Adaptive merging method and system based on multiple monomer BIM models

Abstract

Embodiments of the present application disclose a method and system for adaptive merging of multiple monomer BIM models. The method projects monomer BIM models from three dimensions to two dimensions to obtain a two-dimensional model graph on a two-dimensional plane, and extracts feature intersections and feature edges from the two-dimensional model graph. Based on this, a model feature vector matrix is formed by combining the vertices and edges of the monomer BIM models with the feature intersections and feature edges corresponding to the two-dimensional model graph, and the model feature values are calculated to obtain a model feature value set of the multiple monomer BIM models. The model feature value set is subjected to cluster analysis to obtain a target model for fitting and merging, and the merging is performed to realize lightweight of the model. Moreover, the monomer BIM models are automatically processed in the whole process without relying on manual operation of the designer, thereby improving the merging efficiency of the model.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of BIM, in particular to a self-adaptive merging method and system based on multiple single BIM models. BACKGROUND

[0002] As the core technology in the field of digital construction, Building Information Modeling (BIM) integrates three-dimensional geometric data and non-geometric attribute information to build a digital foundation for the whole life cycle management of buildings. The BIM model scene of current large-scale construction projects contains more than one million components, and the capacity of a single BIM model file can reach GB level. For example, high-rise buildings, subway stations of rail transit, etc. The single BIM model composed of the structure of floors / walls / columns, mechanical and electrical equipment, and the material of decoration generates hundreds of millions of data nodes, covering geometric coordinates, material properties, and other multi-dimensional information, which is a very large model data.

[0003] Therefore, lightweight processing of single BIM models is a key technology in the field. Traditional model lightweight schemes mainly use geometric simplification, data compression, and deletion of unnecessary attributes, which reduces the data volume but often sacrifices model accuracy and analysis functions, making it difficult to meet user application requirements for single BIM models.

[0004] Therefore, how to efficiently merge models while preserving key information and achieve lightweight processing of multiple single BIM models has become one of the key technical bottlenecks in the BIM field. SUMMARY

[0005] To solve the above technical problems of lightweight processing of single BIM models, embodiments of the present application provide a self-adaptive merging method and system based on multiple single BIM models.

[0006] According to an aspect of an embodiment of the present application, a self-adaptive merging method based on multiple single BIM models is disclosed, which comprises:

[0007] Step 1: Set up a spatial coordinate system of a BIM model, import a single BIM model, and project the single BIM model from three dimensions to two dimensions to a two-dimensional plane corresponding to the spatial coordinate system, and obtain a two-dimensional model graph;

[0008] Step 2: Obtain feature intersection points in the two-dimensional model graph and feature edges connected to the feature intersection points; the feature intersection points are intersection points formed by projecting the spatial line structure of the single BIM model to the two-dimensional plane;

[0009] Step 3: According to the vertices and edges in the monomer BIM model, and the feature intersection and feature edge corresponding to the two-dimensional model diagram, a model feature vector matrix is generated according to the order of the two-dimensional model diagram obtained by projecting the monomer BIM model;

[0010] Step 4: Calculate the model feature value according to the model feature vector matrix;

[0011] Step 5: Perform clustering analysis according to the model feature value to obtain a fitted target model, and merge the target model.

[0012] According to an aspect of an embodiment of the present application, a self-adaptive merging system based on multiple monomer BIM models is disclosed, comprising:

[0013] A two-dimensional model diagram acquisition module sets a spatial coordinate system of a BIM model, imports a monomer BIM model, and projects the monomer BIM model from three dimensions to two dimensions to a two-dimensional plane corresponding to the spatial coordinate system, and obtains a two-dimensional model diagram;

[0014] A feature acquisition module acquires feature intersections in the two-dimensional model diagram, and feature edges connected to the feature intersections; the feature intersections are intersection points formed by projecting spatial line structures of the monomer BIM model to the two-dimensional plane;

[0015] A model feature vector matrix generation module generates a model feature vector matrix according to the vertices and edges in the monomer BIM model, and the feature intersections and feature edges corresponding to the two-dimensional model diagram, according to the order of the two-dimensional model diagram obtained by projecting the monomer BIM model;

[0016] A model feature value calculation module calculates a model feature value according to the model feature vector matrix;

[0017] A model merging module performs clustering analysis on the model feature value to obtain a fitted target model, and merges the target model.

[0018] The self-adaptive merging method and system of the multiple monomer BIM models are used to obtain vertices, edges, feature intersections, and feature edges to form a model feature vector matrix, and calculate a model feature value through dimension reduction processing of the monomer BIM model. Based on a model feature value set of multiple monomer BIM models, clustering analysis is performed to obtain a fitted and merged target model and perform merging. The present scheme does not need to delete or simplify any key information such as structure data and attribute data, and efficiently merges monomer BIM models to achieve the lightweight goal of the model.

