A construction progress management method and device based on asymmetric bidirectional complementarity

Abstract

This invention provides a construction progress management method and apparatus based on asymmetric bidirectional complementarity, relating to the field of building construction progress management technology. This invention acquires construction semantic graphs, BIM semantic graphs, and component-level mapping relationship sets; performs incremental information identification and construction oriented towards differences to form incremental information units; constructs and determines the complementary direction based on virtual-real asymmetric quantification of complementary indices; performs hierarchical control and complementary calculation based on complementary indices to achieve asymmetric updates; performs unified expression of construction scenarios and progress reasoning driven by incremental information; and performs iterative updates and progress warnings based on difference-driven mechanisms. This application can transform virtual-real differences into engineering semantic increments, utilizes a complementary mechanism to solve the problem of inaccurate progress judgment caused by occlusion, and achieves dynamic closed-loop updates and accurate warnings throughout the construction process, significantly improving the accuracy and automation level of digital twin construction management.

Description

Technical Field

[0001] This invention relates to the field of construction progress management technology, and more specifically, to a construction progress management method and apparatus based on asymmetric bidirectional complementarity. Background Technology

[0002] Existing technical solutions typically acquire the actual state of a building construction site through perception technologies such as computer vision and 3D laser scanning, and then spatially align and semantically match it with Building Information Modeling (BIM) to associate the geometric attributes, topological relationships, and semantic features of the components at both the virtual and physical ends.

[0003] However, existing construction progress management methods still have significant shortcomings in application. First, most existing technologies are based on the assumption of information consistency, mainly focusing on the alignment accuracy of virtual and real information, and treating the difference between the two as observation error or matching residual. This results in them only supporting static verification or simple comparison of progress, making it difficult to form an interpretable calculation basis for progress deviations. Second, in complex construction environments, due to factors such as scaffolding obstruction, interference from temporary construction measures, and limited observation conditions, there is an inherent information asymmetry between virtual models and on-site observation information. This asymmetric difference is not a simple error, but reflects the dynamic evolution of the construction process. Existing technologies lack a systematic analysis and utilization mechanism for these differences, making it difficult to extract the incremental information value contained in the differences. This leads to problems such as missing progress information and discontinuous state expression in digital twin models under complex working conditions.

[0004] More critically, existing methods struggle to distinguish between schedule deviations caused by actual construction delays and apparent differences resulting from incomplete observations. Lacking quantitative judgment mechanisms and weighting strategies for addressing the asymmetry between real and virtual information, multi-source information fusion often employs simple alignment or overlay methods. This fails to effectively adjust the participation levels of different sources when data is missing, conflicting, or locally biased, potentially leading to fusion results that deviate from the actual progress. Furthermore, existing technologies generally focus on static alignment or local fusion, lacking dynamic and computable mechanisms to address the continuously generated incremental information, changes in construction sequence, and actual schedule deviations during construction. This makes it difficult to form a unified model for dynamic control of construction progress, resulting in inaccurate progress judgments and insufficient basis for control.

[0005] In view of the above, this application is hereby submitted. Summary of the Invention

[0006] The present invention aims to provide a construction progress management method and device based on asymmetric bidirectional complementarity, in order to solve the technical defects in existing construction management technologies, such as incomplete expression of single-source information, unsystematic handling of virtual and real differences, inaccurate progress judgment under obstructed conditions, and lack of dynamic closed-loop update mechanism.

[0007] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution:

[0008] A construction schedule management method based on asymmetric bidirectional complementarity includes: S1, acquire a construction semantic map reflecting the real-time observation status of the construction site, a BIM semantic map reflecting the status from the BIM design perspective, and a map reflecting the relationship between the construction semantic map and the BIM semantic map. Figure 1 A set of component-level mapping relationships with one-to-one correspondence; S2, construct a virtual-real unified semantic vector for each pair of mapping components in the component-level mapping relationship set to quantify the virtual-real differences and extract incremental information, thereby obtaining a multi-dimensional difference set and a difference classification result set; S3, based on the virtual and real unified semantic vector, introduce a neighborhood aggregation mechanism to model structural context features, combine the differential classification result set to generate incremental information units, calculate the importance measurement function of each incremental information unit, and filter to obtain a subset of key information increments; S4. Extract complementary features based on the incremental subset of key information and the multidimensional difference set, and calculate the complementary index based on the incremental information unit to determine the complementary direction in the complementary calculation. S5, perform hierarchical control based on the complementary index to achieve asymmetric complementary calculation, and perform difference-driven updates on different complementary intervals according to the complementary direction, and output the updated BIM semantic map; S6, based on the updated BIM semantic map, constructs a unified construction scenario expression model, calculates the completion confidence score of each component node, and infers the construction progress status of each component; S7, introduce a time decay factor to update the complementary index, and iteratively update the unified construction scenario expression model through complementary calculation to generate a dynamic construction progress deviation heat map.

[0009] The present invention also provides a construction progress management device based on asymmetric bidirectional complementarity, comprising: The acquisition unit is used to acquire a construction semantic map reflecting the real-time observation status of the construction site, a BIM semantic map reflecting the status from the BIM design perspective, and a map reflecting the relationship between the construction semantic map and the BIM semantic map. Figure 1 A set of component-level mapping relationships with one-to-one correspondence; The virtual-real difference identification unit is used to construct a virtual-real unified semantic vector for each pair of mapping components in the component-level mapping relationship set, so as to quantify the virtual-real difference and extract incremental information to obtain a multi-dimensional difference set and a difference classification result set. The key information calculation unit is used to model structural context features based on the neighborhood aggregation mechanism introduced by the virtual and real unified semantic vector, generate incremental information units by combining the differential classification result set, calculate the importance measurement function of each incremental information unit, and filter to obtain a subset of key information increments. The complementary direction determination unit is used to extract complementary features based on the incremental subset of key information and the multidimensional difference set, and to calculate the complementary index based on the incremental information unit to determine the complementary direction in the complementary calculation. The difference-driven update unit is used to perform hierarchical control based on the complementarity index to achieve asymmetric complementarity calculation, and to perform difference-driven updates on different complementary intervals according to the complementary direction, and output the updated BIM semantic map. The construction progress reasoning unit is used to build a unified construction scenario expression model based on the updated BIM semantic map, calculate the completion confidence score of each component node, and reason to obtain the construction progress status of each component. The iterative update unit is used to introduce a time decay factor to update the complementary index, and to iteratively update the unified construction scenario expression model through complementary calculation to generate a dynamic construction progress deviation heatmap.

[0010] The present invention also provides a construction progress management device based on asymmetric bidirectional complementarity, including a processor and a memory. The memory stores a computer program that can be executed by the processor to realize the construction progress management method based on asymmetric bidirectional complementarity as described above.

[0011] The present invention also provides a computer-readable storage medium storing computer-readable instructions, which, when executed by a processor of the device on which the computer-readable storage medium is located, implement the construction progress management method based on asymmetric bidirectional complementarity as described above.

[0012] In summary, compared with the prior art, the present invention has the following beneficial effects: First, it realizes the valuable utilization of discrepancies in information. This invention redefines the matching errors eliminated in traditional methods as incremental information units with engineering semantics. Through structured parsing and incremental information unit construction, it effectively preserves key data such as hidden project progress and temporary process deviations. This approach solves the problem of incomplete expression of single-source information from the underlying logic, significantly improving the BIM model's ability to capture dynamic changes at the construction site.

[0013] Secondly, precise coordination of virtual and real information is achieved through complementary indices. This invention utilizes the constructed complementary indices to quantitatively characterize the compensation potential of virtual and real information in geometric, topological, and semantic dimensions. Combined with a complementary direction determination function, it executes asymmetric paths of correcting virtual information from real information and supplementing real information from virtual information, enabling the prior logic of the Building Information Model (BIM) and on-site measured data to adaptively adjust according to environmental conditions. This quantitative control mechanism significantly improves the accuracy of the construction state representation, ensuring that the fusion results do not deviate from the true physical state.

[0014] Third, it effectively solves the problem of inaccurate progress determination under occlusion conditions. This invention utilizes topological relationship compensation and a virtual-to-real mechanism to logically deduce visually occluded areas using structural constraints of the Building Information Model. This technique can identify the true progress under visual occlusion, distinguish between unobserved and non-constructed states, and eliminate false progress delays caused by scaffolding and formwork occlusion at the algorithm level, greatly improving the objectivity of progress calculation.

