BIM model and drawing comparison method and system based on completion archives

Abstract

The present invention relates to the field of construction engineering technology, and in particular to a BIM model and drawing comparison method and system based on completion archives. The system comprises a completion archive data parsing module, a BIM model generation module, a comparison analysis module, a dynamic update module and an optimization feedback module. The data parsing module uses natural language processing technology to extract building data from the completion archives; the BIM model generation module uses a geometric modeling algorithm to generate a BIM model consistent with the actual construction; the comparison analysis module automatically compares the BIM model with the drawings through computer vision and deep learning algorithms to identify differences; the dynamic update module uses the Internet of Things to monitor the construction site in real time and update the BIM model; the optimization feedback module optimizes the system logic according to the feedback. The present invention not only improves the accuracy and efficiency of BIM model generation and drawing comparison, but also realizes dynamic monitoring and updating of the construction site, providing strong support for digital management and lean construction in the construction industry.

Description

Technical Field

[0001] The present invention relates to the technical field of construction engineering, and in particular to a BIM model and drawing comparison method and system based on completion archives. Background Art

[0002] With the widespread application of building information modeling (BIM) technology, the design, construction and operation and maintenance processes of construction projects have been significantly improved. As a digital form of building information expression, BIM models can provide accurate geometric data, time schedule, cost information and other key data throughout the life cycle of the building. However, in the actual construction process, the completion status is often inconsistent with the initial design BIM model due to site conditions, design changes, construction errors and other reasons. Therefore, in order to ensure the consistency between the actual completion status of the building and the design model, it is particularly important to compare the BIM model of the completion archive with the design drawings in a timely manner. At present, there are still the following problems: traditional completion archives mainly exist in the form of text and images, and lack effective technical means for automated data analysis and extraction; the generation of traditional BIM models is often based on initial design data, which fails to fully reflect the changes and actual conditions during the construction process, resulting in a large deviation between the model and the actual construction, affecting subsequent management and maintenance; the existing comparison work usually relies on manual or simple image comparison tools, which is not only time-consuming and labor-intensive, but also has the problem of insufficient accuracy in identifying and marking differences. Summary of the invention

[0003] To solve the above problems, the present invention provides a BIM model and drawing comparison method and system based on as-built archives. It solves the problem of how to realize the automated data analysis and extraction of as-built archives, and generate a BIM model reflecting the actual construction situation, realizes the efficient comparison of BIM models and design drawings through automation technology, accurately identifies and marks the differences generated during the construction process, and improves the accuracy and efficiency of BIM model generation and drawing comparison.

[0004] To achieve the above object, the technical solution adopted by the present invention is:

[0005] On the one hand, a BIM model and drawing comparison system based on as-built archives includes an as-built archives data parsing module, a BIM model generation module, a comparison analysis module, a dynamic update module and an optimization feedback module which are sequentially connected in communication;

[0006] The completion archive data parsing module is used to perform semantic analysis and data mining from the completion archive through natural language processing technology to parse out building data; the building data includes key component information, geometric data and material information;

[0007] The BIM model generation module is used to generate a BIM model that is highly consistent with the actual construction situation based on the building data using a geometric modeling algorithm and a rule reasoning mechanism;

[0008] The comparison and analysis module is used to automatically compare the generated BIM model and drawings using computer vision and deep learning algorithms, identify and mark the difference areas, and generate a comparison and analysis report;

[0009] The dynamic update module is used to dynamically monitor the construction site through the Internet of Things, and automatically correct and update the BIM model based on real-time data, and recommend corresponding adjustment measures;

[0010] The optimization feedback module is used to feed back the comparison and dynamic update results to relevant departments or personnel, and optimize the comparison analysis and BIM model generation logic through the optimization algorithm according to the feedback information.

[0011] Furthermore, the BIM model includes the geometric shape, positional relationship and material information of the building components.

[0012] Furthermore, the operation process of the BIM model generation module includes the following steps:

[0013] Based on the received building data, a preliminary BIM model is generated by a geometric modeling algorithm, wherein the preliminary BIM model is used to reflect the basic geometric form of the building components;

[0014] Using the graph-based rule reasoning mechanism, the generated preliminary BIM model is logically verified and corrected, including the adjustment of positional relationships and the assignment of material information;

[0015] The optimized BIM model is iterated multiple times to generate a final BIM model that is highly consistent with the actual building status;

[0016] The final generated BIM model is transferred to the comparison and analysis module.

