Method and system for automatically degrading CIM6-level three-dimensional model to CIM5 based on semantic rule engine

Abstract

The invention relates to the field of CIMs, and discloses a method and system for automatically degrading a CIM6-level three-dimensional model to CIM5 based on a semantic rule engine, and the method comprises the steps: obtaining CIM6-level three-dimensional model data, and constructing a scene hierarchical tree based on the three-dimensional model data; identifying the category, the attribute, the service label and the importance constant of each object in the scene hierarchical tree; calculating geometric complexity, semantic importance and topological dependency of each object, and calculating a comprehensive weight; matching the interval of the comprehensive weight and executing a corresponding processing operation based on the matched interval; performing grid merging and texture baking operation on the reserved object; performing grid instantiation on the proxy object after proxy replacement; collision holes in the model are detected through an octree in combination with a GJK algorithm to complete topological consistency verification, a lightweight CIM5-level model data packet is generated, and a difference report is output. According to the method, the consistency of the key semantic information and the topology is kept while the data volume of the CIM6 model is reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of CIM, in particular to a method and system for automatically downgrading a CIM6-level three-dimensional model to a CIM5 based on a semantic rule engine. BACKGROUND

[0002] With the wide application of City Information Modeling (CIM) technology in smart cities, digital twins, urban planning and operation management, etc., the hierarchical expression system of three-dimensional city models has gradually improved.

[0003] At present, the internationally recognized CIM model classification standard divides city models into multiple levels of precision, among which the CIM6 level represents the highest precision Building Information Modeling (BIM) integrated level, usually corresponding to LOD4 and above, containing building external form, internal structure, functional space division, and detailed components such as equipment pipelines, with high geometric precision and rich semantic information, widely used in design review, construction simulation, and facility operation, etc.

[0004] However, high-precision models of CIM6 level usually have a large amount of data, with single building models reaching GB level and city-level models even reaching TB level. Such models face serious performance bottlenecks when used for large-scale city-level visualization rendering, real-time simulation deduction, lightweight display on Web, or collaborative application on mobile devices. For example, when loading large-scale CIM6 models in WebGL or Unity / Unreal engine, problems such as frame rate drop, memory overflow, and loading delay often occur, seriously affecting user experience and system response efficiency.

[0005] Therefore, how to realize the downgrading from CIM6 to a lower level such as CIM5 while ensuring the key semantic information and spatial topological relationship is a technical problem that needs to be overcome by those skilled in the art. SUMMARY

[0006] The embodiments of the present application provide a method for automatically downgrading a CIM6-level three-dimensional model to a CIM5 based on a semantic rule engine to at least partially solve the above technical problems.

[0007] In order to achieve the above-mentioned purpose, according to the first aspect of the present application, a method for automatically downgrading a CIM6-level three-dimensional model to a CIM5 based on a semantic rule engine is provided, comprising: obtaining three-dimensional model data of CIM6 level, and constructing a scene hierarchy tree based on the three-dimensional model data; identifying the class, attribute, business tag and importance constant of each object in the scene hierarchy tree; calculating the geometric complexity, semantic importance and topological dependency of each object, and calculating the comprehensive weight based on the geometric complexity, semantic importance and topological dependency; matching the interval of the comprehensive weight and performing corresponding processing operations based on the matched interval; wherein if the comprehensive weight is less than a first threshold, the object is deleted; if the comprehensive weight is not less than the first threshold and less than a second threshold, the object is proxy replaced; and if the comprehensive weight is not less than the second threshold, the object is retained; performing mesh merging and texture baking operations on the retained objects; and performing mesh instantiation on the proxy objects after proxy replacement; detecting collision voids in the model through octree combined with GJK algorithm to complete topological consistency verification, generating a lightweight CIM5 level model data package and outputting a difference report.

