CN110807835A - Building BIM model and live-action three-dimensional model fusion method - Google Patents

Abstract

The invention discloses a fusion method of a BIM model of a building and a three-dimensional model of a real scene. The fusion and conversion process mainly includes: model reconstruction, geometric information conversion, three-dimensional model registration and semantic mapping; BIM model, generate the json attribute file and ifc geometry file of each component; then the intermediate format is exchanged, the ifc is converted to obj; b3dm; finally add the data description file tileset.json, and convert b3dm to 3DTiles data. The invention realizes how to fully express the geometric figures while retaining the semantic attribute features in the original BIM model, and how to unify the spatial positions of the two models in the same environment, so as to present the display effect of internal and external integration.

Description

A fusion method of building BIM model and real 3D model

technical field

The invention relates to a fusion method of a building BIM model and a real scene three-dimensional model, and belongs to the field of building information models.

Background technique

With the continuous acceleration of urbanization, the problems faced by cities are gradually increasing, and urban construction and management are facing severe challenges. In order to improve work efficiency, achieve cross-departmental collaboration and refined urban management services, urban informatization is imminent. Smart cities can provide good solutions to problems encountered in urban development through real-time integration of multi-source information and cross-domain information sharing, which is an important development trend of urban informatization.

The smart city is developed from the digital city and is based on the city information model, of which the building information model is an important part. The Building Information Modeling (BIM, Building Information Modeling) technology in the construction field can simulate the required building image through digital information simulation, thereby providing data support for the construction of smart cities. However, BIM essentially focuses on the expression of the building itself and its internal information, and does not include the expression of environmental information around the building. It has certain limitations in the application of smart city spatial location. The real 3D model in the GIS field can complement each other with the BIM model. The model focuses on the macroscopic expression of buildings and surface phenomena, and has the characteristics of high precision, high fidelity, and measurability. Therefore, the fusion of building information model and real 3D model is one of the key technologies for the development of smart cities.

In the prior art, by using WebGL as the container on the Web side, transmitting data based on the network service interface, and developing application programs in combination with network programming knowledge, the fast and efficient loading of BIM data in the 3D GIS environment is realized, which is the key to the project's entire life cycle. The application provides the support environment.

However, there are still some deficiencies in the integration of BIM and real 3D models, which need to be further studied, such as how to fully express the geometry while retaining the semantic attributes of the original BIM model, and how to unify the space of the two models in the same environment. position, so that it presents an integrated display effect of inside and outside.

SUMMARY OF THE INVENTION

The purpose of the present invention is to provide a fusion method of a BIM model of a building and a 3D model of a real scene, so as to realize how to retain the semantic attribute features in the original BIM model while completely expressing the geometric figures, and how to unify the two types in the same environment. The spatial position of the model makes it present an integrated display effect of inside and outside.

The technical scheme adopted by the present invention is: a fusion method of a BIM model of a building and a three-dimensional model of a real scene, which comprises the following steps:

Step 1, split the BIM model of the building, and generate the json attribute file and ifc geometry file of each component;

Step 2, perform intermediate format exchange, convert ifc to obj;

Step 3: Convert obj to glTF to further realize 3D geometry data transfer;

In step 4, attributes and geometric data are connected, and glTF and json generate b3dm; specifically, the glTF and json file groups containing geometric and semantic information are converted into an overall three-dimensional data file, that is, b3dm;

Step 5, add the data description file tileset.json, and convert b3dm to 3DTiles data.

Further, in the step 1, the BIM model in the IFC format is split, which specifically includes the construction of the data processing environment and the splitting and exporting, wherein:

(1) Construction of data processing environment: upload the IFC model to be processed to the BIMServer, save it in the underlying database, and connect to the BIM Server server in the Java environment;

(2) Data splitting and exporting: call the BIMServer API, traverse all the types in the above IFC through the BimServerClient interface, find the components corresponding to the required building types, and query the information; according to the query results, use the client.download function to download Download files in two formats for each sub-component information on the server, namely ifc and json, which are used to store geometric and semantic information, and define naming rules for downloading, that is, ifc and json files describing the same component, and their file names Also the same.

