Building engineering quality safety supervision system and method based on multi-source data

Abstract

The invention relates to the technical field of safety supervision, in particular to a building engineering quality safety supervision system and method based on multi-source data. The method comprises the following steps: acquiring positioning data of engineers and instruments, identifying a movement track, generating a dynamic simulation building engineering model to analyze abnormal construction behaviors, performing incentive backtracking analysis to identify the root of the abnormal behaviors, predicting the development trend of the abnormal behaviors, delimiting a focusing supervision area according to prediction data, and performing focusing supervision on the abnormal behaviors. The method comprises the following steps: carrying out rejection path analysis, forming candidate paths, carrying out space mapping and speed optimal matching on the candidate paths, determining an optimal rejection path, and carrying out engineering repair management according to the optimal path when the safety quality is lower than a threshold value after the engineering safety quality is evaluated. According to the invention, a more efficient and safer construction engineering quality safety supervision method is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of safety supervision, and in particular to a construction engineering quality safety supervision system and method based on multi-source data. Background Art

[0002] With the rapid development of the construction industry, engineering safety and quality issues have become increasingly prominent. Frequent accidents not only affect the progress of the project, but also cause casualties and property losses. Therefore, it is particularly important to improve the quality and safety supervision level of construction projects. However, traditional supervision methods often rely on manual inspections and regular inspections, which makes it difficult to achieve real-time monitoring and dynamic supervision, resulting in many potential risks that cannot be identified and handled in a timely manner. There are still many challenges in the collection and analysis of data such as the movement trajectories of engineering personnel and equipment, and construction behaviors. Especially in the process of multi-source data fusion, the phenomenon of information islands is serious, resulting in the inability to fully grasp the real-time dynamics of the construction site. In addition, the cause analysis of abnormal construction behavior and the establishment of prediction models are relatively lagging, lacking an effective backtracking mechanism and developing prediction methods, making it difficult to provide strong support for decision-making. Summary of the invention

[0003] Based on this, it is necessary to provide a construction project quality and safety supervision system and method based on multi-source data to solve at least one of the above technical problems.

[0004] To achieve the above purpose, the construction project quality and safety supervision method based on multi-source data includes the following steps:

[0005] Step S1: Acquire the positioning data of engineering personnel and the positioning data of engineering equipment; perform motion trajectory recognition on the positioning data of engineering personnel to obtain the movement trajectory of personnel; perform positioning change analysis on the positioning data of engineering equipment to generate the movement trajectory of equipment;

[0006] Step S2: Perform dynamic engineering site simulation based on the movement trajectory of personnel and equipment to generate a dynamic simulation building engineering model; perform abnormal construction behavior analysis on the dynamic simulation building engineering model to generate abnormal construction behavior data;

[0007] Step S3: Perform cause retrospective analysis on the abnormal construction behavior data to generate abnormal behavior causes; perform abnormal development prediction on the abnormal construction behavior data based on the abnormal behavior causes to obtain predicted abnormal development data;

[0008] Step S4: cutting the dynamic simulation building engineering model into independent regions according to the predicted abnormal development data to generate a focused supervision region; performing rejection path analysis on the predicted abnormal development data based on the focused supervision region to obtain a candidate rejection path;

[0009] Step S5: Based on the abnormal construction behavior data, path space mapping is performed on the candidate rejection paths to generate abnormal behavior paths; speed optimization rejection matching is performed on the candidate rejection paths according to the abnormal behavior paths to obtain the preferred rejection paths;

[0010] Step S6: Perform safety quality assessment based on the predicted abnormal development data to obtain the project safety quality; when the project safety quality is lower than the predicted safety quality threshold, perform project repair management based on the preferred exclusion path to implement quality and safety supervision of intelligent building projects.

[0011] The present invention realizes accurate identification of motion trajectories by acquiring positioning data of engineering personnel and equipment. The implementation of dynamic engineering site simulation generates an intuitive construction engineering model. The application of abnormal construction behavior analysis improves the ability to identify potential risks. The inducement retrospective analysis provides a clear perspective on the root causes of abnormal behavior. The generation of abnormal development prediction provides forward-looking information support for construction management. The demarcation of focused supervision areas realizes accurate monitoring of key areas. The implementation of rejection path analysis enhances the ability to respond to abnormal situations. The combination of path space mapping and optimal rejection matching improves the effectiveness of the repair plan. Safety and quality assessment provides a quantitative basis for the overall safety of the project. The ultimately formed intelligent building engineering quality and safety supervision system significantly improves the safety and management efficiency of the construction process.

[0012] Preferably, step S1 comprises the following steps:

[0013] Step S11: Acquire engineering personnel positioning data and engineering equipment positioning data;

[0014] Step S12: marking the personnel positioning data and the engineering equipment positioning data to generate a personnel spatiotemporal point set and an equipment spatiotemporal point set;

[0015] Step S13: Continuing the trajectory of the personnel space-time point set to obtain personnel trajectory segments; performing feature fitting processing on the personnel trajectory segments to generate the personnel movement trajectory;

[0016] Step S14: performing noise filtering on the instrument spatiotemporal point set to obtain a filtered instrument point set; performing motion vector calculation on the filtered instrument point set to generate an instrument motion trajectory.

[0017] The present invention achieves accurate spatiotemporal point set marking by acquiring the positioning data of engineering personnel and equipment, laying the foundation for subsequent analysis. The continuous processing of the trajectory of the personnel's spatiotemporal point set generates a complete motion trajectory, providing a basis for dynamic monitoring. The feature fitting processing improves the accuracy and reliability of the motion trajectory, and filtering the noise in the equipment point set effectively improves the data quality. The equipment motion trajectory generated by motion vector calculation provides key data for monitoring and analysis of the construction process. The overall method improves the safety management efficiency and intelligence level of construction projects.

[0018] Preferably, step S2 comprises the following steps:

[0019] Step S21: Acquire engineering terrain data; perform time-series alignment on the movement trajectory of personnel and the movement trajectory of equipment to obtain synchronous trajectory data;

[0020] Step S22: performing terrain matching processing on the synchronous trajectory data based on the engineering terrain data to obtain terrain constraint data; performing three-dimensional spatial reconstruction on the synchronous trajectory data according to the terrain constraint data to obtain a construction site model;

[0021] Step S23: dynamically evolving the construction site model based on the synchronous trajectory data to generate a dynamic simulation construction engineering model;

[0022] Step S24: performing trajectory behavior pattern recognition on the dynamic simulation building engineering model to obtain fused construction behavior features; performing abnormal construction behavior analysis on the fused construction behavior features to generate construction behavior abnormality data.