[0008] Optionally, the scene hierarchy tree is constructed based on the three-dimensional model data, comprising: parsing the input CIM6 level three-dimensional model data and extracting object instances, geometric representations, hierarchical structures and metadata information therefrom; constructing a scene hierarchy tree with a multi-level node structure according to the spatial organization relationship and semantic classification of the extracted object instances in the original building information model.

[0009] Optionally, the class, attribute, business tag and importance constant of each object in the scene hierarchy tree are identified, comprising: classifying each node in the scene hierarchy tree, extracting the class tag of each object and establishing a mapping relationship between the class and the semantic rule; extracting attribute values from the metadata of the object; the attribute values include geometric attributes, physical attributes, functional attributes and time attributes; matching the business tag of the object according to the business rule library; calculating the importance constant of the object based on a preset weight mapping table.

[0010] Optionally, the geometric complexity of each object is calculated, comprising: counting the number of triangular facets, the number of vertices and the mesh density of the object, wherein the mesh density is defined as the number of facets per unit volume; calculating the surface curvature complexity of the object based on the dihedral angle of adjacent triangular facets; obtaining the geometric redundancy based on the ratio of the volume of the spatial bounding box to the actual geometric volume of the object; calculating the semantic importance of the object comprises: According to the object class label, the parent node class label, the attribute value sequence, and the business label, a semantic similarity between the object and the parent node is calculated; In combination with the semantic category importance weight and the semantic space importance weight, a semantic importance of the object is comprehensively determined; The topological dependency of the object includes: The topological dependency of the object is determined according to the geometric complexity of the parent node, the number of child nodes, and the hierarchical relationship of the parent node by traversing the topological parent node set of the object.

[0011] Optionally, the object with a comprehensive weight less than the first threshold is deleted, including: The geometric representation of the object and all attribute values thereof are deleted; The scene hierarchical tree nodes of the parent object and the child object having a parent-child relationship with the object are deleted by traversing the topological connection relationship of the object; The topological structure of the scene hierarchical tree is updated, and the isolated nodes caused by the deletion operation are removed.

[0012] Optionally, the object with a comprehensive weight not less than the first threshold and less than the second threshold is replaced by a proxy, including: The geometric object instances in the scene hierarchical tree having a topological connection with the object are traversed, and a proxy candidate set meeting the condition is screened; The comprehensive weight of each candidate in the proxy candidate set is calculated, and the proxy candidate with the highest weight is selected; When there are multiple candidate combinations, the candidate combination with the smallest comprehensive weight is selected as the final proxy; The original object and the proxy object instance are deleted, and the attribute values and geometric instances of the candidate combination are retained in the scene hierarchical tree.

[0013] Optionally, the grid merging includes: Based on topological consistency detection, geometric object instances are screened from sibling nodes as merging candidates; Through octree space division technology, the vertex coplanar, the face coplanar, and the common vertex between the candidate geometric object instances are detected to generate a topological connection line and record the detection results; According to the detection results, the candidate objects meeting the following conditions are screened: there are shared vertices or edges between geometric faces, the face normal vector directions are consistent or the included angle is less than a preset threshold, and the geometric surface continuity meets the merging requirements; The screened merging candidate objects are prioritized based on geometric complexity, topological connection density, and semantic similarity, and the candidate objects with high priority are preferentially merged.

[0014] Optionally, the grid instantiation includes: The geometric object instances in the topological object set are traversed; If the geometry instance is a model proxy and the geometry complexity is lower than a preset threshold, the geometry instance is converted into a mesh instance; The original geometry object instance is deleted, and an actual geometry representation instance is added in the scene hierarchy tree.

[0015] Optionally, the texture baking includes: Based on the topological consistency detection, geometry object instances are screened from sibling nodes as baking candidates; The screened baking candidates are prioritized; the prioritization is based on material similarity, texture resolution, lighting consistency, and semantic category matching degree; In order of priority from high to low, UV data replication baking operations are sequentially performed; in the baking process, texture information of a high-priority candidate object is preferentially used as the basis of a main texture atlas, and texture of a low-priority object is aligned and merged into the main atlas; The merged texture atlas is layout-optimized to reduce texture switching times and preserve metadata information of the baked texture; The baked uniform texture is mapped to the corresponding geometry instance, and the scene hierarchy tree is updated.

