Construction progress monitoring methods and systems based on big data

Abstract

This application relates to the field of construction progress monitoring technology, and discloses a construction progress monitoring method and system based on big data. The method includes: forming a spatiotemporally aligned dataset through multi-source data acquisition, filtering, and quality assessment; generating a digital model through feature extraction and registration; forming a completion status table through target detection and classification; achieving progress assessment through component-task mapping; identifying risks and generating prediction results through trend analysis; and providing decision-making references through personalized information filtering and augmented reality display. This application, through multi-source data acquisition, fusion, and intelligent analysis, achieves accurate perception, objective assessment, scientific prediction, and intuitive presentation of the actual state of the construction site, thereby providing a comprehensive, accurate, and forward-looking construction progress monitoring method, effectively reducing the risk of construction delays and improving construction management efficiency.

Description

Technical Field

[0001] This application relates to the field of construction progress monitoring technology, and in particular to a construction progress monitoring method and system based on big data. Background Technology

[0002] With the increasing scale and complexity of construction projects, traditional construction progress monitoring methods are no longer sufficient to meet the needs of modern construction project management. Existing construction progress monitoring technologies mainly include manual inspection and recording, periodic photo comparison, and localized measurements. These methods typically rely on the experience and judgment of project managers, and the collected data is often limited to specific areas and time points. Although BIM (Building Information Modeling) technology has been applied to construction management in recent years, enabling the digital representation of design information, and some projects have begun to explore using single technologies such as photogrammetry and laser scanning to assist in progress monitoring, these applications are still in their initial stages and have not yet formed a complete data collection, analysis, and decision support system.

[0003] However, existing technologies have significant shortcomings: First, data collection methods are singular and fragmented, making it difficult to obtain comprehensive and continuous construction site information; second, there is a lack of effective multi-source data fusion mechanisms, making it impossible to integrate data from different sources and of different types into a unified analytical basis; third, progress assessment mainly relies on manual experience and judgment, lacking objective and quantitative evaluation standards and automated analysis methods; furthermore, existing technologies struggle to predict future progress trends and potential risks, failing to provide forward-looking decision support; finally, the presentation of progress information is simplistic, unable to provide personalized, visualized decision-making references based on the needs of different users. These shortcomings severely restrict the efficiency and quality of construction management, making it difficult to prevent and control risks such as project delays, quality problems, and cost overruns in a timely manner. Summary of the Invention

[0004] This application provides a construction progress monitoring method and system based on big data. It is used to achieve accurate perception, objective evaluation, scientific prediction and intuitive presentation of the actual status of the construction site through multi-source data collection, fusion and intelligent analysis, thereby providing a comprehensive, accurate and forward-looking construction progress monitoring method, effectively reducing the risk of construction delays and improving construction management efficiency.

[0005] Firstly, this application provides a construction progress monitoring method based on big data. The method includes: acquiring construction site images, videos, and point clouds through multi-source devices; filtering and quality-assessing the acquired data to obtain a spatiotemporally aligned original construction dataset; performing feature extraction and registration on image sequences and point cloud data based on the original construction dataset; spatially aligning the measured data with the design model to obtain a construction site digital model containing geometric and temporal information; and performing target detection and classification on site components based on the construction site digital model, analyzing the spatial relationship between components, and obtaining a structure-defined digital model. The system generates a construction completion status table with component identification and attributes. Based on this table, it establishes a mapping relationship between actual components and planned construction tasks, calculates component completion and positional deviation values, and obtains a construction progress assessment report and component-level deviation data. Based on the construction progress assessment report and component-level deviation data, it analyzes progress change trends, identifies abnormal patterns and potential risk points during construction, and obtains progress prediction results and a risk warning list. The system then filters the progress prediction results and risk warning list according to user permissions and priorities, overlaying and displaying progress and deviation information on a real-world site view to provide a visualized construction management decision-making reference.

[0006] Secondly, this application provides a construction progress monitoring system based on big data, the construction progress monitoring system based on big data includes:

[0007] The evaluation module is used to acquire images, videos and point clouds of the construction site through multi-source devices, filter and evaluate the quality of the acquired data, and obtain a spatiotemporally aligned original construction dataset.

[0008] The registration module is used to extract and register features from image sequences and point cloud data based on the original construction dataset, spatially align the measured data with the design model, and obtain a digital model of the construction site containing geometric and temporal information.

[0009] The classification module is used to perform target detection and classification of on-site components based on the construction site digital model, analyze the spatial positional relationship between components, and obtain a construction completion status table with component identification and attributes.

[0010] The mapping module is used to establish a mapping relationship between actual components and construction plan tasks based on the construction completion status table, calculate the component completion degree and position deviation value, and obtain a construction progress assessment report and component-level deviation data.

[0011] The analysis module is used to analyze the progress change trend based on the construction progress assessment report and component-level deviation data, identify abnormal patterns and potential risk points in the construction process, and obtain progress prediction results and risk warning list.

[0012] The filtering module is used to filter the progress prediction results and risk warning list according to user permissions and priorities, and overlay the progress and deviation information on the real-world site to obtain a visualized construction management decision-making reference.

[0013] Thirdly, a computer device is provided, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor invokes the instructions in the memory to cause the computer device to execute the above-described big data-based construction progress monitoring method.

[0014] Fourthly, a computer-readable storage medium is provided, wherein instructions are stored therein, which, when executed on a computer, cause the computer to perform the aforementioned construction progress monitoring method based on big data.

[0015] The technical solution provided in this application acquires images, videos, and point cloud data from construction sites through multi-source devices, achieving comprehensive and continuous perception of the construction site. This overcomes the limitations of traditional single data sources and improves the integrity and diversity of raw data. Filtering and quality assessment of the collected data ensures accuracy and reliability, laying a solid foundation for subsequent analysis. The spatiotemporally aligned raw construction dataset solves the problem of fusion of multi-source heterogeneous data, enabling unified representation and processing of different types of data. Based on feature extraction and registration techniques of the raw dataset, the measured data and design model can be accurately matched, forming a digital model of the construction site containing geometric and temporal information, creating conditions for comparison between the actual and planned data. Compared with traditional BIM models, this digital model not only includes static design information but also incorporates measured data from the dynamic construction process, truly reflecting the actual state and changes of the construction site. Deep learning algorithms are used for target detection and classification of on-site components, automatically identifying component types and attributes, significantly improving the accuracy and efficiency of component identification. Particularly for components in complex environments, the feature extraction capabilities of deep neural networks are significantly superior to traditional computer vision methods. Spatial relationship analysis algorithms clearly reveal the positional constraints and connections between components, forming a construction completion status table with component identifiers and attributes, providing accurate basic data for progress assessment. The method of establishing a mapping relationship between actual components and planned construction tasks realizes the transformation from physical completion status to planned completion ratio, solving the problems of strong subjectivity and poor consistency in traditional progress assessment. The method of quantifying component installation accuracy through bounding box occupancy rate and point-to-surface distance evaluation transforms construction quality assessment from qualitative description to quantitative indicators, providing objective basis for management decisions through progress assessment reports and component-level deviation data. The progress change trend analysis method based on time series analysis and machine learning can mine potential patterns from historical data, identifying abnormal patterns and risk points. The risk propagation network established by combining historical project experience data significantly enhances the accuracy and foresight of risk prediction, and the progress prediction results and risk warning list provide managers with opportunities for proactive planning. Finally, by using augmented reality technology to intuitively overlay and display progress and deviation information on the real-world scene, combined with a personalized information filtering mechanism based on user permissions and focus, complex progress data becomes intuitive and easy to understand. Decision-makers can directly obtain the necessary information and respond on-site, greatly improving decision-making efficiency and accuracy. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of one embodiment of the construction progress monitoring method based on big data in this application.

[0018] Figure 2 This is a schematic diagram of one embodiment of the construction progress monitoring system based on big data in this application.

[0019] Figure 3 This is a schematic block diagram of the structure of the computer device in an embodiment of the present invention. Detailed Implementation

[0020] This application provides a construction progress monitoring method and system based on big data. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0021] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of the construction progress monitoring method based on big data in this application includes:

[0022] Step S101: Acquire construction site images, videos and point clouds through multi-source devices, filter and evaluate the quality of the acquired data to obtain a spatiotemporally aligned original construction dataset;

[0023] Step S102: Based on the original construction dataset, perform feature extraction and registration on the image sequence and point cloud data, and spatially align the measured data with the design model to obtain a digital model of the construction site containing geometric and temporal information.

[0024] Step S103: Based on the digital model of the construction site, target detection and classification of on-site components are performed, the spatial positional relationship between components is analyzed, and a construction completion status table with component identification and attributes is obtained.

[0025] Step S104: Based on the construction completion status table, establish a mapping relationship between the actual components and the construction plan tasks, calculate the component completion degree and position deviation value, and obtain the construction progress assessment report and component-level deviation data.

[0026] Step S105: Based on the construction progress assessment report and component-level deviation data, analyze the progress change trend, identify abnormal patterns and potential risk points in the construction process, and obtain the progress prediction results and risk warning list.

