A feature classification method, device, equipment and storage medium

Abstract

The application provides a feature classification method and device, equipment and a storage medium, relates to the technical field of feature classification, and is used for solving the problem that reliable evaluation indexes cannot be provided for part feature similarity analysis. The method comprises the following steps: performing feature segmentation on a part digital model of a target part to obtain a plurality of to-be-analyzed features; discretizing the plurality of to-be-analyzed features respectively to obtain discrete point clouds corresponding to the plurality of to-be-analyzed features respectively; performing vector calculation on the discrete point clouds corresponding to the plurality of to-be-analyzed features respectively to obtain discrete point cloud feature vectors corresponding to the plurality of to-be-analyzed features respectively; and performing similar feature classification on the plurality of to-be-analyzed features according to the correlation degrees between the discrete point cloud feature vectors corresponding to the plurality of to-be-analyzed features respectively.

Description

Technical Field

[0001] This application relates to the field of feature classification technology, and provides a feature classification method, apparatus, device and storage medium. Background Technology

[0002] Flexible production lines, characterized by automation and high speed, are an important way to increase the production capacity of aerospace structural components. However, the multi-variety, small-batch production mode of aerospace structural components results in small production batches for each product type, a small number of quality data samples, and slow accumulation. Consequently, when using "statistical process control theory" for quality data analysis and management, insufficient sample size can easily lead to a reduced ability to detect anomalies in the production process, increasing the probability of false alarms and making it difficult to meet the quality control requirements of flexible production lines.

[0003] However, although aerospace structural components are produced in small batches, they share many similar characteristics, and flexible production lines have similar processing capabilities for these similar characteristics. Therefore, by statistically analyzing the quality data of similar characteristics, the processing capacity of the flexible production line can be reflected, enabling effective quality control during the processing. Thus, classifying the similar characteristics of parts is of great significance for improving the feedback and support capabilities of the flexible production line's processing.

[0004] Currently, there are many methods for classifying similar parts. These methods often involve extracting and training on key features of the parts to obtain feature data, which is then compared and classified to achieve similarity. Alternatively, mathematical evaluation models are established to assess the similarity between parts based on topological similarity, positional similarity, and geometric similarity. However, these methods all compare similarity at the "part level," without considering the similarity of the parts' processing features themselves, and are therefore unsuitable for evaluating the processing capabilities of flexible production lines for similar features.

[0005] Therefore, how to provide reliable evaluation indicators for the similarity analysis of part features is an urgent problem to be solved. Summary of the Invention

[0006] This application provides a feature classification method, apparatus, and device to solve the problem of not being able to provide reliable evaluation indicators for the feature similarity analysis of parts.

[0007] On the one hand, a feature classification method is provided, the method comprising: Feature segmentation is performed on the digital model of the target part to obtain multiple features to be analyzed; Discretize the multiple features to be analyzed to obtain discrete point clouds corresponding to each of the multiple features to be analyzed. Vector calculations are performed on the discrete point clouds corresponding to each of the multiple features to be analyzed to obtain the discrete point cloud feature vectors corresponding to each of the multiple features to be analyzed. Based on the correlation between the discrete point cloud feature vectors corresponding to the multiple features to be analyzed, the multiple features to be analyzed are classified as similar features.

[0008] Optionally, before performing feature segmentation on the digital model of the target part to obtain multiple features to be analyzed, the method further includes: Based on the digital model of the target part, obtain the basic information of the target part; Then, the digital model of the target part is segmented to obtain multiple features to be analyzed, including: Based on the basic information of the target part, feature segmentation is performed on the digital model of the target part to obtain multiple features to be analyzed.

[0009] Optionally, the step of performing vector calculations on the discrete point clouds corresponding to each of the plurality of features to be analyzed to obtain the discrete point cloud feature vectors corresponding to each of the plurality of features to be analyzed includes: For any discrete point cloud of features to be analyzed, a local coordinate system is established with the centroid of the discrete point cloud as the origin; wherein, the directions of each axis of the local coordinate system are parallel to the directions of each axis of the global coordinate system. Using the origin of the local coordinate system as the starting point and the remaining cloud points in the discrete point cloud as the ending point, vector calculation is performed to obtain multiple point cloud vectors corresponding to any feature to be analyzed; With the origin of the local coordinate system as the center point and a preset distance as the radius, a sphere is drawn, and the spherical space is divided into T subspaces; Calculate the length of the sum of all point cloud vectors within the T subspaces; The length of the sum of the cloud vectors of any feature to be analyzed in the T subspaces is taken as the discrete point cloud feature vector of the feature to be analyzed.

