Car cloud showroom virtual display system and method based on dismounting data

Abstract

The application provides a car cloud showroom virtual display system and method based on disassembly data. It includes a data access module for generating structured evidence data sets; a double graph construction module for constructing disassembly evidence graphs based on structured evidence data sets and constructing display scene graphs based on vehicle three-dimensional scenes; a multi-representation asset generation module for generating geometric assets and texture assets of different accuracies for component identification; a semantic incremental release module for generating baseline snapshots and component identification-oriented incremental update packages according to the display scene graph; a transmission scheduling module for performing semantic priority sorting on the baseline snapshot and the incremental update package and issuing them in channels; and an end-side rendering interaction module for loading geometric assets and texture assets, analyzing incremental update packages to update scene node states, and responding to component interaction queries according to mapping relationships. The application can improve evidence consistency with display, improve incremental update efficiency, and enhance weak network interaction stability.

Description

Technical Field

[0001] This application relates to the fields of digital and network technology, and in particular to a virtual display system and method for a cloud-based car showroom based on dismantled vehicle data. Background Technology

[0002] With the accelerated digital transformation of the automotive aftermarket, cloud showrooms based on online 3D scenarios are gradually becoming a crucial infrastructure for in-store marketing and remote sales guidance. Through 3D modeling, panoramic display, and interactive explanations of vehicles, users can view cars, inspect components, and interact with suppliers online via mobile devices or web pages. Simultaneously, the continuous accumulation of dismantling data, disassembly reports, and testing materials provides a quantifiable data foundation for understanding vehicle structure, describing component conditions, and providing differentiated sales guidance. Deeply integrating dismantling data with virtual exhibitions has become a key direction for further enhancing the professionalism and credibility of automotive cloud showrooms.

[0003] Existing technologies typically focus on 3D rendering, cross-platform access compatibility, or general content distribution, often employing static material arrangement or page-level updates. The lack of a unified identifier between disassembled evidence and 3D scene objects makes it difficult to establish a computable mapping between exhibit content and data sources. In the data update stage, existing solutions lack sufficient support for component-level changes and evidence-level revisions, hindering fine-grained incremental releases based on component identifiers. In the transmission stage, existing solutions largely rely on general bandwidth adaptive strategies, lacking priority scheduling mechanisms that incorporate user interaction semantics, leading to delays in the arrival of critical interactive content in complex network environments. Furthermore, in multi-store collaborative distribution scenarios, the capabilities for authorization verification, source tracing, and call auditing of disassembled data, asset indexes, and update packages are insufficient, failing to meet the needs of reliable data flow and refined operations. Summary of the Invention

[0004] In view of this, this application provides a virtual display system and method for a car cloud showroom based on vehicle dismantling data, in order to solve the problems of inconsistent evidence and scene mapping, coarse incremental release granularity, and disordered distribution of key content under weak network conditions in the prior art.

[0005] The first aspect of this application provides a virtual car showroom system based on vehicle dismantling data, comprising: a data access module for acquiring vehicle dismantling data, dismantling reports, and display materials; performing field mapping, unit normalization, and version marking to generate a structured evidence dataset; a dual-image construction module for constructing a dismantling evidence image based on the structured evidence dataset and a display scene image based on the vehicle's 3D scene, and establishing a mapping relationship between the dismantling evidence image and the display scene image according to component identifiers; a multi-representation asset generation module for generating geometric assets and texture assets of different precision for component identifiers, and generating an asset index associated with the component identifiers; and semantics. The incremental release module generates a baseline snapshot based on the displayed scene graph and generates incremental update packages oriented towards component identifiers based on mapping relationships and version tags. The transmission scheduling module performs semantic priority sorting on the baseline snapshot and incremental update packages based on terminal capability parameters, network status parameters, and user interaction trajectories, and distributes them through different channels. The edge rendering interaction module loads the corresponding geometric assets and texture assets according to the asset index, parses the incremental update packages to update the scene node status, and responds to component interaction queries according to mapping relationships. The authorization audit module performs authorization verification, source recording, and call log association storage on the structured evidence dataset, asset index, and incremental update packages.

[0006] The second aspect of this application provides a virtual display method for a car cloud showroom based on vehicle dismantling data, using the system of the first aspect. The method includes: acquiring vehicle dismantling data, dismantling reports, and display materials; performing field mapping, unit normalization, and version marking to generate a structured evidence dataset; constructing a dismantling evidence map based on the structured evidence dataset and a display scene map based on the vehicle's 3D scene; establishing a mapping relationship between the dismantling evidence map and the display scene map according to component identifiers; generating geometric and texture assets of different precision for component identifiers and generating an asset index associated with the component identifiers; generating a baseline snapshot based on the display scene map and generating an incremental update package for the component identifiers based on the mapping relationship and version marking; performing semantic priority sorting on the baseline snapshot and incremental update package based on terminal capability parameters, network status parameters, and user interaction trajectories, and distributing them through different channels; loading the corresponding geometric and texture assets according to the asset index, parsing the incremental update package to update the scene node status, and responding to component interaction queries according to the mapping relationship; and performing authorization verification, source recording, and call log association storage on the structured evidence dataset, asset index, and incremental update package.

[0007] The above-described technical solutions adopted in the embodiments of this application can achieve the following beneficial effects: The system employs several modules: a data access module to acquire vehicle disassembly data, disassembly reports, and display materials; a field mapping module to perform unit normalization and version marking; a dual-graph construction module to construct disassembly evidence graphs based on the structured evidence dataset and display scene graphs based on the vehicle's 3D scene; and a mapping relationship between the disassembly evidence graphs and display scene graphs based on component identifiers. A multi-representation asset generation module generates geometric and texture assets of varying precision for each component identifier and generates asset indexes associated with those identifiers. A semantic incremental publishing module generates baseline snapshots based on the display scene graphs and generates incremental update packages for each component identifier based on the mapping relationship and version marking. A transmission scheduling module prioritizes the baseline snapshots and incremental update packages based on terminal capability parameters, network status parameters, and user interaction trajectories, and distributes them via different channels. An edge-side rendering interaction module loads corresponding geometric and texture assets based on the asset index, parses the incremental update packages to update scene node states, and responds to component interaction queries according to the mapping relationship. An authorization audit module performs authorization verification, source recording, and associated storage of call logs for the structured evidence dataset, asset index, and incremental update packages. This application can improve the consistency of evidence and presentation, increase the efficiency of incremental updates, and enhance the stability of weak network interactions. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 This is a schematic diagram of the structural composition of the car cloud showroom virtual display system based on vehicle dismantling data provided in the embodiments of this application; Figure 2 This is a flowchart illustrating the virtual display method for a car cloud showroom based on vehicle dismantling data provided in this application embodiment. Detailed Implementation

[0010] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0011] Currently, most automotive cloud showroom solutions revolve around 3D modeling, panoramic roaming, cross-platform access, and universal content distribution, enabling basic online car viewing and interactive navigation. Some solutions support material updates and multi-terminal adaptation, but updates are typically at the page or resource package level. There is a lack of a unified and calculable correlation mechanism between component nodes in the 3D scene and evidence items in the disassembly report. This results in a difficulty in achieving a one-to-one correspondence between exhibition content and disassembly evidence, and makes it difficult to quickly and accurately synchronize version changes on the data side to the scene side.

[0012] Under the aforementioned technical approach, at least three key issues exist. First, the mapping between evidence data and scene objects is inconsistent, creating a consistency gap where data is "displayed but difficult to trace." Second, the update granularity is coarse; when component-level evidence or assets change, large-scale resending is often required, leading to release latency and transmission burden. Third, the distribution strategy in weak network environments mainly relies on general bandwidth adaptation, lacking priority control based on interactive semantics, resulting in unstable arrival order of key nodes and key evidence currently of user interest, leading to significant fluctuations in the interactive experience.

[0013] To address the aforementioned issues, this application employs an integrated technical solution encompassing "data access standardization, dual-graph construction, multi-representation assets, semantic incremental publishing, semantic scheduling and distribution, edge-side parsing and interaction, and authorization auditing." Specifically, the decomposed data, reports, and display materials are first standardized to form a structured evidence dataset. Then, a decomposed evidence graph and a display scene graph are constructed separately, and a cross-graph mapping is established using component identifiers. Subsequently, geometric and texture assets of varying precision, along with their asset indexes, are generated around the component identifiers. During the publishing phase, a baseline snapshot is formed based on the display scene graph, and incremental update packages oriented towards component identifiers are generated by combining mapping relationships and version tags. During the transmission phase, terminal capability parameters, network status parameters, and user interaction trajectories are introduced, semantic priority sorting is performed, and distribution is carried out through different channels. On the edge, hierarchical loading is performed according to the asset index, and node status is updated according to the incremental operation sequence. Finally, authorization verification, source records, and call logs are associated and stored for the structured evidence dataset, asset index, and incremental update packages.

[0014] Through the above-mentioned technical solution of this application, due to the establishment of a dual-graph mapping relationship driven by component identifiers, evidence information and 3D display nodes form a stable correspondence, and the display consistency is improved; due to the adoption of baseline snapshots plus component-level incremental update packages, unnecessary retransmissions are reduced, and update efficiency is improved; due to the introduction of semantic priority-based channel distribution, key content can be delivered first in weak network environments, and the interaction stability is enhanced; due to the addition of authorization and audit links, the traceability and management controllability of data distribution and calling processes are improved.

[0015] Before providing a detailed description of the embodiments of this application, the project background, business introduction, and related technologies involved in the actual scenario will first be explained, which may include the following: I. Project Background Party B is the provider of the "Digital Cloud Showroom" service for physical stores.

[0016] Party A and Party B are long-term strategic partners of the "Digital Cloud Exhibition Hall" project. Party A is a member of the "Car Dismantling Workshop Strategic Cooperation Alliance" initiated by Party B.

[0017] Party B authorizes Party A to provide Party B's "Digital Cloud Showroom" service to physical stores that have a cooperative relationship with Party A, and has the right to use the dismantling data, reports and brand logos of the dismantling workshop for store promotion, product marketing and exhibition display during the cooperation period stipulated in this agreement.