[0027] Step S106: Filter the progress forecast results and risk warning list according to user permissions and priorities, and overlay the progress and deviation information onto the real-world site view to obtain a visualized construction management decision-making reference.

[0028] It is understood that the executing entity of this application can be a construction progress monitoring system based on big data, or it can be a terminal or a server; no specific limitation is made here. This application's embodiment uses a server as an example for illustration.

[0029] Specifically, multi-source data acquisition technology was employed. High-definition cameras were deployed at fixed locations on the construction site to capture image sequences, a drone aerial photography system was used to obtain overhead video streams, and a laser scanner was used to acquire 3D point cloud data. The raw data collected contained noise and redundancy, requiring filtering. Median filtering was used to remove noise and correct distortion in the image data, while outliers were removed from the point cloud data using a statistical outlier filtering algorithm. Subsequently, the data from different sources underwent quality assessment, establishing a quality scoring system that included indicators such as clarity, completeness, and coverage to select high-quality data. Finally, the selected data were indexed using timestamps and spatial coordinates to form a spatiotemporally aligned original construction dataset.

[0030] Feature extraction and spatial registration are performed using the original construction dataset. SIFT feature points and descriptors are extracted from the image sequence to identify matching relationships between images and establish a sparse 3D point cloud skeleton. Geometric features such as planes and edges are extracted from the point cloud data to determine the main spatial structure. These features are compared with structural features in the Building Information Design (BID) model to calculate the spatial transformation matrix between the measured data and the design model, completing the coordinate system alignment transformation. Data from different periods are sorted and correlated using timestamps to construct a temporal record of the construction process, ultimately generating a digital model of the construction site containing both geometric and temporal information.

[0031] Edge detection and segmentation are performed on the component areas in the digital model to determine candidate component regions. Feature vectors for each candidate region are calculated, including shape proportions, texture features, and color distribution, and matched and classified against a pre-defined component type library, labeling them as different types such as walls, columns, beams, doors and windows, and pipes. Three-dimensional coordinates, orientation, and dimensions are measured for the classified components to establish their geometric descriptions. Spatial relationships between components are analyzed, identifying topological relationships such as support, connection, and containment, and constructing a component relationship network. Finally, information such as component type, location, size, and installation status is integrated into a structured data table, forming a construction completion status table with component identifiers and attributes. Construction progress assessment and deviation analysis are implemented. First, construction plan data is imported from the project management system, and the work breakdown structure and key milestones are extracted to construct a construction task dependency network. Component coding establishes a mapping relationship between actual components in the construction completion status table and planned tasks, determining the planned task item corresponding to each component. The completion status of components involved in each task is calculated; for example, the completion rate of wall construction tasks is determined by the ratio of the actual number of installed wall components to the planned number. Simultaneously, deviations between actual components and their designed locations are measured using the bounding box occupancy rate method and the point-to-surface distance method, such as the average distance of pipe installation locations from their designed locations. Combining task weights and critical paths, an overall progress assessment report is generated, along with component-level deviation data including a deviation distribution heatmap. Progress trends are analyzed and predicted. Historical data points are extracted from the progress assessment report to construct a progress-time curve, which is then separated into trend, seasonal, and random components using time series decomposition techniques. The correlation between component-level deviation data and progress delays is analyzed to identify construction bottlenecks; for example, areas with large deviations in electromechanical pipe installation often lead to delays in subsequent processes. Risk patterns in similar construction scenarios are identified through comparison with historical project database data, and risk propagation paths are established; for example, deviations in duct installation may trigger subsequent pipeline conflicts. Based on historical trends and current status, the probability of future construction completion is predicted, generating progress prediction curves with different confidence intervals. A tiered risk warning list is formed based on the impact and urgency of risk factors. Visualized decision support is implemented. Information is filtered according to user role permissions; for example, project managers can view all information, while specialized subcontractors can only view the progress and risks relevant to their specific area. Information is prioritized and filtered based on user focus, real-time camera images are acquired and spatially located, and progress and deviation information are converted into layers and overlaid on the real-world view. For example, red warning markers are displayed at delayed component locations, while green markers are displayed in areas with normal progress. Progress prediction curves and risk alerts are integrated to form an interactive interface with multiple views, allowing users to intuitively view progress status and potential risks on the construction site via tablets or head-mounted devices.

[0032] In this embodiment, images, videos, and point cloud data of the construction site are acquired through multi-source devices, achieving comprehensive and continuous perception of the construction site. This overcomes the limitations of traditional single data sources and improves the integrity and diversity of the raw data. Filtering and quality assessment of the collected data ensures accuracy and reliability, laying a solid foundation for subsequent analysis. The spatiotemporally aligned raw construction dataset solves the problem of fusion of multi-source heterogeneous data, enabling unified representation and processing of different types of data. Feature extraction and registration techniques based on the raw dataset enable precise correspondence between measured data and the design model, forming a digital model of the construction site containing geometric and temporal information, creating conditions for comparing the actual situation with the plan. Compared with traditional BIM models, this digital model not only includes static design information but also incorporates measured data from the dynamic construction process, truly reflecting the actual state and changes of the construction site. Deep learning algorithms are used for target detection and classification of on-site components, automatically identifying component types and attributes, significantly improving the accuracy and efficiency of component identification. Particularly for components in complex environments, the feature extraction capabilities of deep neural networks are significantly superior to traditional computer vision methods. Spatial relationship analysis algorithms clearly reveal the positional constraints and connections between components, forming a construction completion status table with component identifiers and attributes, providing accurate basic data for progress assessment. The method of establishing a mapping relationship between actual components and planned construction tasks realizes the transformation from physical completion status to planned completion ratio, solving the problems of strong subjectivity and poor consistency in traditional progress assessment. The method of quantifying component installation accuracy through bounding box occupancy rate and point-to-surface distance evaluation transforms construction quality assessment from qualitative description to quantitative indicators, providing objective basis for management decisions through progress assessment reports and component-level deviation data. The progress change trend analysis method based on time series analysis and machine learning can mine potential patterns from historical data, identifying abnormal patterns and risk points. The risk propagation network established by combining historical project experience data significantly enhances the accuracy and foresight of risk prediction, and the progress prediction results and risk warning list provide managers with opportunities for proactive planning. Finally, by using augmented reality technology to intuitively overlay and display progress and deviation information on the real-world scene, combined with a personalized information filtering mechanism based on user permissions and focus, complex progress data becomes intuitive and easy to understand. Decision-makers can directly obtain the necessary information and respond on-site, greatly improving decision-making efficiency and accuracy.

[0033] In one specific embodiment, the process of performing step S101 may specifically include the following steps:

[0034] Image sequences, video streams, and 3D point cloud data of the construction site were collected using high-definition cameras, drone aerial photography systems, and laser scanners.

[0035] The acquired image sequences are subjected to distortion correction and enhancement processing to remove image noise and redundant information, and a standardized image set is generated.

[0036] Spatial registration and noise reduction are performed on the collected point cloud data to remove outliers and duplicates, forming structured point cloud data.

[0037] Dynamic task scheduling of acquisition devices is achieved through edge computing nodes, and the data acquisition frequency and range are automatically adjusted according to the on-site environmental conditions.

[0038] The processed data is scored based on multi-dimensional quality indicators, and valid datasets that meet the threshold requirements are selected.

[0039] The selected image sets, point cloud data, and sensor information are aligned and fused according to timestamps and spatial locations to establish a unified indexing mechanism and generate a spatiotemporally aligned original construction dataset.

[0040] Specifically, construction site data is collected using multi-source devices. High-definition cameras are fixedly installed at key locations on the construction site to collect image sequence data, typically at a frame rate of 10-30 seconds, forming a continuous record of the construction process. Drone aerial photography systems fly along preset routes, acquiring overhead video streams of the construction site from different angles and altitudes, usually conducting patrols 1-2 times per day. Laser scanners emit laser beams and receive reflected signals to measure the three-dimensional coordinates of spatial points, generating point cloud data; the scanning density is typically set to several thousand points per square meter. These three types of equipment work together to ensure comprehensive data coverage of the construction site. The collected image sequences require distortion correction and enhancement processing. First, lens distortion correction algorithms are applied to the images to eliminate barrel or pincushion distortion caused by wide-angle lenses, restoring the geometric accuracy of the image. Then, image enhancement is performed, including contrast adjustment, brightness equalization, and sharpening, to improve image clarity. For noise processing, median filtering or Gaussian filtering algorithms are used to remove random noise and interference from the images. Image segmentation technology is used to identify and remove redundant background areas, retaining only the effective areas containing construction activities. The processed images are organized according to their capture time and spatial location to form a standardized image set. Point cloud data processing includes two key steps: spatial registration and noise reduction. Spatial registration integrates multiple sets of point cloud data acquired from different locations into a unified coordinate system. The Iterative Closest Point (ICP) algorithm is used to find the optimal rigid body transformation relationship between two sets of point clouds, calculate the rotation matrix and translation vector, and stitch the point clouds from multiple scans into a complete 3D scene. Noise reduction first identifies outliers through statistical outlier analysis, calculates the average distance from each point to its neighbors, and removes points that significantly deviate from the average. Then, duplicate points are filtered, and when the distance between two points is less than a preset threshold, they are merged into one point to reduce data redundancy. The point cloud data is organized using spatial index structures such as octrees or KD trees to form structured point cloud data with hierarchical query capabilities. Edge computing nodes play a crucial scheduling role in the data acquisition process. These nodes are implemented through embedded computing devices and deployed near the field data acquisition equipment to monitor changes in environmental conditions in real time and adjust the acquisition strategy accordingly. Specifically, edge nodes analyze lighting conditions, weather conditions, and the intensity of construction activities, dynamically adjusting camera exposure parameters, acquisition frequency, and image resolution accordingly. When insufficient light is detected, the camera acquisition frequency is automatically reduced and the exposure time is increased; when areas with intensive construction activity are identified, the data acquisition frequency and resolution for that area are increased; during construction pauses, the acquisition frequency is reduced to save storage space. Simultaneously, edge nodes adjust data compression rates and upload priorities based on data transmission bandwidth conditions, ensuring that important data is transmitted first.

