Conversion method for lightweight three-dimensional tile of component-level contact network three-dimensional model

Abstract

The invention belongs to the technical field of geospatial information systems, and particularly relates to a component-level catenary three-dimensional model lightweight three-dimensional tile conversion method, which comprises the following steps: S1, forming a component-level fine model data set with semantic annotations; s2, constructing a detail level division rule; s3, reconstructing a lightweight linear component model; s4, constructing a comprehensive evaluation function Q, deleting the body-shaped components of which the Q values are smaller than a preset threshold value, and retaining the body-shaped components of which the Q values are greater than or equal to the preset threshold value; s5, obtaining optimized body-shaped component model data; and S6, according to the detail level division rule in the S2, assembling and recombining the lightweight linear component model in the S3 and the optimized body-shaped component model data in the S5, generating a hierarchical component structure, converting each component into a three-dimensional instance model, and outputting a three-dimensional tile data structure. And efficient organization and visualization of the contact network BIM model in the GIS environment are realized.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of geographic spatial information system, and particularly relates to a conversion method of a component-level catenary three-dimensional model lightweight three-dimensional tile. BACKGROUND

[0002] The catenary is a high-voltage transmission line provided for a pantograph to draw current, which is arranged in a zigzag shape above the steel rail in an electrified railway. The catenary is the main framework of a railway electrification project and is a special form of power transmission line for supplying power to electric locomotives. The catenary is composed of a contact suspension, a support device, a positioning device, a support column and a foundation.

[0003] In order to realize the visual management of the geometric structure, node connection and running state of the catenary system, the catenary three-dimensional modeling gradually develops from two-dimensional drawing design to fine and component-level three-dimensional model expression, and is widely applied to multiple stages such as design, construction, operation and maintenance and digital twinning. However, in order to meet the multi-scale level application requirements from the regional scale to the engineering scale, the visualization of the catenary three-dimensional model in the regional and engineering different scale scenes still faces severe challenges.

[0004] The catenary model is usually constructed by using a BIM modeling software, and such a model has high geometric precision and semantic integrity, and can comprehensively reflect the spatial layout and physical properties of the components. However, due to the large number of components, the complex shape, the multiple semantic nesting levels, and the large number of model surface patches and the redundant volume, there are problems such as slow loading, rendering lag, interaction delay and the like when visualizing directly on the Web or in the GIS system.

[0005] In order to improve the visualization efficiency of the model, various three-dimensional model lightweight methods are proposed in the prior art.

[0006] Firstly, the surface patch simplification based on geometric features reduces the number of model surface patches by edge folding, point merging and mesh optimization, so as to reduce the rendering burden. However, this kind of method generally processes the components as a whole, lacks a semantic recognition mechanism, is easy to cause the loss of the model semantic structure or the deviation of the importance judgment of the components, and is not suitable for the catenary model which has high requirements for semantic expression.

[0007] Secondly, the lightweight is realized by a component semantic screening method, for example, the retention conditions are set according to the component purpose, visibility, function level and other attributes, and the components with low importance or invisibility are deleted to reduce the data amount. However, this method does not perform further geometric simplification processing on the retained components while removing the redundant data, resulting in limited overall simplification effect. In addition, this method depends on manual configuration rules, and lacks adaptability and automatic screening capability of component features.

[0008] The third is to reconstruct the model by component templating or instantiation, and to reduce data redundancy by uniformly constructing repeated components and copying by using a transformation matrix. However, only the component geometric information is uniformly stored, and the differentiated detail simplification and optimization processing cannot be performed on each instance, which limits its adaptability in fine management and high-interactive visualization scenarios.

[0009] To support the visualization display of large-scale models, the prior art introduces a three-dimensional tile organization method, which often uses a tile layering mechanism based on a spatial octree, quadtree or view frustum LOD structure to realize the hierarchical loading of model data. The three-dimensional tile standard represented by 3DTiles has been widely used in the fields of buildings, cities and terrains. However, in the catenary scene, the existing three-dimensional tile technology has the following shortcomings: first, there is a lack of independent tile expression of fine-grained components, which is usually divided in units of whole models, and it is difficult to realize component-level loading and interaction, and the organization degree is low; second, there is no corresponding mechanism to model semantic information, and the tile content does not contain semantic tags such as component attributes and types, so semantic-driven visualization scheduling cannot be realized; third, adaptive simplification is not combined with component structure characteristics in the model conversion process, which easily causes problems such as large tile data volume and redundancy of repeated components, thereby affecting the platform running efficiency. SUMMARY

[0010] The technical problem to be solved by the present application is to provide a component-level catenary three-dimensional model lightweight three-dimensional tile conversion method, so that the catenary BIM model can realize efficient organization and visualization in the GIS environment.