[0018] Through strategic cooperation leveraging the strengths of both parties, we will jointly create digitally empowered products with market-leading advantages.

[0019] II. Business Introduction By leveraging digital rendering technology, an online virtual exhibition scene is created, highly replicating the offline showroom environment. This allows users to interactively and immersively browse the exhibition, further enhancing the user's comprehensive sensory experience with a 360-degree panoramic multi-dimensional online visual scene. This helps automotive service providers and dealers better interact with users, increase user stickiness, and improve customer operation efficiency. Based on market development trends and customer needs, scenario-based shopping guidance and immersive experiences are integrated, providing automotive dealerships with a closed-loop, highly efficient customer acquisition method. This breaks down the limitations of time and space for consumers, allowing them to gain a multi-faceted and comprehensive understanding of vehicles without leaving home.

[0020] III. Introduction to Relevant Technologies Virtualization Technology: Virtualization technology is one of the core technologies of the digital cloud exhibition hall, utilizing 3D modeling, virtual reality (VR), and an HTML5-based virtual reality display engine. The combination of these technologies enables the simulation of realistic and lifelike scenes and objects. It provides rich APIs and plugins, greatly facilitating the creation of interactive scenes. The front-end supports cross-platform compatibility, seamlessly connecting with PCs, mobile phones, tablets, large screens, and most VR / AR glasses. Simultaneously, our system is compatible with mainstream operating systems such as Windows, Linux, macOS, Android, and iOS, and supports various mainstream browsers, allowing direct use within WeChat. The app also provides Android and iOS client support, ensuring visitors can enjoy an immersive virtual tour of the exhibition hall.

[0021] Network Technology: Relying on the internet and network technology, efficient transmission and exchange of exhibition content were achieved. By optimizing network transmission control algorithms and reconstructing message transmission, the amount of data transmitted was significantly reduced, improving the user experience in complex network environments. This ensured the stable and smooth operation of the exhibition hall and the real-time updates of information without delay.

[0022] The specific composition and functions of the virtual car showroom system based on dismantled vehicle data provided in this application will be described in detail below with reference to the accompanying drawings and specific embodiments. Figure 1 This is a schematic diagram of the structural composition of the car cloud showroom virtual display system based on vehicle dismantling data provided in the embodiments of this application, as shown below. Figure 1 As shown, the system may specifically include the following components: The data access module 101 is used to acquire vehicle dismantling data, dismantling reports and display materials, perform field mapping, unit of measurement normalization and version marking, and generate a structured evidence dataset. The dual-graph construction module 102 is used to construct a disassembly evidence graph based on a structured evidence dataset and a display scene graph based on a vehicle 3D scene, and to establish a mapping relationship between the disassembly evidence graph and the display scene graph according to the component identifier; The multi-representation asset generation module 103 is used to generate geometric assets and texture assets of different precision for component identifiers, and to generate asset indexes associated with component identifiers; The semantic incremental release module 104 is used to generate a baseline snapshot based on the display scene diagram, and to generate an incremental update package oriented towards component identifiers based on the mapping relationship and version tag; The transmission scheduling module 105 is used to perform semantic priority sorting and channel-based distribution of baseline snapshots and incremental update packets based on terminal capability parameters, network status parameters and user interaction trajectories. The edge rendering interaction module 106 is used to load the corresponding geometric assets and texture assets according to the asset index, parse the incremental update package to update the scene node status, and respond to the component interaction query according to the mapping relationship. The authorization audit module 107 is used to perform authorization verification, source record and call log association storage on structured evidence datasets, asset indexes and incremental update packages.

[0023] In some embodiments, the data access module is specifically used for: Multi-source analysis and object-oriented recombination of vehicle dismantling data, dismantling reports, and presentation materials are performed to form a unified set of evidence objects; Based on pre-defined semantic mapping rules, heterogeneous fields in the unified evidence object set are mapped to a standard field set, and unit of measurement and dimension consistency checks are performed on numerical fields. Evidence version markers are generated based on vehicle identification, component identification, collection time, and source identification, and a version chain oriented towards the evidence object is constructed. The standard field set, evidence version marker, and source identifier are associated and encapsulated into a structured evidence dataset. Each evidence object in the structured evidence dataset is bound to a component identifier that can be called by the dual-graph construction module.

[0024] Specifically, in this embodiment, the data access module is deployed in the data processing domain of the cloud exhibition hall server, performing unified access processing on images, videos, inspection records, text reports, and promotional materials in the vehicle dismantling business. To ensure that the subsequent dual-image construction module can directly call upon these materials, the data objects output by the data access module adopt a component identifier-driven organization method, enabling a traceable association to be established between dismantling evidence, display materials, and component nodes of the same vehicle during the access phase.

[0025] In this embodiment, object-oriented reorganization refers to the process of converting and organizing multi-source heterogeneous raw data into unified evidence objects; unified evidence object set refers to the set of evidence objects aggregated according to a unified object model; pre-set semantic mapping rules refer to the set of rules used to convert heterogeneous fields to standard fields; dimensional consistency verification refers to the verification mechanism for judging the consistency of physical dimensions of numerical fields; evidence version marker refers to the structured identifier used to identify the version status of evidence objects; version chain refers to the chain-like association structure formed between multiple versions of the same evidence object according to time sequence and revision relationship; structured evidence dataset refers to the standardized evidence data set that can be directly called for subsequent graph construction and incremental release.

[0026] First, multi-source parsing and object-oriented reorganization are performed on the vehicle disassembly data, disassembly reports, and presentation materials. Multi-source parsing refers to calling corresponding parsers for data from different sources and in different formats, including field parsing for structured inspection files, paragraph semantic extraction for report documents, metadata extraction for images and videos, and tag extraction for marketing materials. Object-oriented reorganization refers to reorganizing the parsing results into evidence objects according to a unified object model. Each evidence object carries at least an object identifier, vehicle identifier, component identifier, evidence type, content payload, and time information. The unified evidence object set is a collection formed by aggregating multiple evidence objects according to vehicle and component identifiers, and this set serves as the input for subsequent semantic mapping.

[0027] For example, in one instance, a vehicle engine assembly disassembly task generates three types of inputs simultaneously: disassembly photos, torque test records, and technician conclusion text. The data access module extracts the image capture time, the sequence of test values, and the component description statements from the text, respectively. Then, using the same component identifier, the three types of inputs are recombined into multiple evidence objects under the same component, forming a unified set of evidence objects that can be queried in parallel.

[0028] Furthermore, after completing the object-oriented reorganization, the data access module maps heterogeneous fields to a standard field set based on pre-defined semantic mapping rules. These pre-defined semantic mapping rules are a combination of pre-configured field alignment rules, synonym merging rules, and enumeration value standardization rules, used to unify fields from different sources and with different naming conventions to the same semantic standard. The standard field set includes, but is not limited to: vehicle identification field, component identification field, evidence category field, collection time field, source field, numerical field, and unit field. For numerical fields, the data access module performs unit of measurement normalization and dimensional consistency verification. Unit of measurement normalization converts different unit expressions into a unified unit of measurement, while dimensional consistency verification determines the consistency of the dimensions of the same field before and after conversion. If a dimensional conflict exists, an anomaly flag is output, and the field is blocked from entering the valid set.

[0029] For example, in one instance, the brake disc thickness is expressed in mm and cm in different reports. The data access module converts cm to mm before storing it in the database and verifies that the dimensions of this field are all length dimensions. If a record mistakenly writes the temperature field into the thickness field, the dimension consistency check will generate a conflict flag, and the record will enter the pending review queue without participating in structured encapsulation.

[0030] Furthermore, after completing field standardization, the data access module generates evidence version tags based on vehicle identifiers, component identifiers, collection time, and source identifiers, and constructs a version chain oriented towards the evidence object. Evidence version tags are used to identify the version status of the same evidence object at different times, from different sources, or under different revision states. A version chain refers to an ordered chain structure formed by linking multiple versions of the same evidence object according to chronological order and revision relationship. Subsequent versions establish a continuous relationship through references to preceding versions, supporting the difference calculations performed by the subsequent incremental release module.

[0031] For example, in one instance, the crack detection results for the same component are generated into three versions after initial inspection, re-inspection, and verification. The data access module generates version tags for each of the three records and associates the re-inspection version with the initial inspection version, and the verification version with the re-inspection version, forming a continuous version chain. Subsequently, when the dual-graph construction module reads the evidence for this component, it can select the latest valid version according to the version chain, or trace back to historical versions by time.

[0032] Finally, the data access module encapsulates the standard field set, evidence version markers, and source identifiers into a structured evidence dataset. The structured evidence dataset refers to a data set that meets unified field constraints, version constraints, and source constraints. Each evidence object is explicitly bound to a component identifier, and the component identifiers adopt an identifier specification that can be directly called by the dual-graph construction module. The source identifier includes at least a source entity identifier and a source channel identifier, used for subsequent authorized audit module execution of source records and call tracing. After encapsulation, the data access module outputs the structured evidence dataset to the dual-graph construction module, which generates a disassembled evidence graph based on it and establishes a cross-graph mapping relationship with the display scenario graph.

[0033] Through the above technical means, the data access phase can achieve unified semantic merging of multi-source disassembled data, traceable version organization, and component-level callable binding, providing a consistent data foundation for subsequent dual-graph construction, incremental updates, and end-side interaction.

[0034] In some embodiments, a disassembled evidence map is constructed based on a structured evidence dataset, and a display scene map is constructed based on a 3D vehicle scene, including: Based on the structured evidence dataset, vehicle entities, assembly entities, component entities, inspection item entities, and evidence entities are extracted, and disassembly evidence diagrams are generated according to preset relationship templates. Semantic annotation is performed on model nodes, hotspot nodes, interactive nodes, and navigation nodes in the 3D vehicle scene, and a display scene map is generated according to the scene topology. Establish cross-graph anchoring relationships between disassembly evidence diagrams and display scene diagrams based on component identifiers, and configure relationship type tags and time sequence version tags for cross-graph anchoring relationships; Upon receiving a graph update trigger event, synchronous incremental reconstruction is performed on the disassembled evidence graph and the display scene graph based on the time-series version marker to maintain the resolvability of cross-graph anchoring relationships.