[0041] Multi-dimensional quality assessment is a crucial step in ensuring data usability. A comprehensive scoring system is established for processed data, encompassing clarity, completeness, coverage, accuracy, and timeliness. Image data clarity is calculated using edge sharpness and contrast; point cloud data completeness is assessed through spatial distribution density to check for large data gaps; coverage is calculated by comparing the spatial extent of the collected data with the target monitoring area; accuracy is evaluated based on the deviation from baseline points; and timeliness is determined by the interval between the data acquisition time and the current time. A comprehensive quality score is calculated for each dimension according to its weight, and data with scores above a preset threshold are selected as valid datasets, while low-quality data is discarded. Data from different sources are time-aligned according to timestamps to establish a time series relationship. Then, spatial alignment is performed using three-dimensional spatial coordinates to establish a unified spatial reference system for image, point cloud, and sensor data. Images have their corresponding world coordinates calculated using the camera's intrinsic and extrinsic parameters; point cloud data itself contains spatial coordinate information; and sensor data has its spatial location calculated using the device's installation location and measurement parameters. A unified spatiotemporal indexing mechanism is established, using time and spatial coordinates as multidimensional index keys to support rapid data retrieval by time period, spatial region, or comprehensive conditions. These steps generate a spatiotemporally aligned original construction dataset, laying the foundation for subsequent analysis and processing.

[0042] Taking a high-rise building construction project as an example, during the concrete pouring of the third-floor slab, four high-definition cameras were deployed at fixed locations around the site. Drones conducted aerial photography twice daily, at 10:00 AM and 3:00 PM, and a laser scanner performed a comprehensive scan after the end of each workday. Of the 4,500 raw images captured by the cameras, 3,800 were selected as valid after distortion correction and enhancement. The laser scanner acquired approximately 800 million raw point cloud data points, which, after noise reduction and deduplication, retained 240 million valid points. Edge computing nodes detected overexposure caused by direct sunlight between 2:00 PM and 4:00 PM, automatically lowering camera exposure parameters and increasing the acquisition frequency. During rainy days, the system automatically increased the point cloud acquisition density to address reduced visibility. In quality assessment, the data quality in the northeast corner area was lower due to obstruction from construction cranes, so the system supplemented data collection in that area. The final spatiotemporally aligned raw construction dataset contained complete coverage of the entire construction area and established an index structure supporting multi-dimensional queries by floor, region, component type, and time period.

[0043] In one specific embodiment, the process of performing step S102 may specifically include the following steps:

[0044] Feature points and feature descriptors of image sequences are extracted from the original construction dataset, matching relationships between images are established, and sparse reconstructed point clouds are generated.

[0045] The sparse reconstructed point cloud is densified to fill in the missing spatial regions and construct continuous surface geometric information;

[0046] Extract key geometric features such as planes, line segments, and corners from point cloud data to form a structural feature set of the construction scene;

[0047] The structural feature set is paired and compared with the geometric features of the building information design model, the transformation matrix is ​​calculated, and the spatial position is corrected for the measured data.

[0048] Data collected at different times is sorted and organized using timestamps to establish a temporal relationship network of construction progress;

[0049] By integrating the corrected geometric data with the temporal relationship network, and adding time attribute tags to each spatial point, a digital model of the construction site containing geometric and temporal information is generated.

[0050] Specifically, feature points and feature descriptors are extracted from image sequences. Feature points are points in an image with unique properties, such as corner points and edge points, which can maintain stable recognition under different viewpoints. The extraction process uses SIFT (Scale Invariant Feature Transform) or SURF (Accelerated Robust Feature Transform) algorithms. The SIFT algorithm constructs a Gaussian difference pyramid to detect local extrema in space at different scales, calculates their principal directions, and generates descriptors. Descriptors are vectors that characterize the gray-level distribution of pixels around feature points, typically composed of 128-dimensional or 64-dimensional values. The Euclidean distance between feature points is calculated for each pair of images, and the matching point pair with the smallest distance is found. Then, the RANSAC (Random Sample Consensus) algorithm is used to eliminate incorrect matches. Based on the spatial positional relationship of the matching point pairs, the relative pose between cameras is calculated, and the three-dimensional coordinates of the feature points are recovered using triangulation, forming a sparse reconstructed point cloud. Although the number of points in this point cloud is limited, it accurately represents the scene structure. The sparse reconstructed point cloud needs to be densified to fill in the missing spatial regions. This process relies on PMVS (Patch Multiview Stereo) or MVS (Multiview Stereo Vision) techniques. First, dense sampling points are expanded around the sparse point cloud. For each sampling point, a normal vector and a local plane hypothesis are constructed. Then, this hypothesis is validated across multiple images, retaining points with high consistency. For missing regions, a Poisson surface reconstruction algorithm is used to convert the point cloud into an implicit function representation. The Poisson equation is solved to fill the gaps, and isosurfaces are extracted to generate a continuous mesh surface. In severely occluded areas, planar constraints and symmetry assumptions are used for completion; for example, if a wall plane is identified, the missing portion is filled by extending along that plane. Through these steps, the sparse point cloud is transformed into a dense model with continuous surface geometry information.

[0051] Extracting geometric features from point cloud data is crucial for constructing a structural feature set. First, plane extraction is performed. The RANSAC algorithm is used to randomly select three points in the point cloud to determine the plane equation. The distances from other points to this plane are calculated, and the number of points meeting a threshold is counted. This process is iterated multiple times to find the optimal plane parameters. For line segment extraction, Hough transform is applied to detect straight line features at the intersections of identified planes, or edge lines are directly extracted where there are significant changes in point cloud density gradient. Corner points are identified at the intersections of line segments, and the coordinates and angles of the intersections are calculated. The extracted planes, line segments, and corner points are organized into a topological relationship graph, recording the connections between elements, such as which line segments constitute a plane, which corner points connect to which line segments, etc., thus forming a structural feature set representing the main structure of the construction scene. The structural feature set is paired and compared with the Building Information Model (BIM) to align the measured data with the design model. First, the BIM model is converted into a geometric representation, and the plane, line segment, and corner point features are extracted to establish a feature library corresponding to the measured structural feature set. Then, a feature matching algorithm is used to calculate the similarity matrix between structural features and find the best matching pair. A set of control points is determined based on matching pairs. The rigid body transformation matrix, including the rotation matrix R and translation vector T, is then solved. This transformation converts the measured data coordinate system to the BIM model coordinate system. The transformation matrix is ​​calculated using either SVD (Singular Value Decomposition) or ICP (Iterative Closest Point) algorithms, obtaining the optimal solution by minimizing the sum of Euclidean distances between control point pairs. This transformation matrix is ​​then applied to transform the coordinates of all measured data, completing the spatial position correction.

[0052] The time dimension processing involves sorting and organizing data using timestamps. Each data point carries the collection time information, and the data is arranged chronologically to form a time series. For sparse data periods, time interpolation techniques are used to fill gaps and maintain data continuity. Based on the timestamp intervals, the data is divided into time windows of different granularities, such as days and weeks. Changes, such as the addition of new components or changes in location, are calculated within each window. By comparing the data differences between adjacent time windows, construction activities and progress changes are identified, and a temporal relationship network characterizing construction progress is constructed. This network records the appearance time, installation sequence, and interdependencies of different components. The corrected geometric data is integrated with the temporal relationship network, adding time attributes to each point or component in space. Specifically, time tags are added to each geometric element, including the first appearance time, change history, and status update records. The resulting digital model of the construction site not only contains accurate three-dimensional spatial information but also records the temporal evolution of the construction process, becoming a four-dimensional model that supports querying and analysis along both time and spatial dimensions. This digital model can intuitively display the construction status at any given time, track the installation progress of specific components, and analyze whether the construction sequence conforms to the plan.