[0015] Fourth, the robustness and noise resistance of the system are enhanced. The neighborhood consistency constraint mechanism and delayed update queue introduced in this invention can automatically identify and suppress spurious differences caused by temporary material stacking, reflections, or personnel movement in the field environment. By temporarily storing rather than discarding low-confidence information (i.e., the medium / low complementarity index range), the system possesses a temporal self-correction capability, ensuring the robust evolution of the model under complex dynamic disturbances.

[0016] Fifth, it drives the automation and intelligence of construction progress management. This invention achieves a direct mapping from physical perception to management decisions, and by automatically inferring component status and generating progress early warning heat maps, it provides a calculable decision-making basis for construction organization and resource scheduling. This closed-loop iterative mechanism enables the digital twin model to continuously evolve with the construction process, significantly improving the automation level of construction management and the timeliness of control. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of a construction progress management method based on asymmetric bidirectional complementarity provided in Example 1.

[0019] Figure 2 This is a schematic diagram of a construction progress management device based on asymmetric bidirectional complementarity provided in Embodiment 2.

[0020] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0022] Example 1 Embodiment 1 of the present invention provides a construction progress management method based on asymmetric bidirectional complementarity, which can be implemented by a construction progress management device based on asymmetric bidirectional complementarity (hereinafter referred to as management device), specifically, executed by one or more processors within the management device.

[0023] In this embodiment, the management device may be an electronic device equipped with a processor, which carries a computer program for the construction progress management method based on asymmetric bidirectional complementarity and the computer program can be executed, such as a computer, smartphone, smart tablet, workstation, etc., which are not limited here.

[0024] In this embodiment, BIM (Building Information Modeling) is a modeling method that expresses the geometric features, semantic attributes, construction procedures, and full life cycle information of a building in a three-dimensional digital way, and is used to support multi-stage applications such as building design, construction, and operation and maintenance.

[0025] In this embodiment, asymmetry refers to the severe mismatch between the virtual BIM model and the real-time observation status of the construction site.

[0026] like Figure 1 As shown, a construction progress management method based on asymmetric bidirectional complementarity includes steps S1 to S7.

[0027] S1, acquire a construction semantic map reflecting the real-time observation status of the construction site, a BIM semantic map reflecting the status from the BIM design perspective, and a map reflecting the relationship between the construction semantic map and the BIM semantic map. Figure 1A set of component-level mapping relationships with one-to-one correspondence.

[0028] In construction digital twin scenarios, the virtual BIM model reflects the design state, while on-site observation reflects the real-time state. Due to dynamic environmental changes, component obstruction, and limited observation conditions, there is an inherent information asymmetry between the virtual and real ends. This inconsistency is not only reflected in the existence of components but also in multiple dimensions such as geometric deviations, construction status, and structural relationships. Traditional alignment matching usually treats the above-mentioned virtual-real inconsistencies as matching errors that need to be eliminated and removes them, resulting in the loss of a large amount of incremental information with progress accounting value (such as hidden construction progress under scaffolding and deviations from unplanned temporary procedures) during the preprocessing stage.

[0029] This step redefines the virtual-real difference as an incremental information source with engineering semantics, used to characterize deviations, omissions, and additions in the state of components during construction. This step transforms the virtual-real inconsistency from a "matching error" into an "incremental information unit," providing a foundation for subsequent incremental information identification.

[0030] First, the pre-processed virtual-real isomorphic construction semantic diagram, BIM semantic diagram, and component-level mapping relationship are received as the initial input.

[0031] The construction semantic diagram is defined as follows: ; The BIM semantic diagram is as follows: ; The set of component-level mapping relationships is as follows: ; in, For construction semantic diagrams; A collection of construction components; This is the feature matrix of construction topology relationships; This is the feature matrix of construction geometric attributes; The semantic attribute feature matrix for construction; For BIM semantic diagrams; A collection of BIM components; This is the BIM topology feature matrix; This is the BIM geometric attribute feature matrix; This is the BIM semantic attribute feature matrix; A set of component-level mapping relationships; BIM component collection The One component; For construction component assembly The Each component.

[0032] Each pair of mapping components This corresponds to the one-to-one mapping relationship already established between the BIM semantic diagram and the construction semantic diagram.

[0033] S2, construct a virtual-real unified semantic vector for each pair of mapping components in the component-level mapping relationship set to quantify the virtual-real differences and extract incremental information, thereby obtaining a multi-dimensional difference set and a difference classification result set.

[0034] For each pair of mapped components, an incrementally guided unified semantic vector of virtual and real is constructed to quantify the differences between virtual and real and extract incremental information. The unified semantic vector of virtual and real uses a unified encoding function of geometry, attributes and topology to format the heterogeneous information of virtual and real into the same tensor space. By splicing multidimensional residual features, a unified tensor describing the state evolution potential of the mapped components is constructed.

[0035] The expression for the unified virtual-real semantic vector is: ; ; ; ; in, For mapping component pairs A unified semantic vector of virtual and real elements; To standardize the mapping function, ensure consistent vector dimensions, facilitate quantification, and facilitate subsequent analysis; It is a geometric difference vector; This is the attribute deviation vector; This is the topological conflict difference vector; , , These are unified encoding functions for geometry, attributes, and topology, used to format heterogeneous virtual and real information into the same tensor space; This is a vector concatenation operation; , Components , Corresponding geometric property features; , Components , Corresponding semantic attribute features; , Components , The corresponding topological relationship features.

[0036] Specifically, the geometric difference vector can be obtained by calculating the feature difference between the observed location and the BIM design location, the attribute deviation vector can be extracted by the semantic similarity function, and the topological conflict difference vector can be obtained by comparing the differences in the neighborhood connection relationships between the virtual and real ends.

[0037] The unified semantic vector of virtual and real is not only used to represent the degree of difference between virtual and real, but also serves as the basic feature expression for the construction of subsequent incremental information units.

[0038] The virtual-real differences include existence differences, attribute differences, and topological differences; the incremental information includes component additions, component missingness, attribute deviations, and topological conflicts.

[0039] The process of quantifying the difference between virtual and real data and extracting incremental information is as follows: By comparing the component sets in the BIM semantic diagram and the construction semantic diagram, missing and newly added components are extracted, and an existence difference set is constructed, expressed as: ; ; ; in, For the set of existence differences; For the set of missing components; For adding a new set of components; This indicates that it does not exist.

[0040] The missing component set corresponds to components that are nodes in the BIM but missing from on-site observation. Under the constraint of consistent perspective, this type of discrepancy not only represents progress delays but also blind spots caused by scaffolding or formwork obstruction. This provides the triggering condition for the subsequent implementation of the virtual-to-real mechanism. The newly added component set corresponds to components that exist on-site but are not included in the BIM. In essence, these correspond to temporary construction measures on-site that are not designed (such as adding diagonal bracing, ground-supported scaffolding) or material stockpiling, and are the key data source for implementing the real-to-virtual mechanism.

[0041] Mapping component pairs of the component-level mapping relationship set Perform attribute difference and topology difference identification, and construct attribute difference set and topology difference set, expressed as: ; ; in, A set of attribute differences; This is a distance function for the similarity between virtual and real attributes, used to calculate the difference in semantic attributes between BIM and on-site components; This is the tolerance threshold for semantic attribute deviation; It is an L2 norm; This is the tolerance threshold for geometric deviation; OR operation; For topological differences; Functions for extracting BIM topology relationships; This is a function for extracting on-site topology relationships.

[0042] The attribute difference set accurately captures actual construction and installation deviations (such as column positioning offsets) and changes in component progress status attributes. (For example, mapping the "planned commencement" status in BIM to the "reinforcement binding completed" status identified on site). Through... Through quantification, the system transforms the inconsistencies in discrete attributes into a fundamental information source characterizing the evolution of construction progress.

[0043] The topological difference set accurately captures dynamic changes in edge loss or stress logic caused by temporary on-site reinforcement, demolition, or obstruction. The attribute deviations and topological conflicts reflect changes in component states and adjustments in structural relationships during construction, thus also constituting a source of incremental information.

[0044] Combining the existence difference set, attribute difference set, and topological difference set, a multidimensional difference set and a difference classification result set are formed, expressed as: ; ; in, A multidimensional set of differences; For the set of differential classification results; For classification mapping operators; For the difference elements in a multidimensional difference set; Add a difference label to the component, corresponding to Components in; For component missing difference labels, corresponding Components in; For attribute difference tags, corresponding Component pairs in; For topological difference labels, corresponding Component pairs in; It means any.

[0045] Through the above steps, the original virtual-real differences are transformed from "unordered matching residuals" into a multidimensional set of differences and a set of difference classification results with clear construction semantics, which are then used as the input basis for incremental information recognition.