[0035] Specifically, in this embodiment, the dual-graph construction module is deployed in the graph computing domain of the cloud exhibition hall server. It takes the structured evidence dataset output from the previous stage as input and adopts the processing flow of "evidence semantic graph construction + scene semantic graph construction + cross-graph anchoring + synchronous reconstruction" to realize the unified association between disassembled evidence and three-dimensional display objects.

[0036] In this embodiment, the preset relationship template refers to a set of rules used to define entity type, relationship type, relationship direction, and constraints; cross-graph anchoring relationship refers to a bidirectionally resolvable association established between the disassembled evidence graph and the display scene graph using component identifiers; relationship type marker refers to a structured identifier used to distinguish the purpose category of cross-graph association; time-series version marker refers to an identifier used to characterize the version time-series status of entities, nodes, and their associated relationships; synchronous incremental reconstruction refers to a graph update method that performs differentiated reconstruction only on the affected subgraphs while maintaining the continuous resolvability of cross-graph relationships.

[0037] First, in the disassembly evidence graph construction phase, the dual-graph construction module performs entity extraction and relation merging. Entity extraction extracts vehicle entities, assembly entities, component entities, inspection item entities, and evidence entities from the structured evidence dataset. Vehicle entities are used to identify vehicle-level objects, assembly entities are used to identify assembly-level objects such as engine assemblies and brake assemblies, component entities are used to identify specific parts that can be directly displayed and interacted with, inspection item entities are used to identify measurement or judgment dimensions, and evidence entities are used to carry images, videos, text conclusions, and numerical records. Preset relation templates are used to define the types and directions of relationships between entities, such as "vehicles contain assemblies," "assemblies contain components," "components correspond to inspection items," "inspection items correspond to evidence," and "evidence originates from source identifiers." The disassembly evidence graph generated based on this template can support tracing the entire evidence chain by component identifier, and can also trace back to the corresponding component and vehicle by inspection item.

[0038] For example, in one instance, targeting the front braking system of a certain vehicle model, the system extracts the following entities from the structured evidence dataset: the front axle vehicle entity, the brake assembly entity, the brake disc component entity, the thickness measurement item entity, and three evidence entities. These three evidence entities are a vernier caliper reading record, disassembly photographs, and technician review text, respectively. After establishing the relationships between these five types of entities based on a preset relationship template, the module generates a queryable partial disassembly evidence subgraph.

[0039] Furthermore, during the construction of the display scene diagram, the dual-diagram construction module performs semantic annotation on model nodes, hotspot nodes, interactive nodes, and navigation nodes in the vehicle's 3D scene. Model nodes correspond to renderable 3D mesh objects, hotspot nodes correspond to clickable prompts, interactive nodes correspond to operation entry points such as exploded views, sectioning, and hiding / showing, and navigation nodes correspond to scene steps organized according to the shopping guide path. Semantic annotations must include at least the node identifier, node category, the component identifier corresponding to the node, node visibility constraints, and interaction constraints. When generating the display scene diagram based on the scene topology, the module establishes connections between nodes according to assembly level, spatial adjacency, and navigation sequence to ensure that stable node relationships can be resolved on the client side at different perspectives and different navigation stages.

[0040] For example, in one example, the model node of the front brake disc is written into the same topology subgraph as the "wear description" hotspot node, the "disassembly before and after comparison" interactive node, and the "chassis tour step 3" tour node; when the user is in the tour step 3, the system can directly parse the hotspots that should be displayed and the interactive actions that can be triggered.

[0041] Furthermore, in the cross-graph association phase, the dual-graph construction module establishes cross-graph anchoring relationships between the disassembly evidence graph and the display scene graph based on component identifiers. Cross-graph anchoring relationships refer to establishing bidirectionally resolvable mapping links between component entities and their associated detection items and evidence entities in the disassembly evidence graph, and model nodes, hotspot nodes, and interaction nodes in the display scene graph, using the same component identifier as the anchor point. To ensure the stability of the mapping links during incremental updates, the module configures a relationship type tag and a time-series version tag for each cross-graph anchoring relationship. The relationship type tag distinguishes between different uses such as "main display association," "auxiliary explanation association," and "alarm prompt association," while the time-series version tag identifies the data and scene version status corresponding to the anchoring relationship.

[0042] For example, in one instance, the brake disc component identifier is associated with its main display node, with the thickness detection evidence as supplementary information, and with the abnormal wear conclusion as an alarm notification. All three types of relationships are written into a time-series version tag for version consistency verification during subsequent incremental releases and edge-side parsing.

[0043] Furthermore, during the graph synchronization update phase, when the system receives a graph update trigger event, the dual-graph construction module performs synchronous incremental reconstruction of the evidence graph and the display scene graph based on the time-series version markers. Graph update trigger events can be triggered by the addition of evidence objects, revision of detection conclusions, replacement of component models, or changes to the navigation script. The synchronous incremental reconstruction process includes: locating the affected component identifier set, extracting affected entities and nodes, calculating incremental differences according to the time-series version markers, reconstructing affected relationships, and updating cross-graph anchoring relationships. Unaffected entities and nodes are not reconstructed to reduce reconstruction overhead and maintain overall graph stability. If inconsistencies are detected between the versions on both sides of the cross-graph, the module first updates the relationships and markers on the older side, and then submits new cross-graph anchoring relationships to ensure that resolvability is not interrupted.

[0044] For example, in one instance, a re-inspection revised the conclusion on brake disc thickness from "normal" to "critical." After the graph update is triggered, the system only performs incremental reconstruction on the associated detection item entities, evidence entities, hotspot node texts, and alarm prompts for that component, and synchronously updates the corresponding time-series version markers. When the user re-enters that node, the client can directly read the updated evidence chain and prompt content without requiring full-scene reconstruction.

[0045] Through the above technical means, this embodiment realizes dual-graph collaborative modeling of evidence semantics and scene semantics, establishes a stable cross-graph association with component identification as the core, and completes dual-graph reconstruction in a synchronous incremental manner when the version changes, thereby improving the consistent association capability between evidence and display, graph update efficiency, and stability of cross-version query and interaction.

[0046] In some embodiments, the asset generation module is specifically used for: Extract the corresponding 3D model data and material data for the component identifier, and perform mesh regularization, topology repair and parametric unfolding according to a unified coordinate benchmark to generate standard component assets; A multi-representation asset set is constructed based on standard component assets, and hierarchical tags, compression tags, and decoding constraint tags are written for each precision level. The multi-representation asset set includes geometric assets and texture assets at different precision levels. Asset index items are generated based on component identifier, precision level, version tag, and storage location tag, and then aggregated by component identifier to form an asset index; Establish a resolvable association between the asset index and the model nodes in the display scene graph, so that the edge rendering interaction module can call the geometric assets and texture assets of the corresponding precision level according to the asset index.

[0047] Specifically, in this embodiment, the multi-representation asset generation module is deployed in the asset processing domain. It receives component identifiers and display scene graph node constraints output by the dual-graph construction module, and generates an asset system that can be called on demand by the client according to the process of "standardized processing - multi-representation construction - index generation - node association". The input of this module includes component identifiers, original 3D model data, material data and version information, and the output includes standard component assets, multi-representation asset sets and asset indexes.

[0048] In this embodiment, standard component assets refer to component assets that meet unified specifications after undergoing unified coordinate, topology repair, and parametric unpacking processing; multi-representation asset sets refer to the geometric and texture asset sets of the same component at different precision levels; hierarchical markers refer to structured fields that identify the precision level of assets; compression markers refer to fields that identify asset compression strategies and parameters; decoding constraint markers refer to fields that limit the decoding conditions and order on the client side; and asset indexes refer to the set of asset index items aggregated by component identifiers, used to drive on-demand loading on the client side.

[0049] First, the module extracts the corresponding 3D model data and material data based on the component identifier. It then reads the original mesh, normals, material parameters, and texture maps from the asset repository according to the component identifier, and performs mesh normalization, topology repair, and parametric unwrapping according to a unified coordinate reference. A unified coordinate reference is used to unify the coordinate systems, orientations, and scales from different modeling sources to the same reference system, preventing misalignment during subsequent assembly. Mesh normalization eliminates duplicate vertices, abnormal faces, and discontinuous normals. Topology repair handles holes, non-manifold edges, and self-intersecting surfaces, ensuring the model meets rendering and interactive computing requirements. Parametric unwrapping establishes stable texture coordinates, allowing texture assets to be reused at different precision levels. Upon completion, a standard component asset is generated, which includes at least normalized geometry, normalized material parameters, and unified texture coordinates.

[0050] For example, in one instance, the front bumper of a certain car model was provided by two external suppliers with different coordinate axis definitions and significantly different patch densities. After the module completed coordinate alignment and scaling according to a unified coordinate benchmark, it performed mesh normalization on the high-density model and repaired connection gaps. Then, it performed parameter unification on the materials, ultimately obtaining a standard component asset that can be directly attached to the model nodes of the display scene graph.

[0051] Furthermore, a multi-representation asset set is constructed based on standard component assets. The module generates multi-level geometric and texture assets for each standard component asset according to a preset precision strategy, forming an asset set containing at least high-precision, medium-precision, and low-precision layers. Geometric assets are generated into versions with different patch sizes through mesh simplification, detail preservation, and boundary constraints; texture assets are generated into versions with different sampling densities through resolution layering and encoding strategies. For each precision level, the module writes a level marker, a compression marker, and a decoding constraint marker. The level marker identifies the precision level to which the asset belongs, the compression marker identifies the compression method and parameter set, and the decoding constraint marker limits the minimum capability requirements and decoding order constraints for edge decoding.

[0052] For example, in one instance, the engine block component generates three layers of geometric assets, with face sizes used for close-up disassembly viewing, mid-range navigation, and distant overview, respectively; simultaneously, three corresponding layers of texture assets are generated. The module writes high-precision hierarchical markers and high-complexity decoding constraints for the close-up layer, and low-precision hierarchical markers and fast decoding constraints for the distant layer, enabling adaptive invocation on the edge under different network and terminal capabilities.