[0011] The technical scheme adopted by the present application to solve the technical problem is: a component-level catenary three-dimensional model lightweight three-dimensional tile conversion method, comprising the steps of: S1, traversing all components in the catenary BIM model, extracting the geometric information and attribute information of each component, establishing a one-to-one correspondence between each component and semantic information, and forming a component-level fine model data set with semantic annotations; S2, defining four levels of detail hierarchy of system, system, unit and component according to the external control parameters and model semantic hierarchy of the catenary BIM model, establishing a mapping relationship between each component and its upper structure, and constructing a detail level division rule; S3, according to the component-level fine model data set in S1, the detail level division rule in S2, and the component geometric form, filtering out linear units, and reconstructing a lightweight linear component model; S4, according to the component-level fine model data set in S1, the detail level division rule in S2 and the component geometry, screening out the volumetric unit, constructing a comprehensive evaluation function Q, sorting all volumetric components according to the Q value, deleting the volumetric components with Q value less than the preset threshold, and retaining the volumetric components with Q value greater than or equal to the preset threshold; S5, performing topological consistency detection on the retained volumetric components in S4; repairing the topological abnormal area of the retained volumetric components; performing Boolean operation on the relationship of each volumetric component, adjusting the boundary relationship of each volumetric component, and realizing topological closure; optimizing the standard triangular mesh data of the retained volumetric components to obtain the optimized volumetric component model data; S6, according to the detail level division rule in S2, assembling and reorganizing the lightweight linear component model in S3 and the optimized volumetric component model data in S5 to generate hierarchical component structure, converting each component into a three-dimensional instance model, and outputting three-dimensional tile data structure.

[0012] Further, the step S1 specifically includes the following steps: S101, reading the overhead control parameters of the model, including unit information, coordinate system definition and component organization structure; setting key constraint parameters for subsequent processing, including the upper limit of the face number of a single component and the model geometry conversion precision; S102, traversing all components in the overhead line BIM model, extracting the geometry information and attribute information of each component, and converting the geometry information of each component into standard triangular mesh data; S103, according to the component semantic description rule in the overhead line BIM model, mapping the attribute information extracted in step S102 to the component semantic table, classifying combined with the structure type field, the purpose field and the function level field, and generating standardized semantic description data; S104, assigning a globally unique identifier to each component, and binding the standard triangular mesh data of each component with its corresponding semantic information to form a component-level fine model data set with semantic annotation.

[0013] Further, the step S2 specifically includes the following steps: according to the model semantic hierarchy of the overhead line BIM model and the component organization structure obtained in step S101, defining four-level detail levels of system, system, unit and component, establishing the mapping relationship between each component and its upper structure, and constructing the detail level division rule.

[0014] Further, the step S4 specifically includes the following steps: S401, according to the component level fine model data set in S1, the detail level division rule in S2, and the component geometry, filtering out the volumetric unit, obtaining the face number N of each volumetric unit and the face number n of each volumetric component in the corresponding volumetric unit, and calculating the minimum circumscribed bounding box volume v of each volumetric component in each volumetric unit and the minimum circumscribed bounding box volume V of the volumetric unit; S402, according to the structure type, use and function of the volumetric component in the catenary BIM model, all the volumetric components are divided into several functional categories; on this basis, combined with the geometric complexity index and the visualization importance level of each volumetric component, a weight parameter table corresponding to the volumetric component classification is set, and the semantic weight factor k of each volumetric component is extracted; S403, defining a local index m of a component level, and simultaneously, defining a global index M for normalizing the local index; ; ; S404, constructing a comprehensive evaluation function Q for reflecting the importance of a certain volumetric component in its corresponding volumetric component unit and the cost of its simplification; ; S405, calculating the Q value of each volumetric component in each volumetric unit one by one, and sorting all the volumetric components according to the Q value; according to the model accuracy requirement, a threshold Q threshold is preset; when Q threshold , the corresponding volumetric component is deleted; when Q≥Q threshold , the corresponding volumetric component is retained.

[0015] Further, the minimum circumscribed bounding box volume v of the volumetric component in step S401 is calculated, which specifically includes steps S401.1-S401.5: S401.1, performing first three-dimensional convex hull calculation on the vertex three-dimensional coordinate set P of the volumetric component; ; The first three-dimensional convex hull C is defined as the minimum convex set of the vertex three-dimensional coordinate set P of the volumetric component: ; Where λ i is a linear combination coefficient, satisfying the non-negative constraint λ i ≥0 and the normalization condition: ; x represents any point in the three-dimensional volumetric component, and is any position point in the first three-dimensional convex hull C; R 3 represents a three-dimensional real number space, that is, the set of all three-dimensional coordinate points of the volumetric component; m represents the number of vertices in the vertex three-dimensional coordinate set P of the body-shaped component, i.e., the total number of vertices of the body-shaped component; P i represents the i-th vertex in the vertex three-dimensional coordinate set P of the body-shaped component, which is the current processing point in the same vertex set in subsequent steps S401.1.2 and S401.1.4, ; performing a first three-dimensional convex hull calculation, specifically including steps S401.1.1-S401.1.5: S401.1.1 selects 4 points that are not coplanar from the vertex three-dimensional coordinate set P of the body-shaped component to construct an initial tetrahedron as an initial face sheet set of the first three-dimensional convex hull C; S401.1.2 takes the points in the vertex three-dimensional coordinate set P of the body-shaped component as the current point p i , judges whether the current point is located inside the first three-dimensional convex hull C; if the current point is located inside the first three-dimensional convex hull C, the current point is discarded; if the current point is located outside the first three-dimensional convex hull C, the current point is determined to be visible to the face sheet set F of the first three-dimensional convex hull C; S401.1.3 deletes the visible face sheet set F to obtain an open boundary ring; S401.1.4 connects the points located outside the first three-dimensional convex hull C with the open boundary ring obtained in step S401.1.3 to generate a new face sheet and update the first three-dimensional convex hull C; S401.1.5 repeats steps S401.1.2 to S401.1.4 until all vertices of the body-shaped component are processed, and a complete first three-dimensional convex hull C is obtained; S401.2, traverses all face sheets of the first three-dimensional convex hull C, and takes the normal vector of each face sheet as a candidate direction set ; for each candidate direction d j , construct an orthogonal coordinate basis (a j , b j , c j ) perpendicular to d j ; wherein, ; taking a space rectangular coordinate system as a reference system, and projecting the vertex three-dimensional coordinate set P of the body-shaped component to the space rectangular coordinate system, the projection coordinates (u j , v j , h j ) of the j-th vertex p j (x k , y k , z k ) of the body-shaped component in the space rectangular coordinate system are calculated: , , ; wherein, “ ” represents dot multiplication operation; S401.3, calculating the range of the vertex three-dimensional coordinate set P of the body-like member in the spatial rectangular coordinate system: ; wherein, L j , W j , and H j respectively correspond to the edge length of the bounding box of the body-like member in the x-axis, y-axis, and z-axis directions; S401.4, calculating the bounding box volume v j of the body-like member in the candidate direction: ; S401.5, traversing all candidate directions, and taking the minimum value as the minimum circumscribed bounding box volume v of the body-like member: .