[0010] Optionally, calculating the length of the sum of all point cloud vectors within the T subspaces includes: Determine the subspace where each of the multiple point cloud vectors of any feature to be analyzed resides; For any subspace, calculate the sum of all point cloud vectors within that subspace; Calculate the length of the sum of all point cloud vectors within any given subspace.

[0011] Optionally, the step of classifying the multiple features to be analyzed based on the correlation degree between the discrete point cloud feature vectors corresponding to each of the multiple features to be analyzed includes: For any feature to be analyzed, the discrete point cloud feature vector corresponding to the feature to be analyzed is determined as the parent sequence, and the discrete point cloud feature vectors of the remaining features to be analyzed are determined as the sub-sequences. Calculate the correlation degree between the parent sequence and each child sequence; Determine whether any of the calculated association scores constitute a target association score that is not less than a preset classification threshold; If it is determined that there is a target correlation degree not less than a preset classification threshold, then the feature to be analyzed corresponding to the target correlation degree is determined to be a similar feature to the feature to be analyzed corresponding to the parent sequence.

[0012] Optionally, the step of determining the discrete point cloud feature vector corresponding to any feature to be analyzed as the parent sequence and determining the discrete point cloud feature vectors of the remaining features to be analyzed as the sub-sequences includes: The discrete point cloud feature vectors corresponding to each of the multiple features to be analyzed are normalized to obtain the normalized feature vectors corresponding to each of the multiple features to be analyzed. For any feature to be analyzed, the normalized feature vector corresponding to the feature to be analyzed is determined as the parent sequence, and the normalized feature vectors of the remaining features to be analyzed are determined as the child sequences.

[0013] Optionally, the step of calculating the correlation degree between the parent sequence and each child sequence includes: Calculate the absolute value of the sequence difference between the parent sequence and each child sequence; Determine the absolute values ​​of the minimum and maximum sequence differences among multiple sequence differences; For any subsequence, the correlation degree between the parent sequence and any subsequence is obtained based on the sequence difference between the parent sequence and the subsequence, the absolute value of the minimum sequence difference, and the absolute value of the maximum sequence difference.

[0014] On the one hand, a feature classification device is provided, the device comprising: The feature acquisition unit is used to perform feature segmentation on the digital model of the target part to obtain multiple features to be analyzed. The point cloud acquisition unit is used to discretize the plurality of features to be analyzed to obtain discrete point clouds corresponding to each of the plurality of features to be analyzed. The feature vector acquisition unit is used to perform vector calculations on the discrete point clouds corresponding to each of the plurality of features to be analyzed, and to obtain the feature vectors of the discrete point clouds corresponding to each of the plurality of features to be analyzed. The similarity feature classification unit is used to classify the multiple features to be analyzed into similar features based on the correlation degree between the discrete point cloud feature vectors corresponding to each of the multiple features to be analyzed.

[0015] On one hand, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the methods described above.

[0016] On the one hand, a computer storage medium is provided that stores computer program instructions thereon, which, when executed by a processor, implement any of the methods described above.

[0017] In this embodiment, when performing part feature similarity analysis, firstly, the digital model of the target part can be segmented to obtain multiple features to be analyzed; then, each of the multiple features to be analyzed can be discretized to obtain discrete point clouds corresponding to each feature; furthermore, vector calculations can be performed on the discrete point clouds corresponding to each feature to obtain discrete point cloud feature vectors corresponding to each feature; finally, based on the correlation between the discrete point cloud feature vectors corresponding to each feature, the multiple features to be analyzed can be classified as similar features. Therefore, in this embodiment, since the part feature classification is based on the similarity of the part processing features of the same part itself, rather than comparing the "part-level" similarity features of multiple parts, and since the unrepresentable part feature similarity is transformed into representable point cloud feature vector similarity, it can be used to evaluate the processing capability of flexible production lines for similar features, thereby providing reliable evaluation indicators for part feature similarity analysis. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of an application scenario provided by an embodiment of this application; Figure 2 A schematic flowchart of a feature classification method provided in an embodiment of this application; Figure 3 A schematic diagram of a digital model of a part provided in an embodiment of this application; Figure 4 A schematic diagram illustrating the extraction of part features provided in an embodiment of this application; Figure 5 A schematic diagram of a feature discrete point cloud provided in an embodiment of this application; Figure 6A schematic diagram of a process for calculating discrete point cloud vectors provided in an embodiment of this application; Figure 7 A schematic diagram of subspace partitioning provided in an embodiment of this application; Figure 8 A schematic diagram of a process for calculating the degree of correlation provided in an embodiment of this application; Figure 9 A schematic diagram illustrating the classification of similar features provided in this application embodiment; Figure 10 This is a schematic diagram of a feature classification device provided in an embodiment of this application.