[0019] Further, the spatial structure of the BIM model is abstracted by features, and a 3x4 model feature vector matrix composed of three-dimensional feature vectors is used for representation, without complex mesh level calculation, and the processing efficiency is higher. Moreover, the model feature vector matrix only needs dozens of bytes, compared with the original model data MB or G level, and the processing speed of the combined monomer BIM model is more convenient and efficient. BRIEF DESCRIPTION OF DRAWINGS

[0020] The accompanying drawings, which are incorporated herein and form a part of the specification, illustrate embodiments consistent with the present application and, together with the description, further serve to explain the principles of the application. It is to be expressly understood, however, that the drawings are only for the purpose of illustration and are not intended as a definition of the limits of the application. In the drawings:

[0021] Figure 1 is a flow chart of an adaptive merging method of multiple monomer BIM models according to an exemplary embodiment.

[0022] Figure 2 is a three-dimensional space diagram of a certain library BIM model scene according to an exemplary embodiment.

[0023] Figure 3 is a projection of a chair in an XY two-dimensional plane according to an exemplary embodiment.

[0024] Figure 4 is Figure 3 a two-dimensional model diagram of a chair in a front view projection and feature intersection and feature edge.

[0025] Figure 5 is Figure 3 a two-dimensional model diagram of a chair in a top view projection and feature intersection and feature edge.

[0026] Figure 6 is Figure 3 a two-dimensional model diagram of a chair in a side view projection and feature intersection and feature edge.

[0027] Figure 7 is a partial three-dimensional diagram of a certain library BIM model scene according to an exemplary embodiment.

[0028] Figure 8 is Figure 7 a schematic diagram of the BIM model structure of a chair and a table in the XZ plane direction projection.

[0029] Figure 9 is a schematic diagram of a certain library three-dimensional model source file according to an exemplary embodiment.

[0030] Figure 10is a comparative diagram of a model after model merging of a certain library 3D model according to an example embodiment.

[0031] Figure 11 is a flow chart of a model feature vector matrix corresponding row weight processing method according to an example embodiment. DETAILED DESCRIPTION

[0032] The example embodiments will be described in detail herein with reference to the accompanying drawings. In the following description, unless otherwise indicated, like numbers in the different drawings represent similar or identical elements. The following description of example embodiments is not representative of all possible embodiments consistent with the present application. Rather, it is merely an example of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0033] The flow charts shown in the drawings are merely illustrative and do not necessarily include all content and operations / steps, nor are they necessarily executed in the order described. For example, some operations / steps can be further broken down, while some operations / steps can be integrated or partially integrated, so the actual execution order can be changed according to actual conditions.

[0034] It should be noted that “multiple” as mentioned in the present application refers to two or more. The association relationship of “and / or” describes the association of the associated objects, which means that there can be three relationships, for example, A and / or B can represent the three cases of A alone, A and B together, and B alone. The character “ / ” generally represents an “or” relationship between the associated objects before and after it.

[0035] It should be noted that in the specific embodiments of the present application, when the embodiments of the present application are applied to specific products or technologies, the user’s permission or consent needs to be obtained, and the collection, use and processing of related data need to comply with relevant laws, regulations and standards of relevant countries and regions. At the same time, the formulas involved in the embodiments of the present application can be flexibly adjusted, for example, by increasing or reducing corresponding parameters.

[0036] Building Information Modeling (BIM) as one of the core technologies of spatial digitalization, converts the model of the physical world into a digital model, and constructs a large number of single BIM models. The format of the model can be IFC, Revit, OBJ, STL or private format. These single BIM models are expressed by three-dimensional geometric data and non-geometric attribute information.

[0037] In three-dimensional geometric data, such as points, lines, surfaces, and bodies; among them, points represent position information, lines represent paths, boundaries or connection relationships, surfaces represent the surface or range of objects, and bodies describe the volume and internal structure of objects. On this basis, topological relationships are also included, which describe the connection, adjacency, and containment of geometric objects, such as whether two lines intersect and whether a point is inside a polygon. Size and constraint describe the length, width, and height of the object, as well as the size constraints between geometric elements.

[0038] In non-geometric attribute information, the characteristics, properties, status, and classification of objects are described, such as ID / coding, name, type, and status.

[0039] In the face of multiple single BIM models, if the three-dimensional geometric data of each single BIM model is stored separately, the amount of data that needs to be stored is very large. In some specific scenarios, such as different versions of BIM scenarios, different engineers have more or less differences in drawing the same model, resulting in "repeated" storage of BIM models with the same essence, making the data volume of the BIM scenario huge, which is a BIM model that needs to be combined. In other scenarios, such as some BIM models that have partial differences but express the same type of geometric model, the structure data is basically the same, only the brand is different, or the color is different, etc. In scenarios where "precision" requirements are not high, such as BIM visualization large screens, they are also combined and processed geometric model objects.

[0040] Based on this, combined with Figure 1 The flowchart of the adaptive combination method for multiple single BIM models of the present scheme, the method comprises:

[0041] Step 1: Set the spatial coordinate system of the BIM model, import the single BIM model, and project the single BIM model from three-dimensional to two-dimensional to the two-dimensional plane corresponding to the spatial coordinate system, and obtain a two-dimensional model graph.

[0042] Specifically, combined with Figure 2 A three-dimensional schematic diagram of a BIM model scene for a certain library is shown. In the BIM model scene, a large number of BIM models of various types are expressed based on the spatial coordinate system. In the BIM system, the models in the BIM model scene are constructed and managed, usually relying on the spatial coordinate system. Whether importing existing BIM models or creating or editing BIM models on the modeling platform, it is based on a unified three-dimensional spatial coordinate system as a reference framework. In the BIM model scene in the present scheme, a polar coordinate system is used. Of course, according to different technical needs, different coordinate systems can be used, such as cylindrical coordinates, spherical coordinates, etc.