[0053] Furthermore, asset index entries are generated based on component identifier, precision level, version tag, and storage location tag, and aggregated by component identifier to form an asset index. Each asset index entry records at least the component identifier, asset category, precision level, version tag, storage location tag, compression tag, and decoding constraint tag. The version tag is used for co-verification with the snapshot version and update package version of the semantic incremental release module, and the storage location tag is used to locate the access path of the asset in object storage or edge cache. After aggregation by component identifier, a set of optional asset index entries for a single component is formed, supporting layer switching on the edge according to a strategy.

[0054] For example, in one example, the door panel component under the door assembly was updated from version V12 to V13 after a model revision. The module only replaces the index entry for the corresponding version and retains the low-precision old version as a fallback entry. When the edge fails to receive the V13 incremental update packet, it can fall back to the available version according to the fallback flag in the index, thus maintaining the continuous display of the scene.

[0055] Finally, a resolvable association is established between the asset index and the model nodes in the scene graph. Based on the component identifier bound to the model node, the module writes the corresponding asset index into the node association table, forming a resolvable mapping of "model node identifier - component identifier - asset index item set". During rendering or interaction, the client-side rendering and interaction module first parses the model node, then reads the asset index item set through the component identifier, and selects the geometric and texture assets of the target precision level for loading and replacement based on the current terminal capabilities and network status. If a scene node triggers a precision switching event, the client-side can seamlessly switch asset levels without changing the semantic relationships of the nodes within the same mapping link.

[0056] For example, in one instance, when a user views chassis details on a mobile device, the device first loads a medium-precision layer to ensure the latency of the first frame; when the user zooms in and lingers on the area of ​​the suspension components, the device switches to high-precision geometric assets and high-resolution texture assets based on the node association table, while keeping the hotspots and interactive node positions unchanged, thus achieving enhanced detail and continuous interaction.

[0057] Through the above technical means, this embodiment realizes the standardized production and multi-precision organization of component assets, establishes an asset index link that can be directly parsed by scene nodes, and enables the terminal to call the corresponding level assets on demand under different terminal capabilities and network conditions, thereby improving the determinism of asset calling, the continuity of rendering switching and the efficiency of incremental update collaboration.

[0058] In some embodiments, the semantic incremental publishing module is specifically used for: Based on the scene diagram, the initial node state, resource reference relationship and interaction script state of the scene are extracted, and a baseline snapshot is generated according to the preset snapshot encapsulation rules and written to the snapshot version tag. When a change in evidence object, asset index, or scene node associated with a component identifier is detected, the target node set corresponding to the target component identifier is located according to the mapping relationship, and the state difference of the target node set before and after the change is extracted. Incremental operation sequences are generated based on state differences, evidence version markers, and snapshot version markers. The incremental operation sequences include node addition, node replacement, attribute revision, and relationship rebinding operations. The incremental operation sequence, target component identifier, dependency verification information, and rollback pointer are encapsulated into an incremental update package, and a version connection relationship is established between the incremental update package and the baseline snapshot, so that the edge rendering interaction module can parse and execute it in version order.

[0059] Specifically, in this embodiment, the semantic incremental publishing module is deployed in the publishing control domain of the cloud exhibition hall server. It receives the display scene diagram and mapping relationship output by the dual-graph construction module, the evidence version mark output by the data access module, and the asset index version information output by the multi-representation asset generation module. It adopts the process of "baseline snapshot construction - change awareness and location - difference orchestration - update package encapsulation" to realize component-level semantic incremental publishing.

[0060] In this embodiment, the baseline snapshot refers to the basic publishing unit used to initialize the semantic state of the scene and the resource reference relationship on the client side; the state difference refers to the set of differences in node attributes, reference relationships and script states of the same target node set before and after the change; the incremental operation sequence refers to the ordered set of operations formed after the state difference is executed and orchestrated; the dependency verification information refers to the set of version and integrity verifications that constrain the preconditions for the execution of the update package; the rollback pointer refers to the location information used to restore the stable version state when the execution fails; and the version connection relationship refers to the previous and next version dependency links between the baseline snapshot and each update package.

[0061] First, the module extracts the initial node states, resource reference relationships, and interaction script states from the scene graph. It then generates a baseline snapshot according to preset snapshot encapsulation rules and writes a snapshot version tag. The initial node states include, but are not limited to: model node visibility, hotspot node activation, interaction node enablement, and navigation node initial state. Resource reference relationships include, but are not limited to: reference chains between nodes and geometric assets, texture assets, and evidence objects. Interaction script states include, but are not limited to: script entry points, execution conditions, and node jump constraints. Preset snapshot encapsulation rules constrain snapshot content boundaries, field order, dependency declarations, and validation fields. The snapshot version tag identifies the current baseline snapshot's version position in the release chain, serving as the dependency basis for subsequent incremental packages.

[0062] For example, in one instance, a baseline snapshot V100 is generated on the release date of a car model showroom, containing a vehicle overview node, a chassis navigation node, and an engine hotspot script. The module encapsulates the node states, asset references, and script states and writes them into the snapshot version tag V100. When the client first enters the system, it only needs to pull this baseline snapshot to complete the interactive initialization.

[0063] Secondly, the module continuously monitors change events associated with component identifiers. When a change in evidence object, asset index, or scene node is detected, the module locates the target node set corresponding to the target component identifier based on the mapping relationship and extracts the state differences of the target node set before and after the change. Change detection can be triggered by the event bus, with event sources including evidence data revision events, asset version switching events, and scene arrangement adjustment events. The mapping relationship is used to map component identifiers to model nodes, hotspot nodes, interactive nodes, and navigation nodes in the scene diagram. State difference extraction covers differences in node attributes, reference relationships, and script state differences.

[0064] For example, in one instance, the re-inspection conclusion of the front brake disc component was revised from "normal" to "critical wear," while the high-precision texture asset version of this component was updated from T21 to T22. The module locates the brake disc model node, wear description hotspot node, and chassis navigation step node based on the component identifier, extracts three types of differences: hotspot text, evidence references, and texture references, and does not perform any processing on other unaffected nodes.

[0065] Then, the module generates an incremental operation sequence based on state differences, evidence version markers, and snapshot version markers. The incremental operation sequence is a set of executable operations for the client side, including node addition, node replacement, attribute revision, and relationship rebinding operations. Node addition is used to introduce new hotspots or new navigation steps; node replacement is used to replace the asset version referenced by model nodes; attribute revision is used to update text, thresholds, and display parameters; and relationship rebinding is used to rebuild the association between nodes and evidence objects or nodes and asset indexes. When generating the operation sequence, the module writes an operation number and dependency conditions for each operation, ensuring that the client side executes them sequentially and can resume from interruption.

[0066] For example, in one example, for the aforementioned brake disc change, the module generates four steps: first, revise the hotspot alarm level of the attribute; second, replace the texture reference version of the node; third, rebind the evidence object version of the relationship; and finally, add a "Re-inspection Instructions" sub-step to the navigation node. All four steps are bound to the same target component identifier and declare a dependency on snapshot version V100.

[0067] Finally, the module encapsulates the incremental operation sequence, target component identifier, dependency verification information, and rollback pointer into an incremental update package, and establishes a version connection relationship between the incremental update package and the baseline snapshot, so that the end-side can parse and execute it in version order. The dependency verification information is used to declare the minimum snapshot version and the previous update package version required for the update package to execute, and the rollback pointer is used to indicate the stable version position to roll back to when execution fails. The version connection relationship is used to form an ordered "baseline snapshot - incremental update package" link, ensuring that the end-side can still obtain a consistent scenario state when resuming interrupted downloads, retrying, or recovering across sessions.

[0068] For example, in one example, the module publishes update package U100-07, declaring its dependency on snapshot V100 and the preceding update package U100-06. If the client detects that U100-06 is not complete, it will temporarily suspend the execution of U100-07; if the execution of U100-07 is interrupted, it will roll back to the completed state of U100-06 according to the rollback pointer and re-fetch U100-07 to prevent the scene node from entering a half-update state.

[0069] Through the above technical means, this embodiment realizes fine-grained semantic incremental release for component identifiers, enabling evidence changes, asset changes and scene changes to be released collaboratively under a unified version link, reducing unnecessary retransmissions and improving end-side parsing consistency, thereby improving the efficiency of exhibition hall updates and operational stability.

[0070] In some embodiments, the transmission scheduling module is specifically used for: Collect and standardize terminal capability parameters, network status parameters, and user interaction trajectories to generate a scheduling status vector associated with session identifiers. The terminal capability parameters include graphics processing capability parameters and memory capacity parameters, and the network status parameters include latency parameters, jitter parameters, and bandwidth parameters. The semantic priority score corresponding to the component identifier is calculated based on the scheduling state vector. The semantic priority score is jointly determined by viewpoint relevance, interaction intent strength, version dependency and asset decoding cost. Based on semantic priority scores, the baseline snapshot and incremental update packages are sorted to generate a channel-specific delivery plan that includes a critical channel queue and a non-critical channel queue. The critical channel queue is used to carry version control information, relationship rebinding operations, and update packages associated with the current view. The distribution plan is executed in segments, and the acknowledgment feedback is processed. Based on the acknowledgment feedback results, the semantic priority score and channel queue order are iteratively updated to maintain the consistency of version connection and distribution sequence.

[0071] Specifically, in this embodiment, the transmission scheduling module is deployed between the publishing gateway and the session control service. It receives the baseline snapshot and incremental update packets output by the semantic incremental publishing module and executes a dynamic scheduling process of "state awareness - semantic scoring - channel-based orchestration - closed-loop reordering" for different terminals and network environments. This module maintains the session-level scheduling context using the session identifier as the primary key to ensure that the version link, delivery order, and confirmation feedback within the same user session are continuously associated.

[0072] In this embodiment, the scheduling state vector refers to the session-level state representation obtained by fusing terminal capability parameters, network state parameters, and user interaction trajectories; the semantic priority score refers to the comprehensive score oriented towards component identifiers and used to sort and distribute tasks; the critical channel queue refers to the priority transmission queue carrying update items that are strongly related to version consistency and the current interaction; the non-critical channel queue refers to the auxiliary transmission queue carrying update items that can be delayed; the acknowledgment feedback refers to the status feedback from the end side on the arrival, verification, and execution results of fragments; and the iterative update refers to the scheduling closed-loop mechanism that dynamically recalculates the score and reorders the queue based on the acknowledgment feedback.

