A structural geometric deviation analysis method and system suitable for smart construction sites

Abstract

The present invention relates to the field of on-site construction monitoring technology, and in particular to a method and system for analyzing geometric deviations of structures applicable to smart construction sites. The method comprises the following steps: performing parametric modeling on a structure to obtain a three-dimensional model of the structure; obtaining on-site information data of the structure, and establishing a three-dimensional engineering construction model by combining the three-dimensional model and the on-site information data; obtaining real-scene information data of the structure, and constructing a real-scene model of the structure based on the real-scene information data; and performing virtual-real registration by combining the three-dimensional engineering construction model and the real-scene model to obtain the geometric deviation of the structure. The present invention compares and analyzes the three-dimensional engineering construction model and the real-scene model, and quantifies the geometric deviation of the structure based on the evolution analysis of the deviation field. It also captures the dynamic change trend of the deviation in real time, generates correction suggestions, provides predictive warnings for construction quality control, and improves the initiative and accuracy of construction management.

Description

Technical Field

[0001] The present invention relates to the technical field of on-site construction monitoring, and in particular to a method and system for analyzing geometric deviations of structures applicable to smart construction sites. Background Art

[0002] In the construction sector, while Building Information Modeling (BIM) has been widely adopted for design and construction management, actual construction conditions on site often exhibit geometric deviations from the BIM model. These deviations may stem from construction errors, material deformation, or environmental factors. If not promptly detected and corrected, they can compromise project quality and increase rework costs. Traditional deviation detection methods, which primarily rely on manual measurement or point cloud comparison techniques, are not only inefficient but also struggle to rapidly align and intuitively visualize BIM models with field data. In particular, they lack the ability to quantify dynamic deviation trends in real time, failing to meet the real-time and precision requirements of modern construction management. With the development of augmented reality (AR) technology, it has become possible to detect construction deviations by integrating the virtual and real worlds of AR scene models with BIM models. However, existing methods still face challenges such as insufficient matching accuracy, poor environmental adaptability, and low automation. In particular, they are unable to continuously track and analyze the evolution of deviations during construction, limiting their practical application in complex construction scenarios.

[0003] The core problem of current AR-BIM deviation analysis technology lies in how to achieve high-precision dynamic matching and intelligent detection. Due to the complex and changeable construction site environment, factors such as lighting conditions and obstructions will interfere with the spatial positioning of AR equipment, resulting in drift or deviation in the superposition of BIM models and real-life models, making it difficult to meet millimeter-level accuracy requirements. In addition, existing methods mostly rely on manual intervention for deviation identification and evaluation, lack automated data analysis and early warning mechanisms, and lack real-time monitoring and trend prediction functions for dynamic changes in deviations. This not only increases labor costs, but also reduces the timeliness of construction quality control. This severely limits the promotion of AR technology in construction deviation detection. There is a need for an intelligent analysis method that can adapt to dynamic construction environments, has high-precision matching capabilities, and can quantify the dynamic change trends of deviations in real time to improve the efficiency and reliability of construction management. Summary of the Invention

[0004] In response to the deficiencies in the prior art, the present invention provides a method and system for analyzing geometric deviations of structures applicable to smart construction sites.

[0005] To achieve the above objectives, in a first aspect, the present invention provides a method for analyzing geometric deviations of structures suitable for smart construction sites, the method comprising the following steps: performing parametric modeling on a structure to obtain a three-dimensional model of the structure; obtaining on-site information data of the structure, and establishing a three-dimensional engineering construction model by combining the three-dimensional model and the on-site information data; obtaining real-scene information data of the structure, and constructing a real-scene model of the structure based on the real-scene information data; and performing virtual-to-real registration on the three-dimensional engineering construction model and the real-scene model to obtain the geometric deviations of the structure. The present invention aligns the three-dimensional engineering construction model with the real-scene model to achieve accurate analysis of geometric deviations of structures, significantly improving the efficiency of construction quality control. The three-dimensional engineering construction model ensures accurate mapping of design drawings. The real-scene model provides real-world feedback, and the virtual-to-real registration technology effectively eliminates traditional measurement errors, achieves millimeter-level deviation detection, supports dynamic monitoring of the construction process, can provide early warning of deformation risks, optimize adjustment plans, and reduce rework costs. At the same time, the digital results provide reliable data support for later operation and maintenance, promote the full life cycle management of smart construction sites, and enhance the intelligence level and comprehensive benefits of engineering construction.

