Assembly model retrieval method based on shape and topological information fusion quantitative representation

Abstract

The invention provides an assembly model retrieval method based on shape and topological information fusion quantitative representation, and belongs to the technical field of computer aided design. Comprising the steps of obtaining an assembly model library; uniformly sampling all parts in each assembly body model, and constructing a topological information matrix of the corresponding assembly body model; according to all the topological information matrixes, all part pairs with the topological relation in the corresponding assembly body model are obtained, and the matching face and the matching face centroid of each part pair are calculated in sequence; by utilizing all the uniform sampling results and all the matching surface centroids, respectively acquiring a shape topology distance value set of the corresponding assembly body model, and respectively acquiring a vector descriptor of the corresponding assembly body model; and obtaining a retrieval result of the to-be-retrieved model by using the cosine similarity between the vector descriptor of each assembly model and the vector descriptor of the to-be-retrieved model. According to the method, the retrieval accuracy and retrieval efficiency of the assembly model can be improved.

Description

Technical Field

[0001] The present application relates to the field of computer-aided design technology, and in particular to an assembly model retrieval method based on quantitative representation of shape and topology information fusion. Background Art

[0002] In modern manufacturing, assembly models have become a core component of product design and production, playing a crucial role in improving R&D efficiency, shortening production cycles, and reducing costs. As the number of historical assembly models accumulated by companies continues to grow, developing accurate and efficient assembly model retrieval methods has become a crucial means of increasing assembly model reuse and reducing duplication and resource waste.

[0003] To retrieve an assembly model, it is first necessary to quantitatively represent the information of the assembly model, convert it into a descriptor that can be used for similarity comparison, and then perform retrieval based on the descriptor.

[0004] Currently, there are some existing methods for retrieval based on assembly model information representation, but there are the following problems:

[0005] For example, a Chinese invention patent entitled "Assembly Retrieval Method Based on Spatial Connection Skeleton Descriptor," with application number CN201810362028.8 and publication date October 9, 2018, proposes establishing a spatial connection skeleton based on the center positions of parts in an assembly, the center positions of mating surfaces, and the connection relationship between parts and mating surfaces. Based on the spatial connection skeleton, the spatial distance distribution of all randomly sampled points on the surfaces of two parts passing through the skeleton is statistically calculated as a matrix descriptor for assembly retrieval input. An optimal subsequence matching algorithm is then used to calculate the matching distance between parts. Furthermore, an optimization model is used to calculate the overall matching distance between different assemblies. The overall matching distances between the target assembly and the assembly to be compared are sorted in ascending order as the result of the similarity search. In this retrieval method, the skeleton path distance between each pair of parts is fixed, resulting in the spatial connection relationship not being fully reflected in the matrix descriptor obtained based on the statistical distribution. This can result in retrieval results where the component parts are similar but the topological relationships do not match.

[0006] For example, another Chinese invention patent entitled "A Method for Similarity Retrieval of Assembly Models Based on Spatial Point Set Matching", with application number CN202010362580.4 and publication date of August 14, 2020, proposes to calculate the feature attribute similarity matrix of the selected model, and then multiply it with the attribute weight matrix to obtain the attribute comprehensive similarity matrix; use this matrix and the set threshold to find the valid elements in each column of the matrix, and all valid elements are arranged and combined to form a set of parts that can be matched; the spatial position similarity matrix is ​​obtained by the centroid coordinates of the parts; the attribute comprehensive similarity matrix and the spatial position similarity matrix are calculated according to the formula to obtain a comprehensive similarity matrix; the similarity weights when assembling the reference body and each part are multiplied with the elements of each column in the comprehensive similarity matrix in turn to obtain a model similarity matrix; the elements of the corresponding positions of the matching parts set and the model similarity matrix are summed to obtain the similarity value; the value with the maximum similarity of the matching parts set is the similarity between the model and the reference body. Although this retrieval method comprehensively considers the attribute information of the assembly components, it still regards each part as a separate individual and does not consider the topological relationship between the parts in the assembly model. Therefore, the retrieval results may result in assembly models with similar components but different topological relationships.

[0007] Therefore, it is necessary to propose a solution to improve one or more problems existing in the above-mentioned related technical solutions.