Further, in the second step, the ifc file set of the sub-component is stored in the same file directory, and then the loop body is constructed by using the Java language, and the conversion function is repeatedly performed.

Further, in the third step, the spatial registration of the BIM model in the obj format is performed first, and the conversion from the obj format to the glTF format is further implemented.

Among them, the specific process of spatial registration is as follows:

(1) Calculate the rotation parameters;

The BIM/obj data is regarded as the source data, the real 3D model/obj data is regarded as the target data, and the two models are imported into the 3D software Geomagic Studio at the same time; in the registration process, the point with the same name is first selected as the starting data, and then the SVD algorithm is used. Calculate rotation parameters:

The coordinates of BIM/obj data are expressed as X' _i =(x' _i ,y' _i ,z' _i ), and the real 3D model/obj coordinate data is expressed as X _i =( _xi ,y _i ,z _i ); in Matlab Import coordinate data in , and the process of solving rotation parameters is as follows:

①Calculate the center of gravity of the two sets of coordinates

② Carry out the center of gravity processing to obtain {ΔX _i =X _i -P}{ΔX′ _i =X′ _i -Q};

③Calculate 3×3 data matrix

④ Use SVD function to decompose S matrix S _m×n =U _m×m Λ _m×n V ^T _n×n ;

^⑤Calculate the rotation matrix R=VUT;

⑥ Calculate the rotation angles α, β, γ in the directions of X, Y, and Z axes;

The meanings of the above parameters are as follows:

X′ _i = (x′ _i , y′ _i , z′ _i ): the three-dimensional coordinates of the i-th point with the same name in the BIM model;

X _i = (x _i , y _i , z _i ): the three-dimensional coordinates of the i-th point with the same name in the three-dimensional model of the real scene;

n: the total number of points with the same name;

P: The barycentric coordinates of all points with the same name in the BIM model;

Q: The barycentric coordinates of all points with the same name in the real 3D model;

△X _i : the difference between the coordinates of the point with the same name of the real 3D model and the coordinates of the center of gravity, that is, the coordinates after the center of gravity;

△X _i ': the difference between the coordinates of the point with the same name in the BIM model and the coordinates of the center of gravity, that is, the coordinates after the center of gravity;

△X _i ^T : barycentric transposed matrix;

S: singular value data matrix;

U: left singular vector matrix;

∧: singular value diagonal matrix;

V: right singular vector matrix;

α: the angle between the BIM model and the real 3D model in the x-axis direction;

β: the angle between the BIM model and the real 3D model in the y-axis direction;

γ: the angle between the BIM model and the real 3D model in the z-axis direction;

(2) Determining translation parameters: Select a significant marker point on the building BIM as the origin of the local coordinate system, if the coordinates of the point in the local coordinate system are (x0, y0, z0), then the translation amount is (- x0, -y0, -z0);

(3) Realize space transformation: rotate and translate the obj model according to the calculated parameters, and at the same time, in order to realize batch rotation transformation, a macro command with python as the script language is established in Geomagic Studio, the transformation operation command is recorded, and the transformation is completed in batches after running the command. Works; the rotated model is still saved in obj format.

Convert obj to glTF format: save the building vertex coordinates, external normal vector, and texture coordinates in obj, and store the texture image of the building in glTF in png or jpg format; draw triangular patches in obj by vertex index Constitute the geometry, and use the grid to represent a single complex geometry in glTF; set the vertex traversal mode so that the grid is correctly transformed.

Further, in the step 4, first retrieve the json file with the same name according to the glTF geometry file of each subcomponent, then binarize the subcomponent glTF, and write the Binary glTF data block in the b3dm file body, and the subcomponent glTF The corresponding semantic information file is written into the Batch Table data block in the b3dm file body; the ID link between the sub-components is converted into the batchid attribute to achieve mutual mapping.