[0023] The present invention achieves the time alignment of the movement trajectories of personnel and equipment and generates synchronous trajectory data by acquiring engineering terrain data. Terrain matching processing provides terrain constraints for trajectory data. The construction site model generated by three-dimensional space reconstruction provides an intuitive basis for subsequent analysis. The dynamically evolving construction engineering model provides a dynamic view for real-time monitoring and evaluation of the construction process. Trajectory behavior pattern recognition reveals the characteristics of construction behavior. Abnormal data generated by abnormal construction behavior analysis provides an important basis for risk management, thereby improving the safety monitoring and management efficiency of construction projects as a whole.

[0024] Preferably, step S24 comprises the following steps:

[0025] Performing spatiotemporal slicing on the dynamic simulation building engineering model to obtain model slicing data; performing trajectory projection on the model slicing data to generate trajectory mapping data;

[0026] Conduct behavioral pattern recognition on trajectory mapping data to obtain fused construction behavior features;

[0027] Perform trajectory collision identification on the fused construction behavior features to obtain trajectory collision data; perform trajectory development simulation based on the fused construction behavior features to obtain trajectory development simulation data;

[0028] Performing trajectory collision identification on trajectory development simulation data to generate trajectory development collision data;

[0029] Based on the trajectory collision data and trajectory development collision data, the fused construction behavior features are used to screen abnormal construction behaviors and generate abnormal construction behavior data.

[0030] The present invention realizes a detailed analysis of the construction process by performing spatiotemporal slicing on the dynamic simulation construction engineering model and generating model slicing data. The trajectory mapping data generated by trajectory projection processing provides a basis for further identification, behavioral pattern recognition reveals the characteristics of construction behavior, trajectory collision recognition effectively identifies potential safety risks, trajectory development simulation provides a forward-looking perspective for the dynamic changes of construction behavior, secondary collision recognition strengthens the analysis of complex situations, and abnormal construction behavior screening based on collision data provides accurate data support for risk management, thereby improving the safety supervision and management efficiency of construction projects as a whole.

[0031] Preferably, step S3 comprises the following steps:

[0032] Step S31: Perform time series analysis on abnormal construction behavior data to obtain abnormal association chain data;

[0033] Step S32: tracing the abnormal cause of the abnormal association chain data based on a preset standard association network to generate abnormal behavior causes;

[0034] Step S33: performing numerical simulation processing on the abnormal behavior inducement to obtain development simulation data; performing trend extrapolation calculation on the development simulation data to obtain abnormal development trend data;

[0035] Step S34: performing abnormal development prediction based on the abnormal development trend data to obtain predicted abnormal development data.

[0036] The present invention identifies abnormal association chain data by performing time series analysis on abnormal construction behavior data, thereby providing a basis for subsequent analysis. The cause tracing based on the preset standard association network realizes in-depth exploration of the root causes of abnormal behavior. The development simulation data generated by numerical simulation processing provides a quantitative basis for the dynamic changes of abnormal behavior. Trend extrapolation calculation reveals potential abnormal development trends. Predictive analysis based on abnormal development trend data provides forward-looking information for construction management, thereby improving the effectiveness of abnormal behavior identification and risk management as a whole.

[0037] Preferably, step S32 includes the following steps:

[0038] Perform chain deconstruction processing on the abnormal association chain data to obtain the abnormal chain node distribution;

[0039] Project the abnormal chain node distribution onto the preset standard association network to obtain node matching data;

[0040] Calculate the path interaction degree based on the node matching data to generate an abnormal interaction path; extract the interaction inducement node features based on the abnormal interaction path to obtain the characteristic inducement node;

[0041] According to the preset standard association network, an abnormal interaction network is constructed for the characteristic inducement nodes to obtain an abnormal interaction network;

[0042] Based on the abnormal interaction network, the path tracing and inversion of the characteristic inducement nodes are performed to generate the inducements of abnormal behavior.

[0043] The present invention identifies the distribution of abnormal chain nodes by performing chain deconstruction processing on abnormal association chain data, thereby providing a detailed basis for subsequent analysis, and projects the node distribution onto a preset standard association network to realize the generation of node matching data. The interactive abnormal path generated by path interaction calculation reveals the interactive characteristics of abnormal behavior, and the extraction of characteristic inducement nodes provides key data for in-depth analysis. The construction of the abnormal interaction network realizes a comprehensive perspective on abnormal behavior, and the path tracing inversion provides effective support for identifying the inducements of abnormal behavior, thereby improving the accuracy of abnormal behavior analysis and risk management as a whole.

[0044] Preferably, step S4 comprises the following steps:

[0045] Step S41: extracting spatial features from the predicted abnormal development data to obtain spatial feature data; performing coordinate transformation processing on the spatial feature data to obtain transformed spatial data;

[0046] Step S42: performing spatial mapping on the transformed spatial data to obtain abnormal impact domain data; performing regional segmentation calculation on the abnormal impact domain data to obtain supervision sub-region data;

[0047] Step S43: performing risk focusing processing on the supervision sub-areas according to the predicted abnormal development data to generate a focused supervision area;

[0048] Step S44: projecting the predicted abnormal development data onto an abnormal trajectory based on the focused supervision area to generate a supervision area trajectory; extending the supervision area trajectory onto an avoidance trajectory based on the abnormal behavior inducement to generate an abnormal avoidance trajectory;

[0049] Step S45: Perform path planning on the abnormal avoidance trajectory to generate a candidate rejection path.

[0050] The present invention obtains detailed spatial feature data by extracting spatial features of predicted abnormal development data, laying the foundation for subsequent analysis. The transformed spatial data generated by coordinate transformation processing improves the applicability of the data. The abnormal impact domain data obtained by spatial mapping processing provides a spatial perspective for risk assessment. The supervision sub-region data generated by regional segmentation calculation realizes the refined management of risks. The focused supervision area generated by risk focusing processing provides a basis for key monitoring. The abnormal trajectory projection enhances the visual analysis of potential risks. The avoidance trajectory extension provides a specific path for implementing risk prevention and control. The candidate rejection path generated by path planning provides an effective solution for safety management, which improves the risk management and safety supervision efficiency of construction projects as a whole.

[0051] Preferably, step S5 comprises the following steps:

[0052] Step S51: performing spatial feature recognition on abnormal construction behavior data to obtain abnormal spatial feature data;

[0053] Step S52: performing spatial feature mapping on the candidate rejection paths based on the abnormal spatial feature data to generate abnormal behavior paths;

[0054] Step S53: performing similarity matching on the abnormal behavior path and the candidate rejection path to obtain path similarity data;

[0055] Step S54: performing path transformation simulation on the abnormal behavior path and the candidate rejection path according to the path similarity data to obtain a simulated path transformation set;

[0056] Step S55: Perform transformation speed analysis on the simulated path transformation set to obtain path transformation speed parameters; perform speed optimization matching on the candidate rejection paths according to the path transformation speed parameters to obtain the optimized rejection paths.