[0017] Furthermore, the construction process of the geometric modeling algorithm includes the following steps:

[0018] Perform format conversion, normalization and noise filtering on the parsed building data;

[0019] According to the key component information in the building data, the geometric parameters, position coordinates, angles and size ratios of the components are extracted, and these parameters are used as basic inputs for geometric modeling;

[0020] According to the extracted geometric parameters, the preliminary geometric form of the component is generated by using the geometric modeling algorithm based on surface fitting and space segmentation, and a preliminary space layout model is constructed;

[0021] The generated preliminary geometric model is used as the input of the BIM model generation module to generate a preliminary BIM model.

[0022] Furthermore, the operation process of the comparison and analysis module includes the following steps:

[0023] Acquire and preprocess the generated BIM model data and original construction drawing data;

[0024] Based on computer vision technology, key features of BIM models and construction drawings are extracted; the key features include geometric shape features, texture features, boundary features, size and material properties;

[0025] Using the capsule network algorithm, the geometric differences and position deviations between the BIM model and the construction drawings are identified through feature matching, regional overlap analysis and spatial position comparison;

[0026] Automatically annotate the identified difference areas, including the type, size, location and potential cause of the difference, and mark them separately on the BIM model and construction drawings;

[0027] Comprehensively analyze the comparison results, generate a comparative analysis report, and provide a visual display of the comparison results; the comparative analysis report includes difference analysis, error statistics, potential risk assessment and suggestions;

[0028] The comparison analysis report is transmitted to the dynamic update module and the optimization feedback module.

[0029] Furthermore, the formula of the capsule network algorithm is as follows:

[0030] ;

[0031] in, represents the final output vector of the jth output capsule, i.e., the comparison result between a feature or object in the BIM model and the construction drawing; Represents the correlation weight between the features extracted from the BIM model generation module and construction drawings and the current output features; Represents building data parsed from as-built archives; represents the bias vector, which is used to correct the systematic error that may be introduced in the comparison process between the BIM model generation module and the construction drawings; n represents the total number of features extracted in the BIM model generation module; m represents the total number of features in the output capsule.

[0032] Furthermore, the Internet of Things specifically includes sensor networks, drones, edge computing devices, central control devices and data transmission devices.

[0033] Furthermore, the optimization feedback module specifically optimizes the comparison analysis and BIM model generation logic through a differential optimization algorithm, and prioritizes the feedback information based on a multi-criteria decision analysis method, giving priority to feedback with higher technical difficulty, impact scope and urgency.

[0034] On the other hand, a BIM model and drawing comparison method based on completion archives includes the following steps:

[0035] Through natural language processing technology, semantic analysis and data mining are carried out from the completion archives to parse out the building data;

[0036] Based on the building data, a BIM model that is highly consistent with the actual construction situation is generated using a geometric modeling algorithm and a rule reasoning mechanism;

[0037] Using computer vision and deep learning algorithms, the generated BIM models and drawings are automatically compared, the difference areas are identified and marked, and a comparative analysis report is generated;

[0038] Dynamically monitor the construction site through the Internet of Things, automatically correct and update the BIM model based on real-time data, and recommend corresponding adjustment measures;

[0039] The results of comparison and dynamic update are fed back to relevant departments or personnel, and the comparison analysis and BIM model generation logic are optimized through optimization algorithms based on the feedback information.

[0040] The beneficial effects of the present invention are:

[0041] In the present invention, through the completion file data parsing module, natural language processing technology is used to perform in-depth semantic analysis and data mining on the completion file, and key building data is accurately extracted. The BIM model generation module uses geometric modeling algorithms and rule reasoning mechanisms to generate a BIM model that is highly close to the actual construction situation based on the parsed building data, ensuring the accuracy and reliability of the model. Through the comparison and analysis module, the system can use computer vision and deep learning algorithms to automatically compare the generated BIM model with the drawings, quickly identify the differences between the two and mark the difference areas. This not only improves the comparison efficiency, but also reduces the errors that may occur in manual comparison, ensuring the consistency of the building model with the design drawings. The dynamic update module monitors the construction site in real time through the Internet of Things technology, and automatically updates and corrects the BIM model according to the collected real-time data. In this way, the BIM model always keeps synchronization with the actual construction site, thereby avoiding the model lag problem caused by changes in the construction site, and provides timely adjustment suggestions, improving the response speed and accuracy of construction management. The optimization feedback module feeds back the comparison results and dynamic update information to relevant departments or personnel, and continuously optimizes the comparison analysis and BIM model generation logic in the system through optimization algorithms. This closed-loop feedback mechanism can not only continuously improve the accuracy and efficiency of the system, but also provide valuable experience accumulation and technical improvements for subsequent projects. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a module schematic diagram of a BIM model and drawing comparison system based on completion archives of the present invention.