[0016] Further, the step S401 of calculating the minimum circumscribed bounding box volume V of the body-like unit specifically comprises steps S401.6-S401.10: S401.6, regarding the body-like unit as a whole composed of multiple members, performing a second three-dimensional convex hull calculation on the vertex three-dimensional coordinate set P ’ of the body-like unit: ; defining a second three-dimensional convex hull C ’ as the minimum convex set of the vertex three-dimensional coordinate set P ’ of the body-like unit: ; wherein, λ i is a linear combination coefficient, satisfying the non-negative constraint λ i ≥ 0 and the normalization condition: ; x ’ represents any point in the three-dimensional space body-like unit, and is any position point in the second three-dimensional convex hull C ’ ; R 3 represents a three-dimensional real space, i.e. the set of all three-dimensional coordinate points of the body-like unit; m ’ represents the number of vertices in the vertex three-dimensional coordinate set P ’ of the body-like unit, i.e. the total number of vertices of the body-like unit; represents the number of vertices in the vertex three-dimensional coordinate set P ’ of the body-like unit, i.e. the total number of vertices of the body-like unit;the i-th vertex in the set of vertices in the body-like unit, the current processing point in the subsequent steps S401.6.2 and S401.6.4, ; performing a second three-dimensional convex hull calculation, specifically comprising steps S401.6.1-S401.6.5: S401.6.1, selecting four points not in the same plane from the set of vertex three-dimensional coordinates P ’ of the body-like unit, to construct an initial tetrahedron as an initial set of face patches of the second three-dimensional convex hull C ’ ; S401.6.2, taking a point in the set of vertex three-dimensional coordinates P ’ of the body-like unit as a current point in turn, judging whether the current point is inside the second three-dimensional convex hull C ’ ; if the current point is inside the second three-dimensional convex hull C ’ , discarding the current point; if the current point is outside the second three-dimensional convex hull C ’ , determining a set of visible face patches F ’ of the second three-dimensional convex hull C ’ visible to the current point; S401.6.3, deleting the set of visible face patches F ’ to obtain an open boundary ring; S401.6.4, connecting the point outside the second three-dimensional convex hull C ’ to the open boundary ring obtained in step S401.6.3 to generate a new face patch and update the second three-dimensional convex hull C ’ ; S401.6.5, repeating steps S401.6.2-S401.6.4 until all vertices of the body-like unit are processed, to obtain a complete second three-dimensional convex hull C ’ ; S401.7, traversing all face patches of the second three-dimensional convex hull C ’ , taking a normal vector of each face patch as a candidate direction set ; for each candidate direction , constructing an orthogonal coordinate basis perpendicular to ; wherein, = ; taking a spatial rectangular coordinate system as a reference system and projecting the set of vertex three-dimensional coordinates P ’ of the body-like unit to the spatial rectangular coordinate system to obtain, by calculation, a projection coordinate of the j-th vertex of the body-like unit in the spatial rectangular coordinate system : , , ; wherein, represents dot multiplication operation; S401.8, calculate the vertex three-dimensional coordinate set P of the body unit in the space rectangular coordinate system ’ The range of: ; wherein, respectively correspond to the edge length of the bounding box of the body unit in the x-axis, y-axis and z-axis three directions; S401.9, calculate the bounding box volume V of the body unit in the candidate direction j : ; S401.10, traverse all candidate directions, take the minimum value as the minimum circumscribed bounding box volume V of the body unit: .

[0017] Further, the step S5 specifically comprises the following steps: S501, obtain the body-shaped component reserved in S4, perform topology consistency detection, and identify topology abnormal area including opening, gap, intersection and overlap; S502, for the body-shaped component with topology abnormal area, perform inflation and corrosion operation by using three-dimensional Minkowski operator, and repair the topology abnormal area of the body-shaped component; S503, in the local space range, perform Boolean operation on the relationship of each body-shaped component, adjust the boundary relationship of each body-shaped component, and realize topology closure; S504, perform mesh simplification processing, based on the mesh compression algorithm of quadratic error measurement, optimize the standard triangular mesh data of the reserved body-shaped component, recursively combine vertices and triangular faces, and control the error range and the number of facets.

[0018] Further, the step S3 specifically comprises the following steps: according to the component-level fine model data set in S1, the detail level division rule in S2, and the component geometric shape, filtering out the linear unit, performing multi-segment straight line fitting operation on the linear component containing curve feature, converting it into several equal-length straight line segment combination representation, extracting the key coordinate data of the linear component, and reconstructing the lightweight linear component model.