[0020] The diagram is labeled as follows: 10-Feature classification device, 101-Processor, 102-Memory, 103-I / O interface, 104-Database, 70-Feature classification device, 1001-Feature acquisition unit, 1002-Point cloud acquisition unit, 1003-Feature vector acquisition unit, 1004-Similar feature classification unit, 1005-Part information acquisition unit. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. Unless otherwise specified, the embodiments and features in the embodiments of this application can be arbitrarily combined with each other. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than that shown here.

[0022] Flexible production lines, characterized by automation and high speed, are an important way to increase the production capacity of aerospace structural components. However, the multi-variety, small-batch production mode of aerospace structural components results in small production batches for each product type, a small number of quality data samples, and slow accumulation. Consequently, when using "statistical process control theory" for quality data analysis and management, insufficient sample size can easily lead to a reduced ability to detect anomalies in the production process, increasing the probability of false alarms and making it difficult to meet the quality control requirements of flexible production lines.

[0023] However, although aerospace structural components are produced in small batches, they share many similar characteristics, and flexible production lines have similar processing capabilities for these similar characteristics. Therefore, by statistically analyzing the quality data of similar characteristics, the processing capacity of the flexible production line can be reflected, enabling effective quality control during the processing. Thus, classifying the similar characteristics of parts is of great significance for improving the feedback and support capabilities of the flexible production line's processing.

[0024] Currently, there are many methods for classifying similar parts. These methods often involve extracting and training on key features of the parts to obtain feature data, which is then compared and classified to achieve similarity. Alternatively, mathematical evaluation models are established to assess the similarity between parts based on topological similarity, positional similarity, and geometric similarity. However, these methods all compare similarity at the "part level," without considering the similarity of the parts' processing features themselves, and are therefore unsuitable for evaluating the processing capabilities of flexible production lines for similar features.

[0025] Based on this, this application provides a feature classification method. In this method, firstly, the digital model of the target part is segmented to obtain multiple features to be analyzed; then, each of the multiple features to be analyzed is discretized to obtain a discrete point cloud corresponding to each feature; furthermore, vector calculations are performed on the discrete point clouds corresponding to each feature to obtain a discrete point cloud feature vector; finally, the multiple features to be analyzed are classified as similar features based on the correlation between the discrete point cloud feature vectors corresponding to each feature. Therefore, in this application embodiment, since the part feature classification is based on the similarity of the part processing features themselves, rather than comparing the "part-level" similarity features of multiple parts, and since the unrepresentable part feature similarity is transformed into representable point cloud feature vector similarity, it can be used to evaluate the processing capability of flexible production lines for similar features, thereby providing reliable evaluation indicators for part feature similarity analysis.

[0026] like Figure 1 The diagram shown illustrates an application scenario provided by an embodiment of this application. This application scenario may include a feature classification device 10.

[0027] The feature classification device 10 can be used to classify the processing features of the same part, for example, it can be a personal computer (PC), server, or laptop. The feature classification device 10 may include one or more processors 101, memory 102, I / O interfaces 103, and databases 104. Specifically, the processor 101 can be a central processing unit (CPU) or a digital processing unit, etc. The memory 102 can be volatile memory, such as random-access memory (RAM); the memory 102 can also be non-volatile memory, such as read-only memory, flash memory, hard disk drive (HDD), or solid-state drive (SSD); or the memory 102 can be any other medium capable of carrying or storing desired program code in the form of instructions or data structures that can be accessed by a computer, but is not limited thereto. The memory 102 can be a combination of the above-mentioned memories. The memory 102 can store some program instructions of the feature classification method provided in the embodiments of this application. When these program instructions are executed by the processor 101, they can be used to implement the steps of the feature classification device method provided in the embodiments of this application, so as to solve the problem of not being able to provide reliable evaluation indicators for part feature similarity analysis. The database 104 can be used to store data such as part digital models, features to be analyzed, discrete point clouds corresponding to features to be analyzed, discrete point cloud feature vectors corresponding to features to be analyzed, and correlation degrees involved in the solution provided in the embodiments of this application.

[0028] In this embodiment, the feature classification device 10 can acquire the digital model of the target part through the I / O interface 103. Then, the processor 101 of the feature classification device 10 will perform feature classification on the processing features of the same part according to the program instructions of the feature classification method provided in this embodiment in the memory 102, so as to provide a reliable evaluation index for part feature similarity analysis. In addition, the digital model of the part, the feature to be analyzed, the discrete point cloud corresponding to the feature to be analyzed, the discrete point cloud feature vector corresponding to the feature to be analyzed, and the correlation degree can be stored in the database 104.