[0043] Figure 2 It is a flowchart of the operation process of the comparison and analysis module provided by an embodiment of the present invention.

[0044] Figure 3 The present invention is a flowchart of a BIM model and drawing comparison method based on completion archives. DETAILED DESCRIPTION

[0045] See also Figure 1-3 As shown, the present invention relates to a BIM model and drawing comparison method and system based on completion archives.

[0046] Example 1

[0047] A BIM model and drawing comparison system based on completion archives, comprising a completion archive data parsing module, a BIM model generation module, a comparison analysis module, a dynamic update module and an optimization feedback module which are sequentially connected in communication;

[0048] The completion archive data parsing module is used to perform semantic analysis and data mining from the completion archive through natural language processing technology to parse out building data; the building data includes key component information, geometric data and material information;

[0049] It should be noted that the system will pre-process the input completion files. Since completion files usually contain files in multiple formats, such as PDF documents, drawing files, Excel tables, etc., the system needs to uniformly process these files in different formats. For PDF files, the system uses optical character recognition (OCR) technology to convert scanned files into editable text formats. For table files, the system will automatically identify and parse the table structure and extract useful data fields.

[0050] The system will extract valuable information from the completion archives through data mining technology. This information includes but is not limited to:

[0051] Key component information: The system extracts the main components of the building, such as beams, columns, walls, floor slabs, etc., and records their geometric shapes, positional relationships, and material properties.

[0052] Geometric data: For the extracted components, the system further analyzes their geometric information, such as size, shape, volume, etc. This step may be cross-validated with the annotation information in the architectural drawings to ensure the accuracy of the geometric data.

[0053] Material information: The system also extracts the material information of the components, such as concrete grade, steel bar specifications, exterior wall paint type, etc. These data are crucial for the subsequent BIM model generation.

[0054] To improve the comprehensiveness of data analysis, the system supports multimodal data fusion. The completion archive may contain multiple data types such as drawings, tables, and text descriptions. The system will fuse data from these different sources to ensure that information from different data sources can verify and complement each other during data mining. For example, the system may compare the geometric dimensions in the drawings with the dimensions described in the text. If inconsistencies are found, the system will prompt the user or automatically correct the error.

[0055] The BIM model generation module is used to generate a BIM model that is highly consistent with the actual construction situation based on the building data using a geometric modeling algorithm and a rule reasoning mechanism; the BIM model includes the geometric shape, positional relationship and material information of building components.

[0056] The operation process of the BIM model generation module includes the following steps:

[0057] Based on the received building data, a preliminary BIM model is generated by a geometric modeling algorithm, wherein the preliminary BIM model is used to reflect the basic geometric form of the building components;

[0058] Using the graph-based rule reasoning mechanism, the generated preliminary BIM model is logically verified and corrected, including the adjustment of positional relationships and the assignment of material information;

[0059] Specifically, the relative position relationship and geometric form of each component in the model are automatically checked to see if they meet the architectural design specifications and construction standards. For example, verify whether the connection between the wall and the floor meets the construction requirements, and whether the size and position of the beams and columns are reasonable. Based on the results of logical verification, the relative positions of the components in the model are automatically adjusted. For example, if it is found that the position of some components deviates from the design drawings, the rule reasoning mechanism will automatically correct these deviations with reference to the design specifications to ensure that the connection and support relationship between the components meet the actual requirements. According to the material information in the building data, the corresponding material properties are assigned to each component in the model. This material information includes not only the type of material used for the component (such as concrete, steel, wood, etc.), but also the physical properties of the material (such as density, strength, thermal conductivity, etc.), ensuring that the model is consistent with the actual building state both visually and physically.

[0060] The optimized BIM model is iterated multiple times to generate a final BIM model that is highly consistent with the actual building status;

[0061] The final generated BIM model is transferred to the comparison and analysis module.

