Building visualization method and system based on BIM and two-dimensional code technology

Abstract

The invention relates to the technical field of building information data processing, in particular to a building visualization method and system based on BIM and two-dimensional code technologies. A building visualization method based on a BIM and a two-dimensional code technology comprises the following steps: S1, decomposing a whole building construction process into N independent construction links, marking a list of all parts for each independent construction link, and determining cross parts and associated independent construction links based on the list of all parts; s2, obtaining the state attribute and the operation requirement of each cross part in the associated independent construction link, and constructing a construction process logic relation network based on the state attribute and the operation requirement. According to the method, self-adaptive data management is realized by decomposing the building construction process, extracting the BIM component data, constructing the association matrix and construction process logic relation network and dynamically determining the preparation updating and caching object, and the data consistency, the access performance and the construction decision real-time performance are improved.

Description

Technical Field

[0001] This invention relates to the field of building information data processing technology, and in particular to a building visualization method and system based on BIM and QR code technology. Background Technology

[0002] Building Information Modeling (BIM) technology comprehensively describes the attributes and spatial relationships of building components digitally, while QR code technology provides an effective means to quickly link physical components with digital information. The combined approach to building visualization can map dynamic information such as construction progress and component status into a 3D model in real time, providing intuitive data support for construction site management. Existing methods, by establishing component-level data association and status tracking mechanisms, have achieved significant results in improving information acquisition efficiency and reducing construction errors, laying the foundation for digital management of building projects.

[0003] However, existing building visualization methods have significant limitations when dealing with multi-stage collaborative construction scenarios. Traditional data management methods struggle to dynamically identify data dependencies between construction stages, leading to frequent data update conflicts in high-concurrency access environments. Simultaneously, fixed caching strategies cannot adapt to the dynamic changes in on-site operational needs, and mobile devices often experience data loading delays and display inconsistencies in weak network environments, severely impacting the real-time performance and accuracy of construction decisions.

[0004] Therefore, a dynamic management method is needed that can intelligently perceive the data correlation characteristics of the construction process. By establishing a data dependency model and adaptive caching mechanism between construction stages, the data consistency guarantee and access performance of the building visualization system in complex construction environments can be improved. Summary of the Invention

[0005] To overcome the shortcomings of poor data collaborative management, this invention provides a building visualization method and system based on BIM and QR code technology.

[0006] The technical implementation scheme of the present invention is: a building visualization method based on BIM and QR code technology, comprising the following steps: S1: Decompose the overall building construction process into N independent construction stages, identify the complete component list for each independent construction stage, and determine the cross-components and associated independent construction stages based on the complete component list; S2: Obtain the status attributes and operational requirements of each cross component in the associated independent construction stage, and construct a logical relationship network of the construction process based on the status attributes and operational requirements; S3: Represent the logical relationship network of the construction process as an adjacency matrix, and calculate the internal and external correlation of each independent construction link based on the adjacency matrix; S4: Determine the objects to be updated and the objects to be cached based on the internal and external correlations, and formulate a building visualization method based on the objects to be updated and the objects to be cached.

[0007] Preferably, the step of decomposing the overall construction process into N independent construction stages, identifying a complete list of components for each independent construction stage, and determining overlapping components and associated independent construction stages based on the complete list of components includes: Based on the complete component list for each independent construction phase, extract the complete BIM data objects of all corresponding components from the BIM model; The complete BIM data objects are used to construct a dedicated dataset corresponding to each independent construction stage; By comparing the dedicated datasets of all independent construction stages, identify all components that belong to at least two dedicated datasets simultaneously and define them as cross components. At the same time, define independent construction stages that share the same cross component as associated independent construction stages of the cross component.

[0008] Preferably, the step of obtaining the status attributes and operational requirements of each intersecting component in the associated independent construction stages, and constructing a logical relationship network of the construction process based on the status attributes and operational requirements, includes: The state attributes include not started, in progress, and completed states; the operation requirements include preceding operation categories, subsequent operation categories, and mutually exclusive operation categories. When the associated independent construction link of the subsequent operation category requires the cross component to be in a completed state, and the associated independent construction link of the preceding operation category is responsible for making the cross component reach the completed state, then a sequential dependency relationship is established for the associated independent construction link. When two related independent construction stages have mutually exclusive operation requirements for the same cross component, or when two related independent construction stages have conflicting requirements for the target state attribute of the same cross component, a mutually exclusive dependency relationship is established for the related independent construction stages. Integrate the sequential and mutual dependencies of all cross-components to establish a set of logical constraints between independent construction stages; Based on the set of logical constraints, a logical relationship network for the construction process is constructed; the logical relationship network for the construction process is represented as a directed graph, where nodes represent independent construction steps and directed edges represent state dependencies.

[0009] Preferably, the step of representing the logical relationship network of the construction process as an adjacency matrix, and calculating the internal and external correlation degrees of each independent construction stage based on the adjacency matrix, includes: The rows and columns of the matrix represent independent construction stages. A matrix element value of 1 indicates the existence of a corresponding state dependency, while a matrix element value of 0 indicates the absence of a corresponding state dependency. Based on the adjacency matrix, the internal correlation degree of each independent construction stage is calculated. The internal correlation degree is measured by degree centrality and is defined as the sum of the in-degree and out-degree of a node. Based on the adjacency matrix, the external correlation degree of each independent construction stage is calculated, and the external correlation degree is measured by eigenvector centrality.