[0043] Combined with Figure 3, the projection of the chair in the embodiment in XY two-dimensional plane. The spatial geometry BIM model in the three-dimensional coordinate system contains information related to the model, such as geometry, attributes, and other information, and simulates the monomer BIM model of the entity in the physical world through digital simulation. Based on this, the position, orientation and size of the monomer BIM model are defined in the three-dimensional coordinate system, and the monomer BIM model is projected to a specific two-dimensional plane (such as projection in XY, YZ, XZ plane), to obtain the two-dimensional model graph of the monomer BIM model in the two-dimensional plane. The two-dimensional model graph contains the geometric contour information of the monomer BIM model projected in the two-dimensional plane, such as the number of geometric points, lines and surfaces, and other two-dimensional plane feature information, so as to realize the simplified expression of the monomer BIM model information.

[0044] Step 2: Obtain the feature intersection in the two-dimensional model graph, and the feature edge connected with the feature intersection; the feature intersection is the intersection point formed by the projection of the spatial line structure of the monomer BIM model to the two-dimensional plane.

[0045] Specifically, in combination with Figures 4-6 The two-dimensional model graphs obtained by projecting the chair in the embodiment in three directions, as well as the feature intersections and feature edges (since the number of feature points / feature edges is large, only part of the feature points / feature edges are marked and used for illustration for clear expression). In the BIM model scene, the spatial line structures that do not intersect in the three-dimensional space of the monomer BIM model will generate visual intersections in the projected two-dimensional plane when projected to the two-dimensional plane. These visual intersections are not real connection points of the entity geometry, but are focal points formed by the projection of the spatial non-intersecting lines on the two-dimensional plane, which are introduced "image intersections", that is, feature intersections. Correspondingly, the feature edges connected with the feature intersections. Through the geometric edge relationship of the monomer BIM model in space, a new feature data set is formed on the two-dimensional plane, which is used to reflect the spatial geometric complexity of the monomer BIM model, both dimension reduction conversion and difficulty reduction of the geometric spatial monomer BIM model, and convenient for feature data acquisition and statistics, providing support for subsequent monomer BIM model merging.

[0046] Further, the feature intersection includes the direct feature intersection formed by the projection of the vertex of the monomer BIM model to the two-dimensional model graph, and the projection feature intersection formed by the projection of the edge of the monomer BIM model to the two-dimensional model graph.

[0047] Specifically, for the spatial geometric complexity of different monomer BIM models, the projection feature intersection is the projection feature intersection generated by projecting the edge of the monomer BIM model to the two-dimensional model graph, which is used to reflect the spatial geometric complexity and is a feature data with high value. On the other hand, the direct feature intersection is the inherent feature of the monomer BIM model, such as vertex, edge and the like. In a specific case, the inherent feature of the monomer BIM model is projected to the two-dimensional model graph, and the vertex coincides with a point or an edge, so that the value of the direct feature intersection is weaker than that of the projection feature intersection, and the corresponding is that the ability of the direct feature intersection to reflect the spatial geometric complexity of the monomer BIM model is weaker.

[0048] The intersection ratio of the projection feature intersection and the feature intersection is calculated, and the reciprocal of the intersection ratio is the number weight of the projection feature intersection.

[0049] Specifically, since the projection feature intersection and the direct feature intersection have different abilities to express the spatial geometric complexity of the monomer BIM model, in order to highlight the contribution of the projection feature intersection to the spatial geometric complexity, the projection feature intersection is given a weight. Based on this, the feature intersection is equal to the sum of the direct feature intersection and the projection feature intersection, the proportion of the projection feature intersection in the total number of intersections is calculated, and the corresponding reciprocal is the weight, that is, the weight given to the projection feature intersection, which is used to highlight the spatial geometric complexity of the monomer BIM model.

[0050] The number of feature intersections is updated, which is the product of the number weight and the projection feature intersection, and then the direct feature intersection is added.

[0051] Specifically, based on the above direct feature intersection and the projection feature intersection with a given weight, the number of new feature intersections is obtained, the dimension reduction of the complex geometric space model is converted into a two-dimensional model graph, the weight of the projection feature intersection is highlighted through differentiation, which is convenient for differentiated statistics of feature data and provides support for subsequent monomer BIM model merging.

[0052] Step 3: According to the vertices and edges in the monomer BIM model, and the feature intersections and feature edges corresponding to the two-dimensional model graph, a model feature vector matrix is generated according to the order of the two-dimensional model graph projected by the monomer BIM model.

[0053] Specifically, the monomer BIM model is projected from three directions (such as projection on XY, YZ, and XZ planes) to obtain corresponding two-dimensional planes, and features in each projected two-dimensional model diagram are extracted, which are vertices, edges, feature intersection points, and feature edges in turn. The data of vertices and edges can be directly extracted from the attribute information of the monomer BIM model, and the data of feature intersection points and feature edges are obtained through the above step 2. It should be noted that the order composed of vertices, edges, feature intersection points, and feature edges can be set autonomously. The order of the two-dimensional model diagram obtained by projecting the monomer BIM model is consistent as a whole, that is, the model feature vector matrix obtained by each monomer BIM model is consistent, and the model "feature vector matrix" is thus formed for similarity analysis.