[0062] The construction process of the geometric modeling algorithm includes the following steps:

[0063] Perform format conversion, normalization and noise filtering on the parsed building data;

[0064] According to the key component information in the building data, the geometric parameters, position coordinates, angles and size ratios of the components are extracted, and these parameters are used as basic inputs for geometric modeling;

[0065] Specifically, the geometric parameters of each component are extracted from the building data. These parameters usually include length, width, height, radius of curvature (for curved components), etc. These geometric parameters are the basic characteristics of the component and are used to determine its shape in three-dimensional space. In order to correctly position each component in the model, the spatial position coordinates of each component are extracted from the building data, usually expressed in an XYZ coordinate system. These coordinates may be given in the form of relative positions or absolute positions, and the system will convert them according to actual needs. For example, the position of a beam may be given relative to a column, which is converted into absolute coordinates in the model. The system also extracts the direction information and angle of each component (such as the angle with the horizontal plane, the angle with other components). For some components (such as inclined beams or stairs), angle information is a key factor in determining their geometric shape and spatial position. This information is used to determine the correct orientation of the component in the model. For components that need to maintain a specific ratio, their size ratio is calculated according to the description in the building data and applied to the generation of the geometric model. For example, the system may need to maintain the aspect ratio of certain columns, or ensure that the radius of curvature of certain decorative components matches that of other parts.

[0066] According to the extracted geometric parameters, the preliminary geometric form of the component is generated by using the geometric modeling algorithm based on surface fitting and space segmentation, and a preliminary space layout model is constructed;

[0067] Specifically, a set of sampling points representing the component surface is generated based on the extracted geometric parameters. These sampling points can be discrete coordinate points, representing various key positions on the component surface (such as vertices, edges, midpoints, etc.). Based on the sampling points, a preliminary surface model is generated using algorithms such as polynomial surface fitting and spline curves (such as B-spline or NURBS curves). Polynomial surface fitting generates a smooth surface that meets the curvature requirements by calculating the weighted average of each sampling point; while spline curve fitting generates a continuous and smooth surface by connecting the sampling points. After the preliminary surface is generated, the error between the surface and the original geometric parameters is evaluated. If the error exceeds a preset threshold, the parameters of the surface fitting algorithm are automatically adjusted to regenerate the surface. For example, the system may increase the density of sampling points or use a higher-order polynomial for fitting to reduce the error and improve the accuracy of the surface. After multiple iterations, the system generates the final surface model. The surface must not only meet the requirements of the geometric parameters, but also be consistent with the actual construction process. For example, for some materials (such as steel plates or concrete), the smoothness and curvature radius of the surface need to meet construction specifications to ensure the feasibility of the model.

[0068] The system uses space segmentation algorithms (such as octree segmentation and KD tree segmentation) to divide the space of the entire building into multiple independent units. Each unit corresponds to a specific area in the building (such as a room, corridor, stairwell, etc.), and these units will serve as the basis for component integration. The basis for space segmentation mainly includes the functional layout, structural characteristics and construction sequence of the building. According to the segmentation results, the geometric form of the component generated by surface fitting is integrated into the corresponding space unit. During the integration process, the matching degree between the component and the space unit is verified to ensure that the position and form of the component are consistent with the spatial restrictions of the unit. After completing space segmentation and component integration, a preliminary space layout model is generated. This model reflects the overall structure of the building, the spatial position and relative relationship of each component. Although the model at this time is still a preliminary version, it already contains the basic framework of the building, laying the foundation for subsequent logical verification and material assignment.

[0069] The generated preliminary geometric model is used as the input of the BIM model generation module to generate a preliminary BIM model.

[0070] The comparison and analysis module is used to automatically compare the generated BIM model and drawings using computer vision and deep learning algorithms, identify and mark the difference areas, and generate a comparison and analysis report;

[0071] The operation process of the comparison and analysis module includes the following steps:

[0072] Acquire and preprocess the generated BIM model data and original construction drawing data;

[0073] Based on computer vision technology, key features of BIM models and construction drawings are extracted; the key features include geometric shape features, texture features, boundary features, size and material properties;

[0074] Specifically, the key feature extraction process includes the following aspects:

[0075] Geometric shape feature extraction: Use edge detection algorithms (such as Canny edge detection) to extract geometric shape features in BIM models and construction drawings. Geometric shape features include the outline, surface morphology, cross-sectional shape, etc. of components. These features are the basis for comparing the consistency between the model and the drawings.