[0019] Compared with the prior art, the beneficial effects of the present application are: the present application provides a kind of component-level contact network three-dimensional model lightweight three-dimensional tile conversion method, makes contact network BIM model realize efficient organization and visualization in GIS environment.Effectively solve the difficulty that contact network BIM model is difficult to efficiently organize and visualize in GIS environment. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 is a general flowchart of the present application; Figure 2 is a generation flowchart of the component-level fine model dataset; Figure 3 is a flowchart of linear unit lightweight; Figure 4 is a flowchart of body-shaped unit lightweight after screening; Figure 5 is a contact network BIM model data organization structure tree diagram; Figure 6 is a railway contact network system hierarchy diagram; Figure 7 is a railway contact network system hierarchy diagram; Figure 8 is a railway contact network unit hierarchy diagram; Figure 9 is a railway contact network component hierarchy diagram. DETAILED DESCRIPTION

[0021] The present application is further illustrated below in conjunction with the accompanying drawings Figure 1 , 2 , 3, 4, 5, 6, 7, 8 and 9, and the embodiments.

[0022] The conversion method of the component-level contact network three-dimensional model lightweight three-dimensional tile is shown in Figure 1 , 2 , 3, 4 and 5, which includes the following steps: S1, read the contact network BIM model file, obtain the external control parameters of the model, set the key constraint parameters for subsequent processing; traverse all components in the contact network BIM model, extract the geometric information and attribute information of each component, convert the geometric information of each component into standard triangular mesh data; generate a unique identifier for each component, and establish a one-to-one correspondence between each component and semantic information, forming a component-level fine model dataset with semantic annotations.

[0023] Specifically, the step S1 includes the following steps: S101, read the contact network BIM model file, obtain the external control parameters of the model, the external control parameters including unit information, coordinate system definition and component organization structure; set the key constraint parameters for subsequent processing, the key constraint parameters including the upper limit of the face number of a single component and the model geometry conversion precision.

[0024] The catenary BIM model file is derived from the railway planning and design stage of a major railway, the model format is.dgn format file generated by Bentley Company, the file size is 208 MB, the model unit is meter, the national geodetic coordinate system CGCS2000 is adopted, the projection mode is 3-degree zone Gauss-Kruger projection, the central meridian is 114°, the upper limit of the face number of a single component is set to 50000, and the model geometry conversion precision is set to 0.1 meter.

[0025] S102, all components in the catenary BIM model are traversed, the geometric information and attribute information of each component are extracted, and the geometric information of each component is converted into standard triangular mesh data. A uniform mesh basis is provided for subsequent processing.

[0026] S103, according to the component semantic description rule in the catenary BIM model, the attribute information extracted in step S102 is mapped to the component semantic table, classified by combining the structure type field, the purpose field and the function level field, and standardized semantic description data is generated. The consistency of the model semantic structure is improved.

[0027] S104, a global unique identifier UUID is assigned to each component, and the standard triangular mesh data of each component is bound with the corresponding semantic information to form a component-level fine model data set with semantic annotation.

[0028] The component-level fine model data set with semantic annotation includes an.obj format standard triangular mesh model file corresponding to each component and a UUID corresponding semantic attribute information table stored in.csv format. In this embodiment, the overall data amount of the component-level fine model data set with semantic annotation is about 1.79 GB.

[0029] S2, as shown in Table 1, according to the external control parameters of the catenary BIM model and the model semantic hierarchy, the four-level detail levels of system, system, unit and component are defined, the mapping relationship between each component and the upper structure to which it belongs is established, and the detail level division rule is constructed. For guiding the subsequent model screening, organization and output.

[0030] Table 1

[0031] Specifically, the step S2 specifically includes the steps of defining the four-level detail levels of system, system, unit and component according to the model semantic hierarchy of the catenary BIM model and the component organization structure obtained in step S101, establishing the mapping relationship between each component and the upper structure to which it belongs, and constructing the detail level division rule.

[0032] S3, as Figure 3As shown, according to the component-level fine model data set in S1, the detail level division rule in S2, and the component geometry, linear units are screened out, multi-segment straight line fitting operation is performed on the linear components containing curve features, and the linear components are converted into several equal-length straight line segment combinations, key coordinate data of the linear components are extracted, and a lightweight linear component model is reconstructed.

[0033] The key coordinate data of the linear components includes three-dimensional coordinate values (x line ,y line ,z line ) of each node in the linear components, a point number identifier automatically assigned to each node, and a node connection relationship defined according to the point number sequence.

[0034] In S4, according to the component-level fine model data set in S1, the detail level division rule in S2, and the component geometry, volumetric units are screened out, a comprehensive evaluation function Q is constructed, all volumetric components are sorted according to the Q value, volumetric components with a Q value less than a preset threshold are deleted, and volumetric components with a Q value greater than or equal to the preset threshold are retained.