[0054] In combination with the application example of the "feature vector matrix", the monomer BIM model is projected towards XY, YZ, and XZ planes to obtain corresponding vertices, edges, feature intersection points, and feature edges in turn, for example: XY plane direction projection: vertex number V1=18, edge number E1=26, feature intersection point number I1=32, and feature edge number S1=43; YZ plane direction projection: vertex number V2=12, edge number E2=18, feature intersection point number I2=39, and feature edge number S2=51; and XZ plane direction projection: vertex number V3=20, edge number E3=28, feature intersection point number I3=29, and feature edge number S3=38. That is, the following feature vector data is obtained, as follows:

[0055] Projection plane Number of vertices (V) Number of edges (E) Number of characteristic intersections (I) Number of characteristic edges (S) XY (1) 18 26 32 43 YZ (2) 12 18 39 51 XZ (3) 20 28 29 38

[0056] Based on the obtained feature data, a unified 3x4 model feature vector matrix is arranged, that is:

[0057]

[0058] Based on this expression, the feature vectors of XY, YZ, and XZ planes are obtained, and the model feature vector matrix is composed of the feature vectors. As can be seen, the feature vectors are feature data of points and edges, that is, the expression of the model feature vector matrix for the monomer BIM model is not limited to the size of the model, and the focus is on the spatial structure of the model.

[0059] In an embodiment, if the spatial structures of different monomer BIM models are consistent, and only the sizes are inconsistent, that is, the spatial proportions are different, the edges of the monomer BIM model can be selected, and the lengths of the edges are recorded. At the same time, the lengths of the target model edges to be fitted are obtained, the scaling ratio is obtained by comparing the lengths, and the size of the original model in space can be restored in the model restoration process.

[0060] Step 4: Calculate the model feature value according to the model feature vector matrix.

[0061] The obtained model feature vector matrix, i.e. the number of vertices, edges, feature intersection points and feature edges obtained by projecting in the XY, YZ and XZ three plane directions, is used to form the corresponding feature vectors, and the corresponding matrix is formed by the feature vectors. The corresponding numerical value "model feature value" is obtained by mathematical conversion and calculation of the matrix, which is used to represent a single BIM model. That is, the complex three-dimensional model is abstracted into a measurable structure "digital signature" for structural analysis, merging and BIM model lightweight processing.

[0062] Further, the model feature value method obtained by numerical calculation of the feature vector matrix specifically includes:

[0063] Step 41: The model feature vector matrix is an MxN matrix, M is three planes formed by the spatial coordinate system, and N is four columns of vertices, edges, feature intersection points and feature edges;

[0064] Step 42: Convert the model feature vector matrix into a 1x(MxN) matrix;

[0065] Step 43: Calculate the value of the 1x(MxN) matrix as the model feature value.

[0066] The specific application example is described as follows:

[0067] The model feature vector matrix f of MxN (i.e. 3x4) is

[0068] The 1x12 matrix F is converted to [18, 26, 32, 43, 12, 18, 39, 51, 20, 28, 29, 38]

[0069] The model feature value of the matrix, i.e. the Euclidean norm (i.e. the vector module or length) of the matrix, is calculated as ||F|| = 109.23 (the last two decimal places are saved). ≈109.23 (the last two decimal places are saved), which is used to represent the single BIM model.

[0070] Step 5: Perform clustering analysis according to the model feature value to obtain a fitted target model and merge the target model.

[0071] Specifically, the model characteristic values of the imported monomer BIM model are calculated, that is, the monomer BIM model of all targets in the BIM model space is calculated according to the above steps respectively, and the corresponding model characteristic values are obtained, thereby forming a set of model characteristic values. According to the set of model characteristic values, clustering analysis is performed, such as k-means, k-means++, bi-kmeans, kernelk-means, and the like, thereby obtaining the target model to be combined, combining the target model to further reduce the number of models, and realizing the lightweight of multiple monomer BIM models.

[0072] By using the adaptive merging method of the above multiple monomer BIM models, a two-dimensional model graph on a two-dimensional plane is obtained by projecting the monomer BIM model from three dimensions to two dimensions, and a feature intersection is extracted from the two-dimensional model graph. Based on this, the vertex and edge in the monomer BIM model are combined with the feature intersection and feature edge corresponding to the two-dimensional model graph to form a model feature vector matrix, and the model characteristic values are calculated, and a set of model characteristic values of multiple monomer BIM models is obtained. The target model to be combined is obtained by performing clustering analysis on the set of model characteristic values and performing merging. As can be seen, the present scheme does not need to delete or simplify any key information such as structure data and attribute data, and realizes efficient merging of monomer BIM models, and achieves the lightweight goal of the model.