[0008] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of this application, and therefore may include information that does not constitute prior art known to ordinary technicians in this field. Summary of the Invention

[0009] This application provides an assembly model retrieval method based on the fusion of shape and topology information and quantitative representation, the method comprising the following steps:

[0010] Acquire an assembly model library, wherein the assembly model library includes a plurality of assembly models, and each of the assembly models includes a plurality of parts;

[0011] Uniformly sampling all the parts contained in each assembly model, and using all uniform sampling results to construct the topological information matrix of the corresponding assembly model;

[0012] According to all the topological information matrices, all the part pairs with topological relationships in the corresponding assembly model are respectively obtained, and the mating surfaces and the mating surface centroids of each of the part pairs are respectively calculated in sequence;

[0013] Using all the uniform sampling results and all the matching surface centroids, respectively obtain a corresponding shape topology distance value set of the assembly model, and using all the shape topology distance value sets to respectively obtain a corresponding vector descriptor of the assembly model;

[0014] The cosine similarity between the vector descriptor of each assembly model and the vector descriptor of the model to be retrieved is used to obtain the retrieval result of the model to be retrieved.

[0015] Furthermore, the steps of uniformly sampling all the parts contained in each assembly model and constructing the topological information matrix of the corresponding assembly model using all uniform sampling results include:

[0016] uniformly sampling all the parts in each assembly model according to the same sampling interval to obtain a point cloud corresponding to each part;

[0017] constructing an initial topological information matrix of each assembly model respectively;

[0018] In each pair of parts consisting of any two parts of the assembly model, the Euclidean distances between all sampling points contained in the point cloud corresponding to one part and all sampling points contained in the point cloud corresponding to the other part are calculated in sequence, and a set of Euclidean distances of the sampling points of the corresponding pair of parts is obtained;

[0019] If there is a sampling point in the Euclidean distance set of the sampling points of the part pair whose Euclidean distance is smaller than the sampling interval, it is considered that there is a topological relationship between the corresponding two parts;

[0020] All the topological relationships are stored in the corresponding initial topological information matrix respectively to obtain the topological information matrix of each assembly model.

[0021] Furthermore, the step of respectively obtaining all the part pairs having topological relationships in the corresponding assembly model based on all the topological information matrices, and respectively calculating the mating surfaces and the mating surface centroids of each of the part pairs in sequence includes:

[0022] Traversing all the topological information matrices to obtain all the part pairs with topological relationships in the corresponding assembly models;

[0023] Calculating the mating surfaces of each pair of parts respectively;

[0024] According to the mating surfaces of all the part pairs, the corresponding mating surface centroids are calculated.

[0025] Furthermore, the expression of the mating surface of the part pair is:

[0026] (1)

[0027] in, Indicates the Parts and The mating surfaces of the parts, Indicates the The point cloud corresponding to each part, Indicates the The point cloud corresponding to each part, Indicates the The set of sampling points in the point cloud corresponding to each part that satisfies the Euclidean distance of the sampling points is less than the sampling interval, Indicates the The set of sampling points in the point cloud corresponding to each part that satisfies the Euclidean distance of the sampling points is less than the sampling interval, Represents a set of sampling points The sampling points, , Represents a set of sampling points The sampling points, , Represents a set of sampling points The number of all sampling points in , Represents a set of sampling points The number of all sampling points in , Indicates any one, Indicates at least one, represents the sampling interval, Represents a set of sampling points The Sampling points and sampling point sets The The Euclidean distance between the sampling points, Represents a set of sampling points The Sampling points and sampling point sets The The Euclidean distance between the sampling points.

[0028] Furthermore, the expression of the centroid of the mating surface of the part pair is:

[0029] (2)

[0030] in, Indicates the Parts and The centroid of the mating surface of the parts, Indicates the Parts and The centroid of the mating surface of the parts, Represents a set of sampling points The number of all sampling points in , Represents a set of sampling points The number of all sampling points in , Represents a set of sampling points The sampling points, , Represents a set of sampling points The sampling points, .

[0031] Furthermore, the steps of respectively obtaining a set of shape topological distance values ​​of the corresponding assembly model using all the uniform sampling results and all the matching surface centroids, and respectively obtaining a vector descriptor of the corresponding assembly model using all the shape topological distance value sets include:

[0032] Randomly selecting from the point clouds corresponding to all the parts of each assembly model, respectively, to obtain a first random sampling point and a second random sampling point corresponding to each part, and calculating the midpoint of the first random sampling point and the second random sampling point corresponding to each part;

[0033] Calculating the shape topological distance values ​​of the corresponding assembly models respectively using the midpoints of all the sampling points and the centroids of all the mating surfaces;

[0034] Repeat all the above steps to obtain the shape topology distance value set of each assembly model;

[0035] Each of the shape topological distance value sets is represented by an isometric histogram to obtain a corresponding vector descriptor of the assembly model.