[0010] Preferably, the step of determining the pre-update object and the pre-cached object based on the internal and external correlation, and formulating the building visualization method based on the pre-update object and the pre-cached object, includes: Select one or more independent construction stages with the highest values ​​in both internal and external correlation as potential update targets; Based on the proposed update object, the corrected proposed update object is obtained using the update behavior prediction model; Select one or more independent construction stages with the lowest values ​​in the internal and external correlation degrees as potential cache objects; The QR code health score and user demand score are obtained based on the prepared cache object, and the actual cache demand is obtained based on the QR code health score and user demand score. The architectural visualization method is determined based on the corrected pre-update object and the actual caching requirements.

[0011] Preferably, obtaining the corrected pre-update object based on the pre-update object using an update behavior prediction model includes: Using a fixed time window as a unit, record the planned update frequency, actual update frequency, update timestamp, number of overlapping components, and construction stage type for each independent construction stage; calculate the absolute value of the deviation between the actual update frequency and the planned update frequency, and use the ratio of the absolute value of the deviation to the planned update frequency as the update prediction factor; combine the planned update frequency, actual update frequency, update prediction factor, update timestamp, number of overlapping components, and construction stage type to form the historical feature vector of the independent construction stage; Based on the long short-term memory network structure, an update behavior prediction model is constructed; historical feature vectors are used as model inputs, and the corresponding actual update frequency is used as the training target to complete model training. The feature vector of the current independent construction stage is input into the pre-trained update behavior prediction model; the update behavior prediction model outputs the predicted update demand value; all independent construction stages are sorted in descending order according to the predicted update demand value, and the independent construction stage ranked first is selected as the corrected pre-update object.

[0012] Preferably, obtaining the QR code health score and user demand score based on the pre-cached object includes: Record the number of scanning attempts and recognition failures for each cross component's corresponding QR code; obtain the QR code health score according to the QR code health score formula, which is: QR code health score = 1 - (number of recognition failures / number of scanning attempts); Collect user interaction feedback data, including the number of active queries, page dwell time, and number of manual refreshes; normalize the user interaction feedback data, and then obtain the user demand score through weighted summation.

[0013] Preferably, obtaining the actual caching requirement based on the QR code health score and the user demand score includes: The actual caching demand is obtained based on the QR code health score and the user demand score. The actual caching demand calculation formula is: actual caching demand value = weight coefficient A × (1 - health score) + weight coefficient B × user demand score, where weight coefficient A and weight coefficient B are preset parameters that satisfy the condition that weight coefficient A plus weight coefficient B equals one. The higher the actual caching demand value, the more urgent the caching demand.

[0014] Preferably, the method for determining architectural visualization based on the corrected pre-update object and actual caching requirements includes: A data index table is constructed to establish a mapping relationship between cross-component identifiers and complete BIM data objects. When a mobile device scans a QR code, the QR code content is parsed to obtain the cross-component identifier. The corresponding complete BIM data object is retrieved from the data index table based on the cross-component identifier. The validity of the complete BIM data objects in the cache is verified, including checking the data version number and timestamp. When the cached data is invalid, an update request is sent to the server, prioritizing the retrieval of the corrected complete BIM data object to be updated. The mobile device's cache space is managed according to the actual cache demand value, eliminating complete BIM data objects with low actual cache demand values ​​and retaining those with high actual cache demand values. The search results are returned to the mobile device, and the building visualization display is updated.

[0015] Preferably, a building visualization system based on BIM and QR code technology includes: The construction phase decomposition module is used to break down the overall building construction process into independent construction phases, identify the complete component list, extract complete BIM data objects, and define cross-components and associated independent construction phases. The state dependency analysis module is used to obtain the state attributes and operational requirements of cross-components, establish sequential and mutually exclusive dependencies, and construct a logical relationship network of the construction process. The correlation calculation module is used to represent the logical relationship network as an adjacency matrix and calculate the internal and external correlation of each independent construction link. The visualization data management module is used to determine the objects to be updated and the objects to be cached based on their relevance, correct the objects to be updated through an update behavior prediction model, calculate the actual caching requirements based on the health of the QR code and user needs, and execute the building visualization display update.

[0016] Beneficial Effects: This invention decomposes the overall building construction process into independent construction stages and identifies a complete list of components. Based on a BIM model, it extracts complete data objects, accurately defines intersecting components and associated independent construction stages, and constructs a logical relationship network of the construction process centered on state attributes and operational requirements. It quantifies the internal and external correlations between stages using an adjacency matrix, dynamically identifies data dependency characteristics, and determines objects to be updated and cached based on these correlations. An update behavior prediction model built using a Long Short-Term Memory (LSTM) network corrects the update objects, and the actual caching requirements are calculated by combining QR code health scores and user demand scores, achieving adaptive data management and caching optimization. This invention effectively overcomes the limitations of traditional building visualization methods, such as frequent data update conflicts, rigid caching strategies, and data loading delays in weak network environments, in multi-stage collaborative scenarios. It significantly improves system data consistency, access performance, and real-time construction decision-making, enhancing intelligent adaptability in complex construction environments. Attached Figure Description

[0017] Figure 1 This is a flowchart of the building visualization method based on BIM and QR code technology of the present invention; Figure 2 This is a structural diagram of the building visualization system based on BIM and QR code technology of the present invention. Detailed Implementation

[0018] The present invention will be further described below with reference to specific embodiments. The illustrative embodiments and descriptions herein are used to explain the present invention, but are not intended to limit the present invention.