[0073] First, in the state awareness phase, the module collects and standardizes terminal capability parameters, network status parameters, and user interaction trajectories, generating a scheduling state vector associated with the session identifier. Terminal capability parameters include at least graphics processing capability parameters and memory capacity parameters; network status parameters include at least latency parameters, jitter parameters, and bandwidth parameters; and user interaction trajectories include at least viewpoint trajectories, component click sequences, and dwell time sequences. Standardization processing employs a unified dimension and interval normalization strategy, mapping data from different sources and with different sampling frequencies to a unified vector space; short-term outliers are smoothed, and missing values ​​are filled with the most recent valid sample. The scheduling state vector is updated on a rolling time window basis, ensuring that scheduling decisions reflect the current session state rather than static configuration.

[0074] For example, in one instance, a user enters the chassis display area using a mobile device. The system detects that the graphics processing capability is moderate, available memory is low, network latency is increasing, and jitter is amplified. Simultaneously, the user repeatedly clicks on the braking system hotspot and holds the position for an extended period. The module merges these states into a scheduling state vector for the current session, which serves as input for subsequent semantic scoring.

[0075] Furthermore, in the semantic scoring stage, the module calculates the semantic priority score corresponding to the component identifier based on the scheduling state vector. The semantic priority score is jointly determined by viewpoint relevance, interaction intent strength, version dependency, and asset decoding cost. Viewpoint relevance measures the degree of matching between the component and the current viewpoint center and viewing direction; interaction intent strength measures the user's operational activity and sustained attention to the component; version dependency measures the prerequisite of a certain update package for the executability of subsequent packages; and asset decoding cost measures the resources required to decode and render the target asset under the current terminal capabilities. The module integrates the four types of indicators into a component-level score according to a preset joint decision rule and generates a candidate distribution sequence based on the score. For update items with strong version dependencies, the module adds dependency constraints to prevent high-scoring items from breaking the version prerequisite relationship.

[0076] For example, in one instance, the brake disc component receives a high score because it is located at the center of the current view, has a high interaction frequency, and its update package carries rebinding operations; the exhaust pipe component, although slightly modified, has a low correlation with the current interaction and a high decoding cost, resulting in a lower score. Based on this, the module prioritizes brake disc-related update items at the beginning of the candidate sequence.

[0077] Furthermore, during the channel orchestration phase, the module sorts the baseline snapshot and incremental update packages according to semantic priority scores, generating a channel-specific delivery plan that includes critical channel queues and non-critical channel queues. The critical channel queue carries version control information, relationship rebinding operations, and update packages associated with the current viewport; the non-critical channel queue carries low-time-sensitivity texture replacements, far-viewport component updates, and deferable script updates. To ensure continuous parsing on the client side, the critical channel queue prioritizes version chain integrity, while the non-critical channel queue delivers in parallel using remaining bandwidth without disrupting dependencies. The module writes a shard number, its assigned queue identifier, and dependency verification fields to each shard so that the client side can perform reorganization according to queue semantics.

[0078] For example, in one instance, for the update sequence following snapshot V120, the module places version control snippets, brake disc relationship rebinding snippets, and current view hotspot text revision snippets into the critical channel queue; and places backup high-resolution textures and remote node navigation prompt updates into the non-critical channel queue. Even in the event of a network outage, the critical queue can still be prioritized and maintain continuous interaction.

[0079] Furthermore, during the closed-loop reordering phase, the module executes fragmented transmission and acknowledgment feedback processing according to the channel-specific delivery plan, and iteratively updates the semantic priority score and channel queue order based on the acknowledgment feedback results. Acknowledgment feedback includes fragment arrival status, end-side verification status, and execution completion status. If a critical channel fragment acknowledgment times out, the module triggers priority retransmission and temporarily increases the blocking weight of its dependent subsequent fragments; if non-critical channels remain congested, the module reduces its concurrency and postpones lower-scoring items. After each iteration, the module recalculates the semantic priority score of affected components and reorders the two types of queues to maintain version continuity and delivery sequence consistency. For consecutively failed fragments, the module can fall back to a confirmed stable version according to the rollback strategy, and resume transmission of subsequent fragments after network recovery.

[0080] For example, in one instance, the first transmission of a critical channel relationship rebinding fragment in a user session is lost. After receiving a timeout feedback, the module performs a critical fragment retransmission and suspends subsequent update items that depend on that fragment. After successful confirmation, subsequent fragment transmission is resumed, and high-cost texture updates in non-critical queues are moved to the back, ultimately maintaining the continuity of the client-side version chain and ensuring that the current interaction is not interrupted.

[0081] Through the above technical means, this embodiment realizes dynamic semantic scheduling and priority delivery of key information based on session state, so that the baseline snapshot and incremental update package maintain version chain consistency and execution order stability under weak network and heterogeneous terminal conditions, thereby improving the timeliness of key interactive content arrival, transmission resource utilization efficiency and overall exhibition hall operation stability.

[0082] In some embodiments, the client-side rendering interaction module is specifically used for: Based on the precision level, compression mark and decoding constraint mark corresponding to the component identifier in the asset index parsing, hierarchical loading of geometric assets and texture assets is performed, and runtime binding relationship between component identifier and scene node identifier is established; The incremental update package is checked for version and dependency. When the check passes, the node addition, node replacement, attribute revision and relationship rebinding operations are executed in sequence according to the incremental operation sequence to update the scene node status and maintain the version connection relationship with the baseline snapshot. When a component interaction query is received, the query target is mapped from the scene node identifier to the component identifier according to the mapping relationship. The evidence object associated with the component identifier is retrieved and a structured interaction response is generated. The structured interaction response includes an evidence version marker and a source identifier. The structured interactive response is checked for consistency with the current scene node state. When the check passes, the corresponding scene node's visual feedback and interactive content are updated.

[0083] Specifically, in this embodiment, the edge rendering interaction module is deployed in the user terminal runtime environment. It receives the baseline snapshot, incremental update package and asset index issued by the transmission scheduling module. It adopts the execution link of "index parsing and hierarchical loading - incremental verification and sequential execution - interactive query and evidence backfilling - consistency verification and visual feedback" to realize component-level traceable interactive display.

[0084] In this embodiment, runtime binding relationship refers to the dynamic mapping relationship between component identifiers and scene node identifiers maintained by the client side during the session; version verification refers to the verification of the consistency between the update package and the local snapshot version; dependency verification refers to the verification of the integrity of the pre-execution conditions of the update package; structured interaction response refers to the interaction return object organized according to unified fields, including evidence content, evidence version marker and source identifier; consistency verification refers to the semantic and version compatibility determination of the response data with the current node state.

[0085] First, the module parses the precision level, compression flags, and decoding constraint flags corresponding to the component identifiers in the asset index, performs hierarchical loading of geometric and texture assets, and establishes runtime binding relationships between component identifiers and scene node identifiers. During session initialization, the module reads the node set from the baseline snapshot and queries the asset index item according to the component identifier bound to the node. It selects the target precision level based on terminal computing power, memory, and network status, prioritizing the loading of geometric and texture assets that meet the constraints of the first frame, and then gradually increasing to higher levels based on available resources. During decoding, the module selects the corresponding decoding process based on the compression flags and controls concurrency and order according to the decoding constraint flags to avoid high-cost resources blocking the display of critical nodes. The runtime binding relationships are written to a local mapping table for subsequent update package execution and interactive queries.

[0086] For example, in one instance, when the mobile terminal first enters the engine compartment scene, the module first loads medium-precision geometric assets and medium-resolution textures to ensure startup speed; when the user continues to focus on the fuel injection system area, the component is switched to a high-precision level according to the mapping table, while keeping the semantics of hotspot nodes and interactive nodes unchanged.

[0087] Secondly, the module performs version and dependency checks on the incremental update package. If the checks pass, it executes node addition, node replacement, attribute revision, and relationship rebinding operations sequentially according to the incremental operation sequence to update the scene node state and maintain version compatibility with the baseline snapshot. Version checks verify whether the snapshot version the update package depends on is consistent with the current local version, while dependency checks verify whether the preceding update package has been fully executed. The module executes the update sequentially only when both types of checks pass. During execution, operation logs and intermediate states are recorded. If an anomaly occurs, the module rolls back to the most recent stable state based on the rollback pointer to prevent the scene from entering an inconsistent state.

[0088] For example, in one instance, the local snapshot is V120, and an update package U120-04, which depends on V120 and also depends on package U120-03, is received. The module first confirms that U120-03 is complete, and then sequentially executes "Replace brake disc texture reference", "Revise hotspot warning text", and "Rebind evidence object reference". If a dependency is missing during the relationship rebinding stage, the module rolls back the first two steps and requests a retransmission of the missing fragment to maintain version chain continuity.

[0089] Furthermore, upon receiving a component interaction query, the module maps the query target from the scene node identifier to the component identifier based on the mapping relationship, retrieves the evidence object associated with the component identifier, and generates a structured interactive response. This structured interactive response includes an evidence version marker and a source identifier. Interactive queries can be triggered by clicking a model node, triggering a hotspot, or using a navigation step. The module first locates the component identifier using a runtime mapping table, then queries the local evidence cache or remote evidence service based on the component identifier, aggregating image evidence, detection conclusions, and text descriptions, and returning them according to a unified response structure. The evidence version marker identifies the version status of the displayed evidence, and the source identifier identifies the evidence's source entity and source channel.

[0090] For example, in one instance, after a user clicks on the "Front Brake Disc" hotspot, the module maps the hotspot node identifier to the brake disc component identifier and returns "Thickness Inspection Record V37, Re-inspection Conclusion V37, Disassembly Photo V37 and Source Identifier". Users can view the evidence summary and source information in the same pop-up without having to navigate to an external page.