Further, in the step 5, the coordinates of the origin of the local coordinate system in the global coordinate system are (longitude, latitude, height), and then the parameters of the coordinate system transformation are calculated by the Cesium.Transforms.eastNorthUpToFixedFrame function and assigned to the transform property to realize The spatial mapping of the local coordinate system to the world coordinate system.

Compared with the prior art, the present invention adopts the above technical scheme, and has the following technical effects:

(1) After the BIM model is converted from the IFC standard to 3DTiles to realize the fusion, the present invention still retains the characteristics of the original model, not only can fully express the geometric figures of each component of the building, but also can query the attribute semantic information corresponding to each component, including the name , type, unique identifier, size, material, etc.

(2) Aiming at the problem that the BIM model itself lacks the coordinate information and cannot unify the position with the real 3D model with the global coordinate system, the present invention proposes a spatial registration scheme, and combines the relevant algorithms to obtain the registration parameters efficiently and quickly to achieve The spatial relationship mapping from the local coordinate system to the global coordinate system.

(3) The present invention realizes the fusion of cross-domain heterogeneous data, and realizes the display of BIM in the field of construction and the three-dimensional model of the real scene in the field of GIS in the same three-dimensional environment. Responsible for the detailed expression of buildings in the scene, the integration of the two provides reliable data support for smart city management and development.

Description of drawings

Figure 1 is the fusion framework of building BIM and real 3D model.

Figure 2 is a schematic diagram of the reconstruction process of the split building BIM model.

FIG. 3 is a mapping diagram of the spatial relationship from the local coordinate system to the global coordinate system.

Figure 4 is a geometric information conversion diagram.

Figure 5 is the model semantic information fusion diagram

Figure 6 is a data structure diagram of 3DTiles.

Fig. 7 is the BIM model of the building in the third embodiment.

Detailed ways

The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

Example 1.

Figure 1 is a fusion framework of building BIM and real 3D model. In this framework, the IFC standard building BIM model needs to go through a series of transformation processes to form a new 3DTiles model building. The fusion and transformation process mainly includes the following four Sections:

(1) Rebuild the model. The BIM model contains rich building semantic information, such as walls, columns, beams, etc. The overall conversion of the building will lead to the loss of construction semantic information. Therefore, it is necessary to reconstruct the building BIM model first, and decompose an overall building BIM model into a BIM model collection of multiple sub-components. The data format of each sub-component model uses IFC files to store geometric and spatial information, such as beam 1.ifc, beam 2.ifc, column 1.ifc, etc., and stores the semantic attribute information corresponding to the sub-component in the attribute file of the same name , in which the property file is stored in JSON (JavaScript Object Notation) format. JSON is a lightweight data exchange format, which is more suitable for the transmission and exchange of text data.

(2) Conversion of geometric information. After reconstruction, the geometry and spatial information of building sub-components are stored in IFC format, which cannot be directly converted into 3D Tiles data format. Therefore, this paper adopts the indirect conversion method, and firstly selects the OBJ format as the intermediate exchange format. The OBJ file format is a standard 3D model file format developed by Alias Wavefront. This format is recorded in text form and is more suitable for mutual guidance and exchange between 3D software models. Then convert the OBJ format to the glTF format in the real 3D model, then convert the glTF format to the b3dm format, and create the tile file in the 3D tiles through the b3dm file, so as to realize the conversion of model geometric information.

(3) Spatial position registration. The 3D Tiles model is a real three-dimensional model, which is essentially a geographic information system model. The biggest feature of the GIS model is the coordinate system it carries, usually a spatial position coordinate system, such as the WGS84 world coordinate system. The coordinate system in the IFC model corresponding to BIM is usually a relative coordinate system, which is a local coordinate system. This coordinate system often expresses the scale and size of the building model. In order to ensure that the BIM model of the building can be displayed in the correct position in the real 3D GIS, the spatial position registration of the two needs to be carried out. The detailed registration process will be described in detail in Section 2.3.