[0073] On this basis, the dimensionality reduction simplification processing of each monomer BIM model from three dimensions to two dimensions is performed, the feature abstraction of the spatial structure of the BIM model is performed, and a 3x4 model feature vector matrix composed of only three-dimensional feature vectors is used for representation, without complex mesh level calculation, and the processing efficiency is higher. Moreover, the model feature vector matrix only needs dozens of bytes, compared with the original model data MB or G level, the processing speed of the monomer BIM model merging is more convenient and efficient.

[0074] Each monomer BIM model is represented by a model characteristic value (i.e. vector norm mapping), and is represented by a "point on a number line" through model visualization. For more complex monomer BIM models, the obtained model characteristic value is larger, and it is farther away from the origin on the number line, which provides convenience for subsequent clustering analysis of the model. As can be seen, the present scheme abstracts the complex three-dimensional model into a measurable structure "digital signature", which is used for classification and merging of monomer BIM models, and serves the lightweight processing of monomer BIM models, and has high accuracy and efficiency.

[0075] The conversion of the monomer BIM model to the vector distance is simple and intuitive. In addition, by setting the threshold range, the redundancy can be set, that is, the range of lightening the monomer BIM model is improved, and the BIM model with high similarity can also be merged, which not only reduces the number of monomer BIM models, but also reduces the number of model objects, so that the BIM model space is more lightweight. Compared with the high-cost mesh comparison of the original model, the scheme has high speed and efficiency.

[0076] By using the method, the monomer BIM model is automatically processed in the whole process without relying on manual operation of the designer, and the merging efficiency of the model is improved.

[0077] The technical logic of the scheme is illustrated by combining specific cases and Figures 7-8 The technical logic of the scheme is illustrated by combining specific cases and Figure 7 A partial perspective view in a certain BIM model scene, Figure 8 The BIM model structure of the chair and the table in the partial perspective view is projected in the XZ plane direction. Figure 7 The BIM model structure of the chair and the table in the partial perspective view is projected in the XZ plane direction. According to the scheme, the monomer BIM model in the BIM model scene, such as a chair, a table, etc., is projected in a two-dimensional plane to obtain a two-dimensional model graph, and then the corresponding feature intersection and feature edge are extracted, and a model feature vector matrix is formed with the vertex and edge of the monomer BIM model, and the model feature value is calculated. Based on this, the target model is obtained according to the model feature value, and the model is merged to realize lightening.

[0078] Further, the technical effect of the scheme is illustrated by combining Figures 9-10 Further, the technical effect of the scheme is illustrated by combining Figure 9 A schematic view of a certain library three-dimensional model source file is shown, and a large number of monomer BIM models of chairs and tables are set in the three-dimensional BIM model scene. Figure 10 A comparison schematic view of a certain library three-dimensional model after merging a plurality of monomer BIM models is shown, and a large number of monomer BIM models of chairs and tables are merged, the number of monomer BIM models is greatly reduced, the size of the BIM model scene is about 55% of the size of the source file, and the BIM model scene shown has almost no change. As can be seen, the scheme has outstanding and significant advantages for large three-dimensional BIM scenes and multiple monomer BIM models.

[0079] In an embodiment, the adaptive merging method based on a plurality of monomer BIM models further comprises:

[0080] Step 11: obtaining attribute information of the monomer BIM model;

[0081] Step 12: determining the type of the monomer BIM model according to the type information in the attribute information;

[0082] Step 13: Projecting the monomer BIM model of the same type from three dimensions to the two-dimensional plane corresponding to the spatial coordinate system, and obtaining a two-dimensional model graph.

[0083] Specifically, for the imported monomer BIM model, the format includes IFC, Revit or custom structured model format, and contains three-dimensional geometric data and non-geometric attribute information. According to the type information obtained from the attribute information of the monomer BIM model, such as extracting the model type information from the field in the attribute information, the information records the type of the model, such as table, chair, door and window, etc. The attribute information of the monomer BIM model is pre-classified, and on this basis, the corresponding two-dimensional model graph is obtained by further projecting the monomer BIM model from three dimensions to two dimensions, which provides support for subsequent model feature recognition. Therefore, the same or similar monomer BIM model is queried in the same type of model, and the efficiency of model recognition and merging is further improved.

[0084] In an embodiment, in combination with Figure 11 the flowchart, the adaptive merging method based on multiple monomer BIM models, step 3 further includes:

[0085] determining whether the minimum bounding box of the monomer BIM model intersects with the minimum bounding box of any BIM model;

[0086] if yes, obtaining the intersecting face of the monomer BIM model projected into the corresponding two-dimensional plane, and calculating the projection area of the intersecting face in the corresponding two-dimensional plane;

[0087] calculating the area ratio of the projection area of the intersecting face to the area of the monomer BIM model in the corresponding two-dimensional plane, and the reciprocal of the area ratio is the weight of the corresponding row of the model feature vector matrix.

[0088] Specifically, for the space model, there are a large number of monomer BIM models, some of which are mutually overlapped, and the overlapping part is also a key "feature", that is, introducing projection weight can enhance the expression ability of "spatial distribution semantics", which is used to identify whether the models belong to the mergable models.