[0057] The present invention obtains abnormal spatial feature data by performing spatial feature recognition on abnormal construction behavior data, which provides a basis for path analysis. Based on these feature data, spatial feature mapping is performed on candidate rejection paths to generate abnormal behavior paths. Path similarity matching reveals the association between abnormal behavior and candidate paths, which provides a basis for path optimization. The transformation set generated by path transformation simulation provides diversified solutions for dynamic management. The speed parameters obtained by transformation speed analysis provide quantitative support for path optimization. Speed ​​optimization matching realizes the refined selection of candidate rejection paths, which improves the safety management and risk avoidance capabilities in the construction process of construction projects as a whole.

[0058] Preferably, step S6 comprises the following steps:

[0059] Step S61: Perform risk quantification processing on the predicted abnormal development data to obtain risk indicator data;

[0060] Step S62: Perform engineering safety quality assessment based on risk indicator data to obtain engineering safety quality;

[0061] Step S63: Perform threshold comparison analysis on the engineering safety quality. When the engineering safety quality is lower than the predicted safety quality threshold, perform engineering repair scheme mapping based on the preferred exclusion path to generate an engineering repair strategy.

[0062] Step S64: Use the engineering repair strategy to manage the construction project to implement quality and safety supervision of the intelligent building project.

[0063] The present invention generates risk indicator data by performing risk quantification processing on the predicted abnormal development data, thus providing a basis for subsequent evaluation. The engineering safety quality assessment realizes a comprehensive analysis of the overall safety of the project. The threshold comparison analysis effectively identifies situations where the safety quality is lower than expected. The engineering repair plan mapping based on the preferred exclusion path provides a specific strategy for risk management. The implementation of engineering repair management ensures the quality and safety of the construction project, and improves the risk response capability and management efficiency of the construction project as a whole.

[0064] The present invention also provides a construction project quality and safety supervision system based on multi-source data, which is used to execute the construction project quality and safety supervision method based on multi-source data as described above. The construction project quality and safety supervision system based on multi-source data includes:

[0065] The safety monitoring module is used to obtain the positioning data of engineering personnel and engineering equipment; identify the motion trajectory of the engineering personnel positioning data to obtain the personnel motion trajectory; analyze the positioning changes of the engineering equipment positioning data to generate the equipment motion trajectory;

[0066] The abnormal behavior analysis module is used to simulate the dynamic engineering site based on the movement trajectory of personnel and equipment, and generate a dynamic simulation building engineering model; analyze the abnormal construction behavior of the dynamic simulation building engineering model and generate abnormal construction behavior data;

[0067] The development prediction module is used to conduct a cause-tracing analysis on the abnormal construction behavior data to generate abnormal behavior inducements; based on the abnormal behavior inducements, the abnormal development of the abnormal construction behavior data is predicted to obtain predicted abnormal development data;

[0068] A supervision focus module is used to cut independent regions of the dynamic simulation building engineering model according to the predicted abnormal development data to generate focused supervision regions; based on the focused supervision regions, an exclusion path analysis is performed on the predicted abnormal development data to obtain candidate exclusion paths;

[0069] The abnormal avoidance analysis module is used to map the candidate rejection paths in path space based on the abnormal construction behavior data to generate abnormal behavior paths; perform speed optimization rejection matching on the candidate rejection paths according to the abnormal behavior paths to obtain the optimal rejection paths;

[0070] The engineering repair management module is used to conduct safety quality assessment based on the predicted abnormal development data to obtain the engineering safety quality; when the engineering safety quality is lower than the predicted safety quality threshold, engineering repair management is performed based on the preferred exclusion path to implement quality and safety supervision of intelligent building projects.

[0071] The present invention obtains the positioning data of engineering personnel and equipment in real time through the safety monitoring module, realizes the accurate identification of motion trajectory, the dynamic simulation construction engineering model generated by the abnormal behavior analysis module improves the monitoring ability of construction behavior, the development prediction module reveals the root cause of abnormal behavior through inducement backtracking analysis, and provides a basis for subsequent decision-making, the independent area cutting of the supervision focus module realizes the accurate supervision of key areas, the preferred rejection path generated by the abnormal avoidance analysis module improves the response ability to potential risks, and the safety quality assessment of the engineering repair management module provides a quantitative standard for construction quality. The overall system significantly enhances the safety management efficiency and intelligence level of construction projects. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 A schematic diagram of the steps of a construction project quality and safety supervision method based on multi-source data;

[0073] Figure 2 Detailed implementation flow chart of step S2;

[0074] Figure 3 Detailed implementation flow chart of step S3.

[0075] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0076] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0077] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0078] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0079] To achieve this, please refer to Figures 1 to 3 , a construction project quality and safety supervision method based on multi-source data, comprising the following steps:

[0080] Step S1: Acquire the positioning data of engineering personnel and the positioning data of engineering equipment; perform motion trajectory recognition on the positioning data of engineering personnel to obtain the movement trajectory of personnel; perform positioning change analysis on the positioning data of engineering equipment to generate the movement trajectory of equipment;

[0081] Step S2: Perform dynamic engineering site simulation based on the movement trajectory of personnel and equipment to generate a dynamic simulation building engineering model; perform abnormal construction behavior analysis on the dynamic simulation building engineering model to generate abnormal construction behavior data;

[0082] Step S3: Perform cause retrospective analysis on the abnormal construction behavior data to generate abnormal behavior causes; perform abnormal development prediction on the abnormal construction behavior data based on the abnormal behavior causes to obtain predicted abnormal development data;

[0083] Step S4: cutting the dynamic simulation building engineering model into independent regions according to the predicted abnormal development data to generate a focused supervision region; performing rejection path analysis on the predicted abnormal development data based on the focused supervision region to obtain a candidate rejection path;

[0084] Step S5: Based on the abnormal construction behavior data, path space mapping is performed on the candidate rejection paths to generate abnormal behavior paths; speed optimization rejection matching is performed on the candidate rejection paths according to the abnormal behavior paths to obtain the preferred rejection paths;

[0085] Step S6: Perform safety quality assessment based on the predicted abnormal development data to obtain the project safety quality; when the project safety quality is lower than the predicted safety quality threshold, perform project repair management based on the preferred exclusion path to implement quality and safety supervision of intelligent building projects.