[0016] According to a second aspect of the present application, a system for automatically downgrading a CIM6-level three-dimensional model to CIM5 based on a semantic rule engine is provided, including: A system for automatically downgrading a CIM6-level three-dimensional model to CIM5 based on a semantic rule engine, characterized by comprising: A first processing module configured to: acquire three-dimensional model data of CIM6 level, and construct a scene hierarchy tree based on the three-dimensional model data; A second processing module configured to: identify the category, attribute, business label, and importance constant of each object in the scene hierarchy tree; A third processing module configured to: calculate the geometry complexity, semantic importance, and topological dependence of each object, and calculate a comprehensive weight based on the geometry complexity, semantic importance, and topological dependence; A fourth processing module configured to: match the interval of the comprehensive weight and perform a corresponding processing operation based on the matched interval; wherein if the comprehensive weight is less than a first threshold, the object is deleted; if the comprehensive weight is not less than the first threshold and less than a second threshold, the object is replaced by a proxy; and if the comprehensive weight is not less than the second threshold, the object is retained; A fifth processing module configured to: perform mesh merging and texture baking operations on the retained object; and perform mesh instantiation on the proxy object after the proxy replacement; The sixth processing module is configured to detect collision voids in the model by using an octree combined with a GJK algorithm to complete a topological consistency check, generate a lightweight CIM5-level model data packet, and output a difference report.

[0017] To sum up, the embodiments of the present application reduce the volume of CIM6 model data while maintaining key semantic information and topological consistency by using the above technical solutions.

[0018] Other features and advantages of the present application will be described in detail in the following specific embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor on the basis of these drawings.

[0020] Figure 1 is a step flow chart of a method for automatically downgrading a CIM6-level three-dimensional model to a CIM5-level based on a semantic rule engine provided in an exemplary embodiment of the present application; Figure 2 is a system schematic diagram of a system for automatically downgrading a CIM6-level three-dimensional model to a CIM5-level based on a semantic rule engine provided in an exemplary embodiment of the present application; Explanation of reference numerals: 01, first processing module; 202, second processing module; 203, third processing module; 204, fourth processing module; 205, fifth processing module; 206, sixth processing module. DETAILED DESCRIPTION

[0021] The technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the protection scope of the present application.

[0022] The present application provides a method for automatically downgrading a CIM6-level three-dimensional model to a CIM5-level based on a semantic rule engine, please refer to Figure 1 The method for automatically downgrading a CIM6-level three-dimensional model to a CIM5-level based on a semantic rule engine provided in the embodiments of the present application includes the following steps: Step 101, obtain a three-dimensional model data of a CIM6 level, and construct a scene hierarchy tree based on the three-dimensional model data.

[0023] Step 102, identify the class, attribute, business tag and importance constant of each object in the scene hierarchy tree.

[0024] Step 103, calculate the geometric complexity, semantic importance and topological dependency of each object, and calculate the comprehensive weight based on the geometric complexity, semantic importance and topological dependency.

[0025] Step 104, match the interval of the comprehensive weight and execute the corresponding processing operation based on the matched interval; if the comprehensive weight is less than the first threshold, the object is deleted; if the comprehensive weight is not less than the first threshold and less than the second threshold, the object is replaced by a proxy; if the comprehensive weight is not less than the second threshold, the object is retained.

[0026] Step 105, perform mesh merging and texture baking operations on the retained objects; perform mesh instantiation on the proxy objects after proxy replacement.

[0027] Step 106, detect collision voids in the model by octree combined with GJK algorithm to complete topological consistency verification, generate lightweight CIM5 level model data package and output difference report.