[0019] Example 1: A building visualization method based on BIM and QR code technology, such as Figure 1 As shown, it includes the following steps: S1: Decompose the overall construction process into N independent construction stages, identify a complete list of components for each independent construction stage, and determine the intersecting components and related independent construction stages based on the complete list of components, including: Based on the complete component list for each independent construction phase, extract the complete BIM data objects of all corresponding components from the BIM model; The complete BIM data objects are used to construct a dedicated dataset corresponding to each independent construction stage; By comparing the dedicated datasets of all independent construction stages, identify all components that belong to at least two dedicated datasets simultaneously and define them as cross components. At the same time, define independent construction stages that share the same cross component as associated independent construction stages of the cross component.

[0020] It should be noted that during the implementation of a building construction project, the overall construction process typically consists of multiple interrelated construction stages, such as pipe installation, electrical wiring, and structural reinforcement. These stages are further broken down into independent construction links, each corresponding to specific tasks. Taking the construction of a water supply and drainage system as an example, it is broken down into four independent construction links: pipe prefabrication, pipe installation, valve installation, and water pressure testing. For each independent construction link, by scanning the design drawings and construction plan, a complete list of all components required for that link is identified. For example, the component list for the pipe installation link includes various types of pipes, fittings, and connectors.

[0021] Complete BIM data objects for corresponding components are extracted from the BIM model. These objects contain complete parameters such as geometric information, material properties, and construction status. These data objects are then assembled into a dedicated dataset for each independent construction phase, forming the digital foundation for that phase. By comparing the dedicated datasets from different phases, components that exist simultaneously in multiple datasets are identified; these components are defined as cross-components. For example, the main pipeline in a water supply and drainage system appears in both the pipeline installation and water pressure testing phases, thus being identified as a cross-component. Independent construction phases sharing the same cross-component are defined as associated independent construction phases; this association reveals the data dependency characteristics between construction phases.

[0022] Traditional building visualization methods often lead to data update conflicts and display inconsistencies because they fail to recognize such data relationships. This solution establishes a correspondence between intersecting components and related independent construction stages, laying a solid foundation for subsequent collaborative data management and effectively solving the data consistency problem in multi-stage collaborative operations.

[0023] It should be further clarified that the "associated independent construction stages" refers to a specific cross-component. The number of associated independent construction stages corresponds to the number of dedicated datasets containing a cross-component. Therefore, the number of associated independent construction stages is at least two. However, in complex construction processes, a key cross-component can be shared by three or more construction stages, forming an associated set composed of multiple stages. The logical relationship network of the construction process constructed in this solution is precisely designed to characterize and manage such diverse and complex inter-stage dependencies.

[0024] S2: Obtain the status attributes and operational requirements of each cross component in the associated independent construction stages, and construct a logical relationship network of the construction process based on the status attributes and operational requirements, including: The state attributes include not started, in progress, and completed states; the operation requirements include preceding operation categories, subsequent operation categories, and mutually exclusive operation categories. When the associated independent construction link of the subsequent operation category requires the cross component to be in a completed state, and the associated independent construction link of the preceding operation category is responsible for making the cross component reach the completed state, then a sequential dependency relationship is established for the associated independent construction link. When two related independent construction stages have mutually exclusive operation requirements for the same cross component, or when two related independent construction stages have conflicting requirements for the target state attribute of the same cross component, a mutually exclusive dependency relationship is established for the related independent construction stages. Integrate the sequential and mutual dependencies of all cross-components to establish a set of logical constraints between independent construction stages; Based on the set of logical constraints, a logical relationship network for the construction process is constructed; the logical relationship network for the construction process is represented as a directed graph, where nodes represent independent construction steps and directed edges represent state dependencies.

[0025] It should be noted that in construction projects, the state attributes and operational requirements of intersecting components form the core foundation of the logical relationship network of the construction process. State attributes include "not started," "in progress," and "completed," representing the initial preparation, execution process, and final completion status of a component in the construction phase, respectively. Operational requirements are divided into prerequisite operations, subsequent operations, and mutually exclusive operations, corresponding to the initiation prerequisites, follow-up actions, and conflicting operations of a construction phase, respectively. These state attributes and operational requirements are obtained through the following methods: parsing the project's construction schedule document, extracting the planned start and end times of each construction phase, and comparing them with real-time collected on-site construction progress reports to determine the real-time state attributes of the components; simultaneously, by parsing the construction organization design and process specification documents, extracting the operational logic and constraints of each construction phase on the same component, and defining its operational requirement category. The state dependencies refer to sequential dependencies and mutually exclusive dependencies.

[0026] The establishment of sequential dependencies is based on the logical connection of operation categories: specifically, when a subsequent operation requires a cross-component to be in a completed state, and a preceding operation is responsible for achieving that state, a sequential dependency is automatically established. Taking pipeline installation and water pressure testing as an example, water pressure testing, as a subsequent operation, requires pipeline installation to be completed as a prerequisite, thus forming a dependency chain from installation to testing. Mutually exclusive dependencies arise from operational conflicts or contradictory target states. For example, if two processes make mutually exclusive operation requests for the same component (such as simultaneously requesting exclusive modification permissions), then a mutually exclusive dependency exists between them. This mechanism effectively solves the problem of out-of-order data updates caused by the lack of dependencies in traditional methods.