[0086] The present invention realizes accurate identification of motion trajectories by acquiring positioning data of engineering personnel and equipment. The implementation of dynamic engineering site simulation generates an intuitive construction engineering model. The application of abnormal construction behavior analysis improves the ability to identify potential risks. The inducement retrospective analysis provides a clear perspective on the root causes of abnormal behavior. The generation of abnormal development prediction provides forward-looking information support for construction management. The demarcation of focused supervision areas realizes accurate monitoring of key areas. The implementation of rejection path analysis enhances the ability to respond to abnormal situations. The combination of path space mapping and optimal rejection matching improves the effectiveness of the repair plan. Safety and quality assessment provides a quantitative basis for the overall safety of the project. The ultimately formed intelligent building engineering quality and safety supervision system significantly improves the safety and management efficiency of the construction process.

[0087] In the embodiment of the present invention, refer to Figure 1 , is a schematic diagram of the steps of the construction project quality and safety supervision method based on multi-source data of the present invention. In this example, the construction project quality and safety supervision method based on multi-source data includes the following steps:

[0088] Step S1: Acquire the positioning data of engineering personnel and the positioning data of engineering equipment; perform motion trajectory recognition on the positioning data of engineering personnel to obtain the movement trajectory of personnel; perform positioning change analysis on the positioning data of engineering equipment to generate the movement trajectory of equipment;

[0089] In this embodiment, the positioning data of engineering personnel and engineering equipment are obtained. The positioning data of personnel and equipment are collected in real time through high-precision RTK-GNSS receivers (real-time dynamic differential global navigation satellite system) and Bluetooth beacon devices deployed on the engineering site. The positioning data is timestamped and converted into spatial coordinates using a data acquisition module. The Kalman filter algorithm is used to dynamically process the data to reduce noise and errors. The acquired personnel positioning data is input into a trajectory recognition algorithm through a distributed data processing system. The algorithm performs segmented fitting of motion trajectories and classification of behavioral patterns based on a multi-time-scale path detection model to generate motion trajectories of engineering personnel. By performing differential analysis on the equipment positioning data, the continuous change information of the equipment displacement is extracted. The path dynamic optimization algorithm is used to model and fit the equipment motion trajectory to generate the equipment motion trajectory.

[0090] Step S2: Perform dynamic engineering site simulation based on the movement trajectory of personnel and equipment to generate a dynamic simulation building engineering model; perform abnormal construction behavior analysis on the dynamic simulation building engineering model to generate abnormal construction behavior data;

[0091] In this embodiment, based on the movement trajectory of personnel and equipment, the multi-body dynamics simulation tool MSCAdams software (mechanical simulation analysis tool) is used to model the dynamic engineering behavior of the construction site, and the two sets of trajectory data are loaded into the three-dimensional site simulation module as input parameters. The motion process is simulated in real time by using discrete time steps to iteratively simulate the dynamic interaction between personnel and equipment, and a dynamic simulation construction engineering model is generated that includes dynamic interaction between personnel and equipment. Through the anomaly detection algorithm integrated in the model, abnormal behavior of the simulation data stream is identified based on logistic regression and support vector machine (SVM) analysis models, and all data marked as abnormal are grouped and output to generate construction behavior abnormal data.

[0092] Step S3: Perform cause retrospective analysis on the abnormal construction behavior data to generate abnormal behavior causes; perform abnormal development prediction on the abnormal construction behavior data based on the abnormal behavior causes to obtain predicted abnormal development data;

[0093] In this embodiment, a cause retrospective analysis is performed on the abnormal construction behavior data, and an abnormal cause relationship diagram is constructed using a causal analysis model based on a Bayesian network. Each group of abnormal data is input into the analysis model as a leaf node, and the cause of the abnormality is traced through the backward reasoning method (Backward Reasoning), and abnormal behavior inducement data is generated. Time series prediction is performed on the generated abnormal behavior inducement data, and a long short-term memory network (LSTM, long short-term memory network) is used to model and predict the future development trend of the abnormal inducement, and the predicted abnormal development data is output.

[0094] Step S4: cutting the dynamic simulation building engineering model into independent regions according to the predicted abnormal development data to generate a focused supervision region; performing rejection path analysis on the predicted abnormal development data based on the focused supervision region to obtain a candidate rejection path;

[0095] In this embodiment, the dynamic simulation construction engineering model is cut into independent areas according to the predicted abnormal development data, and a zoning algorithm is used to mark the abnormal development intensive areas in the model as supervision priorities. The abnormal intensive areas are divided using a spatial clustering algorithm (such as DBSCAN density clustering), and the division results are output as focused supervision areas. The predicted abnormal development data are loaded based on the focused supervision areas, and an exclusion path analysis tool based on the A* path algorithm is used to generate multiple candidate exclusion paths in the supervision area. The candidate paths are optimized with the minimum distance to the abnormal points and the minimum detour cost of obstacles as the optimization goals.

[0096] Step S5: Based on the abnormal construction behavior data, path space mapping is performed on the candidate rejection paths to generate abnormal behavior paths; speed optimization rejection matching is performed on the candidate rejection paths according to the abnormal behavior paths to obtain the preferred rejection paths;

[0097] In this embodiment, based on the abnormal construction behavior data, the candidate rejection paths are mapped in path space, and the interaction area between the candidate paths and the engineering model is mapped in detail using a three-dimensional space geometric model. The details of the candidate paths are supplemented and optimized through a local path reconstruction algorithm to generate an abnormal behavior path. According to the abnormal behavior path, the candidate rejection paths are optimally matched in speed based on the path parameters (such as length, detour time, and engineering cost). An optimization tool based on a genetic algorithm is used to screen the optimal path combination through a path cost function to finally generate a preferred rejection path.

[0098] Step S6: Perform safety quality assessment based on the predicted abnormal development data to obtain the project safety quality; when the project safety quality is lower than the predicted safety quality threshold, perform project repair management based on the preferred exclusion path to implement quality and safety supervision of intelligent building projects.

[0099] In this embodiment, the safety quality of the project is comprehensively evaluated based on the predicted abnormal development data and the preferred rejection path. The fuzzy analytic hierarchy process is used to construct a safety quality evaluation index system, including indicators such as personnel safety, equipment stability, and abnormal development control effect. Each indicator is weighted and a safety score is calculated. The score result is compared with a preset safety quality threshold. When the score is lower than the threshold, an engineering repair management plan is designed based on the preferred rejection path of step S5. The repair project area is divided and parameters are adjusted through a three-dimensional model editing tool to ensure the feasibility and efficiency of the repair management plan.