(4) Semantic mapping. BIM includes the semantic description information of a large number of sub-components, such as beam material, concrete label, etc. Through the previous model reconstruction steps, the semantic information in the sub-components is stored in the corresponding JSON files respectively. Semantic mapping is to map the JSON semantic information of multiple sub-components in BIM to semantic attribute information in the 3D model of the reality, and to correspond the semantic information of each sub-component to the glTF geometry file one-to-one. On this basis, the converted b3dm file is obtained, and finally integrated with the descriptive file tileset.json in the 3D model of the real scene into 3D Tiles, so that the integration of the BIM model of the building and the 3D model of the real scene can be realized.

To sum up, the above fusion framework mainly includes: model reconstruction, geometric information transformation, three-dimensional model registration and semantic mapping.

Example two.

Based on the framework of the above-mentioned embodiment, a fusion method of the building information model (BIM) of the IFC standard and the 3D model of the real scene of the 3DTiles standard is as follows: firstly, the BIM model of the building is split, and the json attribute file and the ifc geometry of each component are generated. file; then the intermediate format is exchanged, ifc is converted to obj; obj is converted to glTF, which further realizes the transfer of 3D geometry data; then attributes and geometric data are connected, glTF and json generate b3dm; finally add the data description file tileset.json, and b3dm Converted to 3DTiles data.

The specific operation steps are as follows:

Step 1: Building BIM Model Split

The purpose of this step is to split the BIM model in IFC format, including data processing environment construction and model splitting.

1) Construction of data processing environment: Download the well-known open source software BIM Server in the BIM field, run the server locally after successful installation, and provide user name and password to register and log in. Upload the pending IFC model to BIMServer to save it in the underlying database. To connect to the BIM Server server in the Java environment, the connection password is the user name and password during registration.

2) Data splitting and exporting: The IFC model consists of many component types, among which the main types of component types related to buildings include architectural columns, structural columns, walls (such as building partition walls, curtain walls), stair railings, ramps, doors , windows, etc. In order to avoid the loss of some information during the fusion process, it is necessary to split the overall building model into many sub-building components. Call the BIMServer API, traverse all the types in the above IFC through the BimServerClient interface, find the component corresponding to the required building type, and query the information in it. According to the geometric characteristics of the components, the query statements are also different, which are divided into three ways:

1. Conventional components, the simplest component types: ifcWindow (window), ifcDoor (door), ifcColumn (column), ifcBeam (beam), IfcCovering (ceiling), IfcRailing (railing), ifc FurnishingElement (furnishing element).

2. Component types with holes: eg ifcWall (wall), ifcSlab (floor).

3. Objects that contain aggregated components: Some complex objects are composed of multiple small components, for example, ifcCurtainWall (curtain wall), contains slabs and members; ifcStair (stairs) contains railings, stair runs, and floors; ifcRamp (ramp) contains boards and ramps.

According to the query result, the client.download function is used to download files in two formats of each subcomponent information from the server, namely ifc and json, which are used to store geometric and semantic information. At the same time, it defines the naming rules for the downloaded files, that is, ifc and json files describing the same component have the same file names. The specific implementation process is shown in Figure 2.

Step 2, from ifc to obj

The file structure of ifc expressing geometric information is relatively complex. In order to facilitate the integration and conversion, a file format with a simple structure is selected as the intermediate format transition. This step will read the sub-component ifc files one by one, convert them into obj intermediate exchange files, and transfer the geometric information.

Store the ifc file collection of subcomponents in the same file directory. Use the Java language to construct the loop body and execute the transformation function repeatedly. The conversion function relies on the open source tool IfcConvert (http://ifcopenshell.org/ifcconvert.html), which can be called directly through the command line.

Step 3, from obj to glTF

BIM uses a local coordinate system, while a real-world 3D model uses a global coordinate system. In order to realize the accurate mapping of the spatial relationship, this step firstly performs spatial registration on the BIM model in obj format, and then further realizes the conversion from obj to glTF format.