[0006] Optionally, the parametric modeling of the structure to obtain the three-dimensional model of the structure includes: establishing a component coding management system and a data quality control process for the structure, and obtaining a parameter set for the structure; verifying the integrity of the parameter set, and obtaining a state change result of the structure based on the time dimension attribute of the parameter set; and parametrically modeling the structure according to the parameter set in combination with the design drawings to obtain the three-dimensional model. The present invention ensures the standardization of the modeling process and the accuracy and reliability of the data by establishing a component coding management system and a data quality control process; parameter set integrity verification and time dimension analysis can accurately track the state evolution of the structure and provide a scientific basis for dynamic modeling; combined with parametric modeling of design drawings, the three-dimensional model can highly restore the design intent and improve modeling efficiency and accuracy; it effectively optimizes the modeling process, lays a solid foundation for structural analysis, construction simulation and quality control, and helps smart construction sites achieve refined and digital management.

[0007] Optionally, the acquiring of the on-site information data of the structure and the establishment of a three-dimensional engineering construction model in combination with the three-dimensional model and the on-site information data include: acquiring the on-site information data based on a drone and a laser scanner; constructing a fitting scene according to the on-site information data, and optimizing the fitting scene to obtain an optimized fitting scene; importing the three-dimensional model into the optimized fitting scene, and constructing the three-dimensional engineering construction model in combination with spatial coordinates. The present invention collects on-site information data through the fusion of drones and laser scanners to achieve multi-source information complementarity, significantly improving data acquisition efficiency and accuracy; constructing an optimized fitting scene to automatically correct equipment errors and environmental interference, ensuring the geometric authenticity and topological integrity of the scene; embedding the parameterized three-dimensional model into the optimized scene and aligning it based on spatial coordinates to achieve accurate mapping between the model and reality, so as to construct a three-dimensional engineering construction model; providing a high-credibility benchmark model for structural construction deviation analysis, supporting visual control and dynamic quality assessment of the construction process, and helping smart construction sites achieve millimeter-level precision control.

[0008] Optionally, the 3D engineering building model includes: lightweighting the 3D engineering building model and uploading the lightweighted 3D engineering building model to a cloud server. This lightweighting significantly reduces the amount of model data, improving transmission efficiency and terminal loading speed. Cloud server deployment enables centralized model storage and multi-terminal sharing, breaking spatial and temporal constraints, facilitating remote collaborative management, and providing a unified benchmark platform for cross-operation, thereby helping to reduce costs and increase efficiency at smart construction sites.

[0009] Optionally, the real-scene information data of the structure is obtained, and the real-scene model of the structure is constructed based on the real-scene information data, including: obtaining control point marks of the structure, wherein the control point marks are distributed at key positions of the construction area; in combination with the control point marks, the construction area is dynamically scanned based on a binocular camera and a depth sensor to obtain the real-scene information data; spatial anchor points of the construction area are obtained according to the real-scene information data, and the real-scene model is constructed based on the spatial anchor points. The present invention provides a high-precision spatial positioning benchmark for dynamic scanning by setting control point marks at key positions of the construction area, significantly improving the coordinate consistency of real-scene data acquisition; in combination with the control point marks, a binocular camera and a depth sensor are used to perform multi-perspective dynamic scanning to comprehensively capture the geometric details and spatial relationships of the structure, ensuring the integrity of the real-scene information data; and a real-scene model is constructed based on the spatial anchor points extracted based on the real-scene data; the deformation distortion problem of traditional real-scene modeling is effectively solved, and it helps smart construction sites to achieve millimeter-level quality control of dynamic construction scenes.

[0010] Optionally, the spatial anchor point of the construction area is obtained based on the real-scene information data, including: pre-filtering the dynamic interference area of ​​the construction area to obtain the static area of ​​the construction area; extracting feature points from the static area to obtain the first key frame of the static area, and filtering the first key frame to obtain the second key frame; combining the second key frame and the inertial measurement unit trajectory through tight coupling optimization to obtain the spatial anchor point. The present invention effectively eliminates the interference of dynamic factors such as construction machinery and personnel flow on spatial positioning by pre-filtering the dynamic interference area, significantly improving the stability of real-scene data acquisition; the key frame extraction and screening mechanism can eliminate redundant data, reduce the computational load while retaining structural features, and improve modeling efficiency; the tightly coupled optimization algorithm fuses the key frame and the inertial measurement unit trajectory data, effectively corrects the scale drift problem, improves the positioning accuracy of the spatial anchor point, effectively improves the analysis accuracy of the geometric deviation of the structure, and supports the smart construction site to achieve accurate digital mapping of dynamic construction scenes.