[0100] Preferably, step S1 comprises the following steps:

[0101] Step S11: Acquire engineering personnel positioning data and engineering equipment positioning data;

[0102] Step S12: marking the personnel positioning data and the engineering equipment positioning data to generate a personnel spatiotemporal point set and an equipment spatiotemporal point set;

[0103] Step S13: Continuing the trajectory of the personnel space-time point set to obtain personnel trajectory segments; performing feature fitting processing on the personnel trajectory segments to generate the personnel movement trajectory;

[0104] Step S14: performing noise filtering on the instrument spatiotemporal point set to obtain a filtered instrument point set; performing motion vector calculation on the filtered instrument point set to generate an instrument motion trajectory.

[0105] In this embodiment, GNSS (Global Navigation Satellite System) receivers and IMU (Inertial Measurement Unit) are combined with RTK (Real-Time Kinematic Differential) positioning technology to collect data. The location information of engineering personnel is recorded in real time through the deployed GNSS base stations and mobile stations. The data points are recorded at a sampling frequency of 5 Hz to ensure high accuracy and continuity. At the same time, the positioning module installed on the engineering equipment transmits the location information of the equipment to the central data processing system. The positioning data is parsed and stored through the data receiving module. Based on the personnel positioning data and the engineering equipment positioning data, a spatiotemporal point marking algorithm is used to process each data stream, and the data points are associated with the spatial coordinates according to the timestamp to generate a marked data set. The spatiotemporal point sets of personnel and equipment are stored in GeoJSON (Geography Markup Language) format, and the KD tree (k-dimensional tree) spatial indexing technology is used to accelerate point search, generate personnel space-time point sets and equipment space-time point sets containing time, space, and tags, and use multi-threaded processing technology to store data in slices. For the personnel space-time point set, a trajectory reconstruction algorithm is used to process the trajectory continuity. The discrete space-time point set is simplified by the DP (Douglas-Peucker) algorithm, and redundant points are eliminated to generate more accurate personnel trajectory fragments. The fragment data is feature extracted and fitted by the trajectory fitting algorithm. The motion characteristic parameters (such as speed, acceleration, and angle change) of each trajectory point are calculated by the cubic spline interpolation method to generate the personnel motion trajectory. The noise of the equipment space-time point set is filtered, and the point set data is smoothed by the Gaussian filtering algorithm. The abnormal points are eliminated by setting the 3σ threshold (standard deviation threshold) to generate the filtered instrument point set. The motion vector of the filtered instrument point set is calculated, and the Euler angle (Euler The motion vector parameters at each moment are calculated by combining the time step and position change, and the motion trajectory of the instrument is finally generated and stored as a two-dimensional vector graphics file (SVG, vector graphics format) to support subsequent analysis and optimization processes.

[0106] Preferably, step S2 comprises the following steps:

[0107] Step S21: Acquire engineering terrain data; perform time-series alignment on the movement trajectory of personnel and the movement trajectory of equipment to obtain synchronous trajectory data;

[0108] Step S22: performing terrain matching processing on the synchronous trajectory data based on the engineering terrain data to obtain terrain constraint data; performing three-dimensional spatial reconstruction on the synchronous trajectory data according to the terrain constraint data to obtain a construction site model;

[0109] Step S23: dynamically evolving the construction site model based on the synchronous trajectory data to generate a dynamic simulation construction engineering model;

[0110] Step S24: performing trajectory behavior pattern recognition on the dynamic simulation building engineering model to obtain fused construction behavior features; performing abnormal construction behavior analysis on the fused construction behavior features to generate construction behavior abnormality data.

[0111] In this embodiment, the terrain data of the construction site is obtained by combining a LiDAR (light detection and ranging) device with a GNSS (global navigation satellite system), and a point cloud data processing tool (such as PDAL, point cloud data abstraction library) is used to denoise the acquired point cloud data. After the noise points are removed, a KD tree (k-dimensional tree) algorithm to index the point cloud and generate engineering terrain data files. Based on the coordinate system of engineering terrain data, a spatial reference system consistent with the movement trajectory of personnel and equipment is established. The two types of trajectory data are aligned in time series by the timestamp matching algorithm. Each position point in the movement trajectory data is bound to the corresponding time point to generate synchronous trajectory data. The synchronous trajectory data is matched with the engineering terrain data. The elevation information and inclination angle information in the terrain data are extracted by the terrain constraint analysis algorithm based on the geographic information system (GIS). The position points in the synchronous trajectory data are spatially superimposed with the terrain data. It is marked whether the trajectory points are located in the feasible area or the restricted area to obtain the terrain constraint data. The terrain constraint data is reconstructed in three-dimensional space by the triangular mesh reconstruction algorithm. The terrain and trajectory data of the construction site are integrated into a three-dimensional construction site model. The model is based on Industry Foundation Classes (IFC, Industrial Foundation Classes) standard storage to ensure compatibility, a dynamic simulation algorithm is used to dynamically evolve the three-dimensional construction site model, combined with the time series information of the synchronized trajectory data, the hourly update of the construction site model is realized through an event-driven dynamic evolution method, the PyChrono library of the Python language is used to dynamically simulate the motion trajectory, the model is divided into multiple dynamically updated sub-areas, each sub-area is independently simulated and then summarized to generate a dynamic simulation construction engineering model, the motion data in the dynamic simulation construction engineering model is analyzed through the trajectory behavior pattern recognition technology, the trajectory data is clustered using DBSCAN (density-based spatial clustering algorithm), the motion pattern characteristics of the construction behavior are extracted, and a fusion construction behavior feature data set is formed, based on the fusion construction behavior features, the abnormal construction behavior is analyzed through a classification model based on support vector machine (SVM, support vector machine), and the characteristic parameters such as the velocity change rate and angle change rate of the trajectory point are combined to determine whether there is an anomaly, the analysis results generate construction behavior abnormal data and store it as a CSV file for subsequent analysis and optimization processing.

[0112] Preferably, step S24 comprises the following steps:

[0113] Performing spatiotemporal slicing on the dynamic simulation building engineering model to obtain model slicing data; performing trajectory projection on the model slicing data to generate trajectory mapping data;

[0114] Conduct behavioral pattern recognition on trajectory mapping data to obtain fused construction behavior features;

[0115] Perform trajectory collision identification on the fused construction behavior features to obtain trajectory collision data; perform trajectory development simulation based on the fused construction behavior features to obtain trajectory development simulation data;

[0116] Performing trajectory collision identification on trajectory development simulation data to generate trajectory development collision data;

[0117] Based on the trajectory collision data and trajectory development collision data, the fused construction behavior features are used to screen abnormal construction behaviors and generate abnormal construction behavior data.