[0027] By integrating the sequential and mutually exclusive dependencies of all intersecting components, a set of logical constraint relationships is constructed. Specifically, the "integration" process refers to: traversing each intersecting component and adding all its sequential and mutually exclusive dependencies to the initially empty set of logical constraint relationships in the form of triplets of (component A, component B, relationship type). If multiple relationships of the same type exist between the same pair of components due to different intersecting components, only one is retained in the set to ensure the uniqueness of the set elements. For example, suppose a construction process includes four independent components: pipe installation (A), valve installation (B), water pressure testing (C), and system registration (D). There are two intersecting components: Intersecting component 1 (main pipe): associated with components A and C. Component C requires the main pipe to be in a "completed" state, and component A is responsible for achieving this state, thus generating a sequential dependency relationship: (A, C, sequence). Intersecting component 2 (control valve): associated with components B and C. Component C requires the valve to be in a "completed" state, and component B is responsible for achieving this state, thus generating another sequential dependency relationship: (B, C, sequence). Meanwhile, links A and B have conflicting operational requirements for controlling the valves (e.g., both require exclusive installation permissions), thus generating a mutually exclusive dependency: (A, B, mutually exclusive). Integrating all these dependencies, the resulting set of logical constraints is: {(A, C, order), (B, C, order), (A, B, mutually exclusive)}. Subsequently, when constructing the construction process logical relationship network (directed graph), this set is used: for each (X, Y, order) triple in the set, a directed edge is added to the network from node X to node Y. For each (X, Y, mutually exclusive) triple in the set, a special label is added to the network (e.g., adding a "mutually exclusive" attribute to the corresponding node), or a specific value (e.g., -1) is used in the adjacency matrix to record its mutual exclusion constraint, rather than a direct flow dependency. In this way, the set of logical constraints acts as a clear and unambiguous bridge from cross-component dependencies to the networked model, thereby generating the construction process logical relationship network. The network is presented as a directed graph, with nodes representing independent construction stages and directed edges representing the direction of state dependencies. For example, by integrating the dependencies between pipeline installation, valve commissioning, and system testing, a network structure with clear temporal constraints is formed. This network modeling enables a visual representation of the dynamic relationships between construction stages, providing a structured foundation for subsequent collaborative data management and improving the ability to maintain data consistency in multi-stage concurrent scenarios.

[0028] S3: Represent the logical relationship network of the construction process as an adjacency matrix, and calculate the internal and external correlation degrees of each independent construction link based on the adjacency matrix, including: The rows and columns of the matrix represent independent construction stages. A matrix element value of 1 indicates the existence of a corresponding state dependency, while a matrix element value of 0 indicates the absence of a corresponding state dependency. Based on the adjacency matrix, the internal correlation degree of each independent construction stage is calculated. The internal correlation degree is measured by degree centrality and is defined as the sum of the in-degree and out-degree of a node. Based on the adjacency matrix, the external correlation degree of each independent construction stage is calculated, and the external correlation degree is measured by eigenvector centrality.

[0029] It should be noted that the logical relationship network of the construction process is mathematically modeled using an adjacency matrix. The matrix's row and column indices correspond to independent construction stages, and the values ​​of the elements depend on the existence of state dependencies. Taking pipeline installation, valve commissioning, and water pressure testing as examples, if the installation stage depends on the commissioning stage, the corresponding position in the matrix is ​​assigned a value of 1; otherwise, it is 0. The adjacency matrix transforms the network topology into a discrete data structure, providing a foundation for subsequent quantitative analysis.

[0030] Internal connectivity is measured using degree centrality, which calculates the sum of a node's in-degree and out-degree. This value reflects the connection density of a node in the local network; a high value indicates that the node undertakes more data flow functions, while a low value suggests a marginal status. External connectivity is evaluated using eigenvector centrality, calculated based on the eigenvectors of the adjacency matrix. A higher value means that the node is associated with multiple core nodes and has a radiating influence in the global network.

[0031] Traditional methods often lead to inaccurate data update priorities due to the lack of quantitative assessment of the relationships between construction stages. By using adjacency matrix modeling and centrality calculation, key stages can be accurately identified, providing decision support for the dynamic allocation of data resources and thus improving system response efficiency in multi-task concurrent environments.

[0032] After the logical relationship network of the construction process is constructed, it is represented as an adjacency matrix, and the internal and external correlation degrees of each independent construction stage are calculated accordingly. The following uses a simplified construction network as an example for illustration: Assume the construction process comprises four independent stages: pipeline installation (A), valve commissioning (B), water pressure testing (C), and system registration (D). The adjacency matrix M formed by their logical relationship network is as follows: In this matrix, both rows and columns are represented in the order [A,B,C,D], and matrix elements... This indicates that the i-th step has a state dependency on the j-th step (i.e., j depends on the completion of i). For example, This indicates that there is a state dependency relationship between stage A (row 0) and stage B (column 1), that is, stage B depends on the completion of stage A.

[0033] 1. Calculation of internal correlation (degree centrality): The internal correlation is defined as the sum of the in-degree and out-degree of a node. The in-degree is the sum of the corresponding elements in the column (indicating how many links depend on it), and the out-degree is the sum of the corresponding elements in the row (indicating how many links it depends on). For link A: in-degree = 0, out-degree = 1 + 1 + 0 + 0 = 2, internal correlation = 0 + 2 = 2. For link B: in-degree = 1 (from A), out-degree = 1 (pointing to C), internal correlation = 1 + 1 = 2. For link C: in-degree = 1 (from A) + 1 (from B) = 2, out-degree = 1 (pointing to D), internal correlation = 2 + 1 = 3. For link D: in-degree = 1 (from C), out-degree = 0, internal correlation = 1 + 0 = 1. The calculation results show that link C has the largest value of internal correlation in this network.