[0011] Optionally, the method of combining the three-dimensional engineering building model and the real-scene model to perform virtual-real registration to obtain the geometric deviation of the structure includes: performing virtual-real registration on the three-dimensional engineering building model and the real-scene model using a spatial mapping algorithm to obtain a model registration result; based on the model registration result, quantifying the deviation between the three-dimensional engineering building model and the real-scene model in real time using a deviation field to obtain a three-dimensional geometric deviation; constructing a dynamic color mapping function, and performing color mapping on the three-dimensional geometric deviation according to the dynamic color mapping function to obtain the geometric deviation. The present invention achieves high-precision virtual-real registration of the three-dimensional engineering model and the real-scene model through a spatial mapping algorithm, eliminating errors in multi-source data fusion and ensuring a unified deviation analysis benchmark; real-time quantification of the deviation field converts the geometric deviation into three-dimensional spatial distribution data, breaking through the limitations of traditional two-dimensional detection and achieving millimeter-level deformation positioning; the dynamic color mapping function establishes a direct correlation between the deviation value and visual perception, intuitively presenting the deviation level of each area of ​​the structure; and converting the geometric data into interactive decision-making information to support construction parties in quickly locating problem points, optimizing adjustment plans, significantly improving quality rectification efficiency, and promoting the evolution of smart construction sites to the predictive maintenance stage.

[0012] Optionally, based on the model registration result, quantifying the deviation between the three-dimensional engineering building model and the real scene model in real time through a deviation field to obtain a three-dimensional geometric deviation includes:

[0013] ;

[0014] in, is the three-dimensional geometric deviation, is the coordinate of the scanned feature point in the real scene model, represents minimization, is the coordinate of any point on the surface point set of the three-dimensional engineering building model, is the surface point set of the 3D engineering building model, is a symbolic function, is the normal vector of the 3D construction model at a point. This invention achieves high-precision quantification of the deviation between the 3D construction model and the real-life model by calculating 3D geometric deviations. This method breaks through the limitations of traditional 2D detection by establishing a 3D spatial deviation field, synchronously capturing the deformation amplitude and directional characteristics of the structure. The results are quantified in real time to support millimeter-level construction quality control. Positive and negative deviations are distinguished by signs, providing precise spatial guidance for corrective measures and significantly improving the intelligent level of geometric quality control at smart construction sites.

[0015] Optionally, constructing a dynamic color mapping function includes:

[0016] ;

[0017] in, is the function value of the dynamic color mapping function, is the three-dimensional geometric deviation, Indicates red, is a conditional branch statement, is the permissible threshold, Indicates green, Represents gradient yellow. The present invention converts three-dimensional geometric deviations into intuitive visual codes by constructing a dynamic color mapping function, thereby realizing visual control of construction quality. When the deviation exceeds the allowable threshold, red and green are used to warn of negative and positive over-limit deviations respectively; when the deviation is within the threshold, a gradient yellow linear mapping of the deviation amplitude is used to form a visualization effect; breaking through the limitations of traditional numerical display, a direct correlation between the deviation value and human visual perception is established through color gradients, quickly capturing millimeter-level deformation information, significantly improving the response speed of quality rectification, and promoting the evolution of smart construction site quality management towards intelligence and visualization.

[0018] In a second aspect, the present invention provides a structure geometric deviation analysis system applicable to smart construction sites. The system executes the structure geometric deviation analysis method applicable to smart construction sites provided by the present invention, and is characterized in that the system includes an input device, an output device, a processor, and a memory, wherein the input device, output device, processor, and memory are interconnected, wherein the memory is used to store a computer program, the computer program includes program instructions, and the processor is configured to call the program instructions. The structure geometric deviation analysis system provided by the present invention realizes automated and intelligent analysis of structure geometric deviations; improves detection efficiency, reduces human error, and ensures the accuracy and reliability of analysis results; provides strong protection for project quality and safety, and improves construction site management efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a flow chart of a method for analyzing geometric deviations of structures applicable to smart construction sites according to an embodiment of the present invention;

[0020] Figure 2 Flowchart of the improved visual-inertial SLAM algorithm according to an embodiment of the present invention;

[0021] Figure 3 A flow chart of spatial mapping and deviation field calculation according to an embodiment of the present invention;

[0022] Figure 4 This is a framework diagram of a structure geometric deviation analysis system applicable to smart construction sites according to an embodiment of the present invention;

[0023] Figure 5 4 is an execution flow chart of the structure geometric deviation analysis system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0024] Specific embodiments of the present invention will be described in detail below. It should be noted that the embodiments described herein are for illustrative purposes only and are not intended to limit the present invention. In the following description, numerous specific details are set forth to provide a thorough understanding of the present invention. However, it will be apparent to one of ordinary skill in the art that these specific details are not necessarily required to practice the present invention. In other instances, well-known circuits, software, or methods are not specifically described to avoid obscuring the present invention.