[0118] In this embodiment, by performing spatiotemporal segmentation processing on the three-dimensional data of the dynamic simulation building engineering model, the model is decomposed into multiple independent slice data according to the time series and the spatial region, the time axis is divided into equal intervals by using a technical method based on time step division, and the three-dimensional space of the building model is refined and sliced ​​in combination with a spatial grid division tool (such as VTK, a visualization toolkit). Each slice contains timestamp information and a spatial position range, and finally the model slice data is generated. The generated model slice data is subjected to trajectory projection processing, and the motion trajectory in the slice is mapped to the two-dimensional plane coordinate according to the geographic coordinate system (WGS84). The timestamp of the trajectory point is matched with the slice time range, and the time stamp of the trajectory point is matched with the slice time range. The GeoPandas library in Python language is used to perform spatial projection transformation on the trajectory data. The trajectory mapping data is generated by calculating the projection coordinates of the trajectory points. At the same time, the attribute information of each trajectory, such as speed, direction and acceleration, is recorded. The behavior pattern recognition algorithm based on deep learning is applied to the trajectory mapping data. The trajectory point sequence is input through the pre-trained LSTM (long short-term memory network) model, and the feature vector of the motion trajectory is extracted. Different trajectory patterns are classified, and the classification results are associated with the trajectory mapping data to obtain the fusion construction behavior characteristics. The trajectory collision detection is performed on the fusion construction behavior characteristics through the trajectory collision recognition method based on spatial relationship analysis. The R-tree (R-tree) is used to identify the trajectory collision of the fusion construction behavior characteristics. The tree spatial index is used to perform a fast neighborhood search on the trajectory points, calculate the minimum distance between the trajectory points and determine whether the collision conditions are met. For the trajectory points that meet the conditions, the trajectory collision data is generated. The collision data includes time, position, trajectory ID and collision type information. The time series of the trajectory points is simulated based on the fusion of construction behavior characteristics. The future position of the trajectory is predicted using the trajectory simulation algorithm based on physical rules. The development state of the trajectory under different time steps is simulated through the PyChrono library. The predicted data of the trajectory development process is stored as a dynamic data frame. Finally, the trajectory development simulation data is obtained. The trajectory collision recognition is performed on the trajectory development simulation data. The same method as before is used. R-tree spatial index and neighborhood search technology are used to perform collision analysis on the simulated trajectory data to determine whether there is a collision event in the trajectory at the future moment, generate trajectory development collision data and attach timestamp and trajectory attribute information, and combine trajectory collision data and trajectory development collision data to screen abnormal construction behaviors based on the fused construction behavior characteristics. By performing statistical analysis on the frequency, spatial distribution, time interval and other characteristics of trajectory collision events, a decision tree-based classification algorithm is used to distinguish abnormal construction behaviors, and the discrimination results are output as construction behavior abnormal data. The construction behavior abnormal data contains the time, location, trajectory and classification label of the abnormal event, and is finally stored in CSV file format for further data analysis.

[0119] Preferably, step S3 comprises the following steps:

[0120] Step S31: Perform time series analysis on abnormal construction behavior data to obtain abnormal association chain data;

[0121] Step S32: tracing the abnormal cause of the abnormal association chain data based on a preset standard association network to generate abnormal behavior causes;

[0122] Step S33: performing numerical simulation processing on the abnormal behavior inducement to obtain development simulation data; performing trend extrapolation calculation on the development simulation data to obtain abnormal development trend data;

[0123] Step S34: performing abnormal development prediction based on the abnormal development trend data to obtain predicted abnormal development data.

[0124] In this embodiment, by performing time series analysis on construction behavior abnormal data, a time slicing technology based on a sliding window is used to segment the abnormal data according to a fixed time step, and the pandas library in Python is used to perform event frequency statistics on the time period after slicing. In combination with the timestamp information of the abnormal events, the time series of the abnormal events is constructed, and the dynamic time warping algorithm (DTW) is used to calculate the time correlation between the abnormal events. The causal relationship chain between the events is further established through a directed graph model (DAG, directed acyclic graph, Directed Acyclic Graph), and abnormal association chain data is generated. Based on a preset standard association network, a network comparison algorithm is used to match the abnormal association chain data with the standard network, and a network similarity algorithm based on graph matching is used to calculate the correlation between the abnormal chain and the standard network. The NetworkX library of Python is used to map the nodes and edges in the standard network to the nodes and edges in the abnormal chain one by one, and the matching results are converted into a derivation basis for abnormal inducements. By identifying the key nodes in the abnormal association chain and their corresponding standard network nodes, the abnormal behavior inducement is generated, and the specific description of the abnormal inducement is recorded in the JSON file format, including Time, event type and correlation analysis, numerical simulation of abnormal behavior inducements, dynamic simulation of abnormal inducement parameters using multivariate nonlinear dynamic model, combined with historical data of abnormal inducements and causal relationships in the correlation chain, and using Pyomo library (Python optimization modeling) to build a numerical simulation model, the behavioral characteristics and correlation relationships of abnormal inducements are input into the model for dynamic solution, and the development simulation data of abnormal inducements are obtained. Trend extrapolation calculation of the development simulation data is performed, and the development simulation data is extrapolated using a time series-based prediction model. Combined with the time series characteristics of abnormal inducements, LSTM (Long Short-Term Memory Network, Long The deep learning model of Short-Term Memory (Short-Term Memory) is used to predict the abnormal inducement behavior in the future time period, and the prediction results output by the model are converted into trend curves to generate abnormal development trend data. The future development of abnormal behavior is predicted based on the abnormal development trend data. The Bayesian network is used to further infer the causal relationship of the abnormal development trend. Combined with the classification characteristics of abnormal inducements, the Bayesian inference model in the Scikit-learn library is used to model the future development state of abnormal behavior. The abnormal development trend data is input to infer and calculate the model, and finally the predicted abnormal development data is generated. The prediction results are jointly analyzed with the development simulation data to obtain the predicted abnormal development data including time, space and abnormal behavior status. All data are stored in nested JSON file format for subsequent analysis and processing.

[0125] Preferably, step S32 includes the following steps:

[0126] Perform chain deconstruction processing on the abnormal association chain data to obtain the abnormal chain node distribution;

[0127] Project the abnormal chain node distribution onto the preset standard association network to obtain node matching data;

[0128] Calculate the path interaction degree based on the node matching data to generate an abnormal interaction path; extract the interaction inducement node features based on the abnormal interaction path to obtain the characteristic inducement node;

[0129] According to the preset standard association network, an abnormal interaction network is constructed for the characteristic inducement nodes to obtain an abnormal interaction network;

[0130] Based on the abnormal interaction network, the path tracing and inversion of the characteristic inducement nodes are performed to generate the inducements of abnormal behavior.