1) Spatial registration:

Cesium, the publisher of the 3DTiles standard, provides a function to calculate the mapping of the local coordinate frame to the global frame in the tileset.json module. The parameters are the Cartesian coordinates of the origin of the local coordinate system in the original coordinate space. The origin of the local coordinate system is on the surface of the 3D ellipsoid with zero elevation. The X-axis points to the due east from the geographic location, the Y-axis points to the due north direction from the geographic location, and the Z-axis points upward from the vertical ellipsoid surface of the location. As shown in Figure 3.

Therefore, in the local coordinate system, the model needs to be rotated and registered in the 3D environment. Secondly, the origin of the local coordinate system is translated to the significant feature points on the building BIM, so as to provide accurate origin parameter values for the above mapping function. The specific process of spatial registration is as follows:

1. Calculate the rotation parameters

The BIM/obj data is regarded as the source data, and the real 3D model/obj data is regarded as the target data, and the two models are imported into the 3D software Geomagic Studio at the same time. The first step of the registration process is to select points with the same name as the starting data. The selection of the points with the same name mainly follows the following three principles: ① The obvious feature points located on the building surface. Since the buildings in the reality model usually only have the outer contour surface information, the points with the same name are not suitable to use the internal points of the building. At the same time, in order to improve the picking accuracy, the obvious landmark points on the building surface should be selected, such as wall corners, window corners, and building surface structural connections. wait. ② evenly distributed. According to the scale of the building group, the positions of the points with the same name are reasonably distributed to avoid excessive concentration and linear arrangement of the points, which will affect the calculation results. ③At least 3 pairs of points with the same name. Using the SVD algorithm to calculate the rotation parameters requires at least three sets of points with the same name as the starting data. In order to improve the calculation accuracy, the logarithm of the points with the same name can be appropriately increased.

The coordinates of the BIM/obj data are expressed as X' _i =(x' _i , y' _i , z' _i ), and the real 3D model/obj coordinate data is expressed as X _i =( _xi , y _i , z _i ). Import coordinate data in Matlab, and the process of solving rotation parameters is as follows:

①Calculate the center of gravity of the two sets of coordinates

② Carry out the center of gravity processing to obtain {ΔX _i =X _i -P}{ΔX′ _i =X′ _i -Q}

③Calculate 3×3 data matrix

④ Use SVD function to decompose the S matrix S _m×n =U _m×m Λ _m×n V ^T _n×n

⑤Calculate the rotation matrix R=VU ^T

⑥ Calculate the rotation angles α, β, γ in the directions of X, Y, and Z axes;

The meaning of each parameter in the formula is as follows:

X′ _i = (x′ _i , y′ _i , z′ _i ): the three-dimensional coordinates of the i-th point with the same name in the BIM model;

X _i = (x _i , y _i , z _i ): the three-dimensional coordinates of the i-th point with the same name in the three-dimensional model of the real scene;

n: the total number of points with the same name;

P: The barycentric coordinates of all points with the same name in the BIM model;

Q: The barycentric coordinates of all points with the same name in the real 3D model;

△X _i : the difference between the coordinates of the point with the same name of the real 3D model and the coordinates of the center of gravity, that is, the coordinates after the center of gravity;

△X _i ': the difference between the coordinates of the point with the same name in the BIM model and the coordinates of the center of gravity, that is, the coordinates after the center of gravity;

△X _i ^T : barycentric transposed matrix;

S: singular value data matrix;

U: left singular vector matrix;

∧: singular value diagonal matrix;

V: right singular vector matrix;

α: the angle between the BIM model and the real 3D model in the x-axis direction;

β: the angle between the BIM model and the real 3D model in the y-axis direction;

γ: The included angle between the BIM model and the real 3D model in the z-axis direction.

2. Determine the translation parameters: select a significant mark point on the building BIM as the origin of the local coordinate system, and follow the principles of building BIM surface inflection points and feature points, such as wall roots, roof corners, etc. If the coordinates of the point in the local coordinate system are (x0, y0, z0), the translation amount is (-x0, -y0, -z0).