[0034] 2. Calculation of external correlation (eigenvector centrality): The external correlation is measured by eigenvector centrality. The core idea is that the importance of a link depends on the importance of its neighbor links. This calculation is achieved by solving the principal eigenvector of the adjacency matrix M. After calculation, the principal eigenvector of matrix M (representing the eigenvector centrality scores of each link) is approximately: ; The calculation results show that the eigenvector centrality score of link C (about 0.71) is the highest, followed by link A (about 0.61). This indicates that link C not only has a dense local connection (large internal correlation), but also the objects it connects (such as link A) itself have high influence, so its radiation influence (external correlation) in the global network is also the largest.

[0035] In summary, based on the quantitative calculation of the adjacency matrix, the key links in the construction network can be accurately identified, providing a decision-making basis for subsequent dynamic data management.

[0036] S4: Determine the preliminary update object and preliminary cache object according to the internal correlation and external correlation, and formulate a building visualization method based on the preliminary update object and preliminary cache object, including: Select one or more independent construction links with the top-ranked values in the internal correlation and external correlation as the preliminary update object; In specific implementation, the quantity of the "one or more" is limited by a preset parameter K, and K is a positive integer. The setting of the parameter K is based on the following method: At the initial stage of system deployment, according to the total number N of independent construction links in the initial project design, determine the initial value of K from the preset mapping relationship (for example: when N ≤ 10, K = 2; when 10 < N ≤ 30, K = 3; when N > 30, K = 4); During the system operation, continuously monitor the key data update conflict rate and average response delay performance indicators. If the conflict rate or delay continuously exceeds the preset threshold, start the dynamic adjustment mechanism, and fine-tune and re-evaluate the K value in steps (such as ±1) until the system performance indicators return to the expected range.

[0037] Based on the proposed update object, the corrected proposed update object is obtained using the update behavior prediction model, including: Using a fixed time window as a unit, record the planned update frequency, actual update frequency, update timestamp, number of overlapping components, and construction stage type for each independent construction stage; calculate the absolute value of the deviation between the actual update frequency and the planned update frequency, and use the ratio of the absolute value of the deviation to the planned update frequency as the update prediction factor; combine the planned update frequency, actual update frequency, update prediction factor, update timestamp, number of overlapping components, and construction stage type to form the historical feature vector of the independent construction stage; Based on the long short-term memory network structure, an update behavior prediction model is constructed; historical feature vectors are used as model inputs, and the corresponding actual update frequency is used as the training target to complete model training. It should be noted that the hyperparameters used for model training include: 128 hidden layer units, a learning rate of 0.001, the Adam optimizer, mean squared error as the loss function, and at least 100 training iterations until the loss function converges on the validation set. The model's training data must cover the entire construction history of at least one complete building project.

[0038] The feature vector of the current independent construction stage is input into the pre-trained update behavior prediction model; the update behavior prediction model outputs the predicted update demand value; all independent construction stages are sorted in descending order according to the predicted update demand value, and the independent construction stage ranked first is selected as the corrected pre-update object.

[0039] It should be noted that in the building visualization data management system, one or more independent construction stages with high internal and external correlation values ​​are selected as potential update targets because these stages occupy a core and pivotal position in the logical relationship network of the construction process. A high internal correlation value indicates that this stage has direct and close data interaction with many other stages in the local network, and changes in its data state will directly and quickly affect multiple downstream stages. A high external correlation value indicates that this stage is closely related to other core pivotal stages in the network, and changes in its data state will have a ripple effect globally, affecting the coordination and data consistency of the entire construction process. Therefore, ensuring the real-time and accuracy of data from these core pivotal stages is crucial for maintaining the data consistency of the entire system and improving the real-time nature of construction decisions; hence, they are prioritized as potential update targets.

[0040] In the building visualization data management system, the objects to be updated refer to the core construction stages that require priority in ensuring data real-time performance, while the objects to be cached correspond to auxiliary stages that need to preload data in weak network environments. Traditional static allocation methods often lead to delays in critical data updates and wasted caching resources because they ignore the dynamic characteristics of the construction process. This solution comprehensively evaluates the structural importance of each stage in the network and selects the independent construction stage with the highest comprehensive evaluation of internal and external correlation as the objects to be updated.

[0041] The update frequency is measured in units of fixed time windows (e.g., 1 hour, 4 hours) that are dynamically adjusted according to the construction schedule or preset. The planned update frequency reflects the planned update density, the actual update frequency reflects the execution status, timestamps record the operational sequence, the number of overlapping components characterizes the complexity of the process relationships, and the type of construction process distinguishes the operational characteristics. These parameters together constitute the historical feature vector of an independent construction process. The absolute deviation between the actual and planned update frequencies indicates execution stability, and its ratio to the planned value serves as an update prediction factor to quantify the intensity of demand fluctuations.

[0042] An update behavior prediction model based on a Long Short-Term Memory (LSTM) network establishes a mapping relationship from multi-dimensional features to actual update requirements by analyzing temporal patterns in historical feature vectors. During model training, the actual update frequency is used as a supervision signal, and the network's gating mechanism captures long-term dependent features. After the predicted update requirement values ​​output by the model are sorted in descending order, the top-ranked independent construction stage is identified as the corrected candidate for update. Compared to the initial static selection based on network topology, the model-corrected target incorporates real-time behavioral features, more accurately reflecting the actual data requirements in a dynamic construction environment.

[0043] This data-driven dynamic optimization mechanism effectively overcomes the rigidity of traditional methods in resource allocation. By continuously learning the time evolution of the construction process, it significantly improves the accuracy of data updates and caching efficiency of the system under complex working conditions.