[0029] Of course, the methods provided in the embodiments of this application are not limited to... Figure 1 The application scenarios shown can also be used in other possible scenarios, and this application embodiment does not impose any limitations. Figure 1The functions that the various devices in the application scenarios shown can achieve will be described in subsequent method embodiments, and will not be elaborated on here. Below, the methods of the embodiments of this application will be described in conjunction with the accompanying drawings.

[0030] like Figure 2 The diagram shown is a flowchart of a feature classification method provided in an embodiment of this application. This method can... Figure 1 The feature classification device 10 is used to perform the process. Specifically, the process flow of this method is described below.

[0031] Step 201: Perform feature segmentation on the digital model of the target part to obtain multiple features to be analyzed.

[0032] In this embodiment of the application, before performing feature segmentation on the digital model of the target part, the digital model of the target part can be read first, such as... Figure 3 The diagram shown is a schematic representation of a digital model of a part provided in an embodiment of this application. Then, based on the digital model of the target part, basic information of the target part can be obtained. Furthermore, based on the basic information of the target part, feature segmentation can be performed on the digital model of the target part to obtain multiple features to be analyzed, such as... Figure 4 The diagram shown is a schematic representation of a method for extracting part features according to an embodiment of this application.

[0033] In this embodiment of the application, the set of machining features of the target part can be represented by the following formula (1), that is, the set of features to be analyzed. : (1) in, For the i-th feature to be analyzed, This represents the maximum number of features to be analyzed.

[0034] like Figure 4 As shown, we can assume M=7, that is, by performing feature segmentation on the digital model of the target part, we can obtain... These are the 7 features to be analyzed.

[0035] Step 202: Discretize the multiple features to be analyzed to obtain the discrete point cloud corresponding to each feature.

[0036] In this embodiment of the application, the set of features to be analyzed of the target part can be... Each feature to be analyzed in the image is discretized to obtain the discrete point cloud corresponding to each feature, such as... Figure 5 The image shown is a schematic diagram of a feature discrete point cloud provided in an embodiment of this application.

[0037] In this embodiment of the application, the set of feature point clouds of the target part to be analyzed can be represented by the following formula (2). : (2) in, Features to be analyzed Discrete point cloud, Discrete point cloud The first in One point, The maximum number of features to be analyzed. This represents the maximum number of discrete point clouds.

[0038] Continuing with the example above, assuming M=7, then the set of feature point clouds to be analyzed... .

[0039] Step 203: Perform vector calculations on the discrete point clouds corresponding to each of the multiple features to be analyzed to obtain the discrete point cloud feature vectors corresponding to each of the multiple features to be analyzed.

[0040] In this embodiment, since the vector calculation process of the discrete point cloud of each feature to be analyzed is the same, the following uses the feature to be analyzed as an example. The process of vector calculation of discrete point clouds will be introduced in detail, for example. Figure 6 The diagram shown is a flowchart illustrating a discrete point cloud vector calculation method provided in an embodiment of this application. The specific process is as follows: Step 601: Target the features to be analyzed Discrete point cloud Discrete point cloud With the center of mass as the origin Establish a local coordinate system .

[0041] In the embodiments of this application, the local coordinate system Each axis can be parallel to the axes of the global coordinate system.

[0042] Step 602: Using the local coordinate system The origin Starting point, discrete point cloud The remaining cloud points Perform vector calculations on the endpoint to obtain the features to be analyzed. Corresponding multiple point cloud vectors .

[0043] In this embodiment of the application, the feature to be analyzed can be represented by the following formula (3). point cloud vector : (3) in, Features to be analyzed The origin of the discrete point cloud.

[0044] Step 603: Using the local coordinate system The origin Center point, preset distance as radius Construct a sphere and divide the spherical space S into: Subspace.

[0045] In this embodiment of the application, the spherical space S can be represented by the following formula (4): (4) in, For the first Individual space, This represents the maximum number of subspaces.

[0046] like Figure 7 The diagram shown illustrates a method for dividing a subspace according to an embodiment of this application. In this embodiment, the spherical space S can be uniformly divided. Assuming T=20, we continue with the above example, focusing on the features to be analyzed. spherical space .

[0047] Step 604: Determine the features to be analyzed one by one. Multiple point cloud vectors Each of them is located in its own subspace.

[0048] In the embodiments of this application, the features to be analyzed can be determined one by one. Each point cloud vector Their respective subspaces, when Falling in Subspace At that time, and marked as .