[0036] Furthermore, the expression for calculating the shape topological distance value is:

[0037] (3)

[0038] in, Represents the shape topological distance value of the assembly model. Represents the shape distance of the assembly model, represents the topological distance of the assembly model, represents the first random sampling point, represents the second random sampling point, represents the Euclidean distance between the first random sampling point and the second random sampling point, Indicates the Parts and The centroid of the mating surface of the parts, represents the midpoint between the first random sampling point and the second random sampling point, Indicates the centroid of the mating surface and the midpoint of the sampling point The Euclidean distance between , Represents an assembly model The topological information matrix, Represents an assembly model Middle Parts and There is a topological relationship between the parts. Represents an assembly model Middle Parts and There is no topological relationship between the parts.

[0039] Furthermore, the expression of the group interval of the equidistant histogram is:

[0040] (4)

[0041] in, represents the bin interval of the equally spaced histogram, Indicates the maximum value of the shape topology distance of the assembly model. Indicates the minimum value of the shape topology distance value of the assembly model. Represents the number of all intervals in the equally spaced histogram;

[0042] The expression for the height of each interval in the equidistant histogram is:

[0043] (5)

[0044] in, Indicates the first The height of the interval, Indicates the first The statistical frequency of an interval.

[0045] Furthermore, the expression of the vector descriptor of the assembly model is:

[0046] v =[ v 1 , v y ,…, v z ] v y = h y (1≤ y ≤ z ) (6)

[0047] in, A vector descriptor representing the assembly model, The first element in the vector descriptor representing the assembly model dimensional vector, The first element in the vector descriptor representing the assembly model dimensional vector, Indicates the The height of the dimensional vector in the equidistant histogram.

[0048] Furthermore, the step of obtaining the retrieval result of the model to be retrieved by respectively utilizing the cosine similarity between the vector descriptor of each assembly model and the vector descriptor of the model to be retrieved comprises:

[0049] Obtaining a vector descriptor of the model to be retrieved;

[0050] respectively calculating the cosine similarity between the vector descriptor of the model to be retrieved and the vector descriptor of each assembly model;

[0051] Arrange all the cosine similarities in descending order to obtain the retrieval result of the model to be retrieved;

[0052] The expression of the cosine similarity is:

[0053] (7)

[0054] in, represents the cosine similarity between the vector descriptor of the assembly model and the vector descriptor of the model to be retrieved, A vector descriptor representing the assembly model, Represents the vector descriptor of the model to be retrieved, The norm of the vector descriptor representing the assembly model, The norm of the vector descriptor representing the model to be retrieved.

[0055] This application provides an assembly model retrieval method based on the fusion of shape and topology information and quantitative representation, which has at least the following beneficial effects:

[0056] (1) This application uses all uniform sampling results and all matching surface centroids to repeatedly calculate the shape topology distance values ​​of the corresponding assembly model to form a shape topology distance value set for each assembly model, thereby constructing a corresponding vector descriptor and realizing the fusion representation of the shape information and topology information of the assembly model;

[0057] (2) This application obtains the model similarity retrieval result of the model to be retrieved by arranging the cosine similarities between the vector descriptors of all assembly models and the vector descriptors of the model to be retrieved in descending order, thereby comprehensively considering the shape information and topological information of the assembly model, making the retrieval result more accurate, and thus improving the reuse rate of the assembly model. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] The accompanying drawings are incorporated into and constitute a part of the specification, illustrate embodiments consistent with the present application, and together with the specification, are used to explain the principles of the present application. Obviously, the drawings described below are only some embodiments of the present application, and those skilled in the art can derive other drawings based on these drawings without inventive effort.