[0044] Select one or more independent construction stages with the lowest values ​​in the internal and external correlation degrees as potential cache objects; Based on the prepared cache object, obtain the QR code health score and user demand score, including: Record the number of scanning attempts and recognition failures for each cross component's corresponding QR code; obtain the QR code health score according to the QR code health score formula, which is: QR code health score = 1 - (number of recognition failures / number of scanning attempts); Collect user interaction feedback data, including the number of active queries, page dwell time, and number of manual refreshes; normalize the user interaction feedback data, and then obtain the user demand score through weighted summation.

[0045] It's important to note that in the caching optimization phase of the building visualization system, one or more independent construction stages with the lowest internal and external correlation values ​​are selected as potential cache targets. This is because these stages are located at the edge of the data interaction network within the construction process's logical relationship network. A low internal correlation value indicates that the stage is sparsely connected in the local network, has weak direct data dependencies with other stages, and its data state tends to be stable with a low update frequency. A low external correlation value indicates that the stage is at the end of the global network's influence, far from the core hub stages, and its data changes will hardly disturb other parts of the network. Therefore, the data from these stages exhibits high stability and low update requirements, making them ideal for long-term caching on mobile devices. This ensures fast data access in weak network environments and optimizes the allocation of server resources and network bandwidth.

[0046] The number of QR code scanning attempts and recognition failures are obtained through mobile scanning logs. A higher number of recognition failures results in a lower health score, reflecting a risk of decreased QR code readability. User interaction data is collected through event tracking technology, including the number of active queries, page dwell time, and manual refreshes. These metrics respectively represent user attention intensity, perceived content value, and data real-time requirements.

[0047] The original data is processed using a min-max normalization method, linearly transforming each indicator to the zero-to-one interval. When setting weighted summation, the weighting coefficients must consider both hardware status and user experience; typically, user demand scores are given higher weight to reflect business priority. Traditional fixed caching strategies often result in low cache hit rates because they ignore physical identifier overhead and changes in user behavior. This solution dynamically perceives on-site needs through a dual-scoring mechanism, effectively improving data accessibility and system response efficiency in weak network environments.

[0048] When the number of scan attempts is greater than zero, the QR code health score is obtained according to the QR code health score formula; when the number of scan attempts is zero, the QR code health score is 1 by default.

[0049] And based on the QR code health score and user demand score, the actual caching requirements are obtained, including: The actual caching demand is obtained based on the QR code health score and the user demand score. The actual caching demand calculation formula is: actual caching demand value = weight coefficient A × (1 - health score) + weight coefficient B × user demand score, where weight coefficient A and weight coefficient B are preset parameters that satisfy the condition that weight coefficient A plus weight coefficient B equals one. The higher the actual caching demand value, the more urgent the caching demand.

[0050] It's important to note that the calculation mechanism for actual cache requirements comprehensively evaluates both the physical identifier status and user behavior characteristics. A decrease in the QR code health score increases the probability of recognition failure, leading to increased cache requirements; conversely, an increase in the user demand score leads to increased access frequency, similarly driving up cache requirements. Weighting coefficients are allocated according to application scenario characteristics; if the construction environment presents a high risk of wear and tear, the health score weight is increased to enhance fault tolerance. Unlike traditional statically selected candidate cache objects, actual cache requirements dynamically adjust the initial candidate set through quantitative indicators, transforming preliminary screening based on network topology into data-driven, precise decision-making.

[0051] The preliminary cache objects are independent construction stages selected based on their network relevance (internal and external relevance values ​​are ranked last) as initial candidates for cache management. These objects are located at the edge of the network topology, have low data update requirements, and are suitable for long-term cache content. The actual cache requirement is calculated for the preliminary cache objects by comprehensively evaluating the QR code health (reflecting the readability of the physical identifier) ​​and user interaction requirements (reflecting user attention). This requirement value quantifies the urgency of caching the object. The system sorts all preliminary cache objects according to this requirement value, forming a cache priority queue to guide the dynamic management of mobile cache space (e.g., evicting low-requirement objects and retaining high-requirement objects). In other words, the preliminary cache objects are "candidate targets," while the actual cache requirement is a "priority score," and the combination of the two enables adaptive cache optimization.

[0052] This optimization model, which integrates real-time perception, effectively solves the problem of insufficient adaptability of fixed strategies under complex working conditions and significantly improves the efficiency of cache resource allocation. Weight coefficients A and B are typically initially set through historical data analysis and further optimized through A / B testing in a construction test environment: different work areas or teams in the same construction phase are divided into groups A and B, with different combinations of weight coefficients applied to each. By comparing the average data loading latency, cache hit rate, and user satisfaction feedback of the two groups within a fixed period, the parameter combination with better overall performance is selected as the final setting.

[0053] The preliminary cache objects are independent construction stages selected based on their lowest network relevance (both internal and external relevance values). These stages serve as initial candidates for cache management. Located at the edge of the network topology, these stages have low data update requirements and are suitable for long-term caching. The actual cache requirement, on the other hand, is calculated for the preliminary cache objects by comprehensively evaluating QR code health (reflecting the readability of the physical identifier) ​​and user interaction requirements (reflecting user engagement). This requirement value quantifies the urgency of caching the object and guides the dynamic management of mobile cache space (e.g., evicting low-requirement objects and retaining high-requirement objects). In short, the preliminary cache objects are "candidate targets," while the actual cache requirement is a "priority score," and their combination enables adaptive cache optimization.