[0076] Texture feature extraction: Texture analysis techniques (such as Gabor filter or wavelet transform) are used to extract texture features in models and drawings. These features reflect the detailed information of the surface material of the component, such as surface smoothness, granularity, pattern, etc. In material consistency inspection, texture features are an important comparison dimension.

[0077] Boundary feature extraction: Boundary features include the dividing lines between components, feature points at the junction, etc. The system uses methods such as Hough transform to identify and extract these boundary features to ensure that the model and the drawing are consistent in spatial segmentation.

[0078] Dimensional Feature Extraction: Automatically measure critical dimensions in models and drawings (such as beam length, wall height, opening width, etc.) and compare these dimensions to design standards.

[0079] Material property extraction: For BIM models that contain material information, the system extracts the material properties of the components and identifies the material symbols or annotations shown in the drawings through image recognition algorithms. The comparison of material properties can reveal the consistency or difference in material selection between the model and the drawings.

[0080] Using the capsule network algorithm, the geometric differences and position deviations between the BIM model and the construction drawings are identified through feature matching, regional overlap analysis and spatial position comparison;

[0081] Specifically, a capsule network is used to match key features in the BIM model and construction drawings. By comparing feature vectors, the similarities and differences between the two can be identified. For example, match the column outlines in the BIM model with those in the drawings to check whether their shapes, sizes, and positions are consistent. The overlap between the BIM model and the corresponding areas of the construction drawings is calculated to detect deviations in geometric positions. Overlap analysis can reveal differences in the spatial layout between the model and the drawings, such as whether certain components are displaced, misplaced, or missing. Through the spatial position comparison function of the capsule network, the position deviation between the model and the drawings in three-dimensional space can be identified. This includes spatial transformation information such as translation, rotation, and scaling of components, and marks these position deviations and evaluates their impact on the overall structure.

[0082] Automatically annotate the identified difference areas, including the type, size, location and potential cause of the difference, and mark them separately on the BIM model and construction drawings;

[0083] Specifically, each identified difference is classified, such as size deviation, position offset, material inconsistency, etc. The difference type annotation helps users quickly understand the main inconsistencies between the model and the drawing.

[0084] Difference size marking: For geometric differences, calculate the quantitative indicators of the differences, such as the specific value of the size deviation, the distance of the position offset, etc. These quantitative indicators are directly marked at the corresponding positions of the model and drawings.

[0085] Position marking: The specific location of the difference is marked in the corresponding position of the BIM model and the construction drawing. For example, the system will mark the position deviation of a wall in the model and the corresponding area of ​​the drawing, helping users to intuitively see where the problem is.

[0086] Potential cause annotation: By analyzing the contextual information of the difference area, the potential causes that may cause the difference are speculated. For example, the dimensional deviation of a component may be due to a marking error in the design drawing or a problem with the material selection during construction. The system will annotate these speculated causes in the difference report.

[0087] Comprehensively analyze the comparison results, generate a comparative analysis report, and provide a visual display of the comparison results; the comparative analysis report includes difference analysis, error statistics, potential risk assessment and suggestions;

[0088] The comparison analysis report is transmitted to the dynamic update module and the optimization feedback module.

[0089] Furthermore, the formula of the capsule network algorithm is as follows:

[0090] ;

[0091] in, represents the final output vector of the jth output capsule, i.e., the comparison result between a feature or object in the BIM model and the construction drawing; Represents the correlation weight between the features extracted from the BIM model generation module and construction drawings and the current output features; Represents building data parsed from as-built archives; represents the bias vector, which is used to correct the systematic error that may be introduced in the comparison process between the BIM model generation module and the construction drawings; n represents the total number of features extracted in the BIM model generation module; m represents the total number of features in the output capsule.

[0092] The dynamic update module is used to dynamically monitor the construction site through the Internet of Things, and automatically correct and update the BIM model based on real-time data, and recommend corresponding adjustment measures; the Internet of Things specifically includes sensor networks, drones, edge computing devices, central control devices and data transmission equipment.

[0093] It should be noted that the design of the dynamic update module is as follows:

[0094] 1. Sensor Networks

[0095] Laser scanner: used to capture the three-dimensional shape of building structures and monitor the installation accuracy and position changes of components in real time by generating high-density point cloud data.

[0096] Thermal imagers: Used to detect temperature distribution in construction materials and identify potential thermal stress or material performance issues.

[0097] Environmental sensors: monitor on-site temperature, humidity, wind speed, air pressure and other environmental factors. These data are crucial to determine construction progress and material performance.