[0035] Specifically, the step S4 includes the following steps: In S401, according to the component-level fine model data set in S1, the detail level division rule in S2, and the component geometry, volumetric units are screened out, the number of surface patches N of each volumetric unit and the number of surface patches n of each component in the corresponding volumetric unit are obtained, and the minimum circumscribed bounding box volume v of each volumetric component in each volumetric unit and the minimum circumscribed bounding box volume V of the volumetric unit are calculated. The step S401 of calculating the minimum circumscribed bounding box volume v of the volumetric component specifically includes steps S401.1-S401.5: In S401.1, a first three-dimensional convex hull calculation is performed on the vertex three-dimensional coordinate set P of the volumetric component. ; The first three-dimensional convex hull C is defined as the minimum convex set of the vertex three-dimensional coordinate set P of the volumetric component: ; Where λ i is a linear combination coefficient, satisfying the non-negative constraint λ i ≥ 0 and the normalization condition: ; x represents any point in the three-dimensional volumetric component, and is any position point in the first three-dimensional convex hull C. R 3 represents a three-dimensional real number space, i.e., a set of all three-dimensional coordinate points of the volumetric component; m represents the number of vertices in the vertex three-dimensional coordinate set P of the body-shaped component, i.e., the total number of vertices of the body-shaped component; p i represents the i-th vertex in the vertex three-dimensional coordinate set P of the body-shaped component, which is the current processing point in the same vertex set in subsequent steps S401.1.2 and S401.1.4, ; The first three-dimensional convex hull calculation is performed, specifically including steps S401.1.1-S401.1.5: S401.1.1 selects four non-coplanar points from the vertex three-dimensional coordinate set P of the body-shaped component to construct an initial tetrahedron as the initial face set of the first three-dimensional convex hull C; S401.1.2 takes the points in the vertex three-dimensional coordinate set P of the body-shaped component as the current point p i , judges whether the current point is located inside the first three-dimensional convex hull C; if the current point is located inside the first three-dimensional convex hull C, the current point is discarded; if the current point is located outside the first three-dimensional convex hull C, the visible face set F of the current point to the first three-dimensional convex hull C is determined; S401.1.3 deletes the visible face set F to obtain an open boundary ring; S401.1.4 connects the points located outside the first three-dimensional convex hull C with the open boundary ring obtained in step S401.1.3 to generate a new face and update the first three-dimensional convex hull C; S401.1.5 repeats steps S401.1.2 to S401.1.4 until all vertices of the body-shaped component are processed, and the complete first three-dimensional convex hull C is obtained.

[0036] S401.2, traverses all the faces of the first three-dimensional convex hull C, and takes the normal vector of each face as a candidate direction set ; for each candidate direction d j , construct an orthogonal coordinate basis (a j , b j , c j ) perpendicular to d j ; wherein, ; Taking the space rectangular coordinate system as the reference system, and projecting the vertex three-dimensional coordinate set P of the body-shaped component into the space rectangular coordinate system, the projection coordinates (u j , v j , h j ) of the j-th vertex p j (x k , y k , z k ) of the body-shaped component in the space rectangular coordinate system are calculated: , , ; wherein “ ” denotes a dot product operation; S401.3, calculating the range of the vertex three-dimensional coordinate set P of the body-like member in the spatial rectangular coordinate system: ; wherein L j , W j , and H j respectively correspond to the edge length of the bounding box of the body-like member in the x-axis, y-axis, and z-axis directions; S401.4, calculating the bounding box volume v j of the body-like member in the candidate direction: ; S401.5, traversing all candidate directions, and taking the minimum value as the minimum circumscribed bounding box volume v of the body-like member: , in cubic meters; The calculation of the minimum circumscribed bounding box volume V of the body-like unit in step S401 specifically includes steps S401.6-S401.10: S401.6, regarding the body-like unit as a whole composed of multiple members, performing a second three-dimensional convex hull calculation on the vertex three-dimensional coordinate set P ’ of the body-like unit; ; defining the second three-dimensional convex hull C ’ as the minimum convex set of the vertex three-dimensional coordinate set P ’ of the body-like unit: ; wherein λ i is a linear combination coefficient, satisfying the non-negative constraint λ i ≥ 0 and the normalization condition: ; x ’ denotes an arbitrary point in the three-dimensional body-like unit, and is an arbitrary position point within the second three-dimensional convex hull C ’ ; R 3 denotes a three-dimensional real space, i.e., the set of all three-dimensional coordinate points of the body-like unit; m ’ denotes the number of vertices in the vertex three-dimensional coordinate set P ’ of the body-like unit, i.e., the total number of vertices of the body-like unit; denotes the vertex three-dimensional coordinate set P ’ of the body-like unit;the i-th vertex in the set of vertices in the body-like unit, the current processing point in the subsequent steps S401.6.2 and S401.6.4, ; performing a second three-dimensional convex hull calculation, specifically comprising steps S401.6.1-S401.6.5: S401.6.1, selecting four points not in the same plane from the set of vertex three-dimensional coordinates P ’ of the body-like unit to construct an initial tetrahedron as an initial set of face patches of the second three-dimensional convex hull C ’ ; S401.6.2, taking a point in the set of vertex three-dimensional coordinates P ’ of the body-like unit as a current point in turn, determining whether the current point is inside the second three-dimensional convex hull C ’ ; if the current point is inside the second three-dimensional convex hull C ’ , discarding the current point; if the current point is outside the second three-dimensional convex hull C ’ , determining a set of visible face patches F ’ of the second three-dimensional convex hull C ’ visible to the current point; S401.6.3, deleting the set of visible face patches F ’ to obtain an open boundary ring; S401.6.4, connecting the point outside the second three-dimensional convex hull C ’ to the open boundary ring obtained in step S401.6.3 to generate a new face patch and update the second three-dimensional convex hull C ’ ; S401.6.5, repeating steps S401.6.2-S401.6.4 until all vertices of the body-like unit are processed to obtain a complete second three-dimensional convex hull C ’ ; S401.7, traversing all face patches of the second three-dimensional convex hull C ’ to take the normal vector of each face patch as a set of candidate directions ; for each candidate direction , constructing an orthogonal coordinate basis perpendicular to ; wherein, ; taking a spatial rectangular coordinate system as a reference system and projecting the set of vertex three-dimensional coordinates P ’ of the body-like unit to the spatial rectangular coordinate system to obtain the projection coordinates of the j-th vertex of the body-like unit in the spatial rectangular coordinate system : , , ; wherein, represents a dot product operation; S401.8, calculate the set of three-dimensional coordinates P of the vertices of the body-like unit in the spatial rectangular coordinate system ’ ; ; wherein, respectively correspond to the edge length of the bounding box of the body-like unit in the x-axis, y-axis and z-axis directions; S401.9, calculate the volume V of the bounding box of the body-like unit in the candidate direction j : ; S401.10, traverse all candidate directions, and take the minimum value as the minimum circumscribed bounding box volume V of the body-like unit: , unit: cubic meters; S402, according to the structure type, purpose and function of the body-like component in the catenary BIM model, all body-like components are divided into several functional categories as shown in Table 2; On this basis, combined with the geometric complexity index and visualization importance level of each body-like component, a weight parameter table corresponding to the body-like component classification is set, and the semantic weight factor k of each body-like component is extracted. Among them, the weight factor k is numerically set according to the function level of the component in the engineering semantic standard, the exposure degree in the model view, and the interaction frequency in different scenes, and a multi-level scoring mechanism is used to assign values uniformly.