[0046] S3. Based on the virtual-real unified semantic vector, a neighborhood aggregation mechanism is introduced to model the structural context features. Incremental information units are generated by combining the differential classification result set, and the importance measurement function of each incremental information unit is calculated to filter out the key information incremental subset.

[0047] This step achieves "incremental information recognition" by mapping the sources of virtual-real differences into computable information carriers. To this end, this step proposes a unified modeling method for incremental information expression, which transforms the difference information into a unified information unit with semantic expression capabilities, structural correlation capabilities, and quantifiable computational capabilities, namely, the incremental information unit.

[0048] The incremental information unit projects the difference baseline vector, difference category encoding, and structural context features into a unified high-dimensional latent space through the incremental information fusion function, thereby eliminating the dimensional differences of features of different dimensions and ensuring the linear additivity of the incremental information unit at the computational level.

[0049] To enhance structural expressive capabilities, the system introduces a neighborhood aggregation mechanism to model structural context features. Local topological environment information is integrated into incremental information units through a neighborhood aggregation function to capture the ripple effects of construction changes. For example, when a newly added temporary steel pipe support is identified, the system not only extracts the features of the steel pipe itself but also aggregates the beam and slab states supporting its upper and lower ends, assigning this discrepancy a clear management semantic.

[0050] Specifically, the expression for the structural context feature is: ; in, For structural context features; For neighborhood aggregation functions; , The component nodes are the component nodes in the neighborhood set of the component nodes; For neighborhood set; For mapping component pairs A unified semantic vector of virtual and real elements; For mapping component pairs A unified semantic vector of virtual and real elements; BIM component collection The One component; For construction component assembly The One component; The expression for the incremental information unit is: ; in, Represents the difference elements in a multidimensional difference set. The corresponding incremental information unit; For incremental information fusion functions, such as multilayer perceptron (MLP) or mapping matrix, heterogeneous geometric, semantic and topological features are projected onto a unified high-dimensional latent space, thereby eliminating the dimensional differences of features of different dimensions, ensuring the linear additivity of incremental information units at the computational level, and outputting standardized incremental information units. The difference category labels in the difference classification result set; This is a vector concatenation operation.

[0051] For isolated components that do not form a mapping pair (i.e., added or missing), their corresponding virtual-real unified semantic vectors The missing end is the empty set Complete it.

[0052] In this way, the system can integrate local topological environment information into incremental information units, enabling it to reflect not only differences in individual components but also the influence of structural relationships. In construction scenarios, this mechanism aims to capture the "ripple effects of construction changes": for example, when a newly added temporary steel pipe support (difference node) is identified, the system not only extracts the features of the steel pipe itself but also aggregates the beam and slab states supported by its upper and lower ends through neighborhood sets, thereby giving this isolated difference object the management semantics of a "support force system".

[0053] Furthermore, the incremental information units of components with different difference category labels must meet different constraints, specifically: (1) Adding difference labels to components The components, lacking prior design knowledge, rely entirely on index numbers. State modeling is performed based on the corresponding construction site observation characteristics, and incremental information units are generated. Must meet: ; in, Indicates the addition of a new difference label to the component. The corresponding incremental information unit; This represents a unified semantic vector of virtual and real elements that relies solely on the characteristics observed at the construction site.

[0054] (2) For component missing difference labels The components, due to a lack of observational data, rely entirely on index numbers. The corresponding BIM prior features are used for logical placement and progress projection, and their incremental information units must meet the following requirements: ; in, Indicates component missing difference label The corresponding incremental information unit; This represents a unified semantic vector of virtual and real elements that relies solely on prior features of BIM.

[0055] (3) For attribute difference tags Topological difference labels The components are constructed based on the mapping component pairs, and are represented as follows: ; in, Indicates attribute difference tags Topological difference label The corresponding incremental information unit.

[0056] Ultimately, an incremental information unit set is formed. , represented as: ; It is a multidimensional set of differences.

[0057] To enable quantifiable evaluation and subsequent screening of incremental information in the construction progress management dimension, an importance measurement function for each incremental information unit is further defined.

[0058] The importance metric function for each incremental information unit is expressed as follows: ; in, Incremental information unit Importance weights; , , , These are the weighting coefficients; It is a geometric difference vector; It is an L2 norm; This is a distance function for the similarity between virtual and real attributes; , Components , Corresponding semantic attribute features; This is a topological conflict indicator function; it takes the value 1 if a conflict exists, and 0 otherwise. This is the critical path indicator factor for the schedule. It is 1 if the difference node is on the critical path, and 0 otherwise.

[0059] For example, for cases involving the critical path ( Significant geometric displacement of load-bearing components ( (Significant), the system will assign it an extremely high importance weight. This was determined to be a "high-risk anomaly" in progress control; while minor material deviations in non-critical paths and non-load-bearing components ( Smaller values ​​are assigned lower weights by the system. Meanwhile, importance weights... The physical deviation of the information, the confidence level of identification, and the sensitivity to the impact on the project schedule are taken into account as the initial weighting basis for the incremental information to participate in subsequent complementary calculations.

[0060] Based on importance weights, a subset of key information increments is constructed. The expression for this subset of key information increments is: ; in, This is an incremental subset of key information. The importance filtering threshold.

[0061] Based on the importance measurement results, the system filters out environmental noise and redundant interference from the incremental subset of key information to ensure the accuracy of subsequent calculations.

[0062] This step realizes the mapping transformation from "virtual and real difference set" to "incremental information unit", which transforms the difference information from discrete residuals into an information expression form with progress management semantics that is calculable and controllable, thereby completing the incremental information identification process.

[0063] S4. Extract complementary features based on the incremental subset of key information and the multidimensional difference set, and calculate the complementary index based on the incremental information unit to determine the complementary direction in the complementary calculation.

[0064] In actual construction, different types of virtual-real discrepancies have varying values ​​for completing construction progress and updating status. For example, geometric deviations may reflect construction delays or component deformation; topological conflicts may reflect structural connection errors or abnormal construction sequences; and attribute deviations may reflect material or construction quality issues. To further quantify the compensation potential of virtual-real asymmetric information, this step constructs a complementary index model based on multidimensional difference characteristics to quantify the contribution of each incremental piece of information to the construction progress status accounting.

[0065] The “complementary calculation” in this step is defined as follows: based on the difference information between the virtual BIM model and the construction site observation, by constructing incremental information units and calculating their complementarity index, the compensation ability and direction of the difference information between the virtual and real systems are determined, thereby driving the asymmetric update process of the virtual and real states, including two types of calculation mechanisms: “real-to-virtual” and “virtual-to-real”.

[0066] Specifically, the complementary features include geometric correction degree, topological relationship compensation degree, and state confidence degree.

[0067] The geometric correction degree is used to characterize the degree of information loss caused by observation occlusion in the construction semantic map, so as to quantify the ability of on-site construction observation to correct the geometry of the BIM virtual terminal. The topology compensation degree is used to characterize the degree of topology information loss caused by construction obstruction / missed measurement, so as to quantify the topology inference and structural restoration requirements of BIM virtual terminal; The state confidence level is used to measure the uncertainty of attribute deviation on the current construction state, in order to quantify the sensitivity of attribute conflict.

[0068] When extracting complementary features, for any incremental information unit in the subset of key information increments, the complementary features are extracted using branching logic based on the difference classification result to which the incremental information unit belongs, specifically as follows: Scenario 1: When the difference classification result belongs to the attribute difference label Topological difference label When both the virtual and real ends have mapped component pairs, complementary features are calculated using relative deviations: The expression for the geometric correction is: ; in, For geometric correction degree; It is a geometric difference vector; It is an L2 norm; BIM components Corresponding geometric property features; To prevent extremely small positive numbers with a denominator of 0.

[0069] To address the unpredictable deviations in actual construction and installation from the BIM model, the geometric correction degree is used to quantify the ability of on-site observations to correct the geometric state of the virtual end through the calculation of relative deviations. This corresponds to the driving force of "complementing the virtual with the real" and ensures that the progress model reflects the actual construction progress status.

[0070] The expression for the topological compensation degree is: ; in, The degree of topological compensation; For neighborhood set; For components The neighborhood set; BIM component collection The One component; For construction component assembly The Each component.

[0071] When nodes or edges are missing in the construction semantic graph due to occlusion, the algorithm does not simply overlay information. Instead, it extracts strong constraint adjacency relationships from the BIM semantic graph and performs topological deduction and structural reconstruction. When the topological structures of the virtual and physical neighborhoods are completely identical, the topological relationship compensation degree is 0; the greater the difference (i.e., the more logical connections are missing as observed on site), the closer the topological relationship compensation degree is to 1. For severe visual occlusion caused by densely packed scaffolding, formwork, or large equipment on the construction site (manifested as observed topological breaks), the topological relationship compensation degree is used to quantify the need for "pseudo-progress delays" due to occlusion, corresponding to the driving intensity of "filling in the physical from the virtual".