[0098] Structural monitoring sensors: such as strain gauges, accelerometers and displacement sensors, are used to monitor the stress, displacement and vibration of key structural components in real time to ensure the safety and stability of components during construction.

[0099] 2. Drone-assisted monitoring

[0100] Equipped with high-precision photography and laser scanning equipment, drones regularly fly to scan construction sites and generate the latest geographic information and point cloud data. The advantage of drones lies in their flexibility and wide coverage. They can perform all-round scanning of complex construction scenes and capture detailed data that traditional sensors cannot obtain.

[0101] After preliminary processing, these data are combined with data from fixed sensor networks to provide a complete 3D view of the construction site. The scanning frequency of the drone can be dynamically adjusted according to the construction progress to ensure timely model updates.

[0102] 3. Edge computing and data processing

[0103] In order to reduce the delay of data transmission and improve real-time performance, the module introduces edge computing devices. These devices are usually deployed at or near the construction site and are responsible for the preliminary processing of data collected by sensors and drones.

[0104] 4. Central control equipment

[0105] The central control device is responsible for receiving data from the edge computing device and integrating this data into the BIM model. The update process includes the following steps: The central control device will first integrate the transmitted real-time data, and match and calibrate it with the existing BIM model data. Based on the integrated data, the central control system automatically corrects the geometry, positional relationship and material information of the BIM model. For example, if the laser scanning data shows that the position of a component has shifted, the central control system will automatically adjust the corresponding part in the BIM model and recalculate the relationship between adjacent components. Through the real-time updated BIM model.

[0106] It should be noted that the dynamic update module not only updates the BIM model, but also provides construction suggestions based on real-time data. Through the built-in rule engine and machine learning model, the system can analyze the real-time status of the construction site and give the following types of suggestions:

[0107] Adjustment suggestions: For example, based on monitoring data, it is recommended to adjust the installation position of a component or readjust the construction sequence to avoid potential construction errors or risks.

[0108] Material replacement suggestions: Based on changes in the on-site environment (such as high temperature or low humidity), it is recommended to use different construction materials to ensure that the final quality meets the requirements.

[0109] Schedule optimization suggestions: The system can adjust the construction schedule based on real-time data to ensure that the project is completed within the established time frame while reducing possible construction risks.

[0110] The optimization feedback module is used to feed back the results of comparison and dynamic update to relevant departments or personnel, and optimize the comparison analysis and BIM model generation logic through the optimization algorithm according to the feedback information; the optimization feedback module specifically optimizes the comparison analysis and BIM model generation logic through the differential optimization algorithm, and prioritizes the feedback information based on the multi-criteria decision analysis method, giving priority to feedback with higher technical difficulty, impact scope and urgency.

[0111] Specifically, the differential optimization algorithm is divided into the following steps:

[0112] The system uses a clustering algorithm to perform pattern recognition on the feedback data to identify frequently occurring error types and their corresponding conditions (such as specific materials, construction stages or environmental conditions).

[0113] Based on the error patterns, the system generates suggestions for parameter adjustments, such as modifying the error tolerance in the geometry modeling algorithm or adjusting the threshold when comparing the drawing to the model.

[0114] Before the actual application of the adjustment, the system will simulate in a virtual environment to test the effect of the adjustment. Through the simulation, the system can predict the accuracy and efficiency improvement after the adjustment.

[0115] Once the adjustments pass the simulation test, the system will apply the new parameter settings in the actual comparison analysis and model generation. At the same time, the module will continue to monitor the effects of these adjustments to ensure that the improvements they bring are in line with expectations.

[0116] To ensure that feedback items are processed effectively, the optimized feedback module uses a multi-criteria decision analysis (MCDA) method to prioritize feedback information. The MCDA method considers multiple factors, including:

[0117] Technical Difficulty: The technical complexity required to solve the feedback item.

[0118] Scope of impact: The impact of the feedback item on the overall project, especially its impact on safety and functionality.

[0119] Urgency: The timeliness with which the feedback item needs to be addressed, especially the potential impact on the construction schedule.

[0120] The system first assigns a weight value to each feedback item and calculates a priority score for each feedback item based on these weight values. The system supports dynamic adjustment of weight assignment according to different stages or goals of the project. For example, in the early stages of construction, the system may pay more attention to technical difficulty and urgency, while as the project approaches completion, it may pay more attention to the scope of impact. Once the priority score calculation is completed, the system will automatically generate a feedback processing priority list to guide relevant departments or personnel to prioritize the feedback items with the highest scores. The system also supports automatic generation of feedback reports, providing detailed decision-making basis and processing suggestions.