[0028] Specifically, the system loads the input CIM6 level three-dimensional model data through the data import module, supports mainstream BIM data sources such as IFC format and Revit project file; CIM6 level model usually corresponds to the highest precision level in building information model, LOD4 and above, contains building facade, internal structural members, functional space division and equipment pipeline system and other fine-grained geometric and semantic information, and is suitable for fine design, construction simulation and facility operation scenarios. After import, the system analyzes the object instances, geometric representations, hierarchical organization structure and metadata information in the model, and constructs a scene hierarchy tree with multi-level node structure according to its spatial inclusion relationship and semantic classification.

[0029] After the scene hierarchy tree is established, the system starts the semantic analysis module and classifies each node object based on the pre-defined multi-level semantic ontology. The semantic ontology is constructed using OWL2 language, covering seven levels of CIM1 to CIM7, seven categories and 138 subcategories, and labeling semantic importance weight for each object category, such as wall importance constant 0.7. The system extracts metadata information of object class label, geometric attribute, physical attribute, functional attribute and time attribute; matches the business label of the object combined with the external business rule library, such as "fire control key area" and "operation key equipment", to enhance the context adaptability of semantic judgment. The system calculates the importance constant of each object according to the class label and the corresponding weight mapping table.

[0030] The system calculates the geometric complexity, semantic importance, and topological dependency for each object o respectively; the geometric complexity G(o) reflects the density of object geometric details, which is measured by counting the number of triangular facets, the number of vertices, and the density of facets per unit volume; the surface curvature complexity is calculated based on the dihedral angle between adjacent triangular facets, and the geometric redundancy is evaluated by combining the ratio of the bounding box volume to the actual geometric volume of the object, and finally normalized to G(o) ∈ [0, 1]; the semantic importance S(o) is calculated by considering the importance of the object category, the influence of the parent node category, the attribute sequence features, and the business label, and combining the semantic category weight and the spatial distribution weight, S(o) ∈ [0, 1]; The topological dependency T(o) measures the connection dependency of the object in the scene structure, which is obtained by traversing the topological parent node set, counting the number of adjacent objects, and combining the parent node level depth for normalization processing, T(o) ∈ [0, 1]; the system uses linear weighting method to fuse the three indicators to calculate the comprehensive weight W(o) of the object.

[0031] If the comprehensive weight is less than the first threshold value, it is determined as a low-value object, and a deletion operation is performed to remove its geometric and attribute data and update the topological connection in the scene hierarchy tree to avoid isolated nodes; if the comprehensive weight is not less than the first threshold value and less than the second threshold value, it is determined as a medium-value object, and a proxy replacement is performed, that is, a simplified geometric proxy (such as an OBB axis-aligned bounding box or a low-polygon LOD model) is generated to retain the basic space placeholder and key attributes; if the comprehensive weight is not less than the second threshold value, it is determined as a high-value object, and a high-precision retention is performed to maintain the original geometric precision and complete semantic information.

[0032] Through the above scheme, the CIM6 model data volume is reduced while maintaining the key semantic information and topological consistency.

[0033] In some embodiments, a scene hierarchy tree is constructed based on the three-dimensional model data, including; The input CIM6 level three-dimensional model data is parsed and the object instances, geometric representations, hierarchical structure, and metadata information are extracted therefrom; The extracted object instances are constructed into a scene hierarchy tree with a multi-level node structure according to their spatial organization relationship and semantic classification in the original building information model.

[0034] Specifically, the input CIM6 level three-dimensional model data is parsed to extract object instances, geometric representations, hierarchical structures, and accompanying metadata information; the object instances include walls, doors, windows, and equipment; the geometric representations include triangular meshes and NURBS surfaces; the hierarchical structures include the inclusion relationships of buildings, floors, spaces, and components; and the metadata information includes IFC attribute sets and custom labels. The extracted object instances are organized according to their spatial organization relationships and semantic classifications in the original building information model, and are organized in a hierarchical structure of parent and child nodes to construct a scene hierarchical tree with a multi-level node structure. The spatial organization relationships include positions, enclosures, and connections.