[0053] Taking an office building construction project as an example, 15,000 SIFT feature points were extracted from the original construction dataset. These were then matched and triangulated to generate a sparse point cloud of approximately 8,000 spatial points. The PMVS algorithm was used to expand this to a dense point cloud of 2 million points, identifying key structural features, including 120 planes (walls, floors, etc.), 350 line segments (edges, beams, etc.), and 180 corner points (corners, column intersections, etc.). These structural features were matched against the BIM model, and the calculated transformation matrix corrected the measured data to the design coordinate system, with an average alignment error controlled within 3 centimeters. The data was organized according to timestamps, constructing a time-series network for the four-month construction period, recording the construction sequence from the foundation structure to the enclosure system and then to the interior decoration. The final digital model of the construction site contained approximately 500 main components, each with geometric information and a time stamp. This model accurately reflects the construction progress and actual condition.

[0054] In one specific embodiment, the process of executing step S103 may specifically include the following steps:

[0055] Feature maps of component areas are extracted from the digital model of the construction site, and the outlines of the components are delineated through edge detection to form candidate regions for component segmentation.

[0056] The feature vectors of the candidate regions are calculated, and the component types are determined according to the preset component classification criteria to generate a set of components with type labels.

[0057] Based on a set of components with type tags, calculate the three-dimensional coordinates, orientation, and dimensional parameters of each component to create a geometric description document for the component;

[0058] Perform spatial association analysis on the components in the geometric description document to identify the connection relationships and relative positional constraints between components and construct a component topology network diagram;

[0059] Extract component installation standards and quality requirements from the design specification library, compare them with the geometric parameters of the actual components, and calculate the installation deviation value and completion status indicator;

[0060] The component type markings, geometric description documents, component topology network diagrams, installation deviation values, and completion status identifiers are integrated into structured data records and organized into a construction completion status table with component identifiers and attributes.

[0061] Specifically, extracting feature maps of component regions from the digital model of the construction site is a multi-step, meticulous process. First, edge detection technology is used to process the image data in the digital model. An improved Canny algorithm is employed for edge detection. This algorithm first applies Gaussian filtering to reduce noise, then calculates the image gradient magnitude and direction, followed by non-maximum suppression to refine the edges, and finally uses a double thresholding method to determine the edge pixels. In this way, structural features such as column-beam joints and wall panel joints are outlined, forming preliminary component contours. For point cloud data, region growing and the RANSAC algorithm are used for planar segmentation, clustering the point cloud into different planar regions to further refine the component boundaries. This processed contour information is integrated into candidate regions for component segmentation, with each candidate region representing a possible building component.

[0062] Calculating the feature vectors of candidate regions is a core step in component classification. Feature vector calculation employs a deep convolutional neural network to extract high-dimensional features, combined with geometric feature analysis. It can be represented as:

[0063]

[0064] in, The geometric feature vector representing the component includes parameters such as aspect ratio, area-to-volume ratio, and curvature distribution; Represents the texture density feature vector, describing the detailed features of the component surface; The material feature vector is represented by parameters such as reflectance and color distribution. The contextual feature vector describes the spatial relationship between the candidate region and its surrounding environment. The calculated feature vector is matched against a pre-defined component classification standard library for similarity, and classification is determined using a Support Vector Machine (SVM) or Random Forest algorithm. The component classification can be expressed as:

[0065]

[0066] Among them, C t T represents the classification result of the components. k This represents the k-th component type (such as column, beam, wall, slab, etc.), where K represents the preset set of component types. Representing the eigenvector Belongs to type T k The probability of this. In this way, the candidate region is divided into different types such as columns, beams, walls, slabs, doors and windows, and pipes, forming a set of components with type labels.

[0067] In the component 3D parameter calculation stage, precise spatial information is extracted for the identified components. The 3D coordinate point set of the component can be represented as:

[0068] P c =(x v ,y v ,z v )|v∈[1,N p ]

[0069] Among them, P c Describes the set of points of a component, (x v ,y v ,z v ) represents the spatial coordinates of the v-th feature point on the component, N p This represents the total number of feature points. The component orientation is calculated using Principal Component Analysis (PCA), extracting three principal direction vectors from the component's point cloud:

[0070]

[0071] in, The orientation matrix of the component. These represent the unit vectors of the component along the three principal axes. The component's dimensional parameters are calculated using the minimum bounding rectangle method.

[0072] S c =[L x ,L y ,L z ]

[0073] Among them, S c L represents the dimension vector of the component. x L y L z These represent the lengths of the component along the three principal axes. These calculated geometric parameters form the geometric description document of the component, establishing a detailed geometric model for each component.

[0074] The spatial association analysis phase identifies topological relationships among components in the geometric description document. By calculating the shortest distances, overlapping areas, and connection points between components, a spatial relationship matrix is ​​established. Spatial relationship types include support relationships, connection relationships, containment relationships, and adjacency relationships. For example, by determining the spatial positional relationship between columns and beams, a support connection can be identified; by analyzing the relative positions of walls and doors / windows, a containment relationship can be determined. This relationship information is integrated into a component topology network diagram, forming a logical model of the overall building structure.

[0075] The installation deviation calculation process extracts the standard parameters of the corresponding components from the design specification library and compares them with the actual measured geometric parameters. Deviation calculation involves three aspects: positional deviation, angular deviation, and dimensional deviation. Positional deviation is obtained by calculating the Euclidean distance between the actual component's center point and the designed position; angular deviation is calculated by the angle between the actual orientation vector and the designed orientation vector; and dimensional deviation is calculated by the difference between the actual size and the designed size. Based on the comparison of the calculated deviation values ​​with the preset allowable error range, the installation status of the component is determined and marked as different levels such as "qualified," "minor deviation," or "serious deviation," generating a completion status indicator.

[0076] Finally, the component type labels, geometric description documents, component topology network diagrams, installation deviation values, and completion status identifiers are integrated according to a unified data structure format to form a complete component data record. These records are organized using various indexing methods such as component ID, spatial location, and component type, forming a construction completion status table with component identifiers and attributes. This status table records in detail the identification results, location information, connection relationships, and installation quality assessment of each component on the construction site, providing an accurate data foundation for subsequent construction progress assessment.

[0077] Taking a concrete frame structure building as an example, this method first extracts the edge features of components such as walls, columns, beams, and slabs from the digital model, forming hundreds of candidate regions. For each candidate region, a feature vector is calculated. For example, the geometric features of a certain beam component show an aspect ratio of 15:1, a surface texture exhibiting the roughness pattern unique to concrete, and a uniformly distributed gray tone in its material characteristics. By comparing these feature vectors with preset beam component standards, the system classifies it as a "concrete beam." Further analysis of the beam's spatial parameters is performed, calculating its centerline coordinates, principal axis direction vector, and three-dimensional dimensions (length 6000mm, width 300mm, height 500mm), generating a geometric description document. Through spatial correlation analysis, it is identified that the beam has a supporting connection with the end column components and a load-bearing relationship with the upper slab components, forming a local structural network. Comparing the beam's standard parameters with those in the design specifications, the calculated positional deviation is 22mm, the tilt angle is 1.2°, and the width deviation is -5mm, all within the allowable range. Therefore, its completion status is marked as "qualified." Ultimately, this information is integrated into a structured record and added to the construction completion status table to provide accurate data for subsequent progress assessment and decision support.

[0078] In one specific embodiment, the process of executing step S104 may specifically include the following steps:

[0079] Import construction plan data from the project management system, extract the task breakdown structure and key milestones, and generate a construction task dependency network.

[0080] By matching component codes, a correspondence is established between component records in the construction completion status table and planned items in the construction task dependency network, forming a component-task mapping table.

[0081] Based on the component-task mapping table, the completion status of the components involved in each task item is statistically analyzed, and the task completion rate and progress weight value are calculated.

[0082] The difference between the component geometric parameters and design parameters in the construction completion status table is calculated. The component installation accuracy is quantified by the bounding box occupancy rate and point-to-surface distance evaluation method, and a component position deviation matrix is ​​generated.

[0083] Based on the task completion rate and schedule weight value, combined with the project critical path, a comprehensive analysis of the overall progress status and the achievement of key milestones is conducted to prepare a construction progress assessment report.

[0084] The component position deviation matrix is ​​hierarchically summarized according to component type, region, and deviation degree to generate component-level deviation data that includes deviation distribution map and hot spot areas.

[0085] Specifically, importing construction plan data from the project management system is a key step in a big data-based construction progress monitoring method. The process begins by connecting to the project management system via a data interface and extracting raw plan data containing information such as task name, planned start time, planned completion time, prerequisite tasks, and resource allocation, using API calls or direct database queries. After extraction, the raw data undergoes structured processing. A hierarchical decomposition algorithm is used to identify the Work Breakdown Structure (WBS), decomposing the overall project tasks into task units at different levels, such as subsystems and sub-items. Simultaneously, key milestones are identified through time node analysis. These milestones are crucial for project progress control, such as foundation completion, main structure topping out, and equipment installation completion. Building upon this, a topology sorting algorithm is applied to analyze the dependencies between tasks, constructing a directed acyclic graph (DAG)-like construction task dependency network. This network clearly displays the sequence and parallel relationships between tasks, providing a logical basis for progress assessment.