[0025] Throughout this specification, references to "one embodiment," "an embodiment," "an example," or "an example" mean that a particular feature, structure, or characteristic described in connection with the embodiment or example is included in at least one embodiment of the present invention. Therefore, appearances of the phrases "in one embodiment," "in an embodiment," "an example," or "an example" in various places throughout this specification are not necessarily all referring to the same embodiment or example. Furthermore, the particular features, structures, or characteristics may be combined in any suitable combinations and / or subcombinations in one or more embodiments or examples. Furthermore, those of ordinary skill in the art will appreciate that the figures provided herein are for illustrative purposes only and are not necessarily drawn to scale.

[0026] See Figure 1 One embodiment of the present invention provides a method for analyzing geometric deviations of structures applicable to smart construction sites, the method comprising the following steps:

[0027] S1. Performing parametric modeling on a structure to obtain a three-dimensional model of the structure.

[0028] In this embodiment, a component coding management system and data quality control process for the structure are established, and a parameter set for the structure is obtained. The integrity of the parameter set is verified, and the state change results of the structure are obtained based on the time dimension attributes of the parameter set. Based on the parameter set and in combination with the design drawings, parametric modeling of the structure is performed to obtain a three-dimensional model.

[0029] Specifically, the component coding management system adopts a four-level coding structure of "project number-professional code-component type-serial number"; a complete data quality control process needs to be established during the data entry stage, including the development of standardized parameter entry templates and the setting of value range restrictions for key data such as geometric dimensions and material properties; 3D modeling is performed for each structure in the engineering project, and relevant parameters such as the name, number, and shape data of each structure are set as the parameter set of the structure to provide a data basis for subsequent dynamic deviation analysis.

[0030] Furthermore, model checking tools such as Autodesk Navisworks (Navisworks) are regularly used to verify the data integrity of the parameter set. It is important to add time dimension attributes such as the design base date and construction start date to each structure to record the status change results from design to construction.

[0031] Furthermore, based on the parameter set of the structure, according to the design drawings and detailed coordinates of the construction area, the structure is three-dimensionally modeled using modeling software such as Revit Building Information Modeling Software (Autodesk Revit, referred to as Revit) and Inventor 3D Mechanical Design Software (Autodesk Inventor, referred to as Inventor) to obtain a three-dimensional model of the structure; it should be noted that when using Revit and Inventor to perform parametric modeling of the structure with reference to the design drawings, a strict component coding management system must be established.

[0032] S2. Acquire on-site information data of the structure, and establish a three-dimensional engineering building model by combining the three-dimensional model and the on-site information data.

[0033] In this embodiment, on-site information data is acquired based on drones and laser scanners; a fitting scene is constructed based on the on-site information data, and the fitting scene is optimized to obtain an optimized fitting scene; a three-dimensional model is imported into the optimized fitting scene, and a three-dimensional engineering building model is constructed in combination with spatial coordinates; the three-dimensional engineering building model is lightweighted, and the lightweighted three-dimensional engineering building model is uploaded to a cloud server.

[0034] Specifically, when establishing a complete three-dimensional engineering building model; first, a drone is used to obtain on-site information data through oblique photography and a ground laser scanner; secondly, based on the on-site information data, the surrounding scene is drawn using three-dimensional modeling, animation and rendering software (Autodesk 3ds Max, referred to as 3D max), and the model is mapped and rendered using 3D max, so as to obtain a more realistic scene as a fitting scene; then, the fitting scene is optimized to obtain an optimized fitting scene, and the optimization process includes mesh simplification, element instantiation and texture optimization; subsequently, the established three-dimensional model is imported into the optimized fitting scene, and combined together according to spatial coordinates to form a complete three-dimensional engineering building model.

[0035] Furthermore, the 3D engineering building model is lightweighted and uploaded to the cloud server; the lightweight processing enables efficient transmission and rendering of the 3D engineering building model while retaining the accuracy of key features; the 3D engineering building model is uploaded to the cloud server to ensure that different terminal devices can obtain adaptive and appropriate accuracy level 3D engineering building model data in real time.

[0036] Specifically, 3D engineering and architectural models are uploaded to the Glodon BIMFACE platform (bimface) for lightweight processing. Different levels of detail are configured for different terminals, preserving key features while controlling complexity. A comprehensive data management system is established for cloud upload, including version control, permission management, and backup solutions. Adaptive streaming loading technology is used for intelligent transmission, dynamically adjusting loading strategies based on device performance and network conditions. This implementation ensures modeling quality, maintains measurement accuracy, optimizes network transmission, and keeps the model updated synchronously with the site. This ensures high-quality model support for multiple terminals and meets the subsequent requirements of virtual-reality integration.

[0037] It should be noted that the Bimface platform automatically converts three-dimensional engineering building models into lightweight ones, and the information is stored in the cloud and can be downloaded and used on different terminals.