[0035] In some embodiments, the category, attribute, business tag, and importance constant of each object in the scene hierarchical tree are identified, including: Each node in the scene hierarchical tree is classified, the category tag of each object is extracted, and a mapping relationship between the category and the semantic rule is established; Attribute values are extracted from the metadata of the object; the attribute values include geometric attributes, physical attributes, functional attributes, and time attributes; The business tag of the object is matched according to the business rule library; The importance constant of the object is calculated based on a preset weight mapping table.

[0036] Specifically, each node object in the scene hierarchical tree is classified, the category of the object is extracted by matching a predefined semantic ontology, and a mapping relationship between the category and a simplified strategy in the rule engine is established. Multiple types of attribute values are extracted from the metadata of the object, including geometric attributes such as length, area, and volume; physical attributes such as material density and thermal conductivity; functional attributes such as whether it is load-bearing or openable; and time attributes such as construction date and update time.

[0037] The business tag of the object is matched according to an externally configured business rule library, for example, fire passages cannot be deleted, core equipment needs to be preserved with high fidelity, and key facilities and safety-related facilities are matched; the importance constant of each object is calculated by combining the category importance and the business tag weighting based on a preset weight mapping table, the importance constant is used to quantify the key degree of the object in the city function and business application, and ensures that key city function objects are preferentially preserved in the model degradation process.

[0038] In some embodiments, the geometric complexity of each object is calculated, including: The number of triangular facets, the number of vertices, and the mesh density of the object are counted, where the mesh density is defined as the number of facets per unit volume; The surface curvature complexity of the object is calculated based on the dihedral angle of adjacent triangular facets; The geometric redundancy is obtained based on the ratio of the volume of the spatial bounding box to the actual geometric volume. The semantic importance of the object is calculated, including: According to the category label of the object, the category label of the parent node, the attribute value sequence, and the business label, the semantic similarity between the object and the parent node is calculated; The semantic importance of the object is determined comprehensively in combination with the semantic category importance weight and the semantic space importance weight; The topological dependence of the object is calculated, including: The topological parent node set of the object is traversed, and the topological dependence of the object is determined according to the geometric complexity of the parent node, the number of child nodes, and the hierarchical relationship of the parent node.

[0039] Specifically, the number of triangular facets, the number of vertices, and the mesh density of the object are counted; the mesh density is the number of triangular facets in a unit volume, which is used to measure the density of geometric details; the surface curvature complexity of the object is calculated based on the dihedral angle between adjacent triangular facets to reflect the shape fluctuation and smoothness; the geometric redundancy is the ratio of the minimum bounding box volume to the actual geometric volume of the object, and the larger the ratio, the more sparse the space utilization and the higher the redundancy; the above indexes are normalized and weighted to obtain the geometric complexity G(o) of the object.

[0040] According to the category label of the object, the category label of its parent node, the attribute value sequence of itself, and the matched business label, the semantic similarity between the object and its parent node is calculated; in combination with the preset semantic category importance weight and the semantic space importance weight, the semantic importance S(o) of the object is determined comprehensively through a weighted model to ensure that the key function and business related objects obtain higher scores.

[0041] The topological parent node set to which the object belongs is traversed, and the connection relationship in the structure hierarchy is analyzed; according to the geometric complexity of the parent node, the number of child nodes contained by the parent node, and the hierarchical depth of the parent node in the entire scene hierarchy tree, the dependence strength of the object on the upper structure is comprehensively evaluated, and the topological dependence is obtained after normalization, which is used to identify key nodes with high correlation in spatial organization or system logic.

[0042] In some embodiments, deleting the object with a comprehensive weight less than the first threshold value includes: The geometric representation of the object and all attribute values thereof are deleted; The scene hierarchy tree nodes of the child object and the parent object having a parent-child relationship with the object are deleted by traversing the topological connection relationship of the object; The topological structure of the scene hierarchy tree is updated to remove isolated nodes caused by the deletion operation.