[0086] Linking the construction completion status table with the construction task dependency network through component coding matching is a crucial step in mapping physical entities to planned tasks. First, the coding rules for each component in the construction completion status table are analyzed. These codes typically include information such as area identifier, component type, and floor location. Simultaneously, the composition rules of task codes in the construction plan are analyzed to find the correspondence between the two coding systems. Then, a matching algorithm is designed. This algorithm establishes coding mapping rules by extracting common elements from the codes (such as area code, floor number, component type code, etc.). For cases where coding rules are inconsistent, fuzzy matching and semantic analysis methods are used to infer possible correspondences by calculating the similarity between task descriptions and component attributes. In this way, a one-to-one or one-to-many mapping relationship is established between each component record in the construction completion status table and the corresponding planned item in the construction task dependency network, forming a component-task mapping table. This table contains fields such as component ID, task ID, and corresponding weight, becoming the basic data structure for subsequent schedule calculations.

[0087] Statistical analysis of task completion based on the component-task mapping table is a core step in evaluating actual progress. For each task, firstly, all component records associated with the task are selected from the component-task mapping table, and then the completion status identifiers of these components are extracted from the construction completion status table. Different completion weights are assigned to components according to their completion status levels (e.g., not started, in progress, completed), and then a weighted average completion rate is calculated by combining the component's importance weight within the task. Specifically, the completion rate of each component is multiplied by its weight and then divided by the total weight value to obtain the overall task completion rate. Simultaneously, the task's progress weight value is determined based on factors such as the task's proportion of workload in the entire project and its criticality; this weight value reflects the task's impact on the overall project progress. In this way, component-level completion status is quantified and converted into task-level progress indicators, establishing a quantitative relationship between physical completion and planned completion.

[0088] Calculating the difference between the component geometric parameters in the construction completion status table and the design parameters is a crucial step in evaluating construction quality. First, the actual geometric parameters of each component are extracted from the construction completion status table, including spatial coordinates, dimensions, and orientation. Simultaneously, the standard parameters for the corresponding components are extracted from the design model. Then, the differences between the two sets of parameters are calculated, including positional deviation (the Euclidean distance between actual and design coordinates), dimensional deviation (the difference between actual and design dimensions), and angular deviation (the angle between actual and design orientations). Based on this, a bounding box occupancy rate evaluation method is used to calculate the degree of overlap between the actual component's bounding box and the design bounding box; a higher overlap rate indicates more accurate positioning. Simultaneously, a point-to-surface distance evaluation method is used to calculate the average distance from the actual component's point cloud to the surface of the design model; a smaller distance indicates a higher shape matching degree. Through the combined application of these two evaluation methods, the installation accuracy of the components is quantified, and the results are organized into a component positional deviation matrix. This matrix, indexed by component ID, contains various deviation values ​​and accuracy scores.

[0089] A comprehensive analysis based on task completion rates and schedule weights forms the foundation for a construction progress assessment report. First, the completion rate of each task is multiplied by its schedule weight to obtain a weighted completion rate. Then, the weighted completion rates of all tasks are summed to calculate the overall project progress completion percentage. Simultaneously, the Critical Path Method (CPM) is applied to analyze the task dependency network, identifying the task chain with the greatest impact on the overall project duration—the critical path. The completion status of tasks on the critical path is closely monitored, and the schedule deviation value of the critical path is calculated. Furthermore, for critical milestones, the difference between actual completion time and planned time is compared to form a milestone achievement assessment. Based on this, a multi-level analysis method is used to generate a comprehensive assessment result, combining the overall progress completion rate, critical path status, and milestone achievement status, resulting in a construction progress assessment report that includes schedule status, deviation analysis, risk warnings, and trend predictions.

[0090] Hierarchical aggregation of component position deviation matrices is a crucial step in generating component-level deviation data. First, the deviation matrix is ​​grouped and statistically analyzed according to component type (e.g., columns, beams, walls, slabs, etc.), calculating the average deviation value and standard deviation for each type of component to analyze the differences in installation accuracy among different component types. Then, it is grouped according to spatial regions (e.g., floors, functional zones, etc.) to analyze the distribution of construction quality in different areas. Simultaneously, thresholds are set based on the degree of deviation for grading; for example, position deviations are divided into normal range (<10mm), slight deviation (10-20mm), and severe deviation (>20mm). This multi-dimensional hierarchical aggregation generates a visual report including statistical charts and deviation distribution heatmaps, intuitively displaying the spatial distribution and concentrated areas of construction deviations, helping managers quickly identify hotspots of construction quality problems.

[0091] Taking a high-rise building project as an example, this method first imports construction plan data containing 523 task items from the project management system. A hierarchical decomposition algorithm identifies three major systems: main structure, electromechanical installation, and decoration, as well as 15 sub-projects including foundation, main structure construction, and external enclosure. Eight key milestone nodes, such as foundation completion and main structure topping-out, are extracted to construct a complete construction task dependency network. Then, an encoding matching algorithm establishes a correspondence between 3850 component records in the construction completion status table and 523 planned tasks, forming a component-task mapping table. For example, a mapping relationship is established between the component coded "F3-C-042" (representing column 42 in section C on the 3rd floor) and the task "T-157" (construction of the frame column on the 3rd floor). Based on the mapping table, statistics show that task "T-157" is associated with 24 frame column components, of which 22 are completed and 2 are in progress, resulting in a task completion rate of 91.7% calculated by weight. Meanwhile, through geometric parameter analysis, the average positional deviation of these column components was calculated to be 8.3 mm, and the average angular deviation was 0.7°, both within the allowable range. A comprehensive analysis of the completion status of all tasks, combined with the critical path and milestone achievement status, resulted in a detailed progress assessment report, showing an overall project progress completion rate of 78.2% and a critical path delay of 3 days. Furthermore, by hierarchically summarizing the component positional deviations, a deviation distribution heatmap was generated, revealing a significant concentration of installation deviations in the northeast corner area of ​​the 5th floor, providing precise quality control data for construction management.

[0092] In one specific embodiment, the process of executing step S105 may specifically include the following steps:

[0093] Extract historical progress data points from the construction progress assessment report, construct a time series dataset, and plot a progress change curve.

[0094] Trend decomposition was performed on the progress change curve to separate seasonal factors, long-term trends and random fluctuations, thereby obtaining the characteristics of progress fluctuations.

[0095] By combining component-level deviation data, a deviation-schedule correlation matrix is ​​established to identify the correlation between deviation accumulation areas and schedule delays, thus generating a construction bottleneck analysis report.

[0096] By comparing construction bottleneck analysis reports with historical project database data, risk evolution paths for similar scenarios are extracted, and a risk propagation network is established.

[0097] Based on the characteristics of schedule fluctuations and the risk propagation network, the probability distribution of schedule completion within the future time window is calculated, and a schedule prediction result containing multiple confidence intervals is generated.

[0098] Risk factors and their impact weights are extracted from the risk transmission network, and then ranked and sorted according to their urgency and scope of impact to form a risk warning list.

[0099] Specifically, extracting historical progress data points from the construction progress assessment report is the first step in progress forecasting. This process uses data parsing technology to extract the progress completion rate value for each assessment point in the report, converting it into a binary structure containing a timestamp and a progress value. Data extraction employs a structured text parsing algorithm to locate and extract the date-progress correspondence in the report, forming preliminary time series data. Subsequently, the extracted data undergoes cleaning and normalization, including removing outliers, adding missing values, and standardizing the time format to ensure the integrity and consistency of the dataset. The cleaned data is arranged in chronological order, forming a complete time series dataset that reflects the progress trajectory of the project from its inception to the present. Based on this dataset, time series visualization technology is used to plot progress change curves, visually displaying the time evolution characteristics of the project progress.

[0100] Trend decomposition of the schedule change curve is a crucial step in understanding the patterns of schedule fluctuations. Trend decomposition employs a time series decomposition algorithm to break down the raw schedule data into three basic components: long-term trend, seasonal factors, and random fluctuations. The long-term trend reflects the main direction of project progress and is extracted using methods such as moving averages or multinomial fitting. Seasonal factors reflect cyclical patterns, such as regular fluctuations caused by alternating workdays / restdays and differences in resource allocation at the beginning and end of the month, and are extracted using Fourier analysis or seasonal index decomposition. Random fluctuations represent irregular changes, including disturbances caused by various temporary factors. This decomposition allows for a deeper understanding of the inherent patterns of schedule fluctuations, providing a foundation for predicting future schedules.