[0037] Table 2

[0038] S403, define a local index m of a component level, and at the same time, define a global index M to normalize the local index; ; ; S404, construct a comprehensive evaluation function Q for reflecting the importance of a body-like component in its belonging body-like component unit and the cost of its simplification, ; S405, calculate the Q value of each body-like component in each body-like unit one by one, and sort all body-like components according to the Q value; according to the model accuracy requirement, preset a threshold Q threshold ; when Q threshold , delete the corresponding body-like component; when Q threshold , retain the corresponding body-like component.

[0039] In this embodiment, Q threshold =1.00.

[0040] S5, as shown in Figure 4 shown in S4 is performed; the three-dimensional Minkowski operator is used to repair the topological abnormal area of the retained body-shaped component; Boolean operation is performed on the relationship of each body-shaped component to adjust the boundary relationship of each body-shaped component, and topological closure is realized; the standard triangular mesh data of the retained body-shaped component is optimized based on the geometric simplification algorithm of quadratic error metric (QEM), and the optimized body-shaped component model data is obtained.

[0041] Specifically, the step S5 includes the following steps: S501, the retained body-shaped component in S4 is obtained, topological consistency detection is performed, and topological abnormal areas including openings, gaps, intersections and overlaps are identified; S502, for the body-shaped component with topological abnormal areas, the three-dimensional Minkowski operator is used to perform inflation and corrosion operation to repair the topological abnormal area of the body-shaped component; S503, in the local spatial range, Boolean operation is performed on the relationship of each body-shaped component to adjust the boundary relationship of each body-shaped component, and topological closure is realized.

[0042] Boolean operation refers to performing AND, INTERSECT and DIFFERENT operations.

[0043] S504, mesh simplification processing is performed, the mesh compression algorithm based on quadratic error metric is used to optimize the standard triangular mesh data of the retained body-shaped component, and the vertices and triangular faces are recursively merged to control the error range and the number of faces.

[0044] S6, according to the detail level division rule in S2, the lightweight linear component model in S3 and the optimized body-shaped component model data in S5 are assembled and reorganized to generate hierarchical component structure, each component is converted into a three-dimensional instance model conforming to.i3dm format, tileset.json file is used to organize the model, and three-dimensional tile data structure meeting the 3D Tiles standard is output. In this embodiment, the total three-dimensional tile data set storage size is 15.6 MB, the lightweight processing result data amount is shown in Table 3, and the four-level detail level model of system, system, unit and component is shown in Figure 6 、 7 、8 and 9.

[0045] Table 3

[0046] S1-S5 belong to the process of building multi-level detail model, and S6 belongs to the process of three-dimensional tile structure organization.

[0047] The embodiments of the specific implementation are the preferred embodiments of the present application, not to limit the protection scope of the present application. Any equivalent changes made in the structure, shape and principle of the present application should be covered within the protection scope of the present application.

Claims

1. Method for converting lightweight three-dimensional tiles of a component-level catenary three-dimensional model, characterized in that, The method comprises the steps of: S1, traversing all components in the catenary BIM model, extracting the geometric information and attribute information of each component, establishing a one-to-one correspondence between each component and semantic information, and forming a component-level fine model dataset with semantic annotations; S2, defining four-level detail levels of system, system, unit and component according to the external control parameters and model semantic hierarchy of the catenary BIM model, establishing the mapping relationship between each component and the upper structure to which it belongs, and constructing the detail level division rule; S3, according to the component-level fine model dataset in S1, the detail level division rule in S2, and the component geometric form, filtering out linear units, and reconstructing a lightweight linear component model; S4, according to the component-level fine model dataset in S1, the detail level division rule in S2, and the component geometric form, filtering out body units, constructing a comprehensive evaluation function Q, sorting all body components according to the Q value, deleting body components with Q value less than a preset threshold, and retaining body components with Q value greater than or equal to a preset threshold; S5, performing topological consistency detection on the retained body components in S4; repairing the topological abnormal area of the retained body components; performing Boolean operation on the relationship of each body component, adjusting the boundary relationship of each body component, and realizing topological closure; optimizing the standard triangular mesh data of the retained body components, and obtaining the optimized body component model data; S6, according to the detail level division rule in S2, assembling and reorganizing the lightweight linear component model in S3 and the optimized body component model data in S5 to generate hierarchical component structures, converting each component into a three-dimensional instance model, and outputting a three-dimensional tile data structure.