[0059] Figure 1 A schematic diagram showing the steps of an assembly model retrieval method based on the fusion of shape and topology information and quantitative representation in an exemplary embodiment of the present application is shown;

[0060] Figure 2 A schematic diagram illustrating a flow chart of an assembly model retrieval method based on fusion quantitative representation of shape and topology information in an exemplary embodiment of the present application is shown;

[0061] Figure 3 A schematic diagram illustrating obtaining an assembly model library in an exemplary embodiment of the present application is shown;

[0062] Figure 4 Shows the assembly model in the exemplary embodiment of this application Schematic diagram of;

[0063] Figure 5 Shows the assembly model in the exemplary embodiment of this application A schematic diagram of a collection of parts;

[0064] Figure 6 Shows the assembly model in the exemplary embodiment of this application Schematic diagram of the point clouds corresponding to all parts;

[0065] Figure 7 Shows the assembly model in the exemplary embodiment of this application A part pair consisting of a sixth part and a seventh part having a topological relationship, and a schematic diagram of the mating surfaces of the part pair;

[0066] Figure 8 Shows the assembly model in the exemplary embodiment of this application Schematic diagram of the shape topological distance value of the sub-assembly model composed of the 6th and 7th parts with a topological relationship;

[0067] Figure 9 Shows the assembly model in the exemplary embodiment of this application Schematic diagram of the corresponding equidistant histogram;

[0068] Figure 10 A schematic diagram showing the retrieval results of the model to be retrieved in an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0069] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be embodied in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete and will fully convey the concepts of the example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0070] In addition, the accompanying drawings are merely schematic illustrations of the present application and are not necessarily drawn to scale. Identical reference numerals in the figures denote identical or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the blocks shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0071] The following is a more detailed description of an assembly model retrieval method based on the fusion of shape and topology information and quantitative representation proposed in this exemplary embodiment.

[0072] This example embodiment provides a quantitative characterization method for the fusion of shape and topology information for assembly model retrieval, such as Figure 1 and Figure 2 As shown, the method may include the following steps:

[0073] In this embodiment, step S101: obtaining an assembly model library, Figure 3 FIG. 4 is a schematic diagram showing the assembly model library used in this embodiment.

[0074] Furthermore, the assembly model library includes multiple assembly models, and each assembly model includes multiple parts.

[0075] In this embodiment, the process of obtaining the vector descriptors of all assembly models and the process of obtaining the vector descriptors of the models to be retrieved are obtained according to the method proposed in this application. Therefore, in this embodiment, if Figure 4 As shown, a pneumatic auxiliary unit in the assembly model library, namely the assembly model Take this as an example to illustrate.

[0076] In this embodiment, step S102: uniformly sample all parts contained in each assembly model, and use all uniform sampling results to construct the topological information matrix of the corresponding assembly model. In this embodiment, step S102 may include the following sub-steps:

[0077] Sub-step S1021: In each assembly model, all parts are uniformly sampled according to the same sampling interval to obtain a point cloud corresponding to each part.

[0078] Furthermore, in this embodiment, Figure 5 and Figure 6 As shown, the assembly model is a set consisting of 12 parts: ,in, Represents an assembly model , Indicates the parts, , Indicates the parts, ,and . At the same sampling interval , , for parts The surface of parts Corresponding point cloud , , Representing point clouds The sampling points, , Representing point clouds The number of all sampling points in .

[0079] Figure 5 in arrive Represents the 1st to 12th parts respectively.

[0080] Figure 6 in arrive They represent the point clouds corresponding to the 1st to 12th parts respectively.

[0081] Sub-step S1022: constructing the initial topological information matrix of each assembly model respectively.

[0082] Furthermore, the initial topology information matrix refers to a multi-order zero matrix obtained by initializing the topology information matrix of the assembly model. Initialization topology information matrix is a 12th-order zero matrix.

[0083] Sub-step S1023: Calculate the Euclidean distances of the sampling points between all sampling points contained in the point cloud corresponding to one part and all sampling points contained in the point cloud corresponding to the other part in each part pair consisting of any two parts of each assembly model, and obtain the Euclidean distance sets of the sampling points of the corresponding part pairs.

[0084] Furthermore, in this embodiment, the assembly model Parts in and parts The parts pair composed of The corresponding point cloud is ,Component The corresponding point cloud is . Calculate the point cloud separately Each sampling point and point cloud in The Euclidean distance between each sampling point in the sample point is obtained to obtain the Euclidean distance set of the sampling points of the part pair. , , Representing point clouds Middle Sampling points to point cloud Middle The Euclidean distance of the sampling points, Representing point clouds The number of all sampling points in , Representing point clouds The number of all sampling points in .