[0049] Step 605: For subspaces Calculation in subspace All point cloud vectors within and .

[0050] In this embodiment, the subspace can be represented by the following formula (5). All point cloud vectors within and : (5) in, For subspace The maximum number of points within.

[0051] Step 606: Calculate the subspace All point cloud vectors within and The length.

[0052] In this embodiment, the subspace can be represented by the following formula (6). The length of the sum of all point cloud vectors within. : (6) Step 607: Select the features to be analyzed exist The length of the cloud vector sum in each subspace is used as a feature to be analyzed. Discrete point cloud feature vector .

[0053] In this embodiment of the application, the statistical features to be analyzed are... exist The length of the vector sum within each subspace is used as a feature to be analyzed. The discrete point cloud feature vectors. Specifically, the discrete point cloud feature vectors of each feature to be analyzed can be represented by the following formula (7): (7) Step 204: Based on the correlation degree between the discrete point cloud feature vectors corresponding to each of the multiple features to be analyzed, classify the multiple features to be analyzed as similar features.

[0054] In this embodiment, grey relational analysis can be used to classify the features to be analyzed of the target part. Since the correlation degree calculation process is the same for each feature to be analyzed, the following section will focus on the features to be analyzed. This article will take the correlation calculation process as an example to introduce it in detail. Specifically, for example... Figure 8 The diagram shown is a flowchart illustrating a correlation calculation method provided in an embodiment of this application. The specific process is as follows: Step 801: Normalize the discrete point cloud feature vectors corresponding to each of the multiple features to be analyzed to obtain the normalized feature vectors corresponding to each of the multiple features to be analyzed.

[0055] In this embodiment, the discrete point cloud feature vectors corresponding to each feature to be analyzed can be normalized to obtain normalized feature vectors corresponding to multiple features to be analyzed, so as to speed up the classification speed, improve the classification accuracy, eliminate the influence caused by the huge difference between the data quantity and structure, and avoid the occurrence of unreasonable phenomena.

[0056] In this embodiment, the normalized feature vector corresponding to the feature to be analyzed can be represented by the following formula (8). : (8) like Figure 9 The diagram shown is a schematic representation of similar feature classification provided in an embodiment of this application. Continuing with the above example, we assume that the maximum number of features to be analyzed, M, is 7, and the maximum number of subspaces is... =20, which shows that the normalized feature vector corresponding to the feature to be analyzed contains 20 elements, and each element is in [0,1].

[0057] Step 802: Targeting the features to be analyzed The features to be analyzed The corresponding normalized feature vector is determined as the parent sequence, and the normalized feature vectors of the remaining features to be analyzed are determined as the child sequences.

[0058] In practical applications, the m=1th normalized eigenvector can be... Determined as the parent sequence Specifically, the parent sequence can be represented by the following formula (9). : (9) The remaining normalized eigenvectors can be used The subsequences are determined sequentially. Specifically, the subsequence can be represented by the following formula (10). : (10) Step 803: Calculate the correlation between the parent sequence and each child sequence.

[0059] In this embodiment, firstly, the absolute values ​​of the sequence differences between the parent sequence and each child sequence can be calculated. Then, the minimum and maximum absolute values ​​of the sequence differences can be determined. The minimum absolute value of the sequence difference can be represented by the following formula (11). : (11) in, This represents the element corresponding to the k-th subspace in the parent sequence. This represents the element corresponding to the k-th subspace in subsequence i.

[0060] The absolute value of the maximum sequence difference can be expressed by the following formula (12). : (12) Furthermore, for any subsequence, the correlation between the parent sequence and any subsequence can be obtained based on the sequence difference, the absolute value of the minimum sequence difference, and the absolute value of the maximum sequence difference.

[0061] In this embodiment of the application, the correlation between the parent sequence and any child sequence can be represented by the following formula (13). : (13) Among them, the resolution coefficient can be If we set it to 0.3, then, continuing with the example above, we can calculate the parent sequence separately. With the remaining subsequence correlation between as follows:

[0062]

[0063]

[0064]

[0065]

[0066]

[0067] Step 804: Determine whether there is a target correlation degree that is not less than the preset classification threshold among the calculated multiple correlation degrees.

[0068] In this embodiment of the application, a preset classification threshold can be assumed. Continuing with the above example, the parent sequence can be determined. With the remaining subsequence The correlation between them is less than the preset classification threshold. The degree of correlation, that is, All with Dissimilar, that is, None of them Because of their similar characteristics, in practical applications, they can be given separate characteristics. It is divided into 1 category.