[0089] Based on this, for the mutually overlapped monomer BIM models, the overlapping part is calculated in the manner of AABB minimum bounding box, and the projection area of the overlapping part in a certain direction is calculated, the ratio of the overlapping area to the projection area of the current monomer BIM model in the certain direction is calculated, and the reciprocal of the area ratio is taken as the weight of the feature vector in the direction. The meaning of the BIM model space expressed thereby is: the higher the overlap rate between the mutually overlapped monomer BIM models, which can be understood as the "spatial coverage" of the BIM model in a certain direction, the lower the corresponding weight, and the more similar the two models.

[0090] To further illustrate the scheme, for a model A and model B AABB has spatial overlap, then in the projection plane XY, the weight of the algorithm as follows:

[0091]

[0092] wherein, : the projection area of the overlapping area of the AABB of model A and model B in the XY plane; : the projection area of model A in the XY plane; : the weight of the XY row in the model feature vector matrix.

[0093] Similarly, the corresponding weight of the YZ and XZ rows in the model feature vector matrix can also be obtained.

[0094] In other embodiments, based on the above method steps, if no, the corresponding two-dimensional plane in the corresponding row of the model feature vector matrix is 1.

[0095] Specifically, if there is no overlap between model A and model B, that is, no intersection, it means that the models are independently arranged, and the weight of the corresponding row of the feature vector matrix is 1.

[0096] In an embodiment, the clustering analysis of the present scheme adopts an optimized K-means algorithm, including:

[0097] Based on the attribute information obtained from the monomer BIM model, one or more dimensions of the geometric size class, the physical performance class, the functional attribute class, or the management information class in the attribute information are classified, and the maximum value of the number of classifications is determined as the K value;

[0098] Based on the determined K value, K initial center point models are randomly selected;

[0099] The distances between the monomer BIM model and the K initial center point models are calculated respectively, the monomer BIM model closest to the initial center point model is obtained, and is clustered into the same class;

[0100] The average value of the monomer BIM model in the same class is calculated, the new initial center point model is determined according to the average value, and the distance of the new initial center point model is repeatedly calculated until the target center point model is obtained by convergence;

[0101] According to the monomer BIM model in the same class, the target model to be fitted is set, and the target model is merged into the target center point model.

[0102] Specifically, by extracting the attribute information of the monomer BIM, such as geometric size, physical performance, functional attribute or management information, etc., the K value is determined according to the number of classifications. Based on the K value, K center point models are randomly selected, the distance between the monomer BIM model and the K initial center point models is calculated, and the redundant space can be set in this process. The monomer BIM model close to or within the redundant range is aggregated into the same class. Then the average value of the monomer BIM model in the same class is calculated to re-determine the initial center point model, and the target center point model is converged. Finally, the target model is determined and merged, and the target model is merged into the target center point model to realize the automatic merging of the model.

[0103] By this scheme, the attribute is vectorized and K-Means clustering is performed, so as to identify similar monomer BIM models, and on this basis, the merging of monomer BIM models is realized.

[0104] In other embodiments, if the target center point model corresponding to the K value does not converge, the K value is sequentially decreased by the inverse elbow method, including:

[0105] Calculate the error sum of squares of the monomer BIM model and the initial center point model;

[0106] Determine the inflection point of the characteristic curve according to the error sum of squares;

[0107] When the inflection point of the characteristic curve is smooth or convergent within the set range;

[0108] The value corresponding to the inflection point of the characteristic curve is the K value.

[0109] Specifically, the present scheme constructs an innovative combination of "error sum of squares" and "elbow method", and gradually converges and finds the inflection point by sequentially decreasing the K value by the "inverse" elbow method, instead of the traditional "elbow method" which starts from zero and sequentially increases to find the corresponding inflection point. The "error sum of squares" refers to the distance between the monomer BIM model and the initial center point model, which obtains the square sum of the position deviation, i.e. the three-dimensional coordinate deviation.

[0110] The algorithm strategy of the present scheme is to determine the maximum K value according to the number of attribute classifications of the BIM model. However, in fact, the attribute classifications of the BIM model, such as backrest chairs, wheelchairs and armchairs, are regarded as three categories, but the "chair" pointed by the three categories includes these three cases, that is, it is essentially the same category. That is, there is an error in the selection of K value due to the error in these attribute classifications. In order to quickly find the appropriate K value, the K value is set from all model categories, and then the K value is searched downward to convergence, which can improve the calculation efficiency and quickly obtain the corresponding K value.

[0111] In an embodiment, on the basis of the scheme, the surface mapping of the monomer BIM model is also considered. If the mappings are the same or similar, the corresponding monomer BIM model is also the same or similar, which is also the merged object. Based on this, the method further comprises:

[0112] Obtaining the mapping of the monomer BIM model and the UV coordinates of the mapping;

[0113] Extracting the vertex coordinates of the monomer BIM model and the target UV coordinates of the mapping corresponding to the vertex coordinates, and obtaining the target pixel value corresponding to the target UV coordinates;

[0114] Judging whether the target pixel values between the monomer BIM models are equal. If yes, projecting the monomer BIM model to a two-dimensional plane and obtaining the BIM model contour;

[0115] Judging whether the BIM model contours are the same. If yes, merging the BIM model corresponding to the BIM model contour.