[0085] Sub-step S1024: If there is a sampling point in the Euclidean distance set of the sampling points of the part pair whose Euclidean distance is less than the sampling interval, a topological relationship exists between the corresponding two parts.

[0086] Furthermore, in this embodiment, if the assembly model The Euclidean distance set of sampling points Existence , then initialize the topological information matrix The matrix elements in , which means the assembly model Parts in and parts There is a topological relationship between them. On the contrary, if , then initialize the topological information matrix The matrix elements in , which means the assembly model Parts in and parts There is no topological relationship between them.

[0087] Sub-step S1025: storing all topological relationships in the corresponding initial topological information matrix respectively, and obtaining the topological information matrix of each assembly model.

[0088] Furthermore, in this embodiment, the assembly model is obtained by comparing the Euclidean distance of all sampling points with the sampling interval. The topological relationship of all part pairs in the final assembly model is obtained The topological information matrix , as shown in Table 1 below.

[0089] Table 1: Assembly model The topological information matrix

[0090]

[0091] Among them, 0 represents that there is no topological relationship, and 1 represents that there is a topological relationship.

[0092] In this embodiment, step S103: Based on all topological information matrices, all topologically related part pairs in the corresponding assembly model are obtained, and the mating surfaces and mating surface centroids of each part pair are calculated in sequence. In this embodiment, step S103 may include the following sub-steps:

[0093] Sub-step S1031: traverse all topological information matrices to obtain all part pairs with topological relationships in the corresponding assembly model.

[0094] Furthermore, in this embodiment, according to the assembly model The topological information matrix , get the assembly model All the topologically related part pairs in the assembly model Part pair collection , .

[0095] Sub-step S1032: Calculate the mating surfaces of each part pair separately.

[0096] Furthermore, in this embodiment, if Figure 7 As shown, for the parts , calculate parts and parts The mating surface between , Use parts Corresponding point cloud The sampling point set and parts that satisfy the Euclidean distance of the sampling points less than the sampling interval Corresponding point cloud It is represented by the union of the sampling point sets whose Euclidean distance of the sampling points is less than the sampling interval. Figure 7 middle, The value is 6. The value is 7, which means calculating the 6th part and the 7th part The mating surface The expression for calculating the mating surface of the part pair is as follows. The sixth part can be and the 7th part Substitute into the calculation.

[0097] The expression for the mating surface of a part pair is:

[0098] (1)

[0099] in, Indicates the Parts and The mating surfaces of the parts, Indicates the The point cloud corresponding to each part, Indicates the The point cloud corresponding to each part, Indicates the The set of sampling points in the point cloud corresponding to each part that satisfies the Euclidean distance of the sampling points is less than the sampling interval, Indicates the The set of sampling points in the point cloud corresponding to each part that satisfies the Euclidean distance of the sampling points is less than the sampling interval, Represents a set of sampling points The sampling points, , Represents a set of sampling points The sampling points, , Represents a set of sampling points The number of all sampling points in , Represents a set of sampling points The number of all sampling points in , Indicates any one, Indicates at least one, represents the sampling interval, Represents a set of sampling points The Sampling points and sampling point sets The The Euclidean distance between the sampling points, Represents a set of sampling points The Sampling points and sampling point sets The The Euclidean distance between the sampling points.

[0100] Sub-step S1033: Calculate the corresponding mating surface centroids based on the mating surfaces of all part pairs.

[0101] Furthermore, in this embodiment, the calculation of parts and parts The mating surface between The centroid of the mating surface The expression of is as follows, where , , .

[0102] The expression for the centroid of the mating surface of a part pair is:

[0103] (2)

[0104] in, Indicates the Parts and The centroid of the mating surface of the parts, Indicates the Parts and The centroid of the mating surface of the parts, Represents a set of sampling points The number of all sampling points in , Represents a set of sampling points The number of all sampling points in , Represents a set of sampling points The sampling points, , Represents a set of sampling points The sampling points, .

[0105] In this embodiment, step S104: using all uniform sampling results and all matching surface centroids, respectively obtain the corresponding assembly model shape topology distance value sets, and using all shape topology distance value sets to respectively obtain the corresponding assembly model vector descriptors. In this embodiment, step S104 may include the following sub-steps:

[0106] Sub-step S1041: Randomly select from the point clouds corresponding to all parts of each assembly model to obtain the first random sampling point and the second random sampling point corresponding to each part, and calculate the sampling point midpoints of the first random sampling point and the second random sampling point corresponding to each part.