[0101] Establishing a deviation-schedule correlation matrix by combining component-level deviation data is a crucial method for identifying the relationship between quality issues and schedule delays. This process first aggregates component-level deviation data by spatial region and time period, calculating the average deviation value and deviation density for each region over different time periods. Simultaneously, schedule delay information for each region is extracted from the schedule data, including the difference between planned and actual deviations and the rate of change. Then, correlation analysis is used to calculate the correlation coefficient between deviation indicators and schedule delay indicators, forming a two-dimensional correlation matrix. Each element in the matrix represents the strength of the correlation between a specific type of deviation and a specific schedule problem. Through matrix data mining, highly correlated regions are identified—that is, regions and processes where deviation accumulation is highly correlated with schedule delays—resulting in a construction bottleneck analysis report that details the bottleneck location, severity, and causes.

[0102] Comparing construction bottleneck analysis reports with historical project database data is a crucial step in fully utilizing experiential knowledge. This process first parametrically describes the characteristics of the current project's construction bottleneck, including attributes such as bottleneck type, location features, and stage of occurrence. Then, it accesses the historical project database, using similarity calculations to filter historical cases with similar bottleneck characteristics, and employs a case retrieval algorithm to quickly locate relevant records. In-depth analysis is performed on the retrieved historical cases to extract their risk evolution paths, including precursory signals, diffusion processes, impact range, and final outcomes. Based on the common characteristics of multiple similar cases, a risk propagation network model is constructed. This network uses nodes to represent risk states, edges to represent risk transfer paths, and weights to represent transfer probabilities, comprehensively describing the evolutionary patterns of risk in time and space.

[0103] Calculating the probability distribution of schedule completion based on schedule fluctuation characteristics and risk propagation networks is the core of achieving scientific prediction. This calculation process integrates time series forecasting and Monte Carlo simulation methods. First, based on the fluctuation characteristics of historical schedule data, a baseline prediction model for schedule changes is established to predict the expected schedule value at future time points. Then, by combining the risk factors and their probability distributions in the risk propagation network, a large number of possible risk scenarios are generated, and the impact on schedule under each scenario is evaluated. Through multiple simulations, the probability distribution function of schedule completion at different future time points is obtained:

[0104]

[0105] Among them, P(C t ≤α) represents the probability that the degree of completion at time point t does not exceed α. This represents the probability density function of the progress completion rate at time point t. Based on the probability distribution function, progress prediction results with different confidence intervals are generated, such as 50%, 75%, and 90% confidence intervals, providing decision-making references for project management under different risk preferences.

[0106] Extracting risk factors and their impact weights from the risk propagation network is a crucial step in risk management. This process begins with topological analysis of the risk propagation network, calculating centrality indicators for each node, including degree centrality, betweenness centrality, and eigenvector centrality, to identify key risk nodes. Then, sensitivity analysis assesses the impact of each risk factor on schedule delays and calculates its weight. Based on the urgency (proximity of occurrence) and scope of impact (affected workload), risk factors are categorized in a two-dimensional hierarchy, forming a risk matrix. Finally, risk factors are ranked according to their comprehensive risk scores, generating a structured risk alert list containing detailed information such as risk descriptions, potential impacts, warning signals, and response recommendations.

[0107] Taking a large commercial complex project as an example, this method first extracts progress data points from 90 consecutive days of progress assessment reports to construct a complete time-series dataset. Trend decomposition reveals a significant weekend slowdown (seasonal factor) and a long-term trend of overall progress slowdown. Combined with component-level deviation data analysis, a high density of pipe installation deviations is found in the commercial podium section, which is also the area with the most severe progress delays. A correlation coefficient of 0.87 confirms a high correlation between the two. A search of historical project database reveals that similar bottleneck problems in the commercial podium pipe construction area have occurred in five similar projects in the past. The typical evolution path starts with pipe positioning deviations, leading to pipe intersections and collisions, which in turn triggers rework, ultimately causing an overall delay in electromechanical installation. Based on this data, Monte Carlo simulation is used to calculate the probability distribution of the project's target progress within the next 30 days. The results show a 90% confidence level of achieving 85% of the planned progress, but only a 50% confidence level of achieving 92% of the planned progress. Key risk factors extracted from the risk network include "insufficient installation accuracy of pipeline supports" and "frequent design changes." After being sorted by urgency and scope of impact, a targeted risk warning list was formed.

[0108] In one specific embodiment, the process of executing step S106 may specifically include the following steps:

[0109] Read the user role and permission configuration table, determine the information access level based on the user's identity, filter sensitive data in the progress prediction results and risk warning list, and obtain a targeted information set;

[0110] Based on a targeted information set and combined with user-defined key focus parameters, the importance of progress and risk data is scored, and key information elements are selected.

[0111] Acquire real-time image streams from the site, perform spatial positioning and attitude correction on the images, and establish a transformation matrix between the image coordinate system and the construction coordinate system;

[0112] Based on the transformation matrix, key information elements are converted into layer data and then visualized and arranged according to transparency and color coding rules.

[0113] Layer data is overlaid onto a real-time image stream to build an augmented reality scene, displaying progress status and deviation markers at key component locations;

[0114] By integrating augmented reality scenarios with decision-making suggestion texts, and generating a multi-view interactive interface according to information presentation templates, a visualized construction management decision-making reference is formed.

[0115] Specifically, reading the user role and permission configuration table is the first step in achieving personalized information presentation. This process begins by retrieving user permission configuration data, including fields such as user ID, role type, department affiliation, scope of responsibilities, and permission level, through database queries or configuration file parsing. When a user logs into the system and provides their identity credentials, the authentication module verifies the user's identity, extracts the user's unique identifier, and then searches the permission configuration table for a match to obtain the user's role type (e.g., project manager, construction supervisor, quality supervisor, safety manager, etc.) and permission level (e.g., levels 1-5). Based on the obtained permission level, corresponding data filtering rules are executed to filter data items in the progress forecast results and risk warning list. Data filtering employs a tag-based access control mechanism, pre-labeling each piece of data with a sensitivity level and department. By comparing the matching relationship between user permissions and data tags, it determines which data items are visible to the current user. For example, financial risk data in the risk warning list is only displayed to users with permissions at the level of finance manager or higher; detailed probability distribution data in the progress forecast is only shown to project management. Through this refined data filtering process, a targeted information set conforming to the user's permission scope is generated.

[0116] Filtering key information based on a targeted information set is a crucial step in achieving information focus. This process first reads the user's focus parameters set through the interactive interface. These parameters include the area of ​​interest (e.g., specific floor, functional area), project type (e.g., structural, mechanical and electrical, decoration), time window (e.g., near-term, medium-term, long-term), and focus dimensions (e.g., schedule delays, quality deviations, safety risks). Next, the matching degree between these parameters and the attributes of data items in the targeted information set is calculated using a multi-dimensional similarity algorithm, comprehensively considering the degree of matching across spatial, temporal, and content dimensions. Then, the importance of each information item is scored. The scoring formula comprehensively considers four factors: urgency, scope of impact, degree of abnormality, and user attention, assigning weight coefficients to each factor and calculating a weighted total score. Finally, the information items are ranked according to the scoring results, and the top-ranked information is selected as key information elements, ensuring that the displayed content not only aligns with the user's focus but also possesses actual importance.

[0117] Acquiring and spatially locating real-time image streams from the construction site is a fundamental step in realizing augmented reality. This process begins by capturing a real-time video stream of the construction site using a mobile device (such as a tablet or smartphone) or a camera on dedicated augmented reality glasses. To ensure accurate image positioning, multi-source positioning technologies are employed, including GPS positioning (determining the device's geographic coordinates), inertial measurement units (IMUs) (recording the device's attitude angles), and computer vision positioning (determining relative positions by identifying feature points in the scene). Specifically, computer vision positioning utilizes Simultaneous Localization and Mapping (SLAM) technology. This technology identifies feature points in video frames, tracks the changes of these feature points across consecutive frames, and calculates the camera's movement trajectory and the 3D structure of the scene. After identifying a sufficient number of feature points, a feature matching algorithm is used to compare the current viewpoint with a pre-built 3D model of the construction site to determine the positional relationship between the current viewpoint and the standard coordinate system. Finally, a transformation matrix is ​​calculated from the image coordinate system (pixel coordinates) to the global construction coordinate system (actual physical coordinates). This matrix contains translation vectors and rotation matrices, fully describing the mapping relationship between the two coordinate systems.

[0118] Spatial mapping of information using a transformation matrix is ​​the core step in achieving accurate visualization. This process first categorizes key information elements according to their spatial attributes, including point information (such as component nodes and equipment locations), linear information (such as pipeline paths and structural edges), and area information (such as regional progress and wall quality). For each type of information, based on its location data in the construction coordinate system, the previously calculated transformation matrix maps it to the image coordinate system of the current viewpoint, determining its precise display position on the screen. Simultaneously, a field-of-view detection is performed, processing only information elements within the current field of view to reduce computational burden. Then, based on the type and importance of the information, preset visualization rules are applied to generate corresponding layer data. These rules include color coding (e.g., using red for risk, green for normal, and yellow for warning), transparency settings (high opacity for important information, semi-transparent for secondary information), icon selection (different icons for different types of problems), and text annotation (key values ​​and brief descriptions). Through these rules, abstract data is transformed into intuitive visual elements, forming multiple information layers.