[0116] Specifically, in the BIM model scene, a large number of monomer BIM models are included, such as door and window, wall, floor, etc. The monomer BIM model at this time is a white model, that is, a BIM model with only line structure. In order to express the geometric model space for three-dimensional visualization, the monomer BIM model needs to be mapped, which is used to express the surface material and visual features of the monomer BIM model. The loading of the mapping on the monomer BIM model needs to read the image resource, and also needs to be accurately mapped with the surface geometry of the monomer BIM model, that is, the UV coordinates define the unfolding way of the two-dimensional mapping on the surface of the three-dimensional monomer BIM model. Therefore, for the monomer BIM models that are the same or highly similar, the structure or contour of the monomer BIM model is consistent or highly similar, and the mapping of the monomer BIM model and the corresponding UV coordinate information are extracted, which are the basic conditions for the classification and subsequent merging of the monomer BIM model. Figure 1

[0117] Extracting the geometric vertex in the monomer BIM model and the UV coordinates corresponding to the geometric vertex after the two-dimensional unfolding of the monomer BIM model and the mapping. The UV coordinates of the mapping correspond to the corresponding pixel value, that is, the target pixel value extracted at the target position of the mapping.

[0118] ​For the target pixel value of the plurality of single BIM models corresponding to the map, the target pixel value between the BIM models is compared, and it is judged whether the target pixel value is equal. If yes, the single BIM model is projected to a two-dimensional plane along a specific direction, and the contour of the single BIM model is obtained. If no, the target pixel values of other single BIM models are compared until the comparison is completed. In this embodiment, the step of judging whether the target pixel value is equal sets a certain threshold. If the error ratio between the target pixel values is 1% or other ratio, which can be adjusted according to the accuracy of the single BIM model merging, the single BIM model set to be merged can be quickly determined.

[0119] For the comparison between the BIM model contours, the point sets of the BIM model contours can be sequentially connected, and the single BIM model contours can be compared point by point or distance by distance to determine whether the two single BIM models are the same. This method includes scaling the single BIM model by the same ratio, and a certain error range can also be set. If the hit rate of point by point comparison or the distance value is within the preset range, the single BIM models are considered the same.

[0120] In the BIM model scene, the visual deployment process of the single BIM model usually has a large model volume, and there are a large number of models with repeated maps but slightly different geometries. Based on this, the scheme of this embodiment provides an initial filtering of "visual approximation" based on the same target pixel value of the map at the corresponding vertex position, and accurately circulates the single BIM model to be merged. Further, the projection of the circled single BIM model on the two-dimensional plane and the contour of the single BIM model are obtained, the single BIM models with the same geometry are determined, and the significant lightweight effect is achieved.

[0121] In an embodiment, based on the adaptive merging method of the plurality of single BIM models described above, an adaptive merging system based on a plurality of single BIM models is proposed, which comprises:

[0122] A two-dimensional model map acquisition module sets a spatial coordinate system of the BIM model, imports the single BIM model, and projects the single BIM model from three dimensions to two dimensions to the two-dimensional plane corresponding to the spatial coordinate system, and obtains a two-dimensional model map;

[0123] A feature acquisition module acquires feature intersection points and feature edges connected to the feature intersection points in the two-dimensional model map. The feature intersection points are intersection points formed by projecting the spatial line structure of the single BIM model to the two-dimensional plane;

[0124] A model feature vector matrix generation module generates a model feature vector matrix according to the vertices and edges of the single BIM model, and the feature intersection points and feature edges corresponding to the two-dimensional model map, and generates the model feature vector matrix according to the order of the two-dimensional model map obtained by projecting the single BIM model.

[0125] a model characteristic value calculation module: calculating the model characteristic value according to the model characteristic vector matrix;

[0126] a model merging module: clustering analysis of the model characteristic value, obtaining the target model to be fitted and merged, and merging the target model.

[0127] The adaptive merging system based on multiple single BIM models, through a two-dimensional model drawing acquisition module, obtains a two-dimensional model drawing; then through a feature acquisition module, extracts feature intersections and feature edges in the two-dimensional model drawing; then combines a model characteristic vector matrix generation module to generate a model characteristic vector matrix according to the vertices, edges, feature intersections and feature edges in the single BIM model; then a model characteristic value calculation module calculates the model characteristic value according to the model characteristic vector matrix; finally, a model merging module performs clustering analysis on the model characteristic value, obtains the target model to be fitted and merged, and merges the target model, realizing model lightweight and also realizing the lightweight of the BIM model scene containing multiple single BIM models.

[0128] The above is only a preferred exemplary embodiment of the present application, and is not intended to limit the implementation of the present application. Those skilled in the art can easily make corresponding modifications or modifications according to the main concept and spirit of the present application, so the protection scope of the present application should be subject to the protection scope required by the claims.

Claims

1. An adaptive merging method based on multiple individual BIM models, characterized in that, include: Step 1: Set the spatial coordinate system of the BIM model, import the individual BIM model, and project the individual BIM model from three dimensions to two dimensions onto the two-dimensional plane corresponding to the spatial coordinate system to obtain a two-dimensional model drawing. Step 2: Obtain the feature intersection points in the two-dimensional model diagram, and the feature edges connected to the feature intersection points; the feature intersection points are the intersection points formed by projecting the spatial line structure of the single BIM model onto the two-dimensional plane; Step 3: Based on the vertices and edges in the single BIM model, and the feature intersections and feature edges corresponding to the two-dimensional model diagram; generate a model feature vector matrix according to the order of the two-dimensional model diagram obtained by projecting the single BIM model; Step 4: Calculate the model eigenvalues ​​based on the model eigenvector matrix; Step 5: Perform cluster analysis based on the feature values ​​of the model to obtain the fitted target model, and then merge the target models.

