parametric features: Define adjustable parameters for each feature of the part (such as holes, slots, bosses, etc.). You can use the parametric tools of the CAD software to create dimensional variables so that the part can be flexibly adjusted; Step h: Geometric modeling: Based on the extracted geometric features, the shape of the parts is described by constructing corresponding mathematical models (such as points, lines, surfaces, curved surfaces, etc.), and the automobile part model is constructed in the modeling software. According to the positional relationship and assembly relationship between the features, the various features are combined to finally generate a complete three-dimensional model of the automobile part; Step i: Correction and optimization: After modeling is completed, the part model needs to be checked and optimized to correct possible geometric inaccuracies or inaccuracies, such as correcting the hole diameter, uneven thickness, etc. Step j: Accuracy check: Check the accuracy of the model to confirm whether it meets the design requirements and whether it can be successfully implemented in the actual manufacturing process; Step k: Storage format selection: Save the completed part model in a standard format (such as STEP, IGES, STL, etc.) to facilitate subsequent use, processing or data exchange with other software.

[0020] Example 4 In Example 3, the following steps are added: The feature recognition method in step e is, for example, if a set of point cloud data forms a closed cylindrical shape and the interior is empty, it can be determined as a hole feature.

[0021] A method for modeling automobile parts based on feature recognition comprises the following steps: Step a: Design analysis, analyzing the design requirements of parts, including shape, size, assembly requirements, stress conditions, etc.; Step b: Data collection: Scan the auto parts with a 3D scanner and identify the key features of the parts according to the design requirements. The key feature information of the parts is presented in the form of point cloud data, which contains the geometric shape information of the parts. Step c: Add geometric constraints: Add appropriate constraint relationships (such as parallel, vertical, concentric, etc.) to the geometric elements (such as line segments, arcs, etc.) in the sketch to ensure that the geometric constraint relationships of the sketch are clear; Step d: Data preprocessing, denoising the collected data, because the scanning process may be disturbed by environmental factors and generate noise points. After denoising, data simplification is performed. The point cloud data may be very large. The simplification operation can reduce the amount of data and improve the efficiency of subsequent processing without affecting the expression of the main features of the parts; Step e: Feature recognition, rule-based recognition: Identify features based on pre-defined geometric and topological rules, and use existing annotated part data to train the model, such as using a convolutional neural network (CNN). The trained model can classify and identify features in new part data; Step f: Feature parameter extraction. For the identified features, such as bosses, it is necessary to extract the dimensional parameters such as height, length, and width. If it is a slot feature, it is necessary to extract the depth and width of the slot. These parameters are used to accurately describe the geometric properties of the feature. Step g: Define parametric features: Define adjustable parameters for each feature of the part (such as holes, slots, bosses, etc.). You can use the parametric tools of the CAD software to create dimensional variables so that the part can be flexibly adjusted; Step h: Geometric modeling: Based on the extracted geometric features, the shape of the parts is described by constructing corresponding mathematical models (such as points, lines, surfaces, curved surfaces, etc.), and the automobile part model is constructed in the modeling software. According to the positional relationship and assembly relationship between the features, the various features are combined to finally generate a complete three-dimensional model of the automobile part; Step i: Correction and optimization: After modeling is completed, the part model needs to be checked and optimized to correct possible geometric inaccuracies or inaccuracies, such as correcting the hole diameter, uneven thickness, etc. Step j: Accuracy check: Check the accuracy of the model to confirm whether it meets the design requirements and whether it can be successfully implemented in the actual manufacturing process; Step k: Storage format selection: Save the completed part model in a standard format (such as STEP, IGES, STL, etc.) to facilitate subsequent use, processing or data exchange with other software.

[0022] Example 5 In Example 4, the following steps are added: The denoising process in step d includes but is not limited to using a filtering algorithm to remove abnormal points that deviate from the main data distribution; the denoising process keeps the data clear and accurately reflects the true characteristics of the object surface and shape.