[0043] In some embodiments, the object with a comprehensive weight not less than the first threshold value and less than the second threshold value is replaced by a proxy, including: traverse the geometric object instances in the scene hierarchy tree that have topological connections with the object, and filter a set of proxy candidates that meet the conditions; calculate the comprehensive weights of the candidates in the set of proxy candidates, and select the proxy candidate with the highest weight as the original object; when there are multiple candidate combinations, select the candidate combination with the smallest comprehensive weight as the final proxy; delete the original object and the proxy object instance, and retain the attribute values and geometric instances of the candidate combination in the scene hierarchy tree.

[0044] In some embodiments, the mesh merging includes: based on topological consistency detection, filtering geometric object instances from sibling nodes as merging candidates; detecting vertex coplanar, face coplanar, and common vertex between the candidate geometric object instances through octree spatial partitioning technology, generating topological connection lines, and recording the detection results; according to the detection results, filtering the candidate objects that meet the following conditions: there are shared vertices or edges between geometric faces, the directions of face normal vectors are consistent or the included angle is less than a preset threshold, and the geometric surface continuity meets the merging requirements; prioritizing the filtered merging candidate objects based on geometric complexity, topological connection density, and semantic similarity, and preferentially merging high-priority candidate objects.

[0045] In some embodiments, the mesh instantiation includes: traversing the geometric object instances in the topological object set; if the geometric instance is a model proxy and the geometric complexity is lower than a preset threshold, converting it into a mesh instance; deleting the original geometric object instance, and adding the actual geometric representation instance in the scene hierarchy tree.

[0046] In some embodiments, the texture baking includes: based on topological consistency detection, filtering geometric object instances from sibling nodes as baking candidates; prioritizing the filtered baking candidate objects; the sorting basis includes material similarity, texture resolution, lighting consistency, and semantic category matching degree; in order from high to low priority, sequentially performing UV data replication baking operations; during the baking process, preferentially using the texture information of high-priority candidate objects as the basis for the main texture atlas, and aligning and merging the texture of low-priority objects to the main atlas; performing layout optimization on the merged texture atlas to reduce the number of texture switches and retain the metadata information of the baked texture; mapping the baked uniform texture to the corresponding geometric instance and updating the scene hierarchy tree.

[0047] Referring to Figure 2 The second embodiment of the present application provides a system for automatically downgrading a CIM6-level three-dimensional model to CIM5 based on a semantic rule engine, comprising: A first processing module 201 is configured to acquire three-dimensional model data at the CIM6 level, and construct a scene hierarchy tree based on the three-dimensional model data. A second processing module 202 is configured to identify the category, attribute, business tag and importance constant of each object in the scene hierarchy tree. A third processing module 203 is configured to calculate the geometric complexity, semantic importance and topological dependence of each object, and calculate a comprehensive weight based on the geometric complexity, semantic importance and topological dependence. A fourth processing module 204 is configured to match the interval of the comprehensive weight and perform a corresponding processing operation based on the matched interval; if the comprehensive weight is less than a first threshold value, the object is deleted; if the comprehensive weight is not less than the first threshold value and less than a second threshold value, the object is replaced by a proxy; and if the comprehensive weight is not less than the second threshold value, the object is retained. A fifth processing module 205 is configured to perform mesh merging and texture baking operations on the retained object, and perform mesh instantiation on the proxy object after proxy replacement. A sixth processing module 206 is configured to detect collision voids in the model by octree combined with GJK algorithm to complete topological consistency verification, generate a lightweight CIM5-level model data package and output a difference report.