[0107] Furthermore, in this embodiment, the assembly model is traversed Parts , randomly select the first random sampling point from and the second random sampling point , and calculate the midpoint of the two random sampling points .

[0108] Sub-step S1042: using the midpoints of all sampling points and the centroids of all mating surfaces, respectively calculate the shape topological distance values ​​of the corresponding assembly models.

[0109] Furthermore, in this embodiment, if Figure 8 As shown, the expression for calculating the shape topological distance value is:

[0110] (3)

[0111] in, Represents the shape topological distance value of the assembly model. Represents the shape distance of the assembly model, represents the topological distance of the assembly model, represents the first random sampling point, represents the second random sampling point, represents the Euclidean distance between the first random sampling point and the second random sampling point, Indicates the Parts and The centroid of the mating surface of the parts, represents the midpoint between the first random sampling point and the second random sampling point, Indicates the centroid of the mating surface and the midpoint of the sampling point The Euclidean distance between , Represents an assembly model The topological information matrix, Represents an assembly model Middle Parts and There is a topological relationship between the parts. Represents an assembly model Middle Parts and There is a topological relationship between the parts.

[0112] According to expression (3), we can calculate Figure 8 The shape topology distance value of the sub-assembly model composed of the 6th and 7th parts shown in the figure.

[0113] Sub-step S1043: Repeat all the above steps to obtain a set of shape topological distance values ​​for each assembly model.

[0114] Furthermore, in this embodiment, the number of repetitions is Count the shape topology distance values ​​obtained each time to get the assembly model A collection of shape topological distance values ,in, Indicates the The shape topological distance value obtained after repetition is . Figure 2 of middle, .

[0115] Sub-step S1044: Figure 9 As shown, the isometric histogram is used to represent each shape topological distance value set, and the corresponding vector descriptor of the assembly model is obtained.

[0116] Furthermore, the expression of the group interval of the equidistant histogram is:

[0117] (4)

[0118] in, represents the bin interval of the equally spaced histogram, Indicates the maximum value of the shape topology distance of the assembly model. Indicates the minimum value of the shape topology distance value of the assembly model. Represents the number of all bins in an equally spaced histogram.

[0119] Furthermore, the expression for the height of each interval in the equidistant histogram is:

[0120] (5)

[0121] in, Indicates the first The height of the interval, Indicates the first The statistical frequency of an interval.

[0122] Furthermore, the expression of the vector descriptor of the assembly model is:

[0123] v =[ v 1 , v y ,…, v z ] v y = h y (1≤ y ≤ z ) (6)

[0124] in, A vector descriptor representing the assembly model, The first element in the vector descriptor representing the assembly model dimensional vector, The first element in the vector descriptor representing the assembly model dimensional vector, Indicates the The height of the dimension vector in the isometric histogram. Here, the number of all vector dimensions in the vector descriptor of the assembly model is equal to the number of all bins in the isometric histogram, and they correspond one to one.

[0125] Figure 9 Middle, horizontal axis Represents the number of all intervals in the equidistant histogram, the vertical coordinate Indicates the interval height.

[0126] In this embodiment, step S105: using the cosine similarity between the vector descriptor of each assembly model and the vector descriptor of the model to be retrieved, the search results of the model to be retrieved are obtained. In this embodiment, step S1051 may include the following sub-steps:

[0127] Sub-step S1051: Obtain the vector descriptor of the model to be retrieved.

[0128] Furthermore, in this embodiment, the vector descriptor of the model to be retrieved can be obtained by the method mentioned in this embodiment. Sub-step S1052: Calculate the cosine similarity between the vector descriptor of the model to be retrieved and the vector descriptor of each assembly model.

[0129] Furthermore, the expression of cosine similarity is:

[0130] (7)

[0131] in, represents the cosine similarity between the vector descriptor of the assembly model and the vector descriptor of the model to be retrieved, A vector descriptor representing the assembly model, Represents the vector descriptor of the model to be retrieved, The norm of the vector descriptor representing the assembly model, The norm of the vector descriptor representing the model to be retrieved.

[0132] Sub-step S1053: Figure 10 As shown, all cosine similarities are arranged in descending order to obtain the retrieval results of the model to be retrieved.

[0133] Depend on Figure 10 It can be seen that the models to be retrieved are sorted in descending order according to the cosine similarity, and the model to be retrieved is compared with each assembly model in descending order. The first one is the assembly model that has been stored in the assembly model library and is the same as the model to be retrieved. , so the similarity is 1, the cosine similarity of the second place is 0.9614, the cosine similarity of the third place is 0.9484, the cosine similarity of the fourth place is 0.8931, the cosine similarity of the fifth place is 0.8551, and the cosine similarity of the sixth place is 0.8026.

[0134] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of this application, the meaning of "plurality" is two or more, unless otherwise clearly specified.

[0135] In the description of this specification, the reference terms "one embodiment", "some embodiments", "example", "specific example" or "some examples" mean that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification.

[0136] The above is only a specific implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any technician familiar with this technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in this application, and these modifications or replacements should be included in the scope of protection of the present application.

[0137] Those skilled in the art will readily appreciate other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present invention that follow the general principles of this application and include common knowledge or customary techniques in the art that are not disclosed herein.

Claims

1. An assembly model retrieval method based on the fusion of shape and topology information quantitative representation, characterized in that: The method comprises the following steps: Acquire an assembly model library, wherein the assembly model library includes a plurality of assembly models, and each of the assembly models includes a plurality of parts; Uniformly sampling all the parts contained in each assembly model, and using all uniform sampling results to construct the topological information matrix of the corresponding assembly model; According to all the topological information matrices, all the part pairs with topological relationships in the corresponding assembly model are respectively obtained, and the mating surfaces and the mating surface centroids of each of the part pairs are respectively calculated in sequence; Using all the uniform sampling results and all the matching surface centroids, respectively obtain a corresponding shape topology distance value set of the assembly model, and using all the shape topology distance value sets to respectively obtain a corresponding vector descriptor of the assembly model; The cosine similarity between the vector descriptor of each assembly model and the vector descriptor of the model to be retrieved is used to obtain the retrieval result of the model to be retrieved.

2. The assembly model retrieval method based on the fusion and quantitative representation of shape and topology information according to claim 1 is characterized in that: The steps of uniformly sampling all the parts contained in each assembly model and constructing the topological information matrix of the corresponding assembly model using all uniform sampling results include: uniformly sampling all the parts in each assembly model according to the same sampling interval to obtain a point cloud corresponding to each part; constructing an initial topological information matrix of each assembly model respectively; In each pair of parts consisting of any two parts of the assembly model, the Euclidean distances between all sampling points contained in the point cloud corresponding to one part and all sampling points contained in the point cloud corresponding to the other part are calculated in sequence, and a set of Euclidean distances of the sampling points of the corresponding pair of parts is obtained; If there is a sampling point in the Euclidean distance set of the sampling points of the part pair whose Euclidean distance is smaller than the sampling interval, it is considered that there is a topological relationship between the corresponding two parts; All the topological relationships are stored in the corresponding initial topological information matrix respectively to obtain the topological information matrix of each assembly model.

3. The assembly model retrieval method based on the fusion and quantitative representation of shape and topology information according to claim 1 is characterized in that: The step of respectively obtaining all the part pairs having topological relationships in the corresponding assembly model according to all the topological information matrices, and respectively calculating the mating surfaces and the mating surface centroids of each of the part pairs in sequence includes: Traversing all the topological information matrices to obtain all the part pairs with topological relationships in the corresponding assembly models; Calculating the mating surfaces of each pair of parts respectively; According to the mating surfaces of all the part pairs, the corresponding mating surface centroids are calculated.

4. The assembly model retrieval method based on the fusion and quantitative representation of shape and topology information according to claim 3 is characterized in that: The expression of the mating surface of the part pair is: (1) in, Indicates the Parts and The mating surfaces of the parts, Indicates the The point cloud corresponding to each part, Indicates the The point cloud corresponding to each part, Indicates the The set of sampling points in the point cloud corresponding to each part that satisfies the Euclidean distance of the sampling points is less than the sampling interval, Indicates the The set of sampling points in the point cloud corresponding to each part that satisfies the Euclidean distance of the sampling points is less than the sampling interval, Represents a set of sampling points The sampling points, , Represents a set of sampling points The sampling points, , Represents a set of sampling points The number of all sampling points in , Represents a set of sampling points The number of all sampling points in , Indicates any one, Indicates at least one, represents the sampling interval, Represents a set of sampling points The Sampling points and sampling point sets The The Euclidean distance between the sampling points, Represents a set of sampling points The Sampling points and sampling point sets The The Euclidean distance between the sampling points.

5. The assembly model retrieval method based on the fusion and quantitative representation of shape and topology information according to claim 3 is characterized in that: The expression of the centroid of the mating surface of the part pair is: (2) in, Indicates the Parts and The centroid of the mating surface of the parts, Indicates the Parts and The centroid of the mating surface of the parts, Represents a set of sampling points The number of all sampling points in , Represents a set of sampling points The number of all sampling points in , Represents a set of sampling points The sampling points, , Represents a set of sampling points The sampling points, .

6. The assembly model retrieval method based on the fusion and quantitative representation of shape and topology information according to claim 1 is characterized in that: The steps of respectively obtaining a set of shape topological distance values ​​of the corresponding assembly model using all the uniform sampling results and all the matching surface centroids, and respectively obtaining a vector descriptor of the corresponding assembly model using all the shape topological distance value sets include: Randomly selecting from the point clouds corresponding to all the parts of each assembly model, respectively, to obtain a first random sampling point and a second random sampling point corresponding to each part, and respectively calculating the midpoint of the first random sampling point and the second random sampling point corresponding to each part; Calculating the shape topological distance values ​​of the corresponding assembly models respectively using the midpoints of all the sampling points and the centroids of all the mating surfaces; Repeat all the above steps to obtain the shape topology distance value set of each assembly model; Each of the shape topological distance value sets is represented by an isometric histogram to obtain a corresponding vector descriptor of the assembly model.

7. The assembly model retrieval method based on the fusion and quantitative representation of shape and topology information according to claim 6 is characterized in that: The expression for calculating the shape topological distance value is: (3) in, Represents the shape topological distance value of the assembly model. Represents the shape distance of the assembly model, represents the topological distance of the assembly model, represents the first random sampling point, represents the second random sampling point, represents the Euclidean distance between the first random sampling point and the second random sampling point, Indicates the Parts and The centroid of the mating surface of the parts, represents the midpoint between the first random sampling point and the second random sampling point, Indicates the centroid of the mating surface and the midpoint of the sampling point The Euclidean distance between , Represents an assembly model The topological information matrix, Represents an assembly model Middle Parts and There is a topological relationship between the parts. Represents an assembly model Middle Parts and There is no topological relationship between the parts.

8. The assembly model retrieval method based on the fusion and quantitative representation of shape and topology information according to claim 6 is characterized in that: The expression of the group distance of the equidistant histogram is: (4) in, represents the bin interval of the equally spaced histogram, Indicates the maximum value of the shape topology distance of the assembly model. Indicates the minimum value of the shape topology distance value of the assembly model. Represents the number of all intervals in the equally spaced histogram; The expression for the height of each interval in the equidistant histogram is: (5) in, Indicates the first The height of the interval, Indicates the first The statistical frequency of an interval.

9. The assembly model retrieval method based on the fusion and quantitative representation of shape and topology information according to claim 8 is characterized in that: The expression of the vector descriptor of the assembly model is: (6) in, A vector descriptor representing the assembly model, The first element in the vector descriptor representing the assembly model dimensional vector, The first element in the vector descriptor representing the assembly model dimensional vector, Indicates the The height of the dimensional vector in the equidistant histogram.

10. The assembly model retrieval method based on the fusion and quantitative representation of shape and topology information according to claim 1 is characterized in that: The step of obtaining the search result of the model to be searched by respectively using the cosine similarity between the vector descriptor of each assembly model and the vector descriptor of the model to be searched comprises: Obtaining a vector descriptor of the model to be retrieved; respectively calculating the cosine similarity between the vector descriptor of the model to be retrieved and the vector descriptor of each assembly model; Arrange all the cosine similarities in descending order to obtain the retrieval result of the model to be retrieved; The expression of the cosine similarity is: (7) in, represents the cosine similarity between the vector descriptor of the assembly model and the vector descriptor of the model to be retrieved, A vector descriptor representing the assembly model, Represents the vector descriptor of the model to be retrieved, The norm of the vector descriptor representing the assembly model, The norm of the vector descriptor representing the model to be retrieved.

Images