[0072] The formula for calculating the state confidence level is: ; in, State confidence level; This is a distance function for the similarity between virtual and real attributes; , Components , Corresponding semantic attribute features; It is a natural exponential function.

[0073] The use of an exponential decay function aims to nonlinearly amplify significant property conflicts, thereby increasing the complementarity index’s extremely high sensitivity to severe design changes, misuse of materials, or abnormal construction quality conditions.

[0074] Scenario 2: When the difference classification result belongs to the newly added difference label of the component In the absence of a BIM target benchmark, the extreme value rule is triggered, defining the geometric correction degree and state confidence degree as the maximization compensation state. , Meanwhile, topological relationship compensation degree It depends on the extent to which the new component alters the connectivity of the local neighborhood on site.

[0075] This type of information represents temporary facilities (such as supports, ground scaffolding) or unplanned components unique to the site, and has absolute physical incremental value.

[0076] Scenario 3: When the difference classification result belongs to the component missing difference label When there are no on-site observed entities (such as dense scaffolding or formwork), and the model relies entirely on prior BIM simulations, then the definition has no geometric correction, i.e.: However, the topological relationship compensation degree and the state certainty degree are maximized compensation states, that is: , .

[0077] The complementary index introduces an adaptive weight adjustment model based on the Softmax function to dynamically allocate the weights of each factor. Simultaneously, a second-order coupling term is introduced to quantify the superposition effect caused by multiple conflicts, thus increasing the incremental information unit. Complementary index The expression is: ; ; ; in, For dimension Adaptive weighting coefficients; These are complementary features; It is a second-order coupling term; , , Preset scaling adjustment factor; Incremental information unit Importance weights; For data source reliability; This is a sensitivity adjustment factor for difference types, used to optimize the weight allocation tendency for different construction stages (such as the main structure stage or the decoration stage); , , This is the preset coupling strength coefficient.

[0078] In this embodiment, the reliability of the data source It can be set as the product of the neighborhood consistency and the state confidence of the incremental information unit, or a fixed value can be set according to experience to characterize the credibility of the current component observation data.

[0079] The difference type sensitivity adjustment factor can be dynamically set according to the geometric correction degree, topological relationship compensation degree, and state confidence degree.

[0080] When construction is in the critical stage of the main structure acceptance, and when there is a large area of ​​obstruction on site affecting the reliability of the sensing data source. When the value is extremely low, the system automatically increases the weight of the topology dimension. This forces the BIM prior logic to dominate the progress projection for the area, avoiding misjudgments of progress delays; conversely, if the confidence quality of the on-site image recognition features is extremely high and there are significant geometric deviations, then the weight of the geometric dimension is increased. The model is corrected based on on-site measurements.

[0081] In real-world construction scenarios, significant geometric displacements of load-bearing components often lead to consequent changes in the local support stress logic. This second-order coupling term... This accurately captures the complex conflict of "deviation + disconnection", ensuring that high-risk complex increments receive a higher complementarity index evaluation.

[0082] The complementary direction determination function is based on the complementary index, used to determine the direction of action of the incremental information unit in the complementary calculation, so as to realize the calculation path selection of real-to-virtual or virtual-to-real. The expression is: ; in, The function for determining complementary directions; The preset direction determination threshold is used; "Real-to-virtual" means correcting the virtual BIM semantic map by modifying the real-world construction semantic map; "virtual-to-real" means deducing the real-world construction semantic map by modifying the virtual BIM semantic map; "bidirectional complementarity" means simultaneous correction at both ends.

[0083] To ensure the consistency of the overall structure, a global consistency constraint is introduced. By minimizing local evaluation deviations, the continuity of the construction semantic structure is guaranteed, effectively avoiding distortion of individual progress calculations caused by local environmental noise on site. The expression for the global consistency constraint is as follows: ; in, This represents the loss for global consistency; a smaller value indicates better structural consistency. Represents a set of incremental information unit pairs of adjacent related components; For incremental information unit pairs of adjacent related components; , Incremental information unit pairs The corresponding complementarity index.

[0084] Considering that a building structure is a continuously stressed whole, the physical states of adjacent related components should not be drastically disconnected. Global consistency constraints ensure that the progress status (percentage of completion) of adjacent components does not drastically disconnect by minimizing local evaluation bias. For example, if a column node is determined to be "completed," the complementary evaluation of its connected beam nodes should also follow the corresponding physical logic, effectively avoiding distortion of individual progress calculations caused by local environmental noise (such as reflections and shadows).

[0085] A set of high-value incremental information is obtained based on the complementary index. This can be used for subsequent hierarchical regulation to reduce computational load. The expression is: ; in, Represents the difference elements in a multidimensional difference set. The corresponding incremental information unit; It is a set of incremental information units; The threshold value is the preset complementarity index.

[0086] Through the above process, a mapping from "incremental information unit" to "quantification of complementary capabilities and determination of the mode of action" is realized, so that each difference information not only has the ability to evaluate compensation, but also has a clear direction of compensation, thus providing a direct driving basis for subsequent complementary calculations.

[0087] S5, perform hierarchical control based on the complementary index to achieve asymmetric complementary calculation, and perform difference-driven updates on different complementary intervals according to the complementary direction, and output the updated BIM semantic map.

[0088] Different incremental information exhibits significant differences in reliability and compensation value. Moreover, in extremely complex construction sites, reflections, wandering workers, or temporary piles of construction waste often generate a large number of highly deceptive "pseudo-differences." Direct fusion calculations can easily introduce noise or conflicting information, affecting model stability. Therefore, this paper proposes a hierarchical control and difference-driven calculation mechanism based on complementarity index and complementarity direction to achieve orderly participation of incremental information and asymmetric complementary calculation.

[0089] Specifically, the tiered regulation based on the complementary index is as follows: First, all incremental information units in the aforementioned key information increment subset are traversed to extract the distribution of the complementarity index, calculate its mean and standard deviation, and introduce a construction stage correction coefficient to dynamically determine the high complementarity threshold. complementary threshold The expression is: ; ; in, This represents the mean of the complementarity index; The standard deviation of the complementarity index; Significant factor; This is a correction factor for the construction phase; This is the environmental noise correction factor.

[0090] During periods of tight schedule, such as the main structure capping and key node acceptance, the system automatically adjusts the height. Incremental information is screened with stricter confidence standards to ensure that the data entering the core computing layer is free of observation noise; for the complex decoration and renovation stage, dynamic adjustments are made. To optimize Expand the coverage of medium / low complementary zones and use delayed queues to avoid the risk of false progress reports caused by temporary obstructions such as equipment yards.

[0091] Based on dynamically determined complementarity thresholds, incremental information units are divided into three priority intervals: ; ; ; in, This is the range of high complementarity index; The range is for the medium complementarity index. This is the range of low complementarity index; Incremental information unit The complementarity index.

[0092] (1) When the complementary direction is from real to virtual, the BIM model is corrected by observing the construction site, that is, the weighted values ​​are superimposed on the BIM design values. The adjusted construction observation deviation vector is expressed as follows: ; ; in, For the updated BIM geometric attribute features; For the updated BIM semantic attribute features; For components Corresponding BIM geometric attribute features; For components Corresponding BIM semantic attribute features; It is a geometric difference vector; This is the attribute deviation vector; Update the weights to complement each other.

[0093] This step is used to correct geometric deviations in the BIM model caused by ideal design. At the same time, it corrects the status of components through semantic deviations (such as updating the progress status from "planned switch" to "actually completed").

[0094] (2) When the complementary direction is virtual to real, the BIM prior information is used to complete the construction site status, and it is necessary to distinguish between the observed incomplete status and the completely missing component status, that is: For incomplete observations caused by partial obstruction, the construction site condition is corrected using BIM model projection, expressed as follows: ; in, For the revised version ; For components Corresponding construction geometric properties; This represents a geometric mapping function based on a projection model.

[0095] For components that are completely missing and not observed on-site, the BIM model is used for projection simulation to correct the construction site conditions and simultaneously complete the missing topological relationships. The expression is as follows: ; ; in, This is the updated construction topology feature matrix; This refers to the topological connection edge of a BIM component.

[0096] (3) When the complementary directions are bidirectionally complementary, a bidirectional collaborative update is performed, expressed as: ; ; in, For the updated .

[0097] Different participation strategies are defined for different complementary intervals, specifically as follows: (1) For incremental information units in the high complementarity index range, they are directly used for model updates. The complementarity update weight is the normalized value of the complementarity index, expressed as: ; in, This is the normalization function.