[0038] In addition, adaptive streaming loading technology is an intelligent "traffic regulator" that can automatically adjust the transmission method of 3D engineering and construction models according to different device performance and network conditions. For high-configuration devices (such as computers), adaptive streaming loading technology will transmit a detailed model with all details; while for mobile devices such as mobile phones or AR glasses, it will automatically simplify the model to ensure smooth operation. When the network conditions are good, the complete data is transmitted; when the network is poor, the basic framework is transmitted first and then the details are gradually added. In this way, whether it is a high-end computer used by a designer or an ordinary mobile phone used by a field worker, they can smoothly view the same set of 3D engineering and construction models, without lag caused by the large size of the model, and can maintain basic usage functions when the network conditions are poor.

[0039] S3. Obtaining real-scene information data of the structure, and constructing a real-scene model of the structure based on the real-scene information data.

[0040] In this embodiment, control point markers are acquired from the structure, distributed at key locations within the construction area. Combined with these markers, the construction area is dynamically scanned using a binocular camera and depth sensor to obtain real-world information data. Spatial anchor points within the construction area are derived from this real-world information data, and a real-world model is constructed based on these spatial anchor points. The coordinates of the spatial anchor points in the real-world model are aligned with the coordinate system of the 3D engineering building model using a unified transformation matrix to ensure consistent registration benchmarks.

[0041] Furthermore, a spatial anchor point of the construction work area is obtained based on the real-scene information data, including: pre-filtering the dynamic interference area of ​​the construction work area to obtain a static area of ​​the construction work area; extracting feature points of the static area to obtain a first key frame of the static area, and filtering the first key frame to obtain a second key frame; and combining the second key frame and the inertial measurement unit trajectory through tight coupling optimization to obtain a spatial anchor point.

[0042] Specifically, AR glasses equipped with binocular cameras and depth sensors scan the construction site, using an improved Visual-Inertial Simultaneous Localization and Mapping (SLAM) algorithm to accurately identify landmarks and feature points. Environmental perception is achieved through depth sensors. The SLAM algorithm uses semantic segmentation to pre-filter dynamic interference areas. It then extracts feature points in static areas using Oriented FAST and Rotated BRIEF (ORB), ensuring a first keyframe that is scale and rotation invariant. Keyframe management filters keyframe data based on the dual criteria of feature quality and semantic stability to obtain a second keyframe. This second keyframe is then combined with the inertial measurement unit trajectory through tightly coupled optimization. Ultimately, feature recognition is achieved in complex environments, establishing stable spatial anchors and subsequently constructing a real-world model.

[0043] It should be noted that the inertial measurement unit trajectory comes from the inertial measurement unit sensor installed on the AR glasses or mobile scanning device. The inertial measurement unit sensor usually contains a three-axis accelerometer, a three-axis gyroscope, and a three-axis magnetometer, which can measure the motion state and orientation changes of the device in real time. The AR glasses worn by the operator have a built-in inertial measurement unit sensor, which collects the acceleration, angular velocity and direction data of the device in real time, and calculates the precise three-dimensional motion trajectory through the sensor fusion algorithm; based on the above inertial measurement unit data, it works in conjunction with the visual-inertial SLAM algorithm to maintain positioning stability when features are missing or the environment is complex, providing a reliable motion reference for real-scene modeling.

[0044] In this embodiment, when using AR glasses equipped with a depth sensing module for dynamic scanning at the construction site, the operator needs to pay attention to the following key points: First, it should be ensured that a sufficient number of control point markers have been pre-deployed at the construction site. These markers should be evenly distributed at key locations in the work area, including building corners, floor elevations and other feature locations, while retaining sufficient natural feature points as auxiliary references. Before starting the scan, the operator needs to complete the equipment calibration process according to the system prompts, including sensor calibration and ambient light adaptation testing, to ensure that each sensor module is in the best working condition. During the actual scanning process, the operator should maintain a steady movement speed, paying special attention to the complete collection of complex structural areas and hidden parts. When the system prompts that there are insufficient feature points or the tracking is unstable, the scan should be paused immediately, and the data quality should be improved by adding manual markings or adjusting the scanning path.