2. The adaptive merging method based on multiple individual BIM models according to claim 1, characterized in that, Step 1 includes: Step 11: Obtain the attribute information of the single-unit BIM model; Step 12: Determine the type of the individual BIM model based on the type information in the attribute information; Step 13: Project the same type of single-unit BIM model from three dimensions to two dimensions onto the two-dimensional plane corresponding to the spatial coordinate system to obtain a two-dimensional model diagram.

3. The adaptive merging method based on multiple individual BIM models according to claim 1, characterized in that, The feature intersections include direct feature intersections formed by projecting the vertices of the single BIM model onto the two-dimensional model diagram, and projected feature intersections formed by projecting the edges of the single BIM model onto the two-dimensional model diagram. Calculate the intersection ratio between the projected feature intersection point and the feature intersection point, where the reciprocal of the intersection ratio is the quantity weight of the projected feature intersection point; Update the number of feature intersections. The number of feature intersections is the product of the number weight and the projected feature intersections, plus the direct feature intersections.

4. The adaptive merging method based on multiple individual BIM models according to claim 1, characterized in that, Step 3 also includes: Determine whether the minimum bounding box of the single BIM model intersects with the minimum bounding box of any BIM model; If so, project the intersecting surfaces of the individual BIM models onto the corresponding intersecting surfaces in the two-dimensional plane, and calculate the projected area of ​​the intersecting surfaces in the corresponding two-dimensional plane; Calculate the ratio of the projected area of ​​the intersecting surface to the area of ​​the individual BIM model on the corresponding two-dimensional plane, where the reciprocal of the area ratio is the weight of the corresponding row of the model feature vector matrix.

5. The adaptive merging method based on multiple individual BIM models according to claim 4, characterized in that, If not, the weight of the corresponding two-dimensional plane in the corresponding row of the model feature vector matrix is ​​1.

6. The adaptive merging method based on multiple individual BIM models according to claim 1, characterized in that, Step 4 includes: Step 41: The model feature vector matrix is ​​an M×N matrix, where M represents the three faces formed by the spatial coordinate system, and N represents the vertices, edges, feature intersections, and feature edges, for a total of four columns; Step 42: Convert the model feature vector matrix into a 1×(M×N) matrix; Step 43: Calculate the value of the 1×(M×N) matrix as the model eigenvalue.

7. The adaptive merging method based on multiple individual BIM models according to claim 1, characterized in that, The clustering analysis employs an optimized K-means algorithm, including: Attribute information is obtained based on the single BIM model. One or more dimensions of the attribute information, such as geometric dimensions, physical performance, functional attributes, or management information, are classified. The maximum value of the number of classifications is determined as the K value. Based on the determined K value, K initial center point models are randomly selected; Calculate the distance between the individual BIM model and the K initial center point models respectively, obtain the individual BIM models that are closest to the initial center point models, and cluster them into the same category; Calculate the average value of the individual BIM models of the same type, determine the new initial center point model based on the average value, and repeat the calculation of the distance of the new initial center point model until the target center point model is obtained by convergence. The individual BIM models of the same type are set as the target models for fitting and merging, and the target models are merged into the target center point model.

8. The adaptive merging method based on multiple individual BIM models according to claim 7, characterized in that, If the target center point model corresponding to the K value does not converge, the K value is then decreased sequentially using the inverse elbow method, including: Calculate the sum of squared errors between the individual BIM model and the initial center point model; The inflection point of the characteristic curve is determined based on the sum of squared errors; When the inflection point of the characteristic curve is smooth or converges within a set range; The value corresponding to the inflection point of the characteristic curve is the K value.

9. The adaptive merging method based on multiple individual BIM models according to claim 1, characterized in that, The two-dimensional plane is formed by any two mutually perpendicular coordinate axes in the spatial coordinate system.

10. An adaptive merging system based on multiple individual BIM models, characterized in that, include: The 2D model drawing acquisition module sets the spatial coordinate system of the BIM model, imports the individual BIM model, and projects the individual BIM model from three dimensions to two dimensions onto the 2D plane corresponding to the spatial coordinate system to obtain the 2D model drawing. The feature acquisition module acquires the feature intersections in the two-dimensional model diagram, as well as the feature edges connected to the feature intersections; the feature intersections are the intersections formed by projecting the spatial line structure of the single BIM model onto the two-dimensional plane. The model feature vector matrix generation module generates a model feature vector matrix based on the vertices and edges in the single BIM model, as well as the feature intersections and feature edges corresponding to the two-dimensional model diagram; according to the order of the two-dimensional model diagram obtained by projecting the single BIM model. Model eigenvalue calculation module: Calculates model eigenvalues ​​based on the model eigenvector matrix; Model merging module: performs cluster analysis on the feature values ​​of the models to obtain the target models that are fitted and merged.

Images