[0098] This mechanism enables highly reliable incremental information to directly supplement and enhance the model's expression.

[0099] (2) For incremental information units in the medium complementarity index range, calculate neighborhood consistency. When the neighborhood consistency is greater than the preset neighborhood consistency threshold, reduce the complementarity update weight and then update the model; otherwise, postpone participation in the current update process. The expression is: ; ; in, Incremental information unit Neighborhood consistency; For semantic graphs and The neighborhood set of incremental information units that have topological relationships; For semantic graphs and Incremental information units with topological relationships; for The corresponding complementarity index.

[0100] This mechanism avoids interference from moderately reliable information on the model through local consistency verification. From a schedule management perspective, this operator serves as a "process coordination verification": if a component is identified as completed, but the confidence level of its neighboring supporting components with strong dependencies is extremely low (…), then… (For smaller nodes), the algorithm will force a weight reduction process on that component, effectively preventing progress errors caused by isolated false detection nodes. This is achieved by introducing neighborhood consistency. By performing nonlinear scaling on the weights, potential observation noise in the complementary region is automatically suppressed, ensuring the robustness of model updates.

[0101] (3) For incremental information units in the low complementarity index range, a suppression function is set, and they are not currently involved in the update, i.e. The information is then entered into a delayed update queue. This queue mechanism ensures that this information can be reactivated when the conditions are met in subsequent multi-time series observations, thus avoiding information loss due to temporary obscurity.

[0102] The expression for the suppression function is: ; in, Incremental information unit The suppression function; The complementary threshold is used.

[0103] Constructing a delayed update queue instead of directly deleting low-quality increments grants the system strong "temporal robustness." In actual construction, some components may be severely obscured by densely packed formwork or large machinery, resulting in extremely poor measured data (falling into low complementarity zones). However, as construction progresses, formwork is removed, or the field of view is restored, this information may be validated with high confidence in subsequent time-series observations. This queue mechanism ensures that this information can be "reawakened" and participate in calculations when multi-source validation conditions are met, fundamentally avoiding permanent information contamination or loss due to "temporary obscurity." The information in this set will be re-participated in calculations when subsequent multi-time-series observations or multi-source validation conditions are met, thus avoiding information contamination caused by misjudgments. At the management level, this mechanism effectively distinguishes between "field of view loss caused by observation blind spots" and "progress delays caused by inactive processes," eliminating the "false progress reports / missed reports" phenomenon commonly found in traditional automatic progress calculations from the algorithm's underlying layer.

[0104] Thus, the fused geometric attribute feature matrix is ​​obtained, and its expression is: ;

[0105] in, The fused BIM geometric attribute feature matrix; This is the BIM geometric attribute feature matrix; This represents the difference action function based on complementary directions; Output the updated BIM semantic map, expressed as: ; in, The updated BIM semantic diagram; A collection of BIM components; This is the updated BIM topology feature matrix; This is the fused BIM semantic attribute feature matrix.

[0106] Through the above process, differential-driven complementary computing based on incremental information units is realized, enabling the virtual end and the real end to complete the state correction and structural completion under asymmetric conditions, thereby transforming the "incremental information identification result" into the "construction state update result".

[0107] S6, based on the updated BIM semantic map, constructs a unified construction scenario representation model, calculates the completion confidence score of each component node, and infers the construction progress status of each component.

[0108] Currently, incremental information exists only in the form of local corrections and has not yet formed a unified expression of construction scenarios, making it difficult to directly support construction progress analysis, structural safety assessment, and construction scheduling decisions. Therefore, this step proposes a unified expression method for construction scenarios based on complementary computation, realizing the transformation from local completion to overall expression.

[0109] The expression of the unified construction scenario representation model is as follows: ; in, To standardize the representation model of construction scenarios; To unify the set of component nodes, including the original set of BIM components. and newly added components on site; To unify the set of topological edges, including the updated BIM topological relationship feature matrix, the updated construction topological relationship feature matrix, and the virtual-to-real corresponding topological edges established by the component-level mapping relationship, it is used to represent the structural connection relationship between components; To unify the component geometric attribute feature matrix, including the merged BIM geometric attribute feature matrix and the construction geometric attribute features corresponding to the newly added components; To unify the semantic attribute feature matrix of components, including the fused BIM semantic attribute feature matrix and the construction semantic attribute features corresponding to the newly added components; To unify the set of construction dependencies, including the pre-defined construction sequence dependencies in the BIM model, the actual installation logic relationships between components on the construction site, and the schedule dependencies formed by multi-dimensional difference sets and critical path constraints, it is used to express the construction sequence and logical constraints.

[0110] Next, incremental information-driven bidirectional complementary updates are performed on each component node. Geometric deviations are corrected, topological logic is completed, and dependencies are rebuilt through the difference action function. The expression is: ； ; in, For component nodes The unified component geometric property characteristics; For component nodes Unified component semantic attribute features; Indicates the node of the component A set of associated incremental information units; and These are the geometric state correction and semantic attribute correction amounts provided for incremental information, respectively. The weights are updated to complement each other, reflecting the importance and reliability of the incremental information.

[0111] When the incremental information originates from geometric or property deviations observed in the field, It is used to correct the idealized state of the virtual BIM model and reflect the actual construction deviations.

[0112] When incremental information comes from prior BIM knowledge (such as hidden components and design topology logic), Alternatively, hidden edges can be generated through topological deduction to fill in unobservable areas on-site and maintain structural logical continuity. Geometric attribute features of components generated using prior BIM knowledge.

[0113] At the topology level, the unified construction scenario representation model corrects or completes the topology edges through incremental information and uses the hidden edges generated by BIM prior logic to complete the occluded areas.

[0114] At the same time, construction dependencies between components are constructed based on construction logic and topology state: ; in, Construction dependencies between components; for Component nodes in the process.

[0115] Through the above steps, a unified scenario representation model for the construction process is formed. This model enables the transformation of incremental information from local correction to globally consistent expression. It not only includes the geometric and topological attributes after virtual-real fusion, but also fully preserves the confidence features in the complementary calculation process, providing a direct computational carrier for subsequent automated progress state reasoning.

[0116] Next, based on the unified set of component nodes and the state characteristics in the complementary calculation process, the completion confidence score of each component node is calculated, and its expression is as follows: ; in, Score the confidence level for completion; For data source reliability; This is a normalization function used to characterize the effectiveness of differential information; Incremental information unit The complementarity index; Incremental information unit Neighborhood consistency.

[0117] Based on the completion confidence score complementary directions And attribute deviation vectors, to infer the construction progress attributes of each component node. The criteria for judgment are as follows: (1) If the component node Corresponding completion confidence score And its geometric difference vector Deviation vector of attributes If all conditions are within the preset tolerance range of the project, then the progress status of the component is marked as completed; among them, The preset high confidence threshold; (2) If the component node Belongs to the set of missing components Insufficient recognition features lead to The value is low, but the topological relationship compensation degree in its complementarity index reaches the maximum value, and the neighborhood consistency proves that it is in the shading area. Therefore, it is marked as observation-restricted / under construction state, which effectively avoids the misreporting of progress calculation caused by scaffolding shading. (3) If the component node In BIM semantic map The site is within the current construction period and there are no obstructions or interference on site, but If the value approaches 0 and there are no effective observation features in the construction semantic graph, it is marked as an abnormal construction delay, and the progress deviation is recorded.

[0118] (4) Finally, all components that have been determined to be completed can be summarized, their geometric volume parameters can be called, and the physical quantity of the current construction section can be automatically calculated to realize the automatic linkage between the progress plan and the physical output value.

[0119] S7, introduce a time decay factor to update the complementary index, and iteratively update the unified construction scenario expression model through complementary calculation to generate a dynamic construction progress deviation heat map.

[0120] In actual construction, the construction status continuously evolves over time, components are gradually completed, and topological relationships and on-site observations are constantly updated. A single analysis and completion of the virtual-real difference can only reflect the construction status at a specific moment, which is insufficient to meet the needs of dynamic tracking and real-time decision-making throughout the entire process. Therefore, this step proposes a complementary computational iterative update method driven by virtual-real differences to achieve continuous evolution and closed-loop optimization of the construction scenario.

[0121] First, the unified construction scenario expression model mentioned at a previous moment. As input, a new round of calculation of virtual-real differences and incremental information is triggered, generating a subset of key information increments and a complementarity index for the current moment, expressed as: ; ； in, For the current moment A subset of key information increments; For the current moment The newly generated incremental subset of key information; This represents the set union operation; For the current moment Incremental Information Unit The complementarity index; This is a time decay factor used to balance historical complementarity indices with new observations; For the current moment Newly calculated incremental information units The complementarity index.