[0023] A method for modeling automobile parts based on feature recognition comprises the following steps: Step a: Design analysis, analyzing the design requirements of parts, including shape, size, assembly requirements, stress conditions, etc.; Step b: Data collection: Scan the auto parts with a 3D scanner and identify the key features of the parts according to the design requirements. The key feature information of the parts is presented in the form of point cloud data, which contains the geometric shape information of the parts. Step c: Add geometric constraints: Add appropriate constraint relationships (such as parallel, vertical, concentric, etc.) to the geometric elements (such as line segments, arcs, etc.) in the sketch to ensure that the geometric constraint relationships of the sketch are clear; Step d: Data preprocessing, denoising the collected data, because the scanning process may be disturbed by environmental factors and generate noise points. After denoising, data simplification is performed. The point cloud data may be very large. The simplification operation can reduce the amount of data and improve the efficiency of subsequent processing without affecting the expression of the main features of the parts; Step e: Feature recognition, rule-based recognition: Identify features based on pre-defined geometric and topological rules, and use existing annotated part data to train the model, such as using a convolutional neural network (CNN). The trained model can classify and identify features in new part data; Step f: Feature parameter extraction. For the identified features, such as bosses, it is necessary to extract the dimensional parameters such as height, length, and width. If it is a slot feature, it is necessary to extract the depth and width of the slot. These parameters are used to accurately describe the geometric properties of the feature. Step g: Define parametric features: Define adjustable parameters for each feature of the part (such as holes, slots, bosses, etc.). You can use the parametric tools of the CAD software to create dimensional variables so that the part can be flexibly adjusted; Step h: Geometric modeling: Based on the extracted geometric features, the shape of the parts is described by constructing corresponding mathematical models (such as points, lines, surfaces, curved surfaces, etc.), and the automobile part model is constructed in the modeling software. According to the positional relationship and assembly relationship between the features, the various features are combined to finally generate a complete three-dimensional model of the automobile part; Step i: Correction and optimization: After modeling is completed, the part model needs to be checked and optimized to correct possible geometric inaccuracies or inaccuracies, such as correcting the hole diameter, uneven thickness, etc. Step j: Accuracy check: Check the accuracy of the model to confirm whether it meets the design requirements and whether it can be successfully implemented in the actual manufacturing process; Step k: Storage format selection: Save the completed part model in a standard format (such as STEP, IGES, STL, etc.) to facilitate subsequent use, processing or data exchange with other software.

[0024] Example 6 In Example 5, the following steps are added: Step h: For complex features, it may be necessary to first perform shape decomposition and feature fusion, and use feature-driven modeling methods to gradually build the three-dimensional shape of the part. According to the part design requirements, set constraints between each feature, such as size, position, symmetry, etc., to ensure that the feature combination meets the actual design requirements.

[0025] A method for modeling automobile parts based on feature recognition comprises the following steps: Step a: Design analysis, analyzing the design requirements of parts, including shape, size, assembly requirements, stress conditions, etc.; Step b: Data collection: Scan the auto parts with a 3D scanner and identify the key features of the parts according to the design requirements. The key feature information of the parts is presented in the form of point cloud data, which contains the geometric shape information of the parts. Step c: Add geometric constraints: Add appropriate constraint relationships (such as parallel, vertical, concentric, etc.) to the geometric elements (such as line segments, arcs, etc.) in the sketch to ensure that the geometric constraint relationships of the sketch are clear; Step d: Data preprocessing, denoising the collected data, because the scanning process may be disturbed by environmental factors and generate noise points. After denoising, data simplification is performed. The point cloud data may be very large. The simplification operation can reduce the amount of data and improve the efficiency of subsequent processing without affecting the expression of the main features of the parts; Step e: Feature recognition, rule-based recognition: Identify features based on pre-defined geometric and topological rules, and use existing annotated part data to train the model, such as using a convolutional neural network (CNN). The trained model can classify and identify features in new part data; Step f: Feature parameter extraction. For the identified features, such as bosses, it is necessary to extract the dimensional parameters such as height, length, and width. If it is a slot feature, it is necessary to extract the depth and width of the slot. These parameters are used to accurately describe the geometric properties of the feature. Step g: Define parametric features: Define adjustable parameters for each feature of the part (such as holes, slots, bosses, etc.). You can use the parametric tools of the CAD software to create dimensional variables so that the part can be flexibly adjusted; Step h: Geometric modeling: Based on the extracted geometric features, the shape of the parts is described by constructing corresponding mathematical models (such as points, lines, surfaces, curved surfaces, etc.), and the automobile part model is constructed in the modeling software. According to the positional relationship and assembly relationship between the features, the various features are combined to finally generate a complete three-dimensional model of the automobile part; Step i: Correction and optimization: After modeling is completed, the part model needs to be checked and optimized to correct possible geometric inaccuracies or inaccuracies, such as correcting the hole diameter, uneven thickness, etc. Step j: Accuracy check: Check the accuracy of the model to confirm whether it meets the design requirements and whether it can be successfully implemented in the actual manufacturing process; Step k: Storage format selection: Save the completed part model in a standard format (such as STEP, IGES, STL, etc.) to facilitate subsequent use, processing or data exchange with other software.