[0048] It should be noted that the system for automatically downgrading a CIM6-level three-dimensional model to CIM5 based on a semantic rule engine provided by the embodiments of the present application is used to perform all process steps of the method for automatically downgrading a CIM6-level three-dimensional model to CIM5 based on a semantic rule engine of the above-mentioned embodiments, and the working principles and beneficial effects of the two are one-to-one corresponding, thus no longer being described in detail.

[0049] In the description of the present application, the terms "first", "second" are only used for descriptive purposes, and cannot be understood or implied to indicate or imply relative importance or implicitly indicate the number of indicated technical features. Therefore, the features defined as "first", "second" can explicitly or implicitly include one or more features. In the description of the present application, the meaning of "multiple" is two or more, unless otherwise specifically limited.

[0050] In the above embodiments, the description of each embodiment has its own emphasis, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0051] The embodiments, implementation manners and related technical features of the present application can be combined with each other without conflict.

[0052] The above is only a preferred embodiment of the present application, and does not limit the present application in any form. Any simple modification, equivalent change and modification made to the above embodiment according to the technical essence of the present application without departing from the technical solution content of the present application still falls within the scope of the technical solution of the present application.

Claims

1. A method for automatic downgrading of a CIM Level 6 semantic rule engine based 3D model to CIM Level 5, characterized in that, The method comprises the following steps: acquiring three-dimensional model data of a CIM6 level, and constructing a scene hierarchical tree based on the three-dimensional model data; identifying the category, attribute, business tag and importance constant of each object in the scene hierarchical tree; calculating the geometric complexity, semantic importance and topological dependency of each object, and calculating the comprehensive weight based on the geometric complexity, semantic importance and topological dependency; matching the interval of the comprehensive weight and performing corresponding processing operations based on the matched interval; if the comprehensive weight is less than a first threshold value, the object is deleted; if the comprehensive weight is not less than the first threshold value and less than a second threshold value, the object is replaced by a proxy; if the comprehensive weight is not less than the second threshold value, the object is retained; performing mesh merging and texture baking operations on the retained object; and performing mesh instantiation on the proxy object after the proxy replacement; detecting collision voids in the model by using an octree combined with a GJK algorithm to complete topological consistency verification, generating a lightweight CIM5 level model data package and outputting a difference report.

2. The method of claim 1, wherein, The method for constructing a scene hierarchical tree based on the three-dimensional model data comprises the following steps: parsing the input CIM6 level three-dimensional model data and extracting object instances, geometric representations, hierarchical structures and metadata information therefrom; constructing a scene hierarchical tree with a multi-level node structure according to the spatial organization relationship and semantic classification of the extracted object instances in the original building information model.

3. The method of claim 2, wherein, The method for identifying the category, attribute, business tag and importance constant of each object in the scene hierarchical tree comprises the following steps: classifying each node in the scene hierarchical tree, extracting the category tag of each object and establishing a mapping relationship between the category and the semantic rule; extracting attribute values from the metadata of the object; the attribute values include geometric attributes, physical attributes, functional attributes and time attributes; matching the business tag of the object according to a business rule library; calculating the importance constant of the object based on a preset weight mapping table.

4. The method of claim 3, wherein, The method for calculating the geometric complexity of each object comprises the following steps: counting the number of triangular facets, the number of vertices and the mesh density of the object, wherein the mesh density is defined as the number of facets in a unit volume; calculating the surface curvature complexity of the object based on the dihedral angle of adjacent triangular facets; obtaining the geometric redundancy of the object based on the ratio of the volume of the spatial bounding box to the actual geometric volume. The method for calculating the semantic importance of the object comprises the following steps: calculating the semantic similarity between the object and the parent node according to the category tag of the object, the category tag of the parent node, the sequence of attribute values and the business tag; comprehensively determining the semantic importance of the object in combination with the semantic category importance weight and the semantic space importance weight. The method for calculating the topological dependency of the object comprises the following steps: traversing the topological parent node set of the object, and determining the topological dependency of the object according to the geometric complexity of the parent node, the number of child nodes and the hierarchical relationship of the parent node.