[0091] Finally, the module performs a consistency check between the structured interactive response and the current scene node state. If the check passes, it drives the corresponding scene node's visual feedback and interactive content updates. The consistency check includes component identifier consistency, evidence version and node version compatibility, and source identifier availability checks. Upon successful check, the module triggers model highlighting, hotspot text replacement, evidence panel refresh, and navigation prompt updates. If the check fails, the module maintains the current display and initiates a compensation request or reverts to a usable evidence version.

[0092] For example, in one instance, if the node is currently bound to version V36 but the query returns evidence version V37, the module will perform a smooth update if it determines that there is a compatible upgrade path between the two. If the source identifier is missing, the evidence will not be displayed, and a message indicating that synchronization is in progress will be displayed to prevent erroneous evidence from entering the interactive interface.

[0093] Through the above technical means, this embodiment realizes the coordinated operation of hierarchical asset loading and incremental sequential execution, enabling the end-side to maintain the continuity of scene status under heterogeneous terminals and dynamic network conditions; at the same time, through component-level mapping retrieval and consistency verification, it realizes traceable interactive feedback, thereby improving the accuracy of interactive response, version execution stability and the credibility of displayed content.

[0094] In some embodiments, the authorization audit module is specifically used for: For the structured evidence dataset, asset index and incremental update package, extract the object identifier, version tag and source identifier respectively, generate the set of objects to be verified, and load fine-grained authorization rules according to the store identifier, vehicle model identifier, time window identifier and region identifier; Based on fine-grained authorization rules, hierarchical authorization verification is performed on the set of objects to be verified. The hierarchical authorization verification includes object-level verification, version-level verification, and association-level verification, and outputs verification result tags and restricted object identifiers. For objects that pass verification, write them to the source record and generate a chain association tag based on the digest value of the previous record and the digest value of the current record. The source record includes the source subject identifier, the processing stage identifier, the distribution node identifier, and the timestamp identifier. The system collects session identifiers, caller identifiers, call action identifiers, and result status identifiers during the object invocation process, generates invocation logs, and associates and stores the invocation logs with the source records based on the object identifiers and chain-linked association markers to form a traceable audit chain.

[0095] Specifically, in this embodiment, the authorization audit module is deployed in the release governance domain of the cloud exhibition hall platform, located between the semantic incremental release module and the transmission scheduling module, and connected to the log feedback channel of the edge rendering interaction module. This module adopts the execution process of "object aggregation - rule loading - layered verification - source solidification - log association" to implement unified authorization control and full-link auditing for structured evidence datasets, asset indexes, and incremental update packages.

[0096] In this embodiment, the set of objects to be verified refers to the set of evidence objects, asset index items, and incremental update package objects collected according to the release batch; fine-grained authorization rules refer to the set of rules simultaneously constrained by store, vehicle model, time window, and region; hierarchical authorization verification refers to the authorization judgment mechanism executed at three levels: object, version, and association; source record refers to the source and processing metadata record of the object in the processing and distribution chain; chain association mark refers to the chain tracking mark generated by the summary value of the previous record and the summary value of the current record; and traceable audit chain refers to the full-process tracking chain formed after the source record and the call log are associated by object and chain association mark.

[0097] First, during the object aggregation phase, the authorized audit module extracts object identifiers, version tags, and source identifiers from the structured evidence dataset, asset index, and incremental update packages, respectively, to generate a set of objects to be verified. The object identifier uniquely identifies the evidence object, asset index item, and update package fragment; the version tag distinguishes the release version of the object; and the source identifier indicates the source entity and source channel. After performing standardized encapsulation on the three types of objects, the module establishes an index for the set of objects to be verified based on the session identifier and release batch, ensuring that subsequent verification can be replayed and reviewed in batches.

[0098] For example, in one instance, a store requests an update for the engine compartment of an SUV model. The set of objects to be verified contains 42 evidence objects, 18 asset index items, and 3 incremental update packages. The module extracts each object into a triple of "object identifier + version identifier + source identifier" and registers it in the same release batch.

[0099] Furthermore, during the rule loading phase, the module loads fine-grained authorization rules based on store identifiers, vehicle model identifiers, time window identifiers, and region identifiers. These fine-grained authorization rules, issued by the rule center, include object access scope rules, version visibility rules, cross-object dependency rules, and region-time constraint rules. Store identifiers limit the range of operable stores, vehicle model identifiers limit the range of distributable vehicle models, time window identifiers limit the effective time period of the rule, and region identifiers limit the applicable region of the rule. After loading, the module generates a session-level rule snapshot and writes the rule snapshot version to the current verification context.

[0100] For example, in one instance, the rule center stipulates that stores in region A can display medium-to-high precision assets of specified car models during promotional periods, but highly sensitive evidence objects are only allowed to be visible to authenticated accounts during working hours. Based on this, the module generates a session-level rule snapshot and binds it to this release task.

[0101] Furthermore, in the layered verification phase, the module performs object-level verification, version-level verification, and relationship-level verification on the set of objects to be verified based on fine-grained authorization rules. Object-level verification determines whether an object falls within the authorized scope of a store and vehicle model; version-level verification determines whether the object's version is within the visible version range or has been revoked; relationship-level verification determines whether the reference relationships between objects conform to rule constraints, such as whether the asset index item referenced by the incremental update package and the evidence object are both authorized and version compatible. After verification, a verification result marker and a restricted object identifier are output. Restricted objects will be blocked from being issued or replaced with available rollback objects.

[0102] For example, in one instance, an update package references an outdated texture index entry. Object-level validation passes, version-level validation fails, and association-level validation determines the reference chain is invalid. The module outputs a restricted object identifier and switches the update package to an available fallback index entry before proceeding with the delivery chain.

[0103] Furthermore, during the source solidification phase, the module writes a source record to the verified object and generates a chain-linked association marker based on the previous record's summary value and the current record's summary value. The source record includes, but is not limited to: source entity identifier, processing stage identifier, distribution node identifier, and timestamp identifier. The previous record's summary value originates from the previous valid source record for the same object, and the current record's summary value is calculated based on the content of the current source record. The module combines these two to generate a chain-linked association marker, enabling the same object to form a continuously traceable record chain across different processing stages.

[0104] For example, in one instance, evidence object E-301 went through four stages: "access standardization," "graph construction and referencing," "incremental encapsulation," and "distribution and distribution." The module writes a source record for each stage and generates chain-linked tags based on the preceding and following summaries, forming a complete processing trajectory.

[0105] Furthermore, during the log association phase, the module collects session identifiers, caller identifiers, call action identifiers, and result status identifiers during the object call process, generates call logs, and associates and stores the call logs with the source records based on the object identifier and chained association markers to form a traceable audit chain. Call actions include, but are not limited to: query, distribute, load, replace, rollback, and retransmit. Result statuses include, but are not limited to: success, rejection, rollback, and exception. The module supports retrieval by object identifier, store identifier, time range, and result status for subsequent compliance review and operational troubleshooting.

[0106] For example, in one instance, when the client-side performed a query on evidence for a certain component, it returned a "rejection." The log recorded the caller identifier, the rejection status, and the rule snapshot version. Auditors retrieved the source record and chain-related markers corresponding to this call using the object identifier, confirming that the rejection was due to a time window exceeding the limit rather than a system failure.

[0107] Through the above technical means, this embodiment realizes fine-grained authorization control and hierarchical verification for multiple types of objects, ensuring that unauthorized objects and incompatible versions are identified and isolated before distribution; at the same time, by linking chained source records and call logs, an audit link that can be replayed, searched, and verified is formed, thereby improving data distribution compliance, problem location efficiency, and end-to-end governance capabilities.

[0108] In some embodiments, a shopping guide strategy generation module is also included, which is used to: Based on user interaction trajectories, viewpoint dwell features, component operation features, and query semantic features are extracted to generate an interaction intent vector associated with the session identifier; The interaction intent vector is jointly calculated with the correlation degree of evidence entities in the disassembled evidence diagram and the reachability of navigation nodes in the display scene diagram to generate a component-level shopping guide priority sequence. An adaptive navigation script is constructed based on the component-level shopping guide priority sequence, and a script version tag is written. The adaptive navigation script includes a target component identification sequence, an evidence call sequence, and an interactive prompt sequence. The script version tag is associated with the incremental update package, and the semantic incremental publishing module issues the navigation node update instruction to the client-side rendering interaction module according to the script version tag.

[0109] Specifically, in this embodiment, the shopping guide strategy generation module is deployed in the business decision domain, located between the behavior feedback link of the edge rendering interaction module and the semantic incremental release module. It adopts the process of "intent modeling - priority sequence generation - script orchestration - incremental linkage release" to realize the dynamic collaboration between shopping guide content and decomposed evidence and scene nodes.

[0110] In this embodiment, the interaction intent vector refers to the session-level intent representation formed by fusing viewpoint dwell features, component operation features, and query semantic features; the evidence entity correlation degree refers to the correlation strength between the target component and the evidence entity under the current session semantics; the navigation node reachability refers to the executable degree of reaching the target navigation node under the current scene topology and node state; the component-level guide priority sequence refers to the component guide order formed by sorting according to the comprehensive priority value; the adaptive guide script refers to the guide execution script that can be dynamically updated with the session state; and the script version tag refers to the identification field used to identify the guide script version and participate in the update package version connection.

[0111] First, the shopping guide strategy generation module extracts viewpoint dwell features, component operation features, and query semantic features based on user interaction trajectories, generating an interaction intent vector associated with session identifiers. Viewpoint dwell features characterize the intensity of user attention to different component viewpoints, including dwell time, number of revisits, and zoom level; component operation features characterize the depth of user interaction with components, including click frequency, rotation frequency, segmentation trigger frequency, and comparison trigger frequency; query semantic features characterize the matching direction between user questions and evidence topics, including keyword topics, question types, and preference tags. The module aggregates and normalizes these three types of features within a unified time window to form a session-level interaction intent vector.

[0112] For example, in one instance, a user continuously zooms in on the front suspension area in the chassis scene and triggers the "reason for wear" query multiple times. The module extracts the high dwell time, high operation, and "durability explanation" semantic preferences of the session on suspension-related components, generates the corresponding interaction intent vector, and binds it to the session identifier.