[0131] In this embodiment, by performing chain deconstruction processing on the abnormal association chain data, a chain decomposition algorithm based on graph theory is used to deconstruct the nodes in the abnormal association chain according to the time series. Based on the timestamp and correlation of the nodes, a depth first search (DFS) is used to disassemble the order of each node in the chain, and the nodes are allocated according to the weight of each node, and the time and space characteristics of the nodes are extracted. The relative position relationship between the nodes is stored in the form of a matrix, and finally the distribution of the abnormal chain nodes is obtained, wherein the coordinates of each node represent its position and time relationship in the abnormal chain, and the distribution of the abnormal chain nodes is projected onto a preset standard association network. First, a standard network graph is constructed and regarded as an undirected graph. Each node represents an event in the standard association network, and each edge represents the relationship between the events. A node mapping algorithm is used to project the nodes of the abnormal chain onto the standard association network. According to the similarity of the nodes, a measurement method based on Cosine similarity is used to calculate the similarity between each abnormal chain node and each node in the standard network. The node matching data is finally generated by selecting the node with the highest similarity for mapping, wherein Including the position of nodes on the standard network and their association, the path interaction degree is calculated based on the node matching data. First, the Dijkstra algorithm (shortest path algorithm) is used to calculate the path length from one node to another, and the interaction degree is calculated based on the interaction frequency and mobility of the path. By clustering the nodes involved in the path, the cluster center is used to represent the interaction intensity of the path, and an abnormal interaction path is generated. The nodes with higher frequency in the path are further analyzed according to the weight of the path, and the node pairs with high interaction are identified. These nodes are marked as abnormal interaction paths, and the interaction degree of the path is extracted as the key feature. The interaction inducement node feature is extracted according to the abnormal interaction path. Combined with the path interaction degree data, a feature extraction method based on signal processing is adopted, and the nodes in the path are regarded as signal sources. Wavelet transform (Wavelet Transform) to process the signal, extract the frequency domain features, time domain features and relationship features of the nodes, identify the inducement nodes with abnormal behavior, extract the main features of these nodes, including the frequency of event occurrence, duration and interaction characteristics between nodes, form a set of characteristic inducement nodes, and construct an abnormal interaction network for the characteristic inducement nodes according to the preset standard association network. First, the characteristic inducement nodes are regarded as key nodes in the network. The network construction method based on topological sorting is used to sort and connect the nodes according to their importance in the network. The connection weights are calculated according to the similarity between the nodes, and the connection structure in the network is adjusted by the edge weights. Finally, an abnormal interaction network is generated. The network shows the abnormal interaction relationship of the characteristic inducement nodes in the standard network and can reflect the interaction mode between the nodes. The path tracing and inversion of the characteristic inducement nodes are performed based on the abnormal interaction network.The backpropagation algorithm is used to trace the path of the abnormal interaction network, trace the abnormal behavior source of the characteristic inducement node, analyze the causal relationship between nodes, gradually invert the path between nodes, output the traceability data and sort it, identify the maximum probability inducement of abnormal behavior based on the traceability results, generate abnormal behavior inducement data, and visualize the results to support subsequent decision analysis.

[0132] Preferably, step S4 comprises the following steps:

[0133] Step S41: extracting spatial features from the predicted abnormal development data to obtain spatial feature data; performing coordinate transformation processing on the spatial feature data to obtain transformed spatial data;

[0134] Step S42: performing spatial mapping on the transformed spatial data to obtain abnormal impact domain data; performing regional segmentation calculation on the abnormal impact domain data to obtain supervision sub-region data;

[0135] Step S43: performing risk focusing processing on the supervision sub-areas according to the predicted abnormal development data to generate a focused supervision area;

[0136] Step S44: projecting the predicted abnormal development data onto an abnormal trajectory based on the focused supervision area to generate a supervision area trajectory; extending the supervision area trajectory onto an avoidance trajectory based on the abnormal behavior inducement to generate an abnormal avoidance trajectory;

[0137] Step S45: Perform path planning on the abnormal avoidance trajectory to generate a candidate rejection path.

[0138] In this embodiment, spatial feature extraction is performed on the predicted abnormal development data, and a clustering-based spatial feature analysis method is used to perform cluster analysis on the geographic coordinates of the predicted data. The K-means algorithm is selected to cluster the spatial data to obtain the cluster center and cluster radius. The spatial distance between each cluster center and other points is calculated. The spatial feature data of each cluster is obtained by calculating the density, distribution form and other features of the points. The accurate position of each spatial feature is further calibrated according to the coordinate system of the geographic space, and the distribution of each spatial feature is represented by a vector diagram. The spatial feature data is subjected to coordinate transformation processing. A linear transformation method based on coordinate transformation is used to map the points of the original spatial coordinate system to the target coordinate system according to a transformation matrix. The points are adapted to the new spatial reference system through rotation, translation and other operations. Each data point in the original coordinate system is linearly transformed to conform to the transformed spatial coordinate system. The transformed coordinates are calculated through a four-dimensional transformation matrix (including rotation and translation of the X, Y and Z axes) to obtain transformed spatial data. The transformation parameters are set according to the difference between the original coordinate system and the target coordinate system. The transformed spatial data is spatially mapped, and an inverse mapping algorithm is used. Algorithm, reproject the transformed spatial data to the standard two-dimensional or three-dimensional space, and map the transformed spatial coordinates of each data point back to the original space through the reverse projection method. Use the nearest neighbor algorithm (K-Nearest Neighbors, KNN) to map each transformed data point, find the closest standard space node, and finally generate abnormal impact domain data, which reflects the area range of abnormal changes in the transformed space. Perform regional segmentation calculation on the abnormal impact domain data, and use the Voronoi diagram (Voronoi The space is divided into regions using the method of graph diagram. Each transformed node is used as the seed point of the region to calculate the boundary of each region and form the segmented regional data. The regional data is stratified according to the intensity of abnormal impact, and the regions with higher impact are marked. Finally, the regulatory sub-region data is obtained. These data can accurately reflect the potential impact of abnormal regions on each link in the essential oil production process. The regulatory sub-region is risk-focused according to the predicted abnormal development data. A risk assessment model optimized based on genetic algorithm is used to divide the risk level of each regulatory sub-region according to the abnormal trends of historical data and predicted data. The focused regulatory region is generated by simulating the impact of different behavioral patterns on the potential risks in the region. The regions with the most potential risks are further screened out through cross-validation, and finally a regulatory region with a focus on high risks is obtained. The predicted abnormal development data is projected with abnormal trajectories based on the focused regulatory region. The abnormal development data is converted into trajectory data using the trajectory projection method based on dynamic simulation.The Monte Carlo method is used to simulate the trajectory, predict the evolution trend of the trajectory in space, and the trajectory is corrected in real time in combination with the monitoring data, and finally the abnormal trajectory in the supervision area is generated. These trajectories can reflect the potential impact of the abnormal development trend on the equipment and personnel in the supervision area. The trajectory of the supervision area is extended based on the abnormal behavior inducement. The path extension method based on the graph algorithm is used to analyze the impact of the abnormal behavior inducement on the trajectory. The shortest path algorithm is used to extend a new path on the trajectory to simulate the trajectory that avoids the abnormal behavior inducement. The extended path takes into account multiple factors such as environmental factors and equipment accessibility to generate abnormal avoidance trajectories. The abnormal avoidance trajectory is planned. The A* algorithm (A-Star Algorithm) is used to plan the trajectory. By inputting the abnormal avoidance trajectory data and combining the obstacle data in the current area for path planning, an optimal path is planned under the given starting point and end point. Taking into account various constraints such as the layout of production equipment and the width of the transportation channel, candidate rejection paths are finally generated. These paths can minimize the impact of abnormal behavior on the production process while ensuring safety. ,