5. The method of claim 4, wherein, The method for deleting the object with a comprehensive weight less than a first threshold value comprises the following steps: deleting the geometric representation of the object and all attribute values thereof; traversing the topological connection relationship of the object, and deleting the scene hierarchical tree nodes of the child object and the parent object having a parent-child relationship with the object; updating the topological structure of the scene hierarchical tree, and removing isolated nodes caused by the deletion operation.

6. The method of claim 5, wherein, The proxy replacement of the object with a comprehensive weight not less than a first threshold and less than a second threshold comprises: traversing geometric object instances in the scene hierarchy tree that are topologically connected to the object, and screening a proxy candidate set that meets the condition; calculating the comprehensive weight of each candidate in the proxy candidate set, and selecting a proxy candidate with the highest comprehensive weight of the original object; when there are multiple candidate combinations, selecting a candidate combination with the smallest comprehensive weight as the final proxy; deleting the original object and the proxy object instance, and retaining the attribute values and geometric instances of the candidate combination in the scene hierarchy tree.

7. The method of claim 6, wherein, The mesh merging comprises: based on topological consistency detection, screening geometric object instances from sibling nodes as merging candidates; detecting vertex coplanar, face coplanar, and common vertex between the candidate geometric object instances by octree spatial division technology, generating topological connection lines, and recording the detection results; according to the detection results, screening the candidate objects that meet the following conditions: there are shared vertices or edges between geometric faces, the face normal vector directions are consistent or the included angle is less than a preset threshold, and the geometric surface continuity meets the merging requirement; prioritizing the screened merging candidate objects based on geometric complexity, topological connection density, and semantic similarity, and preferentially merging high-priority candidate objects.

8. The method of claim 7, wherein, The mesh instantiation comprises: traversing geometric object instances in the topological object set; if the geometric instance is a model proxy and the geometric complexity is lower than a preset threshold, converting it into a mesh instance; deleting the original geometric object instance, and adding an actual geometric representation instance in the scene hierarchy tree.

9. The method of claim 8, wherein, The texture baking comprises: based on topological consistency detection, screening geometric object instances from sibling nodes as baking candidates; prioritizing the screened baking candidate objects; the sorting basis comprises material similarity, texture resolution, lighting consistency, and semantic category matching degree; in order from high to low priority, sequentially performing UV data copy baking operations; in the baking process, preferentially taking the texture information of a high-priority candidate object as the basis of a main texture atlas, and aligning and merging the texture of a low-priority object to the main atlas; performing layout optimization on the merged texture atlas to reduce the number of texture switches and retain the metadata information of the baked texture; mapping the baked uniform texture to the corresponding geometric instance and updating it to the scene hierarchy tree.

10. A system for automatic downgrading of a CIM Level 6 semantic rule engine based 3D model to CIM Level 5, characterized in that, comprises: a first processing module configured to: acquire CIM6-level three-dimensional model data, and construct a scene hierarchy tree based on the three-dimensional model data; a second processing module configured to: identify the category, attribute, business label, and importance constant of each object in the scene hierarchy tree; a third processing module configured to: calculate the geometric complexity, semantic importance, and topological dependence of each object, and calculate a comprehensive weight based on the geometric complexity, semantic importance, and topological dependence; a fourth processing module configured to: match the interval of the comprehensive weight, and perform a corresponding processing operation based on the matched interval; if the comprehensive weight is less than a first threshold, the object is deleted; if the comprehensive weight is not less than the first threshold and less than a second threshold, the object is replaced by a proxy; if the comprehensive weight is not less than the second threshold, the object is retained. The fifth processing module is configured to perform a mesh merging and a texture baking operation on the retained object; and perform mesh instantiation on the agent object after the agent replacement. The sixth processing module is configured to detect collision voids in the model by using an octree combined with a GJK algorithm to complete a topological consistency check, generate a lightweight CIM5 level model data packet, and output a difference report.

Images