[0069] Step 805: If it is determined that there is a target correlation degree not less than the preset classification threshold, then the feature to be analyzed corresponding to the target correlation degree is determined to be a similar feature to the feature to be analyzed corresponding to the parent sequence.

[0070] In practical applications, since it is necessary to classify all the features to be analyzed of the target part, the first feature to be analyzed is... After classification, it is also necessary to... Feature classification is performed sequentially; therefore, the calculation process in steps 802-805 can be used to sequentially classify the features. To perform correlation calculation and feature classification, specifically, the m=2th normalized feature vector can be... Determined as the parent sequence The remaining normalized eigenvectors The subsequences are determined sequentially. Then, by repeating the calculation process of steps 802-805 above, the mother sequence can be calculated respectively. With the remaining subsequence correlation between as follows:

[0071]

[0072]

[0073]

[0074]

[0075] It can be seen that only and A correlation coefficient greater than 0.8 indicates that the correlation between the 3rd and 7th features to be analyzed and the 2nd feature to be analyzed is greater than the preset classification threshold. Therefore, the third and seventh features to be analyzed are similar features to the second feature to be analyzed.

[0076] Since the 3rd and 7th features to be analyzed are already similar to the 2nd feature to be analyzed, we can continue to classify the 4th to 6th features. Specifically, we will classify the m=4th normalized feature vector. Determined as the parent sequence The remaining normalized eigenvectors and The subsequences are determined sequentially. Then, by repeating the calculation process of steps 802-805 above, the mother sequence can be calculated respectively. With the remaining subsequence correlation between as follows:

[0077]

[0078] It can be seen and That is, as can be seen from the correlation, the correlation between the 5th and 6th features to be analyzed and the 4th feature to be analyzed is greater than the preset classification threshold. Therefore, the 5th and 6th features to be analyzed are similar features to the 4th feature to be analyzed.

[0079] At this point, all features to be analyzed have been classified. The classification results show that features 2, 3, and 7 are all planar web features and are similar; therefore, they can be grouped together. Features 4, 5, and 6 are all rib features in the same direction and are similar; therefore, they can also be grouped together. Additionally, feature 1 can be grouped separately.

[0080] In summary, in this embodiment, part feature classification is performed based on the similarity of the processing features of the same part itself, rather than comparing the "part-level" similarity features of multiple parts. Furthermore, the unrepresentable similarity of part features is transformed into representable point cloud feature vector similarity. Therefore, it can be used to evaluate the processing capability of flexible production lines for similar features, further providing reliable evaluation indicators for part feature similarity analysis. In addition, by setting the number of subspaces in the spherical space, the discrimination capability of the similarity of the target part's analyzed features (processing features) can be controlled, effectively balancing feature classification accuracy and computational load, and improving computational efficiency.

[0081] Based on the same inventive concept, embodiments of this application provide a feature classification device 100, such as... Figure 10 As shown, the device includes: The feature acquisition unit 1001 is used to perform feature segmentation on the digital model of the target part to obtain multiple features to be analyzed. The point cloud acquisition unit 1002 is used to discretize multiple features to be analyzed and obtain discrete point clouds corresponding to each feature to be analyzed. The feature vector acquisition unit 1003 is used to perform vector calculations on the discrete point clouds corresponding to multiple features to be analyzed, and to obtain the feature vectors of the discrete point clouds corresponding to the multiple features to be analyzed. The similarity feature classification unit 1004 is used to classify multiple features to be analyzed based on the correlation between the discrete point cloud feature vectors corresponding to each feature.

[0082] Optionally, the device 100 further includes a part information acquisition unit 1005, which is used for: Based on the digital model of the target part, obtain the basic information of the target part; Then, feature segmentation is performed on the digital model of the target part to obtain multiple features to be analyzed, including: Based on the basic information of the target part, feature segmentation is performed on the digital model of the target part to obtain multiple features to be analyzed.

[0083] Optionally, the feature vector acquisition unit 1003 is also used for: For any discrete point cloud of features to be analyzed, a local coordinate system is established with the centroid of the discrete point cloud as the origin; wherein the axes of the local coordinate system are parallel to the axes of the global coordinate system. Vector calculations are performed with the origin of the local coordinate system as the starting point and the remaining cloud points in the discrete point cloud as the ending point to obtain multiple point cloud vectors corresponding to any feature to be analyzed. Construct a sphere with the origin of the local coordinate system as the center point and a preset distance as the radius, and divide the spherical space into T subspaces; Calculate the length of the sum of all point cloud vectors within T subspaces; The length of the sum of the cloud vectors of any feature to be analyzed in T subspaces is taken as the discrete point cloud feature vector of any feature to be analyzed.