[0122] As construction progresses, the reliability and effectiveness of incremental information will change.

[0123] Based on the rate of change of the complementarity index, the participation weight strategy of incremental information is dynamically adjusted to obtain the updated subset of key information increments and the difference-driven update strategy; the formula for the rate of change of the complementarity index is: ; like Incremental information is elevated to the high complementarity index range, increasing the complementarity update weight; if The index is downgraded to the medium or low complementarity index range, enters the suppression zone, and the complementarity update weight is reduced or the update is temporarily suspended. in, The rate of change of the complementarity index; The threshold for the rate of change is increased; The threshold for rate of change degradation.

[0124] This mechanism forms a closed loop with the delayed update queue. When construction progresses to the formwork removal stage and the field of view is restored, the previously suppressed hidden project information will be automatically awakened and included in the calculation.

[0125] By iteratively updating the unified construction scenario representation model through complementary computation, a dynamic model sequence for the entire construction process is formed. Each element in the dynamic model sequence corresponds to a unified construction scenario representation model at a specific moment, expressed as: ; in, It represents complementary computation operators, including operations such as real-to-virtual, virtual-to-real, topological completion, and attribute correction; This is an updated unified construction scenario representation model that reflects the current construction status and structural topology.

[0126] When the incremental information change falls below a set threshold in two consecutive iterations, the system is deemed to have reached a stable state, and the iteration can be terminated. This constructs the virtual-real difference-driven incremental complementary computing mechanism. This mechanism enables closed-loop iteration of the construction scenario driven by virtual-real differences, ensuring that the digital twin model continuously reflects the on-site construction status and provides an accurate basis for real-time decision-making.

[0127] In addition, during each iteration cycle, the system automatically executes the quantitative analysis process of schedule deviation and generates a dynamic heat map of construction schedule deviation.

[0128] When generating a dynamic construction progress deviation heatmap, extract the construction progress attributes at the current moment. and link it with the BIM semantic map of the virtual terminal. The pre-set construction schedule time matrix is ​​compared across dimensions to calculate the progress deviation coefficient between the actual construction completion and the planned construction completion. ,expression: ; in, Based on construction progress attributes The calculated percentage of actual completion. This represents the percentage of the construction plan completed.

[0129] like This indicates that the actual progress is lagging behind the planned progress; if This indicates that the actual progress is ahead of the planned progress.

[0130] A dynamic progress deviation heatmap is generated based on the dynamic scene model sequence of the entire construction process to provide real-time warnings to the management end, realizing a deep mapping from "physical state perception" to "management logic decision-making".

[0131] Through this iterative analysis mechanism, the system achieves a deep mapping from physical state perception to management logic decision-making, providing a calculable decision-making basis for construction organization and resource scheduling.

[0132] In practical applications, such as during the main structure construction phase of a high-rise building, the construction site is often obstructed by scaffolding and formwork. Traditional visual recognition methods often misjudge obscured beams and columns as not under construction because they cannot be observed, leading to delays. This invention identifies these discrepancies and calculates a very high topological compensation degree. When the direction of this incremental information is identified as virtual-to-real, the topological connection logic in the BIM prior model is used to complete the missing areas observed on site. When constructing a unified scene representation and calculation, the completion confidence level is calculated to determine that the area is an observation-limited state rather than a true delay, thus ensuring the objectivity of the progress calculation. Simultaneously, for column positioning offsets caused by construction errors on site, the real-to-virtual mechanism uses geometric feature correction to feed back the measured deviation to the digital twin model, achieving dynamic correction of the design-state model. As construction continues, new observation data is constantly received. After the formwork is removed, the component information that was originally in the delay queue is awakened due to the increase of the complementarity index. The system automatically completes the previous progress records and generates a complete progress evolution curve.

[0133] In summary, compared with the prior art, the present invention has the following beneficial effects: This invention transforms the disordered matching errors of virtual and real-world differences into an incremental information source with engineering semantics. By constructing incremental information units and using complementary exponents to drive asymmetric computation, a deep integration of the virtual terminal's design intent and the actual on-site conditions is achieved.

[0134] The hierarchical control mechanism and neighborhood consistency constraints introduced in this invention significantly enhance the anti-interference capability in complex and dynamic construction environments. The synergistic effect of supplementing real work with virtual work and vice versa not only solves the problem of calculating the progress of concealed works but also enables real-time monitoring of construction deviations. Ultimately, through automated progress reasoning and quantity calculation, this invention provides complete technical support for smart construction site construction, from underlying data analysis to high-level management decision-making, significantly improving the automation level and scientific rigor of construction progress management.

[0135] Example 2 like Figure 2As shown, the second embodiment of the present invention also provides a construction progress management device based on asymmetric bidirectional complementarity, comprising: The acquisition unit is used to acquire a construction semantic map reflecting the real-time observation status of the construction site, a BIM semantic map reflecting the status from the BIM design perspective, and a map reflecting the relationship between the construction semantic map and the BIM semantic map. Figure 1 A set of component-level mapping relationships with one-to-one correspondence; The virtual-real difference identification unit is used to construct a virtual-real unified semantic vector for each pair of mapping components in the component-level mapping relationship set, so as to quantify the virtual-real difference and extract incremental information to obtain a multi-dimensional difference set and a difference classification result set. The key information calculation unit is used to model structural context features based on the neighborhood aggregation mechanism introduced by the virtual and real unified semantic vector, generate incremental information units by combining the differential classification result set, calculate the importance measurement function of each incremental information unit, and filter to obtain a subset of key information increments. The complementary direction determination unit is used to extract complementary features based on the incremental subset of key information and the multidimensional difference set, and to calculate the complementary index based on the incremental information unit to determine the complementary direction in the complementary calculation. The difference-driven update unit is used to perform hierarchical control based on the complementarity index to achieve asymmetric complementarity calculation, and to perform difference-driven updates on different complementary intervals according to the complementary direction, and output the updated BIM semantic map. The construction progress reasoning unit is used to build a unified construction scenario expression model based on the updated BIM semantic map, calculate the completion confidence score of each component node, and reason to obtain the construction progress status of each component. The iterative update unit is used to introduce a time decay factor to update the complementary index, and to iteratively update the unified construction scenario expression model through complementary calculation to generate a dynamic construction progress deviation heatmap.

[0136] Example 3 The third embodiment of the present invention also provides a construction progress management device based on asymmetric bidirectional complementarity, which includes a memory and a processor. The memory stores a computer program, which can be executed by the processor to realize the construction progress management method based on asymmetric bidirectional complementarity as described above.

[0137] Example 4 The fourth embodiment of the present invention also provides a computer-readable storage medium storing computer-readable instructions. When the computer-readable instructions are executed by the processor of the device where the computer-readable storage medium is located, the construction progress management method based on asymmetric bidirectional complementarity as described above is implemented.

[0138] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A construction progress management method based on asymmetric bidirectional complementarity, characterized in that, include: S1, obtain the construction semantic map to reflect the real-time observation status of the construction site, the BIM semantic map to reflect the status from the BIM design perspective, and the component-level mapping relationship set to reflect the one-to-one correspondence between the construction semantic map and the BIM semantic map. S2, construct a virtual-real unified semantic vector for each pair of mapping components in the component-level mapping relationship set to quantify the virtual-real differences and extract incremental information, thereby obtaining a multi-dimensional difference set and a difference classification result set; S3, based on the virtual and real unified semantic vector, introduce a neighborhood aggregation mechanism to model structural context features, combine the differential classification result set to generate incremental information units, calculate the importance measurement function of each incremental information unit, and filter to obtain a subset of key information increments; S4. Extract complementary features based on the incremental subset of key information and the multidimensional difference set, and calculate the complementary index based on the incremental information unit to determine the complementary direction in the complementary calculation. S5, perform hierarchical control based on the complementary index to achieve asymmetric complementary calculation, and perform difference-driven updates on different complementary intervals according to the complementary direction, and output the updated BIM semantic map; S6, based on the updated BIM semantic map, constructs a unified construction scenario expression model, calculates the completion confidence score of each component node, and infers the construction progress status of each component; S7, introduce a time decay factor to update the complementary index, and iteratively update the unified construction scenario expression model through complementary calculation to generate a dynamic construction progress deviation heat map.