[0139] Preferably, step S5 comprises the following steps:

[0140] Step S51: performing spatial feature recognition on abnormal construction behavior data to obtain abnormal spatial feature data;

[0141] Step S52: performing spatial feature mapping on the candidate rejection paths based on the abnormal spatial feature data to generate abnormal behavior paths;

[0142] Step S53: performing similarity matching on the abnormal behavior path and the candidate rejection path to obtain path similarity data;

[0143] Step S54: performing path transformation simulation on the abnormal behavior path and the candidate rejection path according to the path similarity data to obtain a simulated path transformation set;

[0144] Step S55: Perform transformation speed analysis on the simulated path transformation set to obtain path transformation speed parameters; perform speed optimization matching on the candidate rejection paths according to the path transformation speed parameters to obtain the optimized rejection paths.

[0145] In this embodiment, the spatial feature recognition of the abnormal construction behavior data is performed, and the multi-dimensional feature extraction of the data is performed using a convolutional neural network (CNN) based on deep learning. First, the abnormal data is projected into a three-dimensional space, and a three-dimensional feature matrix is ​​established to represent the distribution state of the abnormal construction behavior. The spatial features are extracted layer by layer using a convolution kernel, and the key feature points of the abnormal data are screened out. The extracted feature data are compared with the standard spatial model to identify the spatial features of the abnormal behavior, including information such as position, size, and density. Finally, abnormal spatial feature data is generated, and spatial feature mapping is performed on the candidate rejection paths based on the abnormal spatial feature data. A bidirectional spatial mapping algorithm is used. Algorithm) aligns the abnormal spatial feature data and the candidate rejection path data. First, the abnormal spatial feature data is converted into a set of path nodes. The abnormal nodes are projected onto the adjacent nodes of the candidate rejection path using a mapping method based on the shortest path. The abnormal behavior path is generated by calculating the distance and direction vector between the path nodes. The path represents the spatial relationship between the abnormal behavior and the candidate path in the form of a vector diagram, and the deviation in the mapping process is corrected to ensure accuracy. The abnormal behavior path and the candidate rejection path are matched for similarity. The path data are matched and analyzed using a dynamic time warping algorithm (Dynamic Time Warping, DTW). The abnormal behavior path and the candidate rejection path are aligned in time series. The similarity value of the two paths is obtained by calculating the Euclidean distance of the path points. The similarity value is stored in a matrix form to represent the matching result of each path segment. Finally, the path similarity data is generated. The data contains the similarity distribution information of all path points. The path transformation simulation of the abnormal behavior path and the candidate rejection path is performed based on the path similarity data. The Bezier curve-based method is used to calculate the path transformation simulation of the abnormal behavior path and the candidate rejection path. The path simulation algorithm based on the Lagrange Interpolation Method is used to dynamically adjust the path data, and curve fitting is performed on the abnormal behavior path and the candidate rejection path respectively. The smooth transformation of the path is achieved by adjusting the position of the control point. At the same time, the fitting weight value is adjusted according to the path similarity data to generate a simulated path transformation set, which contains the path transformation results under multiple weight conditions. The transformation speed of the simulated path transformation set is analyzed, and the speed of the path transformation set is calculated using a method based on the Lagrange Interpolation Method. First, the key points of each path are interpolated to generate a continuous speed change curve. The transformation speed parameters of the path are calculated by analyzing the slope and curvature of the speed curve, and the speed parameters are stored in array form. Each array element represents the speed change value of a certain section on the path.According to the path change speed parameter, the candidate rejection paths are matched with optimal speed, and the path with the most stable change speed and meeting the production process requirements is selected as the final optimal rejection path.

[0146] Preferably, step S6 comprises the following steps:

[0147] Step S61: Perform risk quantification processing on the predicted abnormal development data to obtain risk indicator data;

[0148] Step S62: Perform engineering safety quality assessment based on risk indicator data to obtain engineering safety quality;

[0149] Step S63: Perform threshold comparison analysis on the engineering safety quality. When the engineering safety quality is lower than the predicted safety quality threshold, perform engineering repair scheme mapping based on the preferred exclusion path to generate an engineering repair strategy.

[0150] Step S64: Use the engineering repair strategy to manage the construction project to implement quality and safety supervision of the intelligent building project.

[0151] In this embodiment, the predicted abnormal development data is subjected to risk quantification processing. By constructing a multi-layer risk quantification model based on risk factors, the predicted abnormal development data is decomposed into multiple sub-risk factor data. The weight value of each risk factor is calculated using the entropy weight method. Matrix calculation is performed by combining the risk factor weight and the predicted data to generate a risk quantification value for each type of abnormality. All risk quantification values ​​are combined to obtain overall risk index data. The risk index data is stored in a CSV file format, and a risk distribution map is generated in combination with a risk chart tool to display the spatial distribution characteristics of the risk quantification results of each region. The engineering safety quality assessment is performed based on the risk index data. The analytic hierarchy process is used to calculate the risk index data. The evaluation model of AHP is used to compare the risk indicator data with the engineering quality standard. First, the total risk score is calculated according to the risk indicator data, and the score is mapped to the engineering safety quality grade table. The overall safety quality of the project is comprehensively analyzed in combination with the risk classification weight and regