BIM component comparison method and device

Abstract

The invention discloses a BIM component comparison method and device, and belongs to the field of building information models. The comparison method comprises the steps that two BIM components to be compared are converted into a BREP expression, control points in the BREP expression are acquired, and a four-dimensional coordinate system is established; calculating measurement tensors and inertia tensors of the two BIM components according to the four-dimensional coordinates of the control points, and constructing an undirected graph by taking the surfaces as nodes, taking the measurement tensors and the inertia tensors as node attributes, taking common edges in a BREP expression as edges of the graph and taking normal vector included angles of adjacent surfaces as edge attributes; the geometric similarity is obtained by matching the undirected graphs corresponding to the two BIM components; the attribute lists of the two BIM components are sorted according to initials, and the attribute similarity is calculated based on semantic comparison; and calculating a comprehensive similarity according to the geometric similarity and the attribute similarity, and judging whether the two BIM components are similar or not. The geometric similarity and the attribute similarity are comprehensively calculated, the result precision of BIM component comparison is improved, fewer calculation points can be used, and the efficiency is improved.

Description

Technical Field

[0001] The present invention belongs to the field of building information modeling, and more specifically, relates to a BIM component comparison method and device. Background Art

[0002] In the past decade, the emergence of the concept of "Building Information Modeling" (BIM), an information technology in the construction field, has played a great role in improving the level of informatization in my country's construction industry and assisted in the overall optimization of the building information management model. The comparison of BIM components is a relatively basic and practical tool.

[0003] At present, the concept of full life cycle management of Building Information Modeling (BIM) is gradually being promoted, emphasizing "one model to the end", that is, using a unified BIM model at all stages of the project. However, the IFC (Industry Foundation Classes) export specifications between different software have not yet achieved full compatibility, which has led to obstacles in information exchange and model sharing. In addition, BIM models usually contain a large number of duplicate components, and whether geometric information can be reused depends largely on the software used and its export mechanism. Therefore, how to effectively retrieve and identify the same components has become a key challenge for geometric reuse. The solution to this problem not only affects the storage efficiency of the model, but is also directly related to the depth and breadth of BIM application in the entire life cycle of the project.

[0004] In the prior art, there are some BIM component comparisons that obtain positional similarity by aligning the coordinates of the components; by selecting random points in the triangular faces of the components, calculating the distance to the center of the components, obtaining the geometric distribution, and thus calculating the similarity. Such comparisons do not take into account the semantic changes of the components. At the same time, the method of calculating the geometric distribution by selecting random points is more accurate the more points are calculated, but it is very time-consuming. Summary of the invention

[0005] In view of the defects of the related art, the purpose of the present invention is to provide a BIM component comparison method and device, aiming to solve the problems that the existing comparison method does not consider both geometric similarity and attribute similarity at the same time, and that a large number of calculation points are selected when calculating the similarity, resulting in low comparison efficiency and low comparison accuracy.

[0006] To achieve the above objectives, in a first aspect, the present invention provides a BIM component comparison method, comprising:

[0007] S1. Convert the two BIM components to be compared into BREP expressions, obtain the control points therein according to the geometric shapes corresponding to the two BREP expressions, and establish a four-dimensional coordinate system for each control point;

[0008] S2. Calculate the metric tensor and inertia tensor of the two BIM components according to the four-dimensional coordinates of the control points, and construct an undirected graph with the faces as nodes, the metric tensor and inertia tensor as node attributes, the common edges in the BREP expression as the edges of the graph, and the angles between the normal vectors of adjacent faces as edge attributes;

[0009] S3, obtaining geometric similarity by matching the undirected graphs corresponding to the two BIM components;

[0010] S4, sorting the attribute lists of the two BIM components by first letters, and calculating the attribute similarity based on semantic comparison;

[0011] S5. Calculate a comprehensive similarity according to the geometric similarity and the attribute similarity. When the comprehensive similarity is greater than a preset similarity, determine that the two BIM components to be compared are similar.

[0012] Optionally, converting the two BIM components to be compared into BREP expressions includes:

[0013] If the BIM component is a tensile body, the points of the geometric body are obtained by analyzing the tensile body, and the set of points is organized in sequence to generate lines and surfaces in BREP expression;

[0014] If the BIM component is CSG, the shape of the geometry is generated by parsing the geometric primitives and Boolean operations in the CSG representation, and the geometry is converted into BREP representation.

[0015] Optionally, obtaining control points according to the geometric shapes corresponding to the two BREP expressions and establishing a four-dimensional coordinate system for each control point includes:

[0016] The control points are selected from the surface of the geometric shape corresponding to the BREP expression, and the weight coefficient of each control point is obtained by principal component analysis. The four-dimensional coordinate system is established as follows:

[0017] A=(X A ,Y A ,Z A ,ω A )

[0018] Among them, X A , Y A , Z A is the coordinate value of the control point in the natural coordinate system, ω A is the weight parameter of the control point.

[0019] Optionally, the calculating the metric tensor and the inertia tensor of the two BIM components according to the four-dimensional coordinates of the control points includes:

[0020] According to the four-dimensional coordinates of the control points, the vector axis of the edge formed by the control points is calculated:

[0021]

[0022] in, is the four-dimensional coordinate of the ith control point, is the vector axis from the i-th control point to the i+1-th control point, 1≤i≤p;

[0023] Arrange the vector axes of all edges in the geometric shape in order, obtain the vector bundles of different faces, and decompose them:

[0024]

[0025] in, is the vector bundle of surface α, V i x for exist The component in the direction, V i y for exist The component in the direction, V i z for exist The component in the direction, V i ω for exist Directional weight; They correspond to x, y, z and weight parameters respectively;

[0026] The metric tensor is calculated by multiplying the vector bundle of the surface by itself, and the calculation formula is:

[0027]

[0028] in, is the metric tensor of face α;

[0029] The centroid matrices of the two BIM components are calculated and the eigenvalues ​​of the centroid matrices are simplified to obtain the inertia tensors of all faces.

[0030] Optionally, S3 specifically includes:

[0031] S31, determining a matching order in an undirected graph by a breadth-first search algorithm;

[0032] S32, traversing the metric tensors of all faces of the two BIM components in the matching order, and determining two faces with equivalent metric tensors as similar faces;

[0033] S33, traversing the inertia tensors between all faces of the two BIM components in the matching order, and determining two faces with equivalent inertia tensors as faces with the same size;

[0034] S34, determining the intersection of the set of similar faces and the set of faces with the same size as the set of identical faces, and calculating the geometric similarity:

[0035]

[0036] Among them, sim G is the geometric similarity of the components, N Gsim is the number of identical faces, N G The number of faces in the BIM component.

[0037] Optionally, S4 specifically includes:

[0038] The attribute sets of the two BIM components are sorted, and the contents of the synonymous attributes are analyzed by natural language processing, and the synonymous attributes with the same contents are determined as the same attributes, and the attribute similarity is calculated;

[0039] Attribute similarity calculation formula:

[0040]

[0041] Among them, sim A is the attribute similarity of the component, N Asim is the number of identical attributes, N A is the number of component attributes.

[0042] Optionally, the calculation formula of the comprehensive similarity is:

[0043] sim=θ*sim G +(1-θ)*sim A

[0044] Among them, sim is the comprehensive similarity of components, sim G is the geometric similarity of the components, and θ is the weight coefficient of the geometric similarity.

[0045] In a second aspect, the present invention further provides a BIM component comparison device, which is used to perform the comparison method as described in any one of the first aspects, comprising:

[0046] A coordinate system establishment module is used to convert the two BIM components to be compared into BREP expressions, obtain the control points therein according to the geometric shapes corresponding to the two BREP expressions, and establish a four-dimensional coordinate system for each control point;

[0047] The metric tensor and inertia tensor calculation module is used to calculate the metric tensor and inertia tensor of two BIM components according to the four-dimensional coordinates of the control points, and construct an undirected graph with faces as nodes, metric tensors and inertia tensors as node attributes, common edges in BREP expressions as graph edges, and normal vector angles between adjacent faces as edge attributes;

[0048] A geometric similarity calculation module is used to obtain geometric similarity by matching the undirected graphs corresponding to two BIM components;

[0049] The attribute similarity calculation module is used to sort the attribute lists of two BIM components and calculate the attribute similarity based on semantic comparison;

[0050] The similarity judgment module is used to calculate the comprehensive similarity according to the geometric similarity and the attribute similarity. When the comprehensive similarity is greater than a preset similarity, it is determined that the two BIM components to be compared are similar.

[0051] Compared with the prior art, the above technical solution conceived by the present invention can achieve the following beneficial effects:

[0052] 1. A BIM component comparison method provided by the present invention. In the prior art, multiple calculation points are randomly obtained in a geometric entity for similarity calculation. This scheme converts the geometric shapes of two different BIM components into BREP. The BREP expression can carry all geometric shapes, and realizes the comparison of BIM components under parameterization, so that the comparison method of this scheme has universality; the shape similarity of the surface in the BREP expression is determined by calculating the metric tensor. The metric tensor is a matrix expression independent of the coordinate system, which can make the comparison method of this scheme separate from the coordinate system to compare the similarity of geometric shapes; the consistency of the size is determined by calculating the inertia tensor of two similar surfaces, and the geometric consistency of the components is more accurately compared; the control points and coordinate values ​​in the BREP expression are known, and fewer calculation points are used when performing similarity calculation compared with the prior art, which improves the comparison efficiency. By sorting the attribute list and comparing the content of the attribute set, the constructed attribute comparison is realized; the attribute similarity calculation is realized by semantic comparison, and the accuracy of the comparison is further improved. The comprehensive calculation of geometric similarity and attribute similarity makes the BIM component comparison method more comprehensive and the result more accurate.

[0053] 2. BREP expression requires a comparison order. There may be a situation where two graphic surfaces are the same but the topological relationships are inconsistent. The present invention provides a BIM component comparison method that avoids this situation by constructing an undirected graph. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is a flow chart of a BIM component comparison method provided by the present invention;

[0055] Figure 2 It is an undirected graph edge attribute calculation;

[0056] Figure 3 It is a replicated rotation diagram of a single complex geometric model;

[0057] Figure 4 It is a model diagram of a residential building;

[0058] Figure 5 This is a model diagram of a standard floor in a residential building;

[0059] Figure 6 is a retrieval result diagram, where (a) is the retrieval result for standard floor windows, (b) is the retrieval result for standard floor beams, and (c) is the retrieval result for standard floor doors. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0061] The contents involved in the above embodiment are described below in conjunction with a preferred embodiment.

[0062] Embodiment 1

[0063] like Figure 1 As shown, the present invention provides a BIM component comparison method, comprising:

[0064] S1. Convert the two BIM components to be compared into BREP expressions, obtain the control points therein according to the geometric shapes corresponding to the two BREP expressions, and establish a four-dimensional coordinate system for each control point;

[0065] S2. Calculate the metric tensor and inertia tensor of the two BIM components according to the four-dimensional coordinates of the control points, and construct an undirected graph with the faces as nodes, the metric tensor and inertia tensor as node attributes, the common edges in the BREP expression as the edges of the graph, and the angles between the normal vectors of adjacent faces as edge attributes;

[0066] S3, obtaining geometric similarity by matching the undirected graphs corresponding to the two BIM components;

[0067] S4, sorting the attribute lists of the two BIM components by first letters, and calculating the attribute similarity based on semantic comparison;

[0068] S5. Calculate a comprehensive similarity according to the geometric similarity and the attribute similarity. When the comprehensive similarity is greater than a preset similarity, determine that the two BIM components to be compared are similar.

[0069] The geometric expressions of the two BIM components to be compared are converted into BREP format, and the geometrical comparison and geometrical similarity are calculated through BREP. The attribute comparison and attribute similarity are realized through semantic comparison, and the geometrical similarity and attribute similarity of the BIM components are comprehensively calculated to accurately judge whether the BIM components are similar. Based on the parametric comparison, the scope of BIM geometric comparison is expanded, making the BIM model comparison faster and the comparison results more accurate.

[0070] The three most commonly used geometric expressions of BIM components include extrusion, CSG and BREP, among which BREP can express almost all geometric expressions; therefore, specifically, converting the two BIM components to be compared into BREP expressions includes:

[0071] If the BIM component is a stretch body, the stretch body representation of the BIM component is composed of the stretch bottom surface, stretch direction and stretch length. Convert the stretch body to BREP expression: obtain the stretch bottom surface point set of the geometric body by parsing the stretch body, and calculate the stretch top surface point set according to the stretch direction and stretch length. Generate the lines and surfaces in the BREP expression by organizing the collection of bottom and top surface points in order.

[0072] If the BIM component is CSG, the CSG expression is relatively complex and is obtained by intersecting several geometric bodies. Convert CSG to BREP expression: By parsing the geometric primitives and Boolean operations in the CSG expression, the shape of the geometric body is generated, and then the geometric body is converted to BREP expression.

[0073] After the geometry is converted, it is stored in the IFC file as BREP representation and can be used later to avoid multiple conversions.

[0074] Optionally, obtaining control points according to the geometric shapes corresponding to the two BREP expressions and establishing a four-dimensional coordinate system for each control point includes:

[0075] The control points are selected from the surface of the geometric shape corresponding to the BREP expression, and the weight coefficient of each control point is obtained by principal component analysis. The four-dimensional coordinate system is established as follows:

[0076] A=(X A ,Y A ,Z A ,ω A )

[0077] Among them, X A , Y A , ZA is the coordinate value of the control point in the natural coordinate system, ω A is the weight parameter of the control point.

[0078] The component under BREP expression is composed of a surface composed of points arranged in sequence. Take a certain surface as an example: select the surface in the BREP expression, obtain the control points from it, and establish a four-dimensional coordinate system. The BREP expression contains the natural coordinates of the control points.

[0079] Optionally, the calculating the metric tensor and the inertia tensor of the two BIM components according to the four-dimensional coordinates of the control points includes:

[0080] According to the four-dimensional coordinates of the control points, the vector axis of the edge formed by the control points is calculated:

[0081]

[0082] in, is the four-dimensional coordinate of the ith control point, is the vector axis from the i-th control point to the i+1-th control point, 1≤i≤p;

[0083] Arrange the vector axes of all edges in the geometric shape in order, obtain the vector bundles of different faces, and decompose them:

[0084]

[0085] in, is the vector bundle of face α, V i x for exist The component in the direction, V i y for exist The component in the direction, V i z for exist The component in the direction, V i ω for exist Directional weight; They correspond to x, y, z and weight parameters respectively, which are used to transform vector groups into matrices. For example, They are all unit vectors, corresponding to (1, 0, 0, 0), (0, 1, 0, 0), (0, 0, 1, 0), (0, 0, 0, 1) respectively.

[0086] The metric tensor is calculated by multiplying the vector bundle of the surface by itself, and the calculation formula is:

[0087]

[0088] in, is the metric tensor of face α;

[0089] The centroid matrices of the two BIM components are calculated and the eigenvalues ​​of the centroid matrices are simplified to obtain the inertia tensors of all faces.

[0090] The centroid matrix calculation formula is as follows:

[0091]

[0092] in, is the centroid matrix of face α, m i is the mass distribution of the i-th control point. For a surface with uniform mass distribution, the value is 1.

[0093] The eigenvalues ​​of the simplified centroid matrix are obtained:

[0094]

[0095] in, is the inertia tensor of surface α, is the centroid matrix The characteristic value of .

[0096] After calculating the metric tensor and inertia tensor, the face is taken as the node, the metric tensor and inertia tensor are taken as the node attributes, the common edge in the BREP expression is the edge of the graph, and the normal vector angle of adjacent faces is the edge attribute (refer to Figure 2 ), construct an undirected graph.

[0097] Optionally, S3 specifically includes:

[0098] S31, determining a matching order in an undirected graph by a breadth-first search algorithm;

[0099] S32, traversing the metric tensors of all faces of the two BIM components in the matching order, and determining two faces with equivalent metric tensors as similar faces;

[0100] S33, traversing the inertia tensors between all faces of the two BIM components in the matching order, and determining two faces with equivalent inertia tensors as faces with the same size;

[0101] All faces of the two BIM components are traversed in matching order, and the metric tensors of the two faces are compared to see if they are equivalent. The two faces with equivalent metric tensors are determined to be similar faces. The consistency of the size is determined by the inertia tensor, and the inertia tensors between the two faces are compared to determine that the two faces with equivalent inertia tensors are faces with the same size.

[0102] S34. Determine the intersection of the set of similar faces and the set of faces with the same size as the set of same faces.

[0103] When the metric tensor and inertia tensor of two faces are equivalent, we can determine that the two corresponding faces are the same, that is, we take the intersection of similar faces and faces of the same size.

[0104] Calculate geometric similarity:

[0105]

[0106] Among them, sim G is the geometric similarity of the components, N Gsim is the number of identical faces, N G is the number of component faces.

[0107] Optionally, S4 specifically includes:

[0108] The attribute sets of the two BIM components are sorted, and the contents of the synonymous attributes are analyzed by natural language processing, and the synonymous attributes with the same contents are determined as the same attributes, and the attribute similarity is calculated;

[0109] Attribute similarity calculation formula:

[0110]

[0111] Among them, sim A is the attribute similarity of the component, N Asim is the number of identical attributes, N A is the number of component attributes.

[0112] In this solution, natural language processing is used to analyze synonymous attributes and judge some synonymous attribute expressions as identical. When the efficiency requirement is low and the accuracy requirement is high, natural language processing in deep learning can be used to realize the recognition of synonymous semantics, and the accuracy can be improved under the condition of reducing efficiency. Furthermore, if the user does not have high requirements for the comparison accuracy of BIM components, a simplified attribute comparison method can be used. For example, after sorting the attribute set, Boolean judgment is performed to divide the same attributes, and the attribute similarity is obtained by the proportion of the same attributes.

[0113] Optionally, the calculation formula of the comprehensive similarity is:

[0114] sim=θ*sim G +(1-θ)*sim A

[0115] Among them, sim is the comprehensive similarity of components, sim Gis the geometric similarity of the components, θ is the weight coefficient of the geometric similarity; the weight coefficient is set according to the focus of the user. If the user pays more attention to the geometric similarity between BIM components, the weight coefficient will be increased; if the user pays more attention to the attribute similarity between BIM components, the weight coefficient will be reduced.

[0116] Furthermore, after calculating the comprehensive similarity of BIM components, it can be used for component retrieval or comparison of the overall BIM model. The similarity can be set according to the user's needs. If you need to retrieve the same BIM components, set the preset similarity to 100%; if you need to retrieve BIM components with the same geometry, increase the geometric similarity weight coefficient and reduce the semantics (applied to geometric reuse); if you want to retrieve components with the same function, you can increase the semantic similarity weight coefficient (for example, retrieve fire-related components to assist in drawing review).

[0117] The embodiment of the present invention converts the geometric shape of the BIM component to be compared into BREP, and determines the shape similarity of the surface in the BREP expression by calculating the metric tensor. The metric tensor is a matrix expression independent of the coordinate system, so that the comparison method can compare the similarity of geometric shapes without the coordinate system; the consistency of the size is determined by calculating the inertia tensor of two similar surfaces, so as to more accurately compare whether the components are geometrically consistent; the constructed attribute comparison is achieved by sorting the attribute list and comparing the contents of the attribute set; the geometric similarity and attribute similarity are calculated comprehensively to make the BIM component comparison method more comprehensive. It solves the technical problems that the existing comparison method does not consider the geometric similarity and attribute similarity at the same time, and that a large number of calculation points are selected when calculating the similarity, resulting in low comparison efficiency. Based on parameterized comparison, a more universal and efficient BIM component comparison method is achieved.

[0118] In order to verify the effectiveness of the BIM geometric component comparison method based on geometric tensor and graph matching proposed in this paper, several experiments were designed and conducted to compare the geometric similarity and attribute similarity of different types of BIM geometric components.

[0119] In a specific embodiment, Experiment 1 is a comparison verification of a single complex geometric model. This experiment performs geometric comparison on a single complex geometric component. A building window containing a large number of geometric details is selected as the experimental object. The window has 392 faces and 758 edges. Figure 3 As shown in the figure, the window is copied and rotated twice, and the IFC file is exported. Then, the BREP expression is generated and the geometric tensor is calculated. The geometric similarity is evaluated and the geometry is reused using the graph matching algorithm. The purpose of this experiment is to verify the accuracy and robustness of the method proposed in this paper when dealing with complex geometric components.

[0120] After geometrically comparing and reusing the building windows in Experiment 1, it can be seen that the geometric comparison method based on geometric tensors and graph matching in this solution can accurately identify the same geometric components without misjudgment when dealing with complex geometric components. It can ignore the displacement, rotation and other transformations of geometric components and only focus on the essential characteristics of the geometric shape, so it can maintain a constant high accuracy when the geometric complexity increases.

[0121] In the case of geometry reuse, the specific data shows that without geometry reuse, the IFC file size of Experiment 1 is 188KB. After geometry reuse, the file size is reduced to 66KB, a reduction of about 65%. This shows that the geometry reuse technology proposed in this paper can significantly reduce storage requirements, especially in complex geometry components. Geometry reuse reduces redundant geometry data and optimizes file storage and loading efficiency.

[0122] In another specific embodiment, Experiment 2 is a comparison and reuse of geometric models of standard floors, top floors, and bottom floors. The experiment selected a 24-story residential building model such as Figure 4 As shown in Figure 1, the model has a total of 24433 geometric components; its standard layer has a total of 1232 geometric components of different shapes and complexities. Figure 5 The experiment conducts geometric comparison and reuse analysis on all geometric models of standard layers, checks the number of repeated components in a single standard layer, and geometrically reuses the repeated components to reduce their data redundancy.

[0123] In Experiment 2, a geometric comparison and reuse analysis was conducted on the geometric model of a standard floor of a 24-story residential building. The standard floor was artificially set with several identical components. The experimental results show that the proposed method can retrieve 100% of the repeated components and reuse the components geometrically. The results are shown in Tables 1 and Figure 6 As shown in Figure 2, the storage space requirement is greatly reduced through geometric reuse technology. Specifically, the data from Experiment 2 shows that without geometric reuse, the file size is 15.1MB, and after reuse, the file size is reduced to 12.3MB, a reduction of about 18.5%.

[0124] Table 1. Results of retrieval and reuse of repeated components at artificially set standard layers

[0125]

[0126] In Experiment 1, it took 41 seconds to retrieve and reuse the same type of window with 392 faces and 758 edges. In Experiment 2, it took 4721 seconds to retrieve and reuse all geometric components in the standard layer.

[0127] Experiments 1 and 2 jointly verified the effectiveness of the BIM component comparison method proposed in this scheme in processing multi-standard-storey building projects. Especially in large-scale BIM models, geometric reuse can greatly reduce the storage of redundant data and improve the processing efficiency of the overall model.

[0128] Embodiment 2

[0129] The present invention further provides a BIM component comparison device, which is used to perform the comparison method as described in any one of the embodiments, including:

[0130] A coordinate system establishment module is used to convert the two BIM components to be compared into BREP expressions, obtain the control points therein according to the geometric shapes corresponding to the two BREP expressions, and establish a four-dimensional coordinate system for each control point;

[0131] The metric tensor and inertia tensor calculation module is used to calculate the metric tensor and inertia tensor of two BIM components according to the four-dimensional coordinates of the control points, and construct an undirected graph with faces as nodes, metric tensors and inertia tensors as node attributes, common edges in BREP expressions as graph edges, and approximate angles between adjacent faces as edge attributes;

[0132] A geometric similarity calculation module is used to obtain geometric similarity by matching the undirected graphs corresponding to two BIM components;

[0133] The attribute similarity calculation module is used to sort the attribute lists of two BIM components and calculate the attribute similarity based on semantic comparison;

[0134] The similarity judgment module is used to calculate the comprehensive similarity according to the geometric similarity and the attribute similarity. When the comprehensive similarity is greater than a preset similarity, it is determined that the two BIM components to be compared are similar.

[0135] A BIM component comparison device provided in an embodiment of the present invention is used to execute a BIM component comparison method provided in any first embodiment of the present invention, and has corresponding functional modules and beneficial effects.

[0136] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A BIM component comparison method, characterized in that: include: S1. Convert the two BIM components to be compared into BREP expressions, obtain the control points therein according to the geometric shapes corresponding to the two BREP expressions, and establish a four-dimensional coordinate system for each control point; S2. Calculate the metric tensor and inertia tensor of the two BIM components according to the four-dimensional coordinates of the control points, and construct an undirected graph with the faces as nodes, the metric tensor and inertia tensor as node attributes, the common edges in the BREP expression as the edges of the graph, and the angles between the normal vectors of adjacent faces as edge attributes; S3, obtaining geometric similarity by matching the undirected graphs corresponding to the two BIM components; S4, sorting the attribute lists of the two BIM components by first letters, and calculating the attribute similarity based on semantic comparison; S5. Calculate a comprehensive similarity according to the geometric similarity and the attribute similarity. When the comprehensive similarity is greater than a preset similarity, determine that the two BIM components to be compared are similar.

2. The comparison method according to claim 1, characterized in that: The step of converting the two BIM components to be compared into BREP expressions includes: If the BIM component is a tensile body, the points of the geometric body are obtained by analyzing the tensile body, and the set of points is organized in sequence to generate lines and surfaces in BREP expression; If the BIM component is CSG, the shape of the geometry is generated by parsing the geometric primitives and Boolean operations in the CSG representation, and the geometry is converted into BREP representation.

3. The comparison method according to claim 1, characterized in that: The step of obtaining the control points according to the geometric shapes corresponding to the two BREP expressions and establishing a four-dimensional coordinate system for each control point includes: The control points are selected from the surface of the geometric shape corresponding to the BREP expression, and the weight coefficient of each control point is obtained by principal component analysis. The four-dimensional coordinate system is established as follows: A=(X A ,Y A ,Z A ,ω A ) Among them, X A , Y A , Z A is the coordinate value of the control point in the natural coordinate system, ω A is the weight parameter of the control point.

4. The comparison method according to claim 1, characterized in that: The method of calculating the metric tensor and the inertia tensor of the two BIM components according to the four-dimensional coordinates of the control points includes: According to the four-dimensional coordinates of the control points, the vector axis of the edge formed by the control points is calculated: in, is the four-dimensional coordinate of the i-th control point, is the vector axis from the i-th control point to the i+1-th control point, 1≤i≤p; Arrange the vector axes of all edges in the geometric shape in order, obtain the vector bundles of different faces, and decompose them: in, is the vector bundle of surface α, V i x for exist The component in the direction, V i y for exist The component in the direction, V i z for exist The component in the direction, V i ω for exist Directional weight; They correspond to x, y, z and weight parameters respectively; The metric tensor is calculated by multiplying the vector bundle of the surface by itself, and the calculation formula is: in, is the metric tensor of face α; The centroid matrices of the two BIM components are calculated and the eigenvalues ​​of the centroid matrices are simplified to obtain the inertia tensors of all faces.

5. The comparison method according to claim 4, characterized in that: S3 specifically includes: S31, determining a matching order in an undirected graph by a breadth-first search algorithm; S32, traversing the metric tensors of all faces of the two BIM components in the matching order, and determining two faces with equivalent metric tensors as similar faces; S33, traversing the inertia tensors between all faces of the two BIM components in the matching order, and determining two faces with equivalent inertia tensors as faces with the same size; S34, determining the intersection of the set of similar faces and the set of faces with the same size as the set of identical faces, and calculating the geometric similarity: Among them, sim G is the geometric similarity of the components, N Gsim is the number of identical faces, N G The number of faces in the BIM component.

6. The method according to claim 5, characterized in that S4 specifically includes: The attribute sets of the two BIM components are sorted, and the contents of the synonymous attributes are analyzed by natural language processing, and the synonymous attributes with the same contents are determined as the same attributes, and the attribute similarity is calculated; Attribute similarity calculation formula: Among them, sim A is the attribute similarity of the component, N Asim is the number of identical attributes, N A is the number of component attributes.

7. The method according to claim 6, characterized in that The calculation formula of the comprehensive similarity is: yes=θ*yes G +(1-θ)*yes A Among them, sim is the comprehensive similarity of components, sim G is the geometric similarity of the components, and θ is the weight coefficient of the geometric similarity.

8. A BIM component comparison device, used to perform the comparison method according to any one of claims 1 to 7, characterized in that: include: A coordinate system establishment module is used to convert the two BIM components to be compared into BREP expressions, obtain the control points therein according to the geometric shapes corresponding to the two BREP expressions, and establish a four-dimensional coordinate system for each control point; The metric tensor and inertia tensor calculation module is used to calculate the metric tensor and inertia tensor of two BIM components according to the four-dimensional coordinates of the control points, and construct an undirected graph with faces as nodes, metric tensors and inertia tensors as node attributes, common edges in BREP expressions as graph edges, and normal vector angles between adjacent faces as edge attributes; A geometric similarity calculation module is used to obtain geometric similarity by matching the undirected graphs corresponding to two BIM components; The attribute similarity calculation module is used to sort the attribute lists of two BIM components and calculate the attribute similarity based on semantic comparison; The similarity judgment module is used to calculate the comprehensive similarity according to the geometric similarity and the attribute similarity. When the comprehensive similarity is greater than a preset similarity, it is determined that the two BIM components to be compared are similar.

Images