[0084] Optionally, the feature vector acquisition unit 1003 is also used for: Determine the subspace where each of the multiple point cloud vectors of any feature to be analyzed resides; For any subspace, calculate the sum of all point cloud vectors within that subspace; Calculate the length of the sum of all point cloud vectors within any subspace.

[0085] Optionally, the similarity feature classification unit 1004 is also used for: For any feature to be analyzed, the discrete point cloud feature vector corresponding to any feature to be analyzed is determined as the parent sequence, and the discrete point cloud feature vectors of the remaining features to be analyzed are determined as the sub-sequences. Calculate the correlation degree between the parent sequence and each child sequence; Determine whether any of the calculated association scores constitute a target association score that is not less than a preset classification threshold; If it is determined that there is a target correlation degree not less than the preset classification threshold, then the feature to be analyzed corresponding to the target correlation degree is determined to be a similar feature to the feature to be analyzed corresponding to the parent sequence.

[0086] Optionally, the similarity feature classification unit 1004 is also used for: The discrete point cloud feature vectors corresponding to each of the multiple features to be analyzed are normalized to obtain the normalized feature vectors corresponding to each of the multiple features to be analyzed. For any feature to be analyzed, the normalized feature vector corresponding to any feature to be analyzed is determined as the parent sequence, and the normalized feature vectors of the remaining features to be analyzed are determined as the child sequences.

[0087] Optionally, the similarity feature classification unit 1004 is also used for: Calculate the absolute value of the sequence difference between the parent sequence and each of the child sequences; Determine the absolute values ​​of the minimum and maximum sequence differences among multiple sequence differences; For any subsequence, the correlation between the parent sequence and any subsequence is obtained based on the sequence difference, the absolute value of the minimum sequence difference, and the absolute value of the maximum sequence difference.

[0088] This device can be used to perform Figures 2-9 The method performed by the feature classification device in the illustrated embodiment can be referenced for the functions that each functional module of the device can achieve. Figures 2-9 The embodiments shown are described in detail below.

[0089] In some possible implementations, various aspects of the methods provided in this application can also be implemented as a program product comprising program code that, when run on a computer device, causes the computer device to perform the steps of the methods according to the various exemplary embodiments of this application described above. For example, the computer device may perform actions such as... Figures 2-9 The method performed by the feature classification device in the illustrated embodiment.

[0090] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks. Alternatively, if the integrated units of the present invention are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

[0091] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0092] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A feature classification method, characterized in that, The method includes: Feature segmentation is performed on the digital model of the target part to obtain multiple features to be analyzed; Discretize the multiple features to be analyzed to obtain discrete point clouds corresponding to each of the multiple features to be analyzed. Vector calculations are performed on the discrete point clouds corresponding to each of the plurality of features to be analyzed to obtain the discrete point cloud feature vectors corresponding to each of the plurality of features to be analyzed. The step of performing vector calculations on the discrete point clouds corresponding to each of the plurality of features to be analyzed to obtain the discrete point cloud feature vectors corresponding to each of the plurality of features to be analyzed includes: for any discrete point cloud of a feature to be analyzed, establishing a local coordinate system with the centroid of the discrete point cloud as the origin; wherein the axes of the local coordinate system are parallel to the axes of the global coordinate system; calculating the vector difference with the origin of the local coordinate system as the starting point and the remaining cloud points in the discrete point cloud as the ending point to obtain multiple point cloud vectors corresponding to any feature to be analyzed; drawing a sphere with the origin of the local coordinate system as the center point and a preset distance as the radius, dividing the spherical space into T subspaces; calculating the length of the sum of all point cloud vectors in each subspace; and using the length of the sum of cloud vectors of any feature to be analyzed in each subspace as the discrete point cloud feature vector of the any feature to be analyzed. Based on the correlation degree between the discrete point cloud feature vectors corresponding to each of the multiple features to be analyzed, the multiple features to be analyzed are classified into similar features. The step of classifying the multiple features to be analyzed based on the correlation degree between the discrete point cloud feature vectors corresponding to each of the multiple features to be analyzed includes: for any feature to be analyzed, determining the discrete point cloud feature vector corresponding to that feature as a parent sequence, and determining the discrete point cloud feature vectors of the remaining features to be analyzed as sub-sequences; calculating the correlation degree between the parent sequence and each sub-sequence; determining whether there is a target correlation degree not less than a preset classification threshold among the calculated correlation degrees; if a target correlation degree not less than the preset classification threshold is determined, then the feature to be analyzed corresponding to the target correlation degree is determined as a similar feature to the feature to be analyzed corresponding to the parent sequence.