3. Realize space transformation: rotate and translate the obj model according to the calculated parameters. At this stage, the model is still composed of fragmented sub-components, and each component is an independent obj file. In order to realize batch rotation transformation, a macro command with python as the script language is established in Geomagic Studio, the transformation operation command is recorded, and the transformation work is completed in batches after running the command. The rotated model is still saved in obj format.

2) From obj to glTF: glTF can reduce redundant data irrelevant to rendering in 3D format, and is a 3D file format that is more suitable for OpenGL cluster loading.

The data structure framework of glTF format is different from that of obj format. When converting as shown in Figure 4, it is necessary to convert information such as vertex coordinates, external normal vectors, texture coordinates, and texture images in the obj file. The bin file of glTF stores the basic data of the geometry, so you need to save the building vertex coordinates, external normal vectors, and texture coordinates in obj here. The texture images of buildings are stored in glTF in png, jpg and other formats. In addition to the transformation of coordinate information geometry, it is also more important to draw triangle patches in obj through vertex index to form geometry. In glTF, meshes are used to represent a single complex geometric shape, and the meshes are composed of primitive arrays. Each primitive contains the corresponding vertex attributes and textures referenced in the bin file. So set the vertex traversal mode so that the mesh can be transformed correctly.

Step 4, from glTF+json to b3dm

The final stage of fusion is to convert the glTF and json file groups containing geometric and semantic information into an overall three-dimensional data file, namely b3dm, whose structure is shown in Figure 5. A b3dm file consists of a file header and a file body. The file header records information such as file type, version, and size. The file body includes a feature table (featureTable), a batch table (BatchTable), and binary glTF, which respectively record the number of features, feature attributes, and geometric data in the file. Therefore, the purpose of this step is to put the previously formed subcomponent glTF and JSON file into the corresponding block, and still maintain the original association.

The conversion process from glTF to b3dm is shown in Figure 6. First, the json file with the same name is retrieved according to the glTF geometry file of each subcomponent, and the corresponding wall0.json is retrieved from wall0.gltf in Figure 4. ID implements the connection. Then the subcomponent glTF (such as wall0.gltf) is binarized and written into the Binary glTF data block in the b3dm file body. Write the semantic information file (such as wall0.json) corresponding to the subcomponent into the Batch Table data block in the b3dm file body. And convert the ID links between subcomponents into batchid attributes to achieve mutual mapping with each other. Each vertex in Binary glTF contains a batchid attribute, and vertices with the same id value belong to the same subcomponent.

Step 5, from b3dm to 3DTiles

3DTiles consists of the geographic data b3dm file and the descriptive data tileset.json file. The tileset.json file contains tile metadata, which is used to organize the spatial structure of 3D tiles. The data structure is shown in Figure 6. The spatial bounding volume is an array representing the spatial extent of the current tile data. The geometric error is used to determine the level that should be loaded under the current viewing angle in the LOD division. Refine means the refinement method of tile data loading, which supports ADD and REPLACE. The content (content) specifies the b3dm data to be loaded through the link (url). The tile metadata also includes a transform attribute, which stores an array of 16 elements, which is used to represent a 4×4 coordinate transformation matrix in column-major order, so as to realize the transformation of the spatial position of the tile. Assuming that the origin of the local coordinate system is (longitude, latitude, height) in the global coordinate system, the Cesium.Transforms.eastNorthUpToFixedFrame function calculates the parameters of the coordinate system transformation and assigns them to the transform property to realize the space from the local coordinate system to the world coordinate system. map.

Example three.

Taking the buildings of Nanjing Architectural Design and Research Institute as an example, the model fusion method explored in this paper is verified. The DjiMatrice600 multi-rotor drone is used, equipped with a five-lens tilt camera, and the equipment parameters are shown in Table 1. Field surveying and mapping was carried out on the building and its surrounding area of about 2.6 hectares, and more than 1,300 aerial photos were obtained. The model format of the experimental area is 3DTlies using the real-world 3D modeling software Smart3DCapture.