[0121] Example 2

[0122] A BIM model and drawing comparison method based on completion archives includes the following steps:

[0123] Through natural language processing technology, semantic analysis and data mining are carried out from the completion archives to parse out the building data;

[0124] Based on the building data, a BIM model that is highly consistent with the actual construction situation is generated using a geometric modeling algorithm and a rule reasoning mechanism;

[0125] Using computer vision and deep learning algorithms, the generated BIM models and drawings are automatically compared, the difference areas are identified and marked, and a comparative analysis report is generated;

[0126] Dynamically monitor the construction site through the Internet of Things, automatically correct and update the BIM model based on real-time data, and recommend corresponding adjustment measures;

[0127] The results of comparison and dynamic update are fed back to relevant departments or personnel, and the comparison analysis and BIM model generation logic are optimized through optimization algorithms based on the feedback information.

[0128] In this embodiment, a BIM model and drawing comparison method based on as-built files is applied to a BIM model and drawing comparison system based on as-built files described in Example 1, which will not be described in detail here.

[0129] In summary, the present invention uses natural language processing technology, and the system can accurately parse the building data in the completion archives, especially when faced with complex sentences and professional terms, to ensure the accuracy of data extraction. This provides a solid data foundation for the subsequent BIM model generation. The BIM model generated by the geometric modeling algorithm and rule reasoning mechanism can highly restore the actual construction situation and ensure the accuracy of the component geometry, positional relationship and material information in the model. This not only reduces the difference between the model and the actual building, but also improves the reliability of the BIM model in subsequent applications.

[0130] The present invention uses computer vision and deep learning algorithms, and the system can automatically compare the generated BIM model with the construction drawings, identify geometric differences and position deviations, and automatically generate a comparative analysis report. This process greatly reduces the workload of manual comparison and improves efficiency and accuracy. Through the Internet of Things technology, the system can monitor the status of the construction site in real time, and dynamically update the BIM model based on these data to ensure that the model is consistent with the actual construction status. This dynamic update mechanism can promptly reflect changes in the construction process and reduce the deviation between the model and the actual building caused by changes in site conditions.

[0131] The system of the present invention feeds back the results of comparison and dynamic update to relevant departments through the optimization feedback module, and continuously optimizes the BIM model generation and comparison analysis logic using the optimization algorithm. Through the differential optimization algorithm and multi-criteria decision analysis method, the system can give priority to high-priority feedback, improve the efficiency and effect of feedback processing, and ensure the continuous improvement of project quality.

[0132] The above implementation modes are merely descriptions of the preferred implementation modes of the present invention, and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary engineering and technical personnel in the field shall fall within the protection scope determined by the claims of the present invention.