[0045] See Figure 2The figure shows a flow chart of the improved visual-inertial SLAM algorithm. During runtime, the visual-inertial SLAM algorithm integrates three-dimensional data streams from visual feature matching, an inertial measurement unit, and depth sensing in real time. The system's front-end processing stage first enhances construction features and filters dynamic objects, eliminating transient interference through an adaptive filtering algorithm. The back-end optimization stage implements data coupling optimization and incorporates BIM constraints to maintain a stable spatial coordinate system. To ensure positioning accuracy, the system continuously monitors tracking quality indicators. When accuracy degrades, it automatically triggers a relocalization process, prompting the operator to rescan specific areas if necessary. Throughout the scanning process, operators should avoid actions that could affect data quality, such as direct sunlight into the lens, vigorous shaking of the device, or prolonged periods of time spent in feature-poor areas. The system's final AR positioning data undergoes rigorous virtual-reality fusion processing, and upon completion of the scan, an automatic quality assessment report is generated, including key metrics such as feature extraction quality and positioning accuracy. Subsequent virtual-reality alignment operations can only be performed after the operator confirms the data's validity. These measures together ensure that a high-precision spatial positioning benchmark can be maintained in a complex and changing construction environment, providing a reliable spatial reference for the accurate alignment of BIM models with real scenes.

[0046] S4. Perform virtual-real registration on the three-dimensional engineering building model and the real-scene model to obtain a geometric deviation of the structure.

[0047] In this embodiment, a spatial mapping algorithm is used to perform virtual-real registration of the three-dimensional engineering building model and the real-scene model to obtain a model registration result; based on the model registration result, the deviation between the three-dimensional engineering building model and the real-scene model is quantified in real time through a deviation field to obtain a three-dimensional geometric deviation; a dynamic color mapping function is constructed, and the three-dimensional geometric deviation is color mapped according to the dynamic color mapping function to obtain a geometric deviation.

[0048] Specifically, in the practical application of virtual-reality registration technology, the three-dimensional engineering building model and the real-scene model are first aligned at the sub-millimeter level through a high-precision spatial mapping algorithm to obtain the model registration result. This process requires special attention to the interference of ambient lighting changes and dynamic obstacles on feature point matching.

[0049] See Figure 3 The figure shows a flow chart of spatial mapping and deviation field calculation. After the virtual-real registration is completed, the deviation field calculation method is used based on the model registration results to quantify the 3D geometric deviation between the 3D engineering building model and the real-scene model in real time. The spatial mapping and deviation field calculation method not only captures the 3D geometric deviation at the current moment, but also continuously analyzes the rate of change and development trend of the deviation over time. The 3D geometric deviation satisfies the following relationship:

[0050]

[0051] in, is the three-dimensional geometric deviation, is the coordinate of the scanned feature point in the real scene model, represents minimization, is the coordinate of any point on the surface point set of the three-dimensional engineering building model, is the surface point set of the 3D engineering building model, is a symbolic function, is the normal vector of the 3D engineering building model at the point.

[0052] In this embodiment, the real-time status and evolution process of the deviation are intuitively displayed in various forms such as dynamic heat maps and trend vector arrows through the AR interface, achieving a technological breakthrough from static detection to dynamic process monitoring. In the dynamic monitoring stage, a heat map presented in a gradient color spectrum can be observed superimposed on the surface of the real component through AR glasses. The system uses dynamic color spectrum mapping technology to intuitively present the deviation analysis results: the red area indicates the negative out-of-limit deviation (δ<-τ), the green area indicates the positive out-of-limit deviation (δ>τ), and the yellow gradient area linearly reflects the degree of deviation within the threshold range through saturation changes. This standardized heat map highlights the key deviations that need to be handled first in red and green, and marks the acceptable error range in yellow. At the same time, the spatial distribution characteristics of the deviation are intuitively displayed through the color gradient, and the semi-transparent trend vector arrow dynamically displays the development direction and evolution rate of the deviation. When the error is within different ranges, constructing a dynamic color mapping function can dynamically map different colors of the three-dimensional geometric deviation. The dynamic color mapping function satisfies the following relationship:

[0053]

[0054] in, is the function value of the dynamic color mapping function, is the three-dimensional geometric deviation, Indicates red, is a conditional branch statement, is the permissible threshold, Indicates green, Indicates a gradient of yellow.

[0055] It should be noted that is the permissible threshold, the default .

[0056] According to Article 8.3.2 of the "Code for Acceptance of Construction Quality of Concrete Structure Engineering" (GB 50204-2015): The allowable deviation of the surface flatness of cast-in-place concrete structures is stipulated to be ±5mm (ordinary formwork) or ±3mm (advanced formwork).

[0057] S5. Obtain a structured detection report of the structure based on the geometric deviation.

[0058] In this embodiment, the detected geometric deviation data and position information are automatically recorded, a structured detection report is generated and synchronized to the cloud.

[0059] Specifically, during the inspection process, all geometric deviation data and its spatial location information are automatically recorded and stored in a structured cloud database. Based on relevant national and industry standards, the system automatically assesses the severity of deviations against engineering quality standards and generates a structured inspection report containing a deviation distribution diagram, statistical analysis, and corrective action recommendations. This structured inspection report, as a standard inspection report, can be exported in various formats, including portable document formats and images.