[0113] Secondly, the module jointly calculates the interaction intent vector with the evidence entity relevance in the decomposed evidence graph and the accessibility of navigation nodes in the display scenario graph to generate a component-level shopping guide priority sequence. Evidence entity relevance measures the relevance and completeness of evidence that can be accessed by a component under the current session intent; navigation node accessibility measures the feasibility of the path to access the target navigation node in the current scenario state. During the joint calculation, the module outputs a comprehensive priority value for each candidate component and sorts them based on version availability constraints and the current session context to obtain the component-level shopping guide priority sequence. This sequence reflects the order of "which component to discuss first, which type of evidence to display first, and which type of prompt to trigger first."

[0114] For example, in one instance, for the same vehicle model, the module calculates a priority sequence of "front brake disc > suspension bushing > tire tread" because the front brake disc has the highest evidence relevance and the shortest navigation path in the current session, making it suitable for priority entry into the shopping guide process.

[0115] Then, the module constructs an adaptive navigation script based on the component-level guide priority sequence and writes a script version tag. The adaptive navigation script includes a target component identification sequence, an evidence retrieval sequence, and an interactive prompt sequence. The target component identification sequence defines the main navigation order; the evidence retrieval sequence defines the evidence display order and switching conditions for each component; and the interactive prompt sequence defines the operation prompts and attention prompts output to the user at each step of the navigation. The script version tag is used to identify the script iteration status and maintain consistency with the version chain of the semantic incremental release module. After the script is generated, the module performs a script consistency check to ensure that the target components, evidence objects, and navigation nodes all have resolvable mappings.

[0116] For example, in one instance, the module generates script S-208 for the session: The first step locates the front brake disc and retrieves thickness detection evidence and re-inspection conclusions; the second step guides the user to the suspension node and retrieves bushing crack image evidence; the third step prompts the user to compare tire wear distribution. The script is written to version tag S-208 and added to the release queue.

[0117] Finally, the module associates the script version tag with the incremental update package, and the semantic incremental publishing module issues the navigation node update instruction to the client-side rendering interaction module according to the script version tag. The association process writes the script version tag, the target navigation node set, and the prerequisite dependency information into the update package metadata. When generating the incremental operation sequence, the semantic incremental publishing module incorporates navigation node addition, navigation node replacement, prompt attribute revision, and evidence reference rebinding operations into the same version connection chain. After receiving the sequence, the client executes them in version order to ensure that the shopping guide script update and the scene state update take effect synchronously.

[0118] For example, in one instance, the client is currently in snapshot V150. After receiving the update package U150-12 of the associated script version S-208, it first updates the order of the navigation nodes, and then updates the prompt text and evidence access points for each node. Users can enter the new shopping guide path without refreshing the session, and the evidence display and prompt content remain consistent.

[0119] Through the above technical means, this embodiment realizes the dynamic generation and versioning of shopping guide scripts based on real interactive behavior, enabling the shopping guide path to be matched with evidence supply capabilities and scenario accessibility, and to be distributed synchronously with the incremental update link, thereby improving the targeting of the shopping guide process, the continuity of script execution, and the ability to support in-session conversion.

[0120] The above embodiments have provided a detailed description of the composition, structure, and functions of the automotive cloud showroom virtual display system based on vehicle dismantling data of this application. The implementation process of the automotive cloud showroom virtual display method based on vehicle dismantling data of this application will be described in detail below with reference to specific embodiments. Figure 2 This is a flowchart illustrating the virtual display method for a car cloud showroom based on vehicle dismantling data provided in this application embodiment, as shown below. Figure 2 As shown, the method may specifically include the following steps: S201: Obtain vehicle dismantling data, dismantling reports, and presentation materials; perform field mapping, unit of measurement normalization, and version marking to generate a structured evidence dataset. S202, construct a disassembly evidence diagram based on a structured evidence dataset, and construct a display scene diagram based on a vehicle 3D scene, and establish a mapping relationship between the disassembly evidence diagram and the display scene diagram according to the component identification; S203, Generate geometric and texture assets of different precision for the component identifier, and generate an asset index associated with the component identifier; S204, Generate a baseline snapshot based on the displayed scene diagram, and generate an incremental update package oriented towards component identifiers based on the mapping relationship and version tag; S205, based on terminal capability parameters, network status parameters and user interaction trajectories, performs semantic priority sorting on baseline snapshots and incremental update packets and distributes them through different channels; S206: Load the corresponding geometric assets and texture assets according to the asset index, parse the incremental update package to update the scene node status, and respond to the component interaction query according to the mapping relationship; S207 performs authorization verification, source records, and call log association storage on structured evidence datasets, asset indexes, and incremental update packages.

[0121] Specifically, the virtual display method for a car cloud showroom based on dismantled vehicle data in this embodiment is executed along the main line of "evidence data standardization - dual-graph mapping - multi-representation assets - semantic incremental release - semantic scheduling and distribution - end-side parsing interaction - authorization audit closed loop". This method is completed collaboratively by cloud services and end-side runtime. Steps S201 to S207 are continuously connected according to version chain and component identifier, and the output of the previous step is continuously reused as the input of the next step.

[0122] First, the system acquires vehicle disassembly data, disassembly reports, and presentation materials, and performs field mapping, unit normalization, and version tagging to generate a structured evidence dataset. In implementation, the system performs multi-source parsing and object-oriented recombination of images, videos, inspection records, and text reports to form a unified set of evidence objects. Then, it maps heterogeneous fields to a standard field set according to predefined semantic mapping rules and performs dimensional consistency checks on numerical fields. Subsequently, it generates evidence version tags and constructs a version chain based on vehicle identifier, component identifier, collection time, and source identifier, ultimately encapsulating it into a structured evidence dataset. In the example, the front brake disc contains three types of data: inspection report, disassembly photos, and technician text. The system uniformly binds these to the same component identifier and forms a traceable version chain, providing direct input for subsequent graph construction.

[0123] Furthermore, a disassembly evidence map is constructed based on the structured evidence dataset, and a display scene map is constructed based on the vehicle's 3D scene. A mapping relationship between the two maps is then established according to component identifiers. In implementation, the disassembly evidence map extracts vehicle entities, assembly entities, component entities, inspection item entities, and evidence entities and constructs a map according to a preset relationship template. The display scene map semantically annotates model nodes, hotspot nodes, interactive nodes, and navigation nodes and constructs a map according to the scene topology. Subsequently, a cross-map anchoring relationship is established using component identifiers, and relationship type and time-series version markers are configured. In the example, the brake disc component is associated with thickness detection and re-inspection conclusions in the evidence map, and with model nodes and wear hotspots in the scene map. Both are bidirectionally resolvable through the same component identifier.

[0124] Furthermore, geometric and texture assets of varying precision are generated for each component identifier, and an asset index is created. In implementation, the system first performs uniform coordinate alignment, mesh normalization, topology repair, and parametric unwrapping on the original model to generate standard component assets. Then, it constructs a multi-representation asset set according to precision levels, writing a level tag, compression tag, and decoding constraint tag for each level. Finally, it generates asset index items based on component identifier, precision level, version tag, and storage location, and aggregates them into an asset index. In the example, the engine block generates high, medium, and low-level assets. The client-side can select the corresponding level by index under different terminal conditions, ensuring both loadability and interactivity.

[0125] Furthermore, a baseline snapshot is generated based on the displayed scene diagram, and an incremental update package oriented towards the component identifier is generated based on the mapping relationship and version tag. In implementation, the system extracts the initial node state, resource reference relationship, and interaction script state from the displayed scene diagram, generates a baseline snapshot according to the snapshot encapsulation rules, and writes a snapshot version tag. When a change in evidence object, asset index, or scene node is detected, the system locates the corresponding node set of the target component based on the mapping relationship and extracts the state differences before and after the change. Then, based on the differences, evidence version tag, and snapshot version tag, an incremental operation sequence is arranged, encapsulated into an incremental update package containing dependency verification information and rollback pointers, and a version connection relationship with the baseline snapshot is established. In the example, after the brake disc re-inspection conclusion changes, only the attribute revision and relationship rebinding operations of the relevant nodes of that component are generated, without the need for a full scene re-transmission.

[0126] Furthermore, based on terminal capability parameters, network status parameters, and user interaction trajectories, the baseline snapshot and incremental update packages are semantically prioritized and distributed across different channels. In implementation, the system standardizes graphics processing capabilities, memory capacity, latency, jitter, bandwidth, and interaction trajectories into scheduling state vectors; it calculates component-level semantic priority scores based on viewpoint relevance, interaction intent strength, version dependency, and asset decoding cost; and according to the scores, version control information, relationship rebinding operations, and current view-related update packages are placed into the critical channel queue, while the rest are placed into the non-critical channel queue, performing fragmented transmission and acknowledgment feedback closed-loop reordering. In the example, under weak network conditions, the critical channel prioritizes the distribution of brake disc updates for the current view, while non-critical textures are delayed to ensure continuous interaction.

[0127] Furthermore, the client-side loads the corresponding geometric and texture assets based on the asset index, parses the incremental update package to update the scene node status, and responds to component interaction queries according to the mapping relationship. In implementation, the client-side first parses the precision level, compression markers, and decoding constraints, performs hierarchical loading, and establishes a runtime binding relationship between component identifiers and scene node identifiers; then, it performs version and dependency checks on the incremental update package, sequentially performing node addition, node replacement, attribute revision, and relationship rebinding; when a user triggers a component query, it maps the scene node identifier to the component identifier, retrieves the associated evidence object, and returns a structured interactive response containing evidence version markers and source identifiers; after passing consistency checks, it drives node highlighting, prompt updates, and evidence panel refreshes. In the example, after the user clicks on the front brake disc hotspot, the client-side returns the corresponding re-inspection evidence and source information, and synchronously updates the alarm prompt content.