Claims

1. A BIM model and drawing comparison system based on completion archives, characterized in that: It includes a completion archive data analysis module, a BIM model generation module, a comparison analysis module, a dynamic update module and an optimization feedback module which are sequentially connected in communication; The completion archive data parsing module is used to perform semantic analysis and data mining from the completion archive through natural language processing technology to parse out the building data; The building data includes key component information, geometric data and material information; The BIM model generation module is used to generate a BIM model that is highly consistent with the actual construction situation based on the building data using a geometric modeling algorithm and a rule reasoning mechanism; The comparison and analysis module is used to automatically compare the generated BIM model and drawings using computer vision and deep learning algorithms, identify and mark the difference areas, and generate a comparison and analysis report; The dynamic update module is used to dynamically monitor the construction site through the Internet of Things, and automatically correct and update the BIM model based on real-time data, and recommend corresponding adjustment measures; The optimization feedback module is used to feed back the comparison and dynamic update results to relevant departments or personnel, and optimize the comparison analysis and BIM model generation logic through the optimization algorithm according to the feedback information; Wherein, the BIM model includes the geometric form, positional relationship and material information of the building components; The operation process of the BIM model generation module includes the following steps: Based on the received building data, a preliminary BIM model is generated by a geometric modeling algorithm, wherein the preliminary BIM model is used to reflect the basic geometric form of the building components; Using the graph-based rule reasoning mechanism, the generated preliminary BIM model is logically verified and corrected, including the adjustment of positional relationships and the assignment of material information; The optimized BIM model is iterated multiple times to generate a final BIM model that is highly consistent with the actual building status; The generated final BIM model is transferred to the comparison and analysis module; The operation process of the comparison and analysis module includes the following steps: Acquire and preprocess the generated BIM model data and original construction drawing data; Based on computer vision technology, key features of BIM models and construction drawings are extracted; the key features include geometric shape features, texture features, boundary features, size and material properties; Using the capsule network algorithm, the geometric differences and position deviations between the BIM model and the construction drawings are identified through feature matching, regional overlap analysis and spatial position comparison; Automatically annotate the identified difference areas, including the type, size, location and potential cause of the difference, and mark them separately on the BIM model and construction drawings; Comprehensively analyze the comparison results, generate a comparative analysis report, and provide a visual display of the comparison results; the comparative analysis report includes difference analysis, error statistics, potential risk assessment and suggestions; Transmitting the comparative analysis report to the dynamic update module and the optimization feedback module; The formula of the capsule network algorithm is as follows: ； in, represents the final output vector of the jth output capsule, i.e., the comparison result between a feature or object in the BIM model and the construction drawing; Represents the correlation weight between the features extracted from the BIM model generation module and construction drawings and the current output features; Represents building data parsed from as-built archives; represents the bias vector, which is used to correct the systematic error that may be introduced in the comparison process between the BIM model generation module and the construction drawings; n represents the total number of features extracted in the BIM model generation module; m represents the total number of features in the output capsule; The construction process of the geometric modeling algorithm includes the following steps: Perform format conversion, normalization and noise filtering on the parsed building data; According to the key component information in the building data, the geometric parameters, position coordinates, angles and size ratios of the components are extracted, and these parameters are used as basic inputs for geometric modeling; According to the extracted geometric parameters, a preliminary geometric shape of the component is generated by using a geometric modeling algorithm based on surface fitting and space segmentation, and a preliminary spatial layout model is constructed; the construction of the spatial layout model includes the following steps: a set of sampling points representing the surface of the component is generated according to the extracted geometric parameters; based on the sampling points, a preliminary surface model is generated by using polynomial surface fitting and spline curve algorithms; polynomial surface fitting generates a smooth surface that meets the curvature requirements by calculating the weighted average of each sampling point; spline curve fitting generates a continuous and smooth surface by connecting the sampling points; after the preliminary surface is generated, the error between the surface and the original geometric parameters is evaluated; if the error When the preset threshold is exceeded, the parameters of the surface fitting algorithm are automatically adjusted to regenerate the surface; after multiple iterations, the final surface model is generated; the space of the entire building is divided into multiple independent units using the space segmentation algorithm; each unit corresponds to a specific area in the building, and the basis for space segmentation mainly includes the functional layout, structural characteristics and construction sequence of the building; according to the segmentation results, the geometric forms of the components generated by the surface fitting are integrated into the corresponding space units; during the integration process, the matching degree between the components and the space units is verified to ensure that the position and form of the components are consistent with the space restrictions of the units; after completing the space segmentation and component integration, a preliminary space layout model is generated; The generated preliminary geometric model is used as the input of the BIM model generation module to generate a preliminary BIM model; The Internet of Things specifically includes sensor networks, drones, edge computing devices, central control devices, and data transmission devices; The optimization feedback module specifically optimizes the comparison analysis and BIM model generation logic through a differential optimization algorithm, and prioritizes the feedback information based on a multi-criteria decision analysis method, giving priority to feedback with higher technical difficulty, impact scope and urgency.

2. A BIM model and drawing comparison method based on completion archives, characterized in that: The method is applied to a BIM model and drawing comparison system based on completion archives as claimed in claim 1, comprising the following steps: Through natural language processing technology, semantic analysis and data mining are carried out from the completion archives to parse out the building data; Based on the building data, a BIM model that is highly consistent with the actual construction situation is generated using a geometric modeling algorithm and a rule reasoning mechanism; Using computer vision and deep learning algorithms, the generated BIM models and drawings are automatically compared, the difference areas are identified and marked, and a comparative analysis report is generated; Dynamically monitor the construction site through the Internet of Things, automatically correct and update the BIM model based on real-time data, and recommend corresponding adjustment measures; The results of comparison and dynamic update are fed back to relevant departments or personnel, and the comparison analysis and BIM model generation logic are optimized through optimization algorithms based on the feedback information.

Images