Table 1 UAV platform and sensor parameters

At the same time, with Revit software, the BIM model of the building group of Nanjing Architectural Design and Research Institute was established, which are three interconnected seven-story buildings, as shown in Figure 7. A total of 11 categories and 10,533 components are included in the BIM model. In order to realize the integration of the two, the BIM model of the entire building is first exported to the IFC format, and reconstructed into 11 types of components as shown in Table 2.

Table 2 Breakdown of sub-components of the case building BIM model

The case BIM model generates 3DTiles data after model reconstruction, geometric transformation, spatial registration, and semantic mapping. Load the model data in the Cesium platform.

Experiments have proved that the 3D reality model of the building and the BIM model are integrated in the 3D geographical environment. The 3D reality model truly restores the appearance of the building and the surrounding environment in the experimental area. The BIM model shows the internal structure of the building in detail. The results achieved indoor and outdoor integrated modeling, which provided a good foundation for the detailed monitoring and management of smart city buildings. In order to further verify the consistency of the information before and after BIM fusion, write element click and view events, in this test platform, zoom in on the fusion model to a certain level, enter the real 3D model, and view the BIM component information. As shown in Figure 7, each component of BIM is independent and consistent with the original model. Click a component to view the attribute Name, Type, etc., indicating that the geometry and attribute information are kept intact, and the expected effect of the experiment is achieved.

The foregoing has shown and described the basic principles, main features and advantages of the present invention. Those of ordinary skill in the art should understand that the above-mentioned embodiments do not limit the protection scope of the present invention in any form, and all technical solutions obtained by means of equivalent replacement and the like all fall within the protection scope of the present invention.

The parts not involved in the present invention are the same as the prior art or can be implemented by using the prior art.

Claims (8)

1. a fusion method of a building BIM model and a real scene three-dimensional model, is characterized in that comprising the steps: Step 1, split the BIM model of the building, and generate the json attribute file and ifc geometry file of each component; Step 2, perform intermediate format exchange, convert ifc to obj; Step 3: Convert obj to glTF to further realize 3D geometry data transfer; In step 4, attributes and geometric data are connected, and glTF and json generate b3dm; specifically, the glTF and json file groups containing geometric and semantic information are converted into an overall three-dimensional data file, that is, b3dm; Step 5, add the data description file tileset.json, and convert b3dm to 3DTiles data. 2. The method for merging a BIM model of a building and a 3D model of a real scene according to claim 1, wherein in the step 1, the BIM model in the IFC format is split, which specifically includes building a data processing environment and According to split export, where: (1) Construction of data processing environment: upload the IFC model to be processed to the BIMServer, save it in the underlying database, and connect to the BIM Server server in the Java environment; (2) Data splitting and exporting: call the BIMServer API, traverse all the types in the above IFC through the BimServerClient interface, find the components corresponding to the required building types, and query the information; according to the query results, use the client.download function to download Download files in two formats for each sub-component information on the server, namely ifc and json, which are used to store geometric and semantic information, and define naming rules for downloading, that is, ifc and json files describing the same component, and their file names Also the same. 3. the fusion method of a kind of building BIM model and real scene three-dimensional model according to claim 1, is characterized in that, in described step 2, the ifc file collection of subcomponent is stored in the same file directory, and then utilizes Java The language constructs the body of the loop, which repeatedly executes the transformation function. 4. The fusion method of a BIM model of a building and a 3D model of a real scene according to claim 1, characterized in that, in the step 3, the BIM model in the obj format is firstly spatially registered, and then the obj to obj to 3D model is further realized. glTF format conversion. 5. the fusion method of a kind of building BIM model and real scene three-dimensional model according to claim 4, is characterized in that, the specific process of space registration is as follows: (1) Calculate the rotation parameters; The BIM/obj data is regarded as the source data, the real 3D model/obj data is regarded as the target data, and the two models are imported into the 3D software Geomagic Studio at the same time; in the registration process, the point with the same name is first selected as the starting data, and then the SVD algorithm is used. Calculate rotation parameters: The coordinates of BIM/obj data are expressed as X' _i =(x' _i ,y' _i ,z' _i ), and the real 3D model/obj coordinate data is expressed as X _i =( _xi ,y _i ,z _i ); in Matlab Import coordinate data in , and the process of solving rotation parameters is as follows: ①Calculate the center of gravity of the two sets of coordinates