[0026] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for modeling automobile parts based on feature recognition, characterized in that: The following steps are involved: Step a: Design analysis, analyzing the design requirements of parts, including shape, size, assembly requirements, stress conditions, etc.; Step b: Data collection: Scan the auto parts with a 3D scanner and identify the key features of the parts according to the design requirements. The key feature information of the parts is presented in the form of point cloud data, which contains the geometric shape information of the parts. Step c: Add geometric constraints: Add appropriate constraint relationships (such as parallel, vertical, concentric, etc.) to the geometric elements (such as line segments, arcs, etc.) in the sketch to ensure that the geometric constraint relationships of the sketch are clear; Step d: Data preprocessing, denoising the collected data, because the scanning process may be disturbed by environmental factors and generate noise points. After denoising, data simplification is performed. The point cloud data may be very large. The simplification operation can reduce the amount of data and improve the efficiency of subsequent processing without affecting the expression of the main features of the parts; Step e: Feature recognition, rule-based recognition: Identify features based on pre-defined geometric and topological rules, and use existing annotated part data to train the model, such as using a convolutional neural network (CNN). The trained model can classify and identify features in new part data; Step f: Feature parameter extraction. For the identified features, such as bosses, it is necessary to extract the dimensional parameters such as height, length, and width. If it is a slot feature, it is necessary to extract the depth and width of the slot. These parameters are used to accurately describe the geometric properties of the feature. Step g: Define parametric features: Define adjustable parameters for each feature of the part (such as holes, slots, bosses, etc.). You can use the parametric tools of the CAD software to create dimensional variables so that the part can be flexibly adjusted; Step h: Geometric modeling: Based on the extracted geometric features, the shape of the parts is described by constructing corresponding mathematical models (such as points, lines, surfaces, curved surfaces, etc.), and the automobile part model is constructed in the modeling software. According to the positional relationship and assembly relationship between the features, the various features are combined to finally generate a complete three-dimensional model of the automobile part; Step i: Correction and optimization: After modeling is completed, the part model needs to be checked and optimized to correct possible geometric inaccuracies or inaccuracies, such as correcting the hole diameter, uneven thickness, etc. Step j: Accuracy check: Check the accuracy of the model to confirm whether it meets the design requirements and whether it can be successfully implemented in the actual manufacturing process; Step k: Storage format selection: Save the completed part model in a standard format (such as STEP, IGES, STL, etc.) to facilitate subsequent use, processing or data exchange with other software.

2. The method for modeling automobile parts based on feature recognition according to claim 1, characterized in that: In the step a, in the actual design process, it is first necessary to obtain the geometric information of the part, which may come from traditional two-dimensional engineering drawings, scanned point cloud data, or existing three-dimensional CAD models.

3. The method for modeling automobile parts based on feature recognition according to claim 1, characterized in that: In step b, the key features of the parts include the supporting ribs, holes, threads, grooves, chamfers, fillets, etc. of the structural parts. In addition to the 3D scanner, you can also use the sketch tool in the CAD software to draw the basic outline of the part (such as a rectangle, circle, etc.). The sketch should contain all necessary dimensions and constraints.

4. The method for modeling automobile parts based on feature recognition according to claim 1, characterized in that: The feature recognition method in step e is, for example, if a set of point cloud data forms a closed cylindrical shape and the interior thereof is empty, it can be determined as a hole feature.

5. The method for modeling automobile parts based on feature recognition according to claim 1, characterized in that: The denoising process in step d includes but is not limited to using a filtering algorithm to remove abnormal points that deviate from the main data distribution; De-noising keeps the data clear and accurately reflects the true characteristics of the object's surface and shape.

6. The method for modeling automobile parts based on feature recognition according to claim 1, characterized in that: In step h, for complex features, it may be necessary to first perform shape decomposition and feature fusion, use feature-driven modeling methods to gradually build the three-dimensional shape of the part, and set constraints between various features according to part design requirements, such as size, position, symmetry, etc., to ensure that the feature combination meets the actual design requirements.

7. The method for modeling automobile parts based on feature recognition according to claim 1, characterized in that: After checking and optimizing the part model in step i, finite element analysis is performed using CAE software to simulate the performance of the part under different working conditions, such as stress, deformation, heat conduction, etc.

8. The method for modeling automobile parts based on feature recognition according to claim 1, characterized in that: In step k, the completed part model is saved in a standard format, and then virtual simulation is performed: the function and performance of the part are verified through the simulation module in the CAD system. The simulation analysis includes stress analysis, thermal analysis, etc. to confirm the reliability of the part in actual use.