[0054] The architectural visualization method is determined based on the corrected pre-update object and the actual caching requirements, including: A data index table is constructed to establish a mapping relationship between cross-component identifiers and complete BIM data objects. When a mobile device scans a QR code, the QR code content is parsed to obtain the cross-component identifier. The corresponding complete BIM data object is retrieved from the data index table based on the cross-component identifier. The validity of the complete BIM data objects in the cache is verified, including checking the data version number and timestamp. When the cached data is invalid, an update request is sent to the server, prioritizing the retrieval of the corrected complete BIM data object to be updated. The mobile device's cache space is managed according to the actual cache demand value, eliminating complete BIM data objects with low actual cache demand values ​​and retaining those with high actual cache demand values. The search results are returned to the mobile device, and the building visualization display is updated.

[0055] It's important to note that the architectural visualization method relies on a structured data indexing mechanism. A hash table is created to construct a data index table, establishing a bidirectional mapping between cross-component identifiers and complete BIM data objects, using cross-component identifiers as keys. This mapping allows mobile devices to quickly and accurately trigger the intelligent retrieval, updating, and visualization of the entire set of digital information for a component when scanning a QR code (where the QR code acts as a precise medium connecting the physical component and the digital BIM model: each QR code encodes a corresponding cross-component identifier, and scanning this graphical code enables the system to quickly and accurately retrieve, update, and visualize the component's complete set of digital information). The embedded cross-component identifier—a unique numerical code—is rapidly extracted using a parsing algorithm. When verifying the validity of cached data, the data version number and timestamp are compared. If they are inconsistent with the server record, the cached data is deemed invalid. The data version number is automatically incremented by the server with each update of the BIM data object and is globally unique. The timestamp uses a standard time synchronized from a unified time server. Update requests sent to the server at this time will prioritize retrieving the corrected, pre-updated object data, as it has been optimized by a predictive model and has the highest real-time requirements.

[0056] Mobile cache space management employs a dynamic eviction strategy. When storage space utilization reaches a preset threshold (e.g., 80% of total capacity), the last-ranked complete BIM data object in terms of actual cache demand is automatically removed, while high-demand data is retained. This mechanism ensures rapid response to frequently accessed data while avoiding storage resource waste. Retrieval results update the building visualization display via the rendering engine, forming a complete closed loop from data request to interface update. Traditional methods, lacking dynamic indexing and priority update mechanisms, often result in data loading delays and display lag. This solution, by establishing an intelligent data flow system, significantly improves the real-time visualization and data consistency in multi-terminal collaborative scenarios.

[0057] Example 2: Based on Example 1, a building visualization system based on BIM and QR code technology, such as... Figure 2 As shown, it includes: The construction phase decomposition module is used to break down the overall building construction process into independent construction phases, identify the complete component list, extract complete BIM data objects, and define cross-components and associated independent construction phases. The state dependency analysis module is used to obtain the state attributes and operational requirements of cross-components, establish sequential and mutually exclusive dependencies, and construct a logical relationship network of the construction process. The correlation calculation module is used to represent the logical relationship network as an adjacency matrix and calculate the internal and external correlation of each independent construction link. The visualization data management module is used to determine the objects to be updated and the objects to be cached based on their relevance, correct the objects to be updated through an update behavior prediction model, calculate the actual caching requirements based on the health of the QR code and user needs, and execute the building visualization display update.

[0058] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A building visualization method based on BIM and QR code technology, characterized in that, Includes the following steps: S1: Decompose the overall building construction process into N independent construction stages, identify the complete component list for each independent construction stage, and determine the cross-components and associated independent construction stages based on the complete component list; S2: Obtain the status attributes and operational requirements of each cross component in the associated independent construction stage, and construct a logical relationship network of the construction process based on the status attributes and operational requirements; S3: Represent the logical relationship network of the construction process as an adjacency matrix, and calculate the internal and external correlation of each independent construction link based on the adjacency matrix; S4: Determine the objects to be updated and the objects to be cached based on the internal and external correlations, and formulate a building visualization method based on the objects to be updated and the objects to be cached.

2. The building visualization method based on BIM and QR code technology according to claim 1, characterized in that, The process of decomposing the overall building construction process into N independent construction stages, identifying a complete list of components for each independent construction stage, and determining overlapping components and associated independent construction stages based on the complete list of components includes: Based on the complete component list for each independent construction phase, extract the complete BIM data objects of all corresponding components from the BIM model; The complete BIM data objects are used to construct a dedicated dataset corresponding to each independent construction stage; By comparing the dedicated datasets of all independent construction stages, identify all components that belong to at least two dedicated datasets simultaneously and define them as cross components. At the same time, define independent construction stages that share the same cross component as associated independent construction stages of the cross component.

3. The building visualization method based on BIM and QR code technology according to claim 1, characterized in that, The step of obtaining the status attributes and operational requirements of each intersecting component in the associated independent construction stages, and constructing a logical relationship network of the construction process based on the status attributes and operational requirements, includes: The state attributes include not started, in progress, and completed states; the operation requirements include preceding operation categories, subsequent operation categories, and mutually exclusive operation categories. When the associated independent construction link of the subsequent operation category requires the cross component to be in a completed state, and the associated independent construction link of the preceding operation category is responsible for making the cross component reach the completed state, then a sequential dependency relationship is established for the associated independent construction link. When two related independent construction stages have mutually exclusive operation requirements for the same cross component, or when two related independent construction stages have conflicting requirements for the target state attribute of the same cross component, a mutually exclusive dependency relationship is established for the related independent construction stages. Integrate the sequential and mutual dependencies of all cross-components to establish a set of logical constraints between independent construction stages; Based on the set of logical constraints, a logical relationship network for the construction process is constructed; the logical relationship network for the construction process is represented as a directed graph, where nodes represent independent construction steps and directed edges represent state dependencies.