[0060] S6. Construct a distributed collaborative platform, and share the structured detection report in real time based on the distributed collaborative platform.

[0061] In this embodiment, a collaborative working mode of multiple pairs of AR glasses is constructed to share detection results and annotation information in real time.

[0062] Specifically, a distributed collaborative platform is constructed, employing a three-tiered architecture: Multiple pairs of AR glasses collect spatial data in real time at the terminal layer, while the edge layer processes spatial modeling and preliminary data screening locally. The cloud integrates global information and dynamically schedules tasks. Leveraging the low latency and high bandwidth of fifth-generation mobile communication technology (5G), this platform utilizes an incremental synchronization protocol to transmit only the modified portions of the deviated data. This protocol uses timestamps to mark changes and a differential compression algorithm to reduce data volume. Low-latency transmission is achieved through the 5G network, enabling efficient multi-terminal collaboration. The incremental synchronization protocol utilizes a bidirectional verification strategy. When edge nodes detect data deviations between devices, they automatically trigger a local resynchronization process, while the cloud continuously optimizes the stability of the global coordinate system. Through layered processing and intelligent priority management, this architecture ensures real-time performance while building a high-precision collaborative perception network shared by multiple terminals.

[0063] It should be noted that the incremental synchronization protocol is an efficient data synchronization technology for multi-terminal collaboration on smart construction sites. It only transmits the changed portion of the deviation data, marking the changes with timestamps, reducing the data volume with a differential compression algorithm, and relying on the 5G network to achieve low-latency transmission. For example, when a certain AR glasses detects a new structural deviation or an update to the design model, the system automatically identifies the changed area, compresses it and synchronizes it to other terminals and the cloud. At the same time, it uses timestamp priority and a two-way verification mechanism to resolve conflicts caused by concurrent modifications on multiple terminals, ultimately enabling all devices to share a consistent engineering status in real time.

[0064] See Figure 4In an optional embodiment, the present invention provides a structure geometry deviation analysis system for smart construction sites. The system includes an input device, an output device, a processor, and a memory, wherein the hardware components are interconnected. The memory is used to store a computer program, which includes program instructions. The processor is configured to call the program instructions and execute the specific steps of the embodiment of the structure geometry deviation analysis method for smart construction sites provided by the present invention. The structure geometry deviation analysis system for smart construction sites provided by the present invention is structurally complete, objective, and stable, enhancing the overall applicability and practical application capabilities of the present invention.

[0065] See Figure 5 The figure is an execution flow chart of the structure geometric deviation analysis system; it shows the analysis steps of the structure geometric deviation analysis system in detail and describes the specific structure geometric deviation analysis process of the system processor.

[0066] This invention integrates BIM models with AR real-world technology to construct a complete intelligent dynamic construction quality monitoring system. First, a visual-inertial SLAM algorithm is used to perform high-precision three-dimensional reconstruction of the construction site. Simultaneously, the lightweight BIM model and the real-world model are spatially registered to form a precise coordinate mapping relationship. By calculating the signed distance field between the BIM model surface and the real-world model, three-dimensional deviation field data is constructed, providing a quantitative basis for quality inspection. A multi-sensor fusion solution is employed to ensure positioning accuracy, with a binocular camera providing visual feature matching and a depth sensor acquiring spatial geometric information. AR glasses, serving as human-computer interaction terminals, not only intuitively present the calculated real-time deviation data in the form of dynamic heat maps in the real scene but also demonstrate the future development direction of deviations through trend prediction algorithms, achieving a technological leap from static detection to process prediction. Construction deviations can also be detected in real time, automatically recording all detection data and generating structured reports. Data sharing across multiple terminals is achieved through cloud synchronization. This improves the quality management level of building construction. Through real-time detection with millimeter-level precision, construction deviations can be detected early and corrected promptly, thereby improving construction quality.

[0067] In summary, the method of the present invention provides a method and system for analyzing the geometric deviation of structures suitable for smart construction sites, which realizes high-precision three-dimensional dynamic matching of construction scenes, enhances the stability of the AR environment by using multi-sensor data, ensures the precise alignment of the three-dimensional engineering building model and the real-scene model, and automatically identifies and quantifies geometric deviations in combination with intelligent algorithms. The dynamic change trend of the deviation is captured in real time through time series analysis, and correction suggestions are generated. The deviation field evolution analysis in the time dimension is introduced, which can monitor the change rate and direction of the deviation in real time, provide predictive warning for construction quality control, and thus significantly improve the initiative and accuracy of construction management. The method of the present invention is easy to understand, simple to calculate, has a small workload, is convenient for engineering application, and provides a theoretical basis and technical support for the further development of on-site construction monitoring technology.