2. The construction progress management method based on asymmetric bidirectional complementarity according to claim 1, characterized in that... The construction semantic diagram is defined as follows: ; The BIM semantic diagram is as follows: ; The set of component-level mapping relationships is as follows: ; in, For construction semantic diagrams; A collection of construction components; This is the feature matrix of construction topology relationships; This is the feature matrix of construction geometric attributes; The semantic attribute feature matrix for construction; For BIM semantic diagrams; A collection of BIM components; This is the BIM topology feature matrix; This is the BIM geometric attribute feature matrix; This is the BIM semantic attribute feature matrix; A set of component-level mapping relationships; BIM component collection The One component; For construction component assembly The One component; Each pair of mapping components This corresponds to the one-to-one mapping relationship already established between the BIM semantic diagram and the construction semantic diagram; The expression for the unified virtual-real semantic vector is then: ； ； ； ； in, For mapping component pairs A unified semantic vector of virtual and real elements; For standardized mapping functions; It is a geometric difference vector; This is the attribute deviation vector; This is the topological conflict difference vector; , , These are unified encoding functions for geometry, attributes, and topology, used to format heterogeneous virtual and real information into the same tensor space; This is a vector concatenation operation; , Components , Corresponding geometric properties; , Components , Corresponding semantic attribute features; , Components , Corresponding topological relationship features; The virtual-real differences include existence differences, attribute differences, and topological differences; the incremental information includes component additions, component missingness, attribute deviations, and topological conflicts. The process of quantifying the difference between virtual and real data and extracting incremental information is as follows: By comparing the component sets in the BIM semantic diagram and the construction semantic diagram, missing and newly added components are extracted, and an existence difference set is constructed, expressed as: ； ； ； in, For the set of existence differences; For the set of missing components; For adding a new set of components; This indicates that it does not exist; Mapping component pairs of the component-level mapping relationship set Perform attribute difference and topology difference identification, and construct attribute difference set and topology difference set, expressed as: ； ； in, For attribute differences; This is a distance function for the similarity between virtual and real attributes, used to calculate the difference in semantic attributes between BIM and on-site components; This is the tolerance threshold for semantic attribute deviation; It is an L2 norm; This is the tolerance threshold for geometric deviation; OR operation; For topological differences; Functions for extracting BIM topology relationships; Functions for extracting on-site topology relationships; Combining the existence difference set, attribute difference set, and topological difference set, a multidimensional difference set and a difference classification result set are formed, expressed as: ； ； in, A multidimensional set of differences; For the set of differential classification results; For classification mapping operators; For the difference elements in a multidimensional difference set; Add a difference label to the component, corresponding to Components in; For component missing difference labels, corresponding Components in; For attribute difference tags, corresponding Component pairs in; For topological difference labels, corresponding Component pairs in; It means any.

3. The construction progress management method based on asymmetric bidirectional complementarity according to claim 2, characterized in that... The expression for the structural context feature is: ； in, For structural context features; For neighborhood aggregation functions; , The component nodes are the component nodes in the neighborhood set of the component nodes; For neighborhood set; For mapping component pairs A unified semantic vector of virtual and real elements; For mapping component pairs A unified semantic vector of virtual and real elements; BIM component collection The One component; For construction component assembly The One component; The expression for the incremental information unit is: ； in, Represents the difference elements in a multidimensional difference set. The corresponding incremental information unit; This is an incremental information fusion function; The difference category labels in the difference classification result set; This is a vector concatenation operation; Furthermore, the incremental information units of components with different difference category labels must meet different constraints, specifically: For adding difference labels to components The components rely entirely on on-site construction observation characteristics for state modeling, with incremental information units. Must meet: ； in, Indicates the addition of a new difference label to the component. The corresponding incremental information unit; Indicates that it depends only on the component index number. The corresponding virtual-real unified semantic vector of construction site observation features; Represents the empty set; For component missing difference labels For components that lack observational data and rely entirely on BIM prior features for logical placement and progress projection, their incremental information units must meet the following requirements: ； in, Indicates component missing difference label The corresponding incremental information unit; Indicates that it depends only on the component index number. The corresponding BIM prior feature's unified virtual and real semantic vector; For attribute difference tags Topological difference labels The components are constructed based on the mapping component pairs, and are represented as follows: ； in, Indicates attribute difference tags Topological difference labels The corresponding incremental information unit; Ultimately, an incremental information unit set is formed. , is represented as: ; A multidimensional set of differences; The importance metric function for each incremental information unit is expressed as follows: ； in, Incremental information unit Importance weights; , , , These are the weighting coefficients; It is a geometric difference vector; It is an L2 norm; This is a distance function for the similarity between virtual and real attributes; , Components , Corresponding semantic attribute features; This is a topological conflict indicator function; it takes the value 1 if a conflict exists, and 0 otherwise. This is the critical path indicator factor for the schedule; it is 1 if the difference node is on the critical path, and 0 otherwise. The expression for the incremental subset of key information is: ； in, This is an incremental subset of key information. The importance filtering threshold.

4. The construction progress management method based on asymmetric bidirectional complementarity according to claim 3, characterized in that... The complementary features include geometric correction degree, topological relationship compensation degree, and state confidence degree; The geometric correction degree is used to characterize the degree of information loss caused by observation occlusion in the construction semantic map, so as to quantify the ability of on-site construction observation to correct the geometry of the BIM virtual terminal. The topology compensation degree is used to characterize the degree of topology information loss caused by construction obstruction / missed measurement, so as to quantify the topology inference and structural restoration requirements of BIM virtual terminal; The state confidence level is used to measure the uncertainty of attribute deviation on the current construction state, so as to quantify the sensitivity of attribute conflict. When extracting complementary features, for any incremental information unit in the subset of key information increments, the complementary features are extracted using branching logic based on the difference classification result to which the incremental information unit belongs, specifically as follows: When the difference classification result belongs to the attribute difference label Topological difference labels When both the virtual and real ends have mapped component pairs, the complementary features are calculated through relative deviation, specifically: The expression for the geometric correction is: ； in, For geometric correction degree; It is a geometric difference vector; It is an L2 norm; BIM components Corresponding geometric properties; To prevent extremely small positive numbers with a denominator of 0; The expression for the topological compensation degree is: ； in, The degree of topological compensation; For neighborhood set; For components The neighborhood set; BIM component collection The One component; For construction component assembly The One component; The formula for calculating the state confidence level is: ； in, State confidence level; This is a distance function for the similarity between virtual and real attributes; , Components , Corresponding semantic attribute features; It is a natural exponential function; When the difference classification result belongs to the newly added difference label of the component In the absence of a BIM target benchmark, the extreme value rule is triggered, defining the geometric correction degree and state confidence degree as the maximization compensation state. , Meanwhile, topological relationship compensation degree It depends on the extent to which the new component alters the local connectivity of the surrounding area; When the difference classification result belongs to the component missing difference label When there are no on-site observed entities and the model relies entirely on prior BIM simulations, then the definition has no geometric correction, i.e.: However, the topological relationship compensation degree and the state certainty degree are maximized compensation states, that is: , .

5. A construction progress management method based on asymmetric bidirectional complementarity according to claim 4, characterized in that... The complementary index introduces an adaptive weight adjustment model based on the Softmax function to dynamically allocate the weights of each factor. Simultaneously, a second-order coupling term is introduced to quantify the superposition effect caused by multiple conflicts. The incremental information unit... Complementary index The expression is: ； ； ； in, For dimension Adaptive weighting coefficients; These are complementary features; It is a second-order coupling term; , , Preset scaling adjustment factor; Incremental information unit Importance weights; For data source reliability; This is a sensitivity adjustment factor for difference types, used to optimize the weight allocation tendency for different construction stages; , , The preset coupling strength coefficient; The complementary direction determination function is based on the complementary index, used to determine the direction of action of the incremental information unit in the complementary calculation, so as to realize the calculation path selection of real-to-virtual or virtual-to-real. The expression is: ； in, The function for determining complementary directions; The preset direction determination threshold is used; "Real-to-virtual" means correcting the virtual BIM semantic map by modifying the real-world construction semantic map; "virtual-to-real" means deducing the real-world construction semantic map by modifying the virtual BIM semantic map; "bidirectional complementarity" means simultaneous correction at both ends.

6. The construction progress management method based on asymmetric bidirectional complementarity according to claim 5, characterized in that... It also includes: introducing global consistency constraints to ensure the continuity of the construction semantic structure and avoid drastic breaks in the progress status of adjacent components; The expression for the global consistency constraint is: ； in, This represents the loss for global consistency; a smaller value indicates better structural consistency. Represents a set of incremental information unit pairs of adjacent related components; For incremental information unit pairs of adjacent related components; , Incremental information unit pairs The corresponding complementarity index; A set of high-value incremental information is obtained based on the complementary index. This is used for hierarchical control to reduce computational load, and its expression is: ； in, Represents the difference elements in a multidimensional difference set. The corresponding incremental information unit; It is a set of incremental information units; The threshold value is the preset complementarity index.