2. The method as described in claim 1, characterized in that, Before performing feature segmentation on the digital model of the target part to obtain multiple features to be analyzed, the method further includes: Based on the digital model of the target part, obtain the basic information of the target part; Then, the digital model of the target part is segmented to obtain multiple features to be analyzed, including: Based on the basic information of the target part, feature segmentation is performed on the digital model of the target part to obtain multiple features to be analyzed.

3. The method as described in claim 1, characterized in that, The calculation of the length of the sum of all point cloud vectors within each subspace includes: Determine the subspace where each of the multiple point cloud vectors of any feature to be analyzed resides; For any subspace, calculate the sum of all point cloud vectors within that subspace; Calculate the length of the sum of all point cloud vectors within any given subspace.

4. The method as described in claim 1, characterized in that, For any feature to be analyzed, determining the discrete point cloud feature vector corresponding to that feature as the parent sequence and determining the discrete point cloud feature vectors of the remaining features to be analyzed as sub-sequences includes: The discrete point cloud feature vectors corresponding to each of the multiple features to be analyzed are normalized to obtain the normalized feature vectors corresponding to each of the multiple features to be analyzed. For any feature to be analyzed, the normalized feature vector corresponding to the feature to be analyzed is determined as the parent sequence, and the normalized feature vectors of the remaining features to be analyzed are determined as the child sequences.

5. The method as described in claim 1, characterized in that, The step of calculating the correlation degree between the parent sequence and each child sequence includes: Calculate the absolute value of the sequence difference between the parent sequence and each child sequence; Determine the absolute values ​​of the minimum and maximum sequence differences among multiple sequence differences; For any subsequence, the correlation degree between the parent sequence and any subsequence is obtained based on the sequence difference between the parent sequence and the subsequence, the absolute value of the minimum sequence difference, and the absolute value of the maximum sequence difference.

6. A feature classification device, characterized in that, The device includes: The feature acquisition unit is used to perform feature segmentation on the digital model of the target part to obtain multiple features to be analyzed. The point cloud acquisition unit is used to discretize the plurality of features to be analyzed and obtain discrete point clouds corresponding to each of the plurality of features to be analyzed. A feature vector acquisition unit is used to perform vector calculations on the discrete point clouds corresponding to each of the plurality of features to be analyzed, and to obtain the discrete point cloud feature vectors corresponding to each of the plurality of features to be analyzed. The step of performing vector calculations on the discrete point clouds corresponding to each of the plurality of features to be analyzed to obtain the discrete point cloud feature vectors corresponding to each of the plurality of features to be analyzed includes: for any discrete point cloud of a feature to be analyzed, establishing a local coordinate system with the centroid of the discrete point cloud as the origin; wherein the axes of the local coordinate system are parallel to the axes of the global coordinate system; calculating the vector difference with the origin of the local coordinate system as the starting point and the remaining cloud points in the discrete point cloud as the ending point to obtain multiple point cloud vectors corresponding to any feature to be analyzed; drawing a sphere with the origin of the local coordinate system as the center point and a preset distance as the radius, and dividing the spherical space into T subspaces; calculating the length of the sum of all point cloud vectors in each subspace; and using the length of the sum of cloud vectors of any feature to be analyzed in each subspace as the discrete point cloud feature vector of the any feature to be analyzed. A similarity feature classification unit is used to classify the plurality of features to be analyzed into similar features based on the correlation degree between the discrete point cloud feature vectors corresponding to each of the plurality of features to be analyzed. The step of classifying the plurality of features to be analyzed into similar features based on the correlation degree between the discrete point cloud feature vectors corresponding to each of the plurality of features to be analyzed includes: for any feature to be analyzed, determining the discrete point cloud feature vector corresponding to that feature as a parent sequence, and determining the discrete point cloud feature vectors of the remaining features to be analyzed as sub-sequences; calculating the correlation degree between the parent sequence and each sub-sequence; determining whether there is a target correlation degree not less than a preset classification threshold among the calculated correlation degrees; if a target correlation degree not less than the preset classification threshold is determined, then determining the feature to be analyzed corresponding to the target correlation degree as a similar feature to the feature to be analyzed corresponding to the parent sequence.

7. An electronic device, characterized in that, The device includes: Memory, used to store program instructions; A processor is configured to invoke program instructions stored in the memory and execute the method described in any one of claims 1-5 according to the obtained program instructions.

8. A storage medium, characterized in that, The storage medium stores computer-executable instructions for causing a computer to perform the method described in any one of claims 1-5.

Images