4. The building visualization method based on BIM and QR code technology according to claim 1, characterized in that, The step of representing the logical relationship network of the construction process as an adjacency matrix, and calculating the internal and external correlation degrees of each independent construction stage based on the adjacency matrix, includes: The rows and columns of the matrix represent independent construction stages. A matrix element value of 1 indicates the existence of a corresponding state dependency, while a matrix element value of 0 indicates the absence of a corresponding state dependency. Based on the adjacency matrix, the internal correlation degree of each independent construction stage is calculated. The internal correlation degree is measured by degree centrality and is defined as the sum of the in-degree and out-degree of a node. Based on the adjacency matrix, the external correlation degree of each independent construction stage is calculated, and the external correlation degree is measured by eigenvector centrality.

5. A building visualization method based on BIM and QR code technology according to claim 1, characterized in that, The step of determining the pre-update objects and pre-cached objects based on the internal and external correlation degrees, and formulating a building visualization method based on the pre-update objects and pre-cached objects, includes: Select one or more independent construction stages with the highest values ​​in both internal and external correlation as potential update targets; Based on the proposed update object, the corrected proposed update object is obtained using the update behavior prediction model; Select one or more independent construction stages with the lowest values ​​in the internal and external correlation degrees as potential cache objects; The QR code health score and user demand score are obtained based on the prepared cache object, and the actual cache demand is obtained based on the QR code health score and user demand score. The architectural visualization method is determined based on the corrected pre-update object and the actual caching requirements.

6. A building visualization method based on BIM and QR code technology according to claim 5, characterized in that, The step of obtaining the corrected pre-update object based on the pre-update object using an update behavior prediction model includes: Using a fixed time window as a unit, record the planned update frequency, actual update frequency, update timestamp, number of overlapping components, and construction stage type for each independent construction stage; calculate the absolute value of the deviation between the actual update frequency and the planned update frequency, and use the ratio of the absolute value of the deviation to the planned update frequency as the update prediction factor; combine the planned update frequency, actual update frequency, update prediction factor, update timestamp, number of overlapping components, and construction stage type to form the historical feature vector of the independent construction stage; Based on the long short-term memory network structure, an update behavior prediction model is constructed; historical feature vectors are used as model inputs, and the corresponding actual update frequency is used as the training target to complete model training. The feature vector of the current independent construction stage is input into the pre-trained update behavior prediction model; the update behavior prediction model outputs the predicted update demand value; all independent construction stages are sorted in descending order according to the predicted update demand value, and the independent construction stage ranked first is selected as the corrected pre-update object.

7. A building visualization method based on BIM and QR code technology according to claim 5, characterized in that, The step of obtaining the QR code health score and user demand score based on the pre-cached object includes: Record the number of scanning attempts and recognition failures for each cross component's corresponding QR code; obtain the QR code health score according to the QR code health score formula, which is: QR code health score = 1 - (number of recognition failures / number of scanning attempts); Collect user interaction feedback data, including the number of active queries, page dwell time, and number of manual refreshes; normalize the user interaction feedback data, and then obtain the user demand score through weighted summation.

8. A building visualization method based on BIM and QR code technology according to claim 7, characterized in that, The step of obtaining the actual caching requirement based on the QR code health score and user demand score includes: The actual caching demand is obtained based on the QR code health score and the user demand score. The actual caching demand calculation formula is: actual caching demand value = weight coefficient A × (1 - health score) + weight coefficient B × user demand score, where weight coefficient A and weight coefficient B are preset parameters that satisfy the condition that weight coefficient A plus weight coefficient B equals one. The higher the actual caching demand value, the more urgent the caching demand.

9. A building visualization method based on BIM and QR code technology according to claim 5, characterized in that, The method for determining architectural visualization based on the corrected pre-update object and actual caching requirements includes: A data index table is constructed to establish a mapping relationship between cross-component identifiers and complete BIM data objects. When a mobile device scans a QR code, the QR code content is parsed to obtain the cross-component identifier. The corresponding complete BIM data object is retrieved from the data index table based on the cross-component identifier. The validity of the complete BIM data objects in the cache is verified, including checking the data version number and timestamp. When the cached data is invalid, an update request is sent to the server, prioritizing the retrieval of the corrected complete BIM data object to be updated. The mobile device's cache space is managed according to the actual cache demand value, eliminating complete BIM data objects with low actual cache demand values ​​and retaining those with high actual cache demand values. The search results are returned to the mobile device, and the building visualization display is updated.

10. A building visualization system based on BIM and QR code technology, used to implement the building visualization method based on BIM and QR code technology as described in any one of claims 1-9, characterized in that, include: The construction phase decomposition module is used to break down the overall building construction process into independent construction phases, identify the complete component list, extract complete BIM data objects, and define cross-components and associated independent construction phases. The state dependency analysis module is used to obtain the state attributes and operational requirements of cross-components, establish sequential and mutually exclusive dependencies, and construct a logical relationship network of the construction process. The correlation calculation module is used to represent the logical relationship network as an adjacency matrix and calculate the internal and external correlation of each independent construction link. The visualization data management module is used to determine the objects to be updated and the objects to be cached based on their relevance, correct the objects to be updated through an update behavior prediction model, calculate the actual caching requirements based on the health of the QR code and user needs, and execute the building visualization display update.

Images