2. The method of claim 1, wherein the method further comprises: The step S1 specifically comprises the following steps: S101, reading the catenary BIM model file, obtaining the external control parameters of the model, and setting key constraint parameters for subsequent processing; S102, traversing all components in the catenary BIM model, extracting the geometric information and attribute information of each component, and converting the geometric information of each component into standard triangular mesh data; S103, according to the component semantic description rule in the catenary BIM model, mapping the attribute information extracted in step S102 to the component semantic table, classifying according to the structure type field, purpose field and function level field, and generating standardized semantic description data; S104, assigning a globally unique identifier to each component, and binding the standard triangular mesh data of each component with its corresponding semantic information to form a component-level fine model dataset with semantic annotations.

3. The method of claim 2, wherein the method further comprises: The step S2 specifically comprises the steps of defining four-level detail levels of system, system, unit and component according to the model semantic hierarchy of the catenary BIM model and the component organization structure obtained in step S101, establishing the mapping relationship between each component and the upper structure to which it belongs, and constructing the detail level division rule.

4. The method of claim 1, wherein the method further comprises: The step S4 specifically comprises the following steps: S401, filtering out the volumetric unit according to the component-level fine model dataset in S1, the detail level division rule in S2, and the component geometry, obtaining the face number N of each volumetric unit and the face number n of each volumetric component in the corresponding volumetric unit, and calculating the minimum circumscribed bounding box volume v of each volumetric component in each volumetric unit and the minimum circumscribed bounding box volume V of the volumetric unit; S402, dividing all the volumetric components into several functional categories according to the structure type, purpose and function of the volumetric components in the catenary BIM model; on this basis, combining the geometric complexity index and the visualization importance level of each volumetric component, setting the weight parameter table corresponding to the volumetric component classification, and extracting the semantic weight factor k of each volumetric component; S403, defining a local index m of a component level, and simultaneously defining a global index M for normalizing the local index; ； ； S404, constructing a comprehensive evaluation function Q for reflecting the importance of a certain volumetric component in its corresponding volumetric component unit and the cost of its simplification; ； S405, calculate the Q value of each volumetric component in each volumetric unit one by one, and sort all volumetric components according to the Q value; according to the model accuracy requirement, preset threshold Q threshold ; when Q threshold , delete the corresponding volumetric component; when Q threshold , keep the corresponding volumetric component.

5. The method of claim 4, wherein the method further comprises: The minimum circumscribed bounding box volume v of the volumetric component in step S401 specifically includes steps S401.1-S401.5: S401.1, performing first three-dimensional convex hull calculation on the vertex three-dimensional coordinate set P of the volumetric component; ； The first three-dimensional convex hull C is defined as the minimum convex set of the vertex three-dimensional coordinate set P of the volumetric component: ； wherein λ i is a linear combination coefficient satisfying the non-negative constraint λ i ≥ 0 and the normalization condition: ; x represents any point in the three-dimensional space volumetric component, and is any position point in the first three-dimensional convex hull C; R 3 denotes the three-dimensional real space, i.e. the set of all three-dimensional coordinate points of the body-like component; m represents the number of vertices in the vertex three-dimensional coordinate set P of the volumetric component, i.e. the total number of vertices of the volumetric component; P i denotes the i-th vertex in the set P of three-dimensional coordinates of the vertices of the body-like member, which is the current processing point in both the subsequent steps S401.1.2 and S401.1.4, ; The first three-dimensional convex hull calculation specifically includes steps S401.1.1-S401.1.5: S401.1.1 selects four non-coplanar points from the vertex three-dimensional coordinate set P of the volumetric component to construct an initial tetrahedron as the initial face set of the first three-dimensional convex hull C; S401.1.2 take the points in the vertex three-dimensional coordinate set P of the body-shaped component in turn as the current point p i , judge whether the current point is located inside the first three-dimensional convex hull C; if the current point is located inside the first three-dimensional convex hull C, discard the current point; if the current point is located outside the first three-dimensional convex hull C, determine the set F of the visible facets of the first three-dimensional convex hull C for the current point; S401.1.3 deletes the visible face set F to obtain an open boundary ring; S401.1.4 connects the points outside the first three-dimensional convex hull C with the open boundary ring obtained in step S401.1.3 to generate new faces and update the first three-dimensional convex hull C; S401.1.5 repeats steps S401.1.2 to S401.1.4 until all vertices of the volumetric component are processed to obtain the complete first three-dimensional convex hull C; S401.2, traverse all facets of the first three-dimensional convex hull C, take the normal vector of each facet as a candidate direction set ; for each candidate direction d j , construct an orthogonal coordinate basis (a j , b j , c j ) perpendicular to d j ; wherein ; With the space rectangular coordinate system as the reference system, and the vertex three-dimensional coordinate set P of the body-shaped component is projected to the space rectangular coordinate system, the projection coordinates (u, v, h) of the jth vertex p of the body-shaped component in the space rectangular coordinate system are obtained by calculation: j (x j ,y j ,z j ) in the space rectangular coordinate system: k ,v k ,h k ). , , ； wherein, denotes a dot product operation; S401.3, calculate the range of the vertex three-dimensional coordinate set P of the body-shaped component in the spatial rectangular coordinate system: ; wherein L j W j H j respectively correspond to the edge length of the bounding box of the body-like member in the x-axis, y-axis and z-axis directions. S401.4, calculate the bounding box volume v of the volumetric component in the candidate direction j : ; S401.5, traverse all candidate directions, take the minimum value as the minimum circumscribed bounding box volume v of the volumetric component: .