7. A construction progress management method based on asymmetric bidirectional complementarity according to claim 2, characterized in that... The specific implementation of tiered regulation based on the aforementioned complementarity index is as follows: First, all incremental information units in the aforementioned key information increment subset are traversed to extract the distribution of the complementarity index, calculate its mean and standard deviation, and introduce a construction stage correction coefficient to dynamically determine the high complementarity threshold. complementary threshold The expression is: ； ； in, This represents the mean of the complementarity index; The standard deviation of the complementarity index; Significant factor; This is a correction factor for the construction phase; This is the environmental noise correction factor; Based on dynamically determined complementarity thresholds, incremental information units are divided into three priority intervals: ； ； ； in, This is the range of high complementarity index; The range is for the medium complementarity index. This is the range of low complementarity index; Incremental information unit The complementarity index; Based on the complementary direction, difference-driven updates are performed on different complementary intervals, specifically as follows: When the complementary direction is from real to virtual, the BIM model is corrected using on-site construction observations, that is, weighted values ​​are superimposed on the BIM design values. The adjusted construction observation deviation vector is expressed as follows: ； ； in, For the updated BIM geometric attribute features; For the updated BIM semantic attribute features; For components Corresponding BIM geometric attribute features; For components Corresponding BIM semantic attribute features; It is a geometric difference vector; This is the attribute deviation vector; To update the weights for complementarity; When the complementary direction is virtual to real, the BIM prior information is used to complete the construction site status, and it is necessary to distinguish between the observed incomplete state and the state of completely missing components, that is: For incomplete observations caused by partial obstruction, the construction site condition is corrected using BIM model projection, expressed as follows: ； in, For the revised version ; For components Corresponding construction geometric properties; Represents a geometric mapping function based on a projection model; For components that are completely missing and not observed on-site, the BIM model is used for projection simulation to correct the construction site conditions and simultaneously complete the missing topological relationships. The expression is as follows: ； ； in, This is the updated construction topology feature matrix; For the topological connection edges of BIM components; When the complementary directions are bidirectionally complementary, a bidirectional collaborative update is performed, expressed as: ； ； in, For the updated ; For incremental information units in the high complementarity index range, they are directly used for model updates. The complementarity update weights are the normalized values ​​of the complementarity index, expressed as: ； in, This is the normalization function; For incremental information units within the medium complementarity index range, neighborhood consistency is calculated. If the neighborhood consistency is greater than a preset neighborhood consistency threshold, the complementarity update weight is reduced before model updating; otherwise, participation in the current update process is temporarily suspended. The expression is: ； ； in, Incremental information unit Neighborhood consistency; For semantic graphs and The neighborhood set of incremental information units that have topological relationships; For semantic graphs and Incremental information units with topological relationships; for The corresponding complementarity index; For incremental information units in the low complementarity index range, a suppression function is set, and they are not currently involved in the update. It enters the delayed update queue; the expression of the suppression function is: ； in, Incremental information unit The suppression function; The complementary threshold is used. Thus, the fused geometric attribute feature matrix is ​​obtained, and its expression is: ； in, The fused BIM geometric attribute feature matrix; This is the BIM geometric attribute feature matrix; This represents the difference action function based on complementary directions; This results in the updated BIM semantic map, expressed as: ； in, The updated BIM semantic diagram; A collection of BIM components; This is the updated BIM topology feature matrix; This is the fused BIM semantic attribute feature matrix.

8. A construction progress management method based on asymmetric bidirectional complementarity according to claim 7, characterized in that... The expression of the unified construction scenario representation model is as follows: ； in, To standardize the representation model of construction scenarios; To unify the set of component nodes, including the original set of BIM components. and newly added components on site; To unify the set of topological edges, used to represent the structural connection relationships between components; To unify the geometric attribute feature matrix of components; To unify the semantic attribute feature matrix of components; To unify the set of construction dependencies, used to express construction sequence and logical constraints; The completion confidence score for each component node is calculated using the following expression: ； in, Score the confidence level for completion; For data source reliability; This is the normalization function; Incremental information unit The complementarity index; Incremental information unit Neighborhood consistency; Based on the completion confidence score complementary directions And attribute deviation vectors, to infer the construction progress attributes of each component node. The criteria for judgment are as follows: If component node Corresponding completion confidence score And its geometric difference vector Deviation vector of attributes If all conditions are within the preset tolerance range of the project, then the progress status of the component is marked as completed; among them, The preset high confidence threshold; If component node Belongs to the set of missing components Insufficient recognition features lead to The value is low, but the topological compensation degree in its complementarity index reaches the maximum value, and it is proven to be in the occlusion area according to the neighborhood consistency. Therefore, it is marked as observation-restricted / under construction. If component node In BIM semantic map The site is within the current construction period and there are no obstructions or interference on site, but If the value approaches 0 and there are no effective observation features in the construction semantic graph, it is marked as an abnormal construction delay, and the progress deviation is recorded.

9. A construction progress management method based on asymmetric bidirectional complementarity according to claim 1, characterized in that... A time decay factor is introduced to update the complementarity index, and the unified construction scenario expression model is iteratively updated through complementarity calculation, specifically as follows: The unified construction scenario expression model described at a previous moment As input, a new round of calculation of virtual-real differences and incremental information is triggered, generating a subset of key information increments and a complementarity index for the current moment, expressed as: ； ； in, For the current moment A subset of key information increments; For the current moment The newly generated incremental subset of key information; This represents the set union operation; For the current moment Incremental Information Unit The complementarity index; This is the time decay factor; For the current moment Newly calculated incremental information units The complementarity index; Based on the rate of change of the complementarity index, the participation weight strategy of incremental information is dynamically adjusted to obtain the updated subset of key information increments and the difference-driven update strategy; the formula for the rate of change of the complementarity index is: ； like Incremental information is elevated to the high complementarity index range, increasing the complementarity update weight; if The index is downgraded to the medium or low complementarity index range, enters the suppression zone, and the complementarity update weight is reduced or the update is temporarily suspended. in, The rate of change of the complementarity index; The threshold for the rate of change is increased; The threshold for downgrading the rate of change; By iteratively updating the unified construction scenario expression model through complementary computation, a dynamic model sequence of the entire construction process is formed, where each element in the dynamic model sequence of the entire construction process corresponds to a unified construction scenario expression model at a certain moment. When generating a dynamic construction progress deviation heatmap, the construction progress attributes of each component's construction progress status at the current moment are extracted. and link it with the BIM semantic map of the virtual terminal. The pre-set construction schedule time matrix is ​​compared across dimensions to calculate the progress deviation coefficient between the actual construction completion and the planned construction completion. ; like This indicates that the actual progress is lagging behind the planned progress; if This indicates that the actual progress is ahead of the planned progress; A dynamic progress deviation heatmap is generated based on the dynamic scene model sequence of the entire construction process to provide real-time warnings to the management.

10. A construction progress management device based on asymmetric bidirectional complementarity, for implementing the construction progress management method based on asymmetric bidirectional complementarity as described in any one of claims 1-9, characterized in that, include: The acquisition unit is used to acquire a construction semantic map that reflects the real-time observation status of the construction site, a BIM semantic map that reflects the status from the BIM design perspective, and a set of component-level mapping relationships that reflects the one-to-one correspondence between the construction semantic map and the BIM semantic map. The virtual-real difference identification unit is used to construct a virtual-real unified semantic vector for each pair of mapping components in the component-level mapping relationship set, so as to quantify the virtual-real difference and extract incremental information to obtain a multi-dimensional difference set and a difference classification result set. The key information calculation unit is used to model structural context features based on the neighborhood aggregation mechanism introduced by the virtual and real unified semantic vector, generate incremental information units by combining the differential classification result set, calculate the importance measurement function of each incremental information unit, and filter to obtain a subset of key information increments. The complementary direction determination unit is used to extract complementary features based on the incremental subset of key information and the multidimensional difference set, and to calculate the complementary index based on the incremental information unit to determine the complementary direction in the complementary calculation. The difference-driven update unit is used to perform hierarchical control based on the complementarity index to achieve asymmetric complementarity calculation, and to perform difference-driven updates on different complementary intervals according to the complementary direction, and output the updated BIM semantic map. The construction progress reasoning unit is used to build a unified construction scenario expression model based on the updated BIM semantic map, calculate the completion confidence score of each component node, and reason to obtain the construction progress status of each component. The iterative update unit is used to introduce a time decay factor to update the complementary index, and to iteratively update the unified construction scenario expression model through complementary calculation to generate a dynamic construction progress deviation heatmap.

Images