② Carry out the center of gravity processing to obtain {ΔX _i =X _i -P}{ΔX′ _i =X′ _i -Q}; ③Calculate 3×3 data matrix

④ Use SVD function to decompose S matrix S _m×n =U _m×m Λ _m×n V ^T _n×n ; ^⑤Calculate the rotation matrix R=VUT; ⑥ Calculate the rotation angles α, β, γ in the directions of X, Y, and Z axes; The meanings of the above parameters are as follows: X′ _i = (x′ _i , y′ _i , z′ _i ): the three-dimensional coordinates of the i-th point with the same name in the BIM model; X _i = (x _i , y _i , z _i ): the three-dimensional coordinates of the i-th point with the same name in the three-dimensional model of the real scene; n: the total number of points with the same name; P: The barycentric coordinates of all points with the same name in the BIM model; Q: The barycentric coordinates of all points with the same name in the real 3D model; △X _i : the difference between the coordinates of the point with the same name of the real 3D model and the coordinates of the center of gravity, that is, the coordinates after the center of gravity; △X _i ': the difference between the coordinates of the point with the same name in the BIM model and the coordinates of the center of gravity, that is, the coordinates after the center of gravity; △X _i ^T : barycentric transposed matrix; S: singular value data matrix; U: left singular vector matrix; ∧: singular value diagonal matrix; V: right singular vector matrix; α: the angle between the BIM model and the real 3D model in the x-axis direction; β: the angle between the BIM model and the real 3D model in the y-axis direction; γ: the angle between the BIM model and the real 3D model in the z-axis direction; (2) Determining translation parameters: Select a significant marker point on the BIM of the building as the origin of the local coordinate system. If the coordinates of the point in the local coordinate system are (x0, y0, z0), the translation amount is (- x0, -y0, -z0); (3) Realize space transformation: rotate and translate the obj model according to the calculated parameters, and at the same time, in order to realize batch rotation transformation, a macro command with python as the script language is established in Geomagic Studio, the transformation operation command is recorded, and the transformation is completed in batches after running the command. Works; the rotated model is still saved in obj format. 6. the fusion method of a kind of building BIM model and real scene three-dimensional model according to claim 4, is characterized in that, obj is converted to glTF format: the building vertex coordinates, outer normal vector, texture coordinates in obj are saved , the texture image of the building is stored in glTF in png or jpg format; in obj, the triangular patch is drawn to form the geometry, and the grid is used to represent a single complex geometric shape in glTF; set the vertex traversal mode to make the grid be converted correctly. 7. the fusion method of a kind of building BIM model and real scene three-dimensional model according to claim 1, is characterized in that, in described step 4, at first according to the glTF geometry file of each subcomponent, retrieve the json file of the same name, then Binaryize the subcomponent glTF and write it into the Binary glTF data block in the b3dm file body, and write the semantic information file corresponding to the subcomponent into the Batch Table data block in the b3dm file body; and link the IDs between the subcomponents Converted to batchid attributes to achieve mutual mapping with each other. 8. The fusion method of a BIM model of a building and a three-dimensional model of a real scene according to claim 1, wherein in the step 5, the coordinates of the origin of the local coordinate system in the global coordinate system are (longitude, latitude, height), and then the Cesium.Transforms.eastNorthUpToFixedFrame function calculates the parameters of the coordinate system transformation and assigns them to the transform property to realize the spatial mapping of the local coordinate system to the world coordinate system.

Images