6. The method of claim 4, wherein the method further comprises: calculating the minimum circumscribed bounding box volume of the body-shaped unit in the step S401 and specifically comprises steps S401.6-S401.10: S401.6, regarding the body unit as a whole composed of multiple components, the vertex three-dimensional coordinate set P of the body unit is calculated ’ performing a second three-dimensional convex hull calculation; ； The second three-dimensional convex hull C ’ is defined as the minimum convex set of the set of vertex three-dimensional coordinates P ’ of the body-shaped unit: ； where λ i is a linear combination coefficient satisfying the non-negative constraint λ i ≥ 0 and the normalization condition: ; x ’ represents an arbitrary point in the three-dimensional space body unit, and is an arbitrary position point in the second three-dimensional convex C ’ inner R 3 denotes the three-dimensional real space, i.e. the set of all three-dimensional coordinate points of the volume elements; m ’ a set of vertex three-dimensional coordinates P representing the volume-like unit ’ the number of vertices in the middle, i.e. the total number of vertices of the volume-like unit; a set of vertex three-dimensional coordinates P representing the body-shaped unit ’ the i-th vertex in the set of vertices, which in both subsequent steps S401.6.2 and S401.6.4 refers to the current processing point in the set of vertices, ; The second three-dimensional convex hull calculation specifically includes steps S401.6.1-S401.6.5: S401.6.1, from the set of vertex three-dimensional coordinates P of the body unit ’ four non-coplanar points are selected to construct an initial tetrahedron as the initial set of surface patches of the second three-dimensional convex hull C ’ ​ S401.6.2, take the vertex three-dimensional coordinate set P of the body-shaped unit in turn as the current point ’ , judge whether the current point is located inside the second three-dimensional convex hull C ’ ; if the current point is located inside the second three-dimensional convex hull C ’ , discard the current point; if the current point is located outside the second three-dimensional convex hull C ’ , determine the visible face sheet set F ’ of the second three-dimensional convex hull C ’ ;​ S401.6.3, delete the visible face sheet set F ’ , get the opening boundary ring; S401.6.4, the point outside is connected with the opening boundary ring obtained in step S401.6.3, a new face is generated and the second three-dimensional convex hull C is updated ’ S401.6.4, the point outside is connected with the opening boundary ring obtained in step S401.6.3, a new face is generated and the second three-dimensional convex hull C is updated ’ ; S401.6.5, repeat steps S401.6.2 to S401.6.4 until all vertexes of the body unit are processed, obtaining a complete second three-dimensional convex hull C ’ ; S401.7, traversing the second three-dimensional convex hull C ’ of all facets of C, take the normal vector of each facet as the candidate direction set ; for each candidate direction , construct an orthogonal coordinate basis ( ) perpendicular to ; wherein = ;​ With the space rectangular coordinate system as the reference system, and the vertex three-dimensional coordinate set P of the body-shaped unit ’ Projecting to the space rectangular coordinate system, the jth vertex of the body-shaped unit is calculated The projection coordinates in the space rectangular coordinate system : , , ； wherein, denotes a dot product operation; S401.8, calculate the vertex three-dimensional coordinate set P of the body unit under the space rectangular coordinate system ’ The range of the application is: ; wherein, respectively correspond to the edge length of the bounding box of the body-shaped unit in the x-axis, y-axis and z-axis directions. S401.9, calculate the bounding box volume V of the volumetric unit in the candidate direction j : ; S401.10, traverse all candidate directions, take the minimum value as the minimum circumscribed bounding box volume V of the volumetric unit: .

7. The method of claim 1, wherein the method further comprises: The step S5 specifically includes the following steps: S501, obtaining the volumetric components retained in S4, performing topological consistency detection to identify topological abnormal areas including openings, gaps, intersections and overlaps; S502, for the volumetric components with topological abnormal areas, performing inflation and corrosion operations using three-dimensional Minkowski operators to repair the topological abnormal areas of the volumetric components; S503, performing Boolean operations on the relationships of the volumetric components in a local spatial range to adjust the boundary relationships of the volumetric components and realize topological closure; S504, performing a mesh simplification process, based on a mesh compression algorithm of a quadratic error metric, optimizing the standard triangular mesh data of the retained volumetric component, recursively merging vertices and triangular faces, controlling the error range and the number of facets.

8. The method of claim 1, wherein the method further comprises: The step S3 specifically comprises the following steps: according to the component-level fine model data set in S1, the detail level division rule in S2, and the component geometry, filtering out linear units, performing a multi-segment straight line fitting operation on the linear component containing curve features, converting it into a combination representation of several equal-length straight line segments, extracting the key coordinate data of the linear component, and reconstructing a lightweight linear component model.

Images