[0119] Overlaying layer data onto a real-time image stream is a key step in constructing an augmented reality scene. This process employs image compositing techniques, layering information elements according to layer priority while maintaining the integrity of the original image. The overlay process considers perspective and occlusion relationships to ensure visual consistency between augmented reality elements and the actual scene. For component information requiring precise positioning, such as installation deviation markers, edge snapping technology is used to ensure the markers are accurately attached to the component edges. For regional information, such as progress completion, a semi-transparent heatmap is used to overlay the corresponding area, and the display effect is adjusted according to actual lighting conditions to ensure readability in different environments. To address occlusion issues in real-time images, a depth estimation algorithm is applied to calculate the relative depth of objects in the scene, ensuring that virtual information elements are correctly occluded by foreground objects, enhancing realism. Simultaneously, layer positions and perspective effects are updated in real-time to adapt to changes in the user's viewpoint, ensuring that augmented reality elements always maintain the correct spatial relationship with the actual scene, maintaining a stable display even as the user moves.

[0120] Integrating augmented reality scenarios with decision-making advice text is the final step in forming a complete decision-making reference. This process begins by selecting a suitable information presentation template from a pre-defined interface template library based on user role characteristics and preferences. The template defines the information layout, interaction methods, and navigation structure. Multiple view modes are designed to address different types of decision-making needs, including a global overview view (displaying overall progress and key indicators), a detailed analysis view (showing in-depth analysis of specific problems), a time-travel view (showing past and predicted future trends), and a comparative evaluation view (showing the differences between the plan and the actual situation side-by-side). Within each view, augmented reality visualizations are integrated with corresponding text explanations, data charts, and decision-making advice to form an information-rich and hierarchically structured interactive interface. To enhance user experience, intuitive interaction mechanisms are designed, such as touch gestures (zoom in, zoom out, rotate, tap), voice commands, and gaze tracking, allowing users to easily switch between different views and explore information of interest in depth. Ultimately, this results in an intuitive, easy-to-understand, and content-rich visualized construction management decision-making reference, providing the necessary decision support for managers at different levels.

[0121] Taking a large public building construction project as an example, when the project supervisor logs into the progress monitoring system via a mobile tablet, the system first reads their user permission configuration information, confirming that they have the role of a supervising engineer with a level 3 permission, responsible for structural and mechanical / electrical supervision. Based on this identity information, the system filters out 1200 data records related to structural and mechanical / electrical engineering from the progress prediction results and risk warning list containing more than 2500 data items. It also removes sensitive information requiring higher permissions, such as contract disputes and cost overruns, forming a targeted information set. The system reads the key focus parameters set by the supervising engineer, including "Area of ​​Focus: Central Hall," "Project Type: Mechanical / Electrical," "Time Window: Recent Two Weeks," and "Focus Dimension: Installation Quality." Based on these parameters, the system calculates the matching degree and importance score for the data items in the targeted information set, filtering out the top 20 key information elements, including specific issues such as "high risk of collision with air conditioning ducts in the central hall." When the supervising engineer carried the tablet to the construction area in the central hall, the system captured a real-time image stream through the device's camera. Simultaneously, using SLAM technology, it identified over 50 feature points in the environment and matched them with a pre-built BIM model, accurately calculating the transformation relationship between the image coordinate system and the construction coordinate system from the current viewpoint. Based on this transformation matrix, the system converted the identified key information elements (such as pipe collision points and support deviations) into layer data, highlighting them with red highlights and flashing icons according to their importance. This layer data was overlaid on the real-time image stream, displaying deviation and risk markers above the actual pipe installation locations, while providing detailed text explanations and handling suggestions at the screen edges. Through this intuitive visualization, the supervising engineer immediately discovered installation deviations at two pipe intersections and, based on the system's decision-making suggestions, guided the construction team to make adjustments on-site, effectively avoiding potential subsequent collisions and rework.

[0122] The above describes the construction progress monitoring method based on big data in the embodiments of this application. The following describes the construction progress monitoring system based on big data in the embodiments of this application. Please refer to [link / reference]. Figure 2 One embodiment of the construction progress monitoring system based on big data in this application includes:

[0123] The evaluation module is used to acquire images, videos and point clouds of the construction site through multi-source devices, filter and evaluate the quality of the acquired data, and obtain a spatiotemporally aligned original construction dataset.

[0124] The registration module is used to extract and register features from image sequences and point cloud data based on the original construction dataset, spatially align the measured data with the design model, and obtain a digital model of the construction site containing geometric and temporal information.

[0125] The classification module is used to perform target detection and classification of on-site components based on the construction site digital model, analyze the spatial positional relationship between components, and obtain a construction completion status table with component identification and attributes.

[0126] The mapping module is used to establish a mapping relationship between actual components and construction plan tasks based on the construction completion status table, calculate the component completion degree and position deviation value, and obtain a construction progress assessment report and component-level deviation data.

[0127] The analysis module is used to analyze the progress change trend based on the construction progress assessment report and component-level deviation data, identify abnormal patterns and potential risk points in the construction process, and obtain progress prediction results and risk warning list.

[0128] The filtering module is used to filter the progress prediction results and risk warning list according to user permissions and priorities, and overlay the progress and deviation information on the real-world site to obtain a visualized construction management decision-making reference.

[0129] Through the collaborative efforts of the aforementioned components, images, videos, and point cloud data from the construction site are acquired via multi-source devices, enabling comprehensive and continuous perception of the construction site. This overcomes the limitations of traditional single data sources and improves the integrity and diversity of the raw data. Filtering and quality assessment of the collected data ensures accuracy and reliability, laying a solid foundation for subsequent analysis. The spatiotemporally aligned raw construction dataset solves the problem of fusing multi-source heterogeneous data, achieving unified representation and processing of different data types. Feature extraction and registration techniques based on the raw dataset enable precise correspondence between measured data and the design model, forming a digital model of the construction site containing geometric and temporal information, facilitating comparison between the actual and planned data. Compared to traditional BIM models, this digital model not only includes static design information but also incorporates measured data from the dynamic construction process, truly reflecting the actual state and changes of the construction site. Deep learning algorithms are used for target detection and classification of on-site components, automatically identifying component types and attributes, significantly improving the accuracy and efficiency of component identification. Particularly for components in complex environments, the feature extraction capabilities of deep neural networks are significantly superior to traditional computer vision methods. Spatial relationship analysis algorithms clearly reveal the positional constraints and connections between components, forming a construction completion status table with component identifiers and attributes, providing accurate basic data for progress assessment. The method of establishing a mapping relationship between actual components and planned construction tasks realizes the transformation from physical completion status to planned completion ratio, solving the problems of strong subjectivity and poor consistency in traditional progress assessment. The method of quantifying component installation accuracy through bounding box occupancy rate and point-to-surface distance evaluation transforms construction quality assessment from qualitative description to quantitative indicators, providing objective basis for management decisions through progress assessment reports and component-level deviation data. The progress change trend analysis method based on time series analysis and machine learning can mine potential patterns from historical data, identifying abnormal patterns and risk points. The risk propagation network established by combining historical project experience data significantly enhances the accuracy and foresight of risk prediction, and the progress prediction results and risk warning list provide managers with opportunities for proactive planning. Finally, by using augmented reality technology to intuitively overlay and display progress and deviation information on the real-world scene, combined with a personalized information filtering mechanism based on user permissions and focus, complex progress data becomes intuitive and easy to understand. Decision-makers can directly obtain the necessary information and respond on-site, greatly improving decision-making efficiency and accuracy.

[0130] Reference Figure 3 This invention also provides a computer device, which can be a server, and its internal structure can be as follows: Figure 3As shown, the computer device includes a processor, memory, display screen, input device, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores the data corresponding to this embodiment. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements the above-described method.

[0131] Those skilled in the art will understand that Figure 3 The structures shown are merely block diagrams of some structures related to the present invention and do not constitute a limitation on the computer devices on which the present invention is applied.