[0128] Furthermore, the system performs authorization verification, source record, and call log association storage on the structured evidence dataset, asset index, and incremental update package. In implementation, the system first extracts object identifiers, version tags, and source identifiers to generate a set of objects to be verified, and loads fine-grained authorization rules constrained by store identifiers, vehicle model identifiers, time window identifiers, and region identifiers. Then, it performs hierarchical verification at the object level, version level, and association level, outputting verification result tags and restricted object identifiers. Objects that pass verification are written to the source record, and a chain-like association tag is generated based on the summary values ​​of the preceding and current records. The session identifier, caller identifier, call action identifier, and result status identifier are then used to form a call log, which is associated with the source record according to the object identifier and the chain-like association tag, forming a traceable audit chain. In the example, if a store requests highly sensitive evidence within an unauthorized time window, the system blocks its issuance and records the audit event during the version-level verification stage, and subsequent traceability verification can be completed according to the object identifier.

[0129] In summary, the above process can achieve component-level evidence association, fine-grained incremental updates, priority delivery of key content under weak network conditions, and full-link auditable management without changing the core display framework.

[0130] It should be understood that the sequence number of each step in the above method embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0131] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although the technical solutions of this application have been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A virtual car showroom system based on vehicle dismantling data, characterized in that, include: The data access module is used to acquire vehicle dismantling data, dismantling reports and display materials, perform field mapping, unit of measurement normalization and version marking, and generate structured evidence datasets. The dual-graph construction module is used to construct a disassembly evidence graph based on the structured evidence dataset and a display scene graph based on the vehicle 3D scene, and to establish a mapping relationship between the disassembly evidence graph and the display scene graph according to the component identifier; The multi-representation asset generation module is used to generate geometric and texture assets of different precision for component identifiers, and to generate asset indexes associated with component identifiers; The semantic incremental release module is used to generate a baseline snapshot based on the displayed scene diagram, and to generate an incremental update package oriented towards component identifiers based on the mapping relationship and version tag; The transmission scheduling module is used to perform semantic priority sorting and channel-specific distribution of the baseline snapshot and the incremental update packet based on terminal capability parameters, network status parameters and user interaction trajectories. The edge rendering interaction module is used to load the corresponding geometric assets and texture assets according to the asset index, parse the incremental update package to update the scene node status, and respond to component interaction queries according to the mapping relationship. The authorization audit module is used to perform authorization verification, source record and call log association storage on the structured evidence dataset, asset index and incremental update package; The step of constructing a disassembled evidence map based on the structured evidence dataset and constructing a display scene map based on the vehicle 3D scene includes: Based on the structured evidence dataset, vehicle entities, assembly entities, component entities, inspection item entities, and evidence entities are extracted, and a disassembly evidence diagram is generated according to a preset relationship template. Semantic annotation is performed on model nodes, hotspot nodes, interactive nodes, and navigation nodes in the 3D vehicle scene, and a display scene map is generated according to the scene topology. Establish cross-graph anchoring relationships between disassembly evidence diagrams and display scene diagrams based on component identifiers, and configure relationship type tags and time sequence version tags for cross-graph anchoring relationships; Upon receiving a graph update trigger event, synchronous incremental reconstruction is performed on the disassembled evidence graph and the displayed scene graph according to the time-series version marker to maintain the resolvability of cross-graph anchoring relationships.

2. The system according to claim 1, characterized in that, The data access module is specifically used for: Multi-source analysis and object-oriented recombination of vehicle dismantling data, dismantling reports, and presentation materials are performed to form a unified set of evidence objects; Based on the pre-set semantic mapping rules, the heterogeneous fields in the unified evidence object set are mapped to the standard field set, and the unit of measurement and dimension consistency verification are performed on the numerical fields. Evidence version markers are generated based on vehicle identification, component identification, collection time, and source identification, and a version chain oriented towards the evidence object is constructed. The standard field set, evidence version marker, and source identifier are associated and encapsulated into a structured evidence dataset, wherein each evidence object in the structured evidence dataset is bound to a component identifier that can be called by the dual-graph construction module.

3. The system according to claim 1, characterized in that, The multi-representation asset generation module is specifically used for: Extract the corresponding 3D model data and material data for the component identifier, and perform mesh regularization, topology repair and parametric unfolding according to a unified coordinate benchmark to generate standard component assets; A multi-representation asset set is constructed based on the standard component assets, and a hierarchical tag, a compression tag, and a decoding constraint tag are written for each precision level. The multi-representation asset set includes geometric assets and texture assets at different precision levels. Asset index items are generated based on component identifier, precision level, version tag, and storage location tag, and then aggregated by component identifier to form an asset index; Establish a resolvable association between the asset index and the model nodes in the display scene diagram, so that the terminal rendering interaction module can call the geometric assets and texture assets of the corresponding precision level according to the asset index.

4. The system according to claim 1, characterized in that, The semantic incremental publishing module is specifically used for: Based on the displayed scene graph, extract the initial node state, resource reference relationship and interaction script state of the scene, generate a baseline snapshot according to the preset snapshot encapsulation rules and write the snapshot version tag; When a change in evidence object, asset index, or scene node associated with a component identifier is detected, the target node set corresponding to the target component identifier is located according to the mapping relationship, and the state difference of the target node set before and after the change is extracted. Based on the state differences, evidence version markers, and snapshot version markers, an incremental operation sequence is generated, which includes node addition, node replacement, attribute revision, and relationship rebinding operations. The incremental operation sequence, along with the target component identifier, dependency verification information, and rollback pointer, is encapsulated into an incremental update package. A version connection relationship is established between the incremental update package and the baseline snapshot, so that the edge rendering interaction module can parse and execute it in version order.

5. The system according to claim 1, characterized in that, The transmission scheduling module is specifically used for: The system collects and standardizes terminal capability parameters, network status parameters, and user interaction trajectories to generate a scheduling status vector associated with a session identifier. The terminal capability parameters include graphics processing capability parameters and memory capacity parameters, and the network status parameters include latency parameters, jitter parameters, and bandwidth parameters. The semantic priority score corresponding to the component identifier is calculated based on the scheduling state vector. The semantic priority score is jointly determined by viewpoint relevance, interaction intent strength, version dependency, and asset decoding cost. The baseline snapshot and the incremental update package are sorted according to the semantic priority score to generate a channel-based delivery plan that includes a critical channel queue and a non-critical channel queue. The critical channel queue is used to carry version control information, relationship rebinding operations and update packages associated with the current view. According to the channel-based delivery plan, the fragmented sending and confirmation feedback processing are performed, and the semantic priority score and channel queue order are iteratively updated based on the confirmation feedback results to maintain the consistency of version connection and delivery sequence.

6. The system according to claim 1, characterized in that, The endpoint rendering interaction module is specifically used for: Based on the precision level, compression mark and decoding constraint mark corresponding to the component identifier in the asset index parsing, hierarchical loading of geometric assets and texture assets is performed, and runtime binding relationship between component identifier and scene node identifier is established; The incremental update package is checked for version and dependency. When the check passes, the node addition, node replacement, attribute revision and relationship rebinding operations are executed in sequence according to the incremental operation sequence to update the scene node status and maintain the version connection relationship with the baseline snapshot. Upon receiving a component interaction query, the query target is mapped from the scene node identifier to the component identifier according to the mapping relationship. Evidence objects associated with the component identifier are retrieved and a structured interaction response is generated. The structured interaction response includes an evidence version marker and a source identifier. The structured interactive response is checked for consistency with the current scene node state. When the check passes, the corresponding scene node's visual feedback and interactive content are updated.

7. The system according to claim 1, characterized in that, The authorization audit module is specifically used for: For the structured evidence dataset, asset index and incremental update package, extract the object identifier, version tag and source identifier respectively, generate the set of objects to be verified, and load fine-grained authorization rules according to the store identifier, vehicle model identifier, time window identifier and region identifier; Based on the fine-grained authorization rules, hierarchical authorization verification is performed on the set of objects to be verified. The hierarchical authorization verification includes object-level verification, version-level verification, and association-level verification, and outputs verification result markers and restricted object identifiers. For objects that pass verification, write them to a source record and generate a chain association tag based on the summary value of the previous record and the summary value of the current record. The source record includes a source subject identifier, a processing stage identifier, a distribution node identifier, and a timestamp identifier. The system collects session identifiers, caller identifiers, call action identifiers, and result status identifiers during the object invocation process, generates invocation logs, and associates and stores the invocation logs with the source records based on the object identifiers and chain-linked association markers to form a traceable audit chain.

8. The system according to claim 1, characterized in that, It also includes a shopping guide strategy generation module, which is used for: Based on user interaction trajectories, viewpoint dwell features, component operation features, and query semantic features are extracted to generate an interaction intent vector associated with the session identifier; The interaction intent vector is jointly calculated with the correlation degree of the evidence entities in the disassembled evidence diagram and the reachability of the guide nodes in the display scene diagram to generate a component-level shopping guide priority sequence. An adaptive navigation script is constructed based on the component-level shopping guide priority sequence, and a script version tag is written in it. The adaptive navigation script includes a target component identification sequence, an evidence call sequence, and an interactive prompt sequence. The script version tag is associated with the incremental update package, and the semantic incremental publishing module issues a navigation node update instruction to the end-side rendering interaction module according to the script version tag.

9. A virtual car showroom display method based on dismantled vehicle data, using a system as described in any one of claims 1 to 8, characterized in that, include: Acquire vehicle dismantling data, dismantling reports, and presentation materials; perform field mapping, unit of measurement normalization, and version marking to generate a structured evidence dataset. Based on the structured evidence dataset, a disassembly evidence diagram is constructed, and based on the vehicle 3D scene, a display scene diagram is constructed. A mapping relationship between the disassembly evidence diagram and the display scene diagram is established according to the component identifier. Generate geometric and texture assets of different precision for each part identifier, and generate an asset index associated with the part identifier; A baseline snapshot is generated based on the displayed scene diagram, and an incremental update package oriented towards component identifiers is generated based on the mapping relationship and the version tag. Based on terminal capability parameters, network status parameters, and user interaction trajectories, the baseline snapshot and the incremental update package are semantically prioritized and distributed through different channels. Load the corresponding geometric assets and texture assets according to the asset index, parse the incremental update package to update the scene node status, and respond to component interaction queries according to the mapping relationship; Authorization verification, source records, and call logs are performed on the structured evidence dataset, the asset index, and the incremental update package.

Images