[0068] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some or all of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and description of the present invention.

Claims

1. A method for analyzing geometric deviations of structures applicable to smart construction sites, characterized in that: The steps include: Performing parametric modeling on the structure to obtain a three-dimensional model of the structure; Acquiring on-site information data of the structure, and establishing a three-dimensional engineering building model by combining the three-dimensional model and the on-site information data; Obtaining real-scene information data of the structure, and constructing a real-scene model of the structure based on the real-scene information data; Performing virtual-real registration on the three-dimensional engineering building model and the real-scene model to obtain a geometric deviation of the structure; The step of performing virtual-real registration on the three-dimensional engineering building model and the real scene model to obtain the geometric deviation of the structure includes: Performing virtual-real registration on the three-dimensional engineering building model and the real scene model using a spatial mapping algorithm to obtain a model registration result; Based on the model registration result, the deviation between the three-dimensional engineering building model and the real scene model is quantified in real time through a deviation field to obtain a three-dimensional geometric deviation; Constructing a dynamic color mapping function, and performing color mapping on the three-dimensional geometric deviation according to the dynamic color mapping function to obtain the geometric deviation; The method of quantifying the deviation between the three-dimensional engineering building model and the real scene model in real time through a deviation field based on the model registration result to obtain a three-dimensional geometric deviation includes: ； in, is the three-dimensional geometric deviation, is the coordinate of the scanned feature point in the real scene model, represents minimization, is the coordinate of any point on the surface point set of the three-dimensional engineering building model, is the surface point set of the 3D engineering building model, is a symbolic function, is the normal vector of the 3D engineering building model at the point; The dynamic color mapping function is constructed, comprising: ； in, is the function value of the dynamic color mapping function, is the three-dimensional geometric deviation, Indicates red, is a conditional branch statement, is the permissible threshold, Indicates green, Indicates a gradient of yellow.

2. The method for analyzing geometric deviations of structures applicable to smart construction sites according to claim 1 is characterized in that: The parametric modeling of the structure to obtain a three-dimensional model of the structure includes: Establishing a component coding management system and data quality control process for the structure, and obtaining a parameter set for the structure; Performing integrity verification on the parameter set, and obtaining a state change result of the structure based on a time dimension attribute of the parameter set; According to the parameter set and in combination with the design drawings, the structure is parametrically modeled to obtain the three-dimensional model.

3. The method for analyzing geometric deviations of structures applicable to smart construction sites according to claim 1 is characterized in that: The step of acquiring the on-site information data of the structure and establishing a three-dimensional engineering building model by combining the three-dimensional model and the on-site information data includes: Acquiring the on-site information data based on drones and laser scanners; Constructing a fitting scene according to the on-site information data, and optimizing the fitting scene to obtain an optimized fitting scene; The three-dimensional model is imported into the optimized fitting scene, and the three-dimensional engineering building model is constructed in combination with the spatial coordinates.

4. The method for analyzing geometric deviations of structures applicable to smart construction sites according to claim 3 is characterized in that: The three-dimensional engineering building model includes: The three-dimensional engineering building model is lightweighted and uploaded to a cloud server.

5. The method for analyzing geometric deviations of structures applicable to smart construction sites according to claim 1 is characterized in that: The obtaining of the real scene information data of the structure and constructing the real scene model of the structure based on the real scene information data includes: Obtaining control point marks of the structure, wherein the control point marks are distributed at key locations in the construction area; In combination with the control point marks, the construction work area is dynamically scanned based on a binocular camera and a depth sensor to obtain the real scene information data; A spatial anchor point of the construction operation area is obtained according to the real scene information data, and the real scene model is constructed based on the spatial anchor point.

6. The method for analyzing geometric deviations of structures applicable to smart construction sites according to claim 5 is characterized in that: The obtaining of the spatial anchor point of the construction operation area according to the real scene information data includes: Pre-filtering the dynamic interference area of ​​the construction operation area to obtain the static area of ​​the construction operation area; Extracting feature points from the static area to obtain a first key frame of the static area, and filtering the first key frame to obtain a second key frame; The second key frame and the inertial measurement unit trajectory are combined through tightly coupled optimization to obtain the spatial anchor point.

7. A structural geometric deviation analysis system suitable for smart construction sites, characterized by: The system includes an input device, an output device, a processor and a memory, and the input device, output device, processor and memory are interconnected, wherein the memory is used to store a computer program, the computer program includes program instructions, and the processor is configured to call the program instructions to execute the structural geometric deviation analysis method applicable to smart construction sites as described in any one of claims 1-6.

Images