[0132] An embodiment of the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method. It is understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0133] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A construction progress monitoring method based on big data, characterized in that, The big data-based construction progress monitoring method includes: By acquiring images, videos, and point clouds of the construction site through multi-source devices, filtering and quality assessment of the collected data, a spatiotemporally aligned original construction dataset is obtained. Based on the original construction dataset, feature extraction and registration are performed on image sequences and point cloud data. The measured data is spatially aligned with the design model to obtain a digital model of the construction site containing geometric and temporal information. This includes: extracting feature points and feature descriptors from the image sequences in the original construction dataset, establishing matching relationships between images, and generating a sparse reconstructed point cloud; densifying the sparse reconstructed point cloud to fill missing spatial regions and construct continuous surface geometric information; extracting planar, line segment, and corner geometric features from the point cloud data to form a structural feature set for the construction scene; pairing and comparing the structural feature set with the geometric features of the building information design model, calculating the transformation matrix, and correcting the spatial position of the measured data; sorting and organizing data collected at different times using timestamps to establish a temporal relationship network for construction progress; integrating the corrected geometric data with the temporal relationship network, adding a time attribute marker to each spatial point, and generating a digital model of the construction site containing both geometric and temporal information. Based on the digital model of the construction site, target detection and classification of on-site components are performed, the spatial positional relationship between components is analyzed, and a construction completion status table with component identification and attributes is obtained. Based on the construction completion status table, a mapping relationship is established between actual components and construction plan tasks. The component completion rate and positional deviation values ​​are calculated to obtain a construction progress assessment report and component-level deviation data. This includes: importing construction plan data from the project management system, extracting the task decomposition structure and key node milestones, and generating a construction task dependency network; establishing a correspondence between component records in the construction completion status table and planned items in the construction task dependency network through component coding matching to form a component task mapping table; based on the component task mapping table, statistically analyzing the component completion status involved in each task item, calculating the task completion rate and progress weight value; calculating the difference between the component geometric parameters and design parameters in the construction completion status table, quantifying the component installation accuracy through bounding box occupancy rate and point-to-surface distance evaluation methods, and generating a component positional deviation matrix; based on the task completion rate and progress weight value, combined with the project critical path, comprehensively analyzing the overall progress status and key node achievement status, and compiling a construction progress assessment report; and hierarchically summarizing the component positional deviation matrix according to component type, region, and deviation degree to generate component-level deviation data including deviation distribution maps and hotspot areas. Based on the construction progress assessment report and component-level deviation data, the progress change trend is analyzed, abnormal patterns and potential risk points in the construction process are identified, and progress prediction results and risk warning list are obtained. The progress prediction results and risk warning list are filtered according to user permissions and priorities, and the progress and deviation information is overlaid and displayed in the real-world scene to obtain a visualized construction management decision-making reference.

2. The construction progress monitoring method based on big data according to claim 1, characterized in that, The process involves acquiring construction site images, videos, and point clouds from multiple sources, filtering and evaluating the quality of the collected data to obtain a spatiotemporally aligned original construction dataset, including: Image sequences, video streams, and 3D point cloud data of the construction site were collected using high-definition cameras, drone aerial photography systems, and laser scanners. The acquired image sequences are subjected to distortion correction and enhancement processing to remove image noise and redundant information, and a standardized image set is generated. Spatial registration and noise reduction are performed on the collected point cloud data to remove outliers and duplicates, forming structured point cloud data. Dynamic task scheduling of acquisition devices is achieved through edge computing nodes, and the data acquisition frequency and range are automatically adjusted according to the on-site environmental conditions. The processed data is scored based on multi-dimensional quality indicators, and valid datasets that meet the threshold requirements are selected. The selected image sets, point cloud data, and sensor information are aligned and fused according to timestamps and spatial locations to establish a unified indexing mechanism and generate a spatiotemporally aligned original construction dataset.

3. The construction progress monitoring method based on big data according to claim 1, characterized in that, Based on the digital model of the construction site, target detection and classification of on-site components are performed, and the spatial relationship between components is analyzed to obtain a construction completion status table with component identifiers and attributes, including: Feature maps of component areas are extracted from the digital model of the construction site, and the outlines of the components are delineated by edge detection to form candidate regions for component segmentation. The candidate regions are used to calculate feature vectors, and the component types are determined according to preset component classification criteria to generate a set of components with type labels. Based on the set of components with type tags, calculate the three-dimensional coordinates, orientation, and size parameters of each component to create a geometric description document for the component; Spatial association analysis is performed on the components in the geometric description document to identify the connection relationships and relative positional constraints between components, and to construct a component topology network diagram; Extract component installation standards and quality requirements from the design specification library, compare them with the geometric parameters of the actual components, and calculate the installation deviation value and completion status indicator; The component type markings, geometric description documents, component topology network diagrams, installation deviation values, and completion status identifiers are integrated into structured data records and organized into a construction completion status table with component identifiers and attributes.

4. The construction progress monitoring method based on big data according to claim 1, characterized in that, The process involves analyzing the progress change trend based on the construction progress assessment report and component-level deviation data, identifying abnormal patterns and potential risk points during construction, and obtaining progress prediction results and a risk warning list, including: Historical progress data points are extracted from the construction progress assessment report to construct a time series dataset and plot a progress change curve. The progress change curve is then decomposed to separate seasonal factors, long-term trends, and random fluctuation components to obtain progress fluctuation characteristics. Based on the component-level deviation data, a deviation-to-schedule correlation matrix is ​​established to identify the correlation between the deviation accumulation area and the schedule delay, thereby generating a construction bottleneck analysis report. By comparing the construction bottleneck analysis report with historical project database data, risk evolution paths of similar scenarios are extracted, and a risk propagation network is established. Based on the aforementioned progress fluctuation characteristics and risk propagation network, the probability distribution of progress completion within future time windows is calculated, generating progress prediction results containing multiple confidence intervals. Risk factors and their impact weights are extracted from the risk propagation network, and then classified and ranked according to their urgency and scope of impact to form a risk warning list.

5. The construction progress monitoring method based on big data according to claim 1, characterized in that, The process involves filtering the progress forecast results and risk warning list according to user permissions and priorities, and overlaying the progress and deviation information onto the real-world site view to obtain a visualized construction management decision-making reference, including: Read the user role and permission configuration table, determine the information access level based on the user's identity, filter sensitive data in the progress prediction results and risk warning list, and obtain a targeted information set; Based on the targeted information set and combined with the user-defined focus parameters, the progress and risk data are scored for importance, and key information elements are selected. Acquire real-time image streams from the site, perform spatial positioning and attitude correction on the images, and establish a transformation matrix between the image coordinate system and the construction coordinate system; based on the transformation matrix, convert the key information elements into layer data, and arrange them for visualization according to transparency and color coding rules; The layer data is overlaid onto the real-time image stream to construct an augmented reality scene, displaying progress status and deviation markers at key component locations; By integrating the augmented reality scenario with the decision-making suggestion text, a multi-view interactive interface is generated according to the information presentation template, forming a visualized construction management decision-making reference.

6. A construction progress monitoring system based on big data, used to implement the construction progress monitoring method based on big data as described in any one of claims 1 to 5, characterized in that, The big data-based construction progress monitoring system includes: The evaluation module is used to acquire images, videos and point clouds of the construction site through multi-source devices, filter and evaluate the quality of the acquired data, and obtain a spatiotemporally aligned original construction dataset. The registration module is used to extract and register features from image sequences and point cloud data based on the original construction dataset, spatially align the measured data with the design model, and obtain a digital model of the construction site containing geometric and temporal information. This includes: extracting feature points and feature descriptors from the image sequences in the original construction dataset, establishing matching relationships between images, and generating a sparse reconstructed point cloud; performing densification processing on the sparse reconstructed point cloud to fill missing spatial regions and construct continuous surface geometric information; extracting planar, line segment, and corner geometric features from the point cloud data to form a structural feature set for the construction scene; pairing and comparing the structural feature set with the geometric features of the building information design model, calculating the transformation matrix, and performing spatial position correction on the measured data; sorting and organizing data collected at different times using timestamps to establish a temporal relationship network for construction progress; integrating the corrected geometric data with the temporal relationship network, adding a time attribute marker to each spatial point, and generating a digital model of the construction site containing geometric and temporal information. The classification module is used to perform target detection and classification of on-site components based on the construction site digital model, analyze the spatial positional relationship between components, and obtain a construction completion status table with component identification and attributes. The mapping module is used to establish a mapping relationship between actual components and construction plan tasks based on the construction completion status table, calculate the component completion degree and position deviation value, and obtain a construction progress assessment report and component-level deviation data. This includes: importing construction plan data from the project management system, extracting the task decomposition structure and key node milestones, and generating a construction task dependency network; establishing a correspondence between component records in the construction completion status table and planned items in the construction task dependency network through component coding matching, forming a component task mapping table; statistically analyzing the component completion status of each task item based on the component task mapping table, calculating the task completion rate and progress weight value; calculating the difference between the component geometric parameters and design parameters in the construction completion status table, quantifying the component installation accuracy through bounding box occupancy rate and point-to-surface distance evaluation methods, and generating a component position deviation matrix; comprehensively analyzing the overall progress status and key node achievement status based on the task completion rate and progress weight value, combined with the project critical path, and compiling a construction progress assessment report; and hierarchically summarizing the component position deviation matrix according to component type, region, and deviation degree to generate component-level deviation data including deviation distribution maps and hotspot areas. The analysis module is used to analyze the progress change trend based on the construction progress assessment report and component-level deviation data, identify abnormal patterns and potential risk points in the construction process, and obtain progress prediction results and risk warning list. The filtering module is used to filter the progress prediction results and risk warning list according to user permissions and priorities, and overlay the progress and deviation information on the real-world site to obtain a visualized construction management decision-making reference.

7. A computer device, characterized in that, It includes a memory and a processor, the memory storing a computer program that can run on the processor, and the processor executing the computer program to implement the construction progress monitoring method based on big data as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, It stores a computer program, which, when run by a processor, causes the processor to execute the big data-based construction progress monitoring method as described in any one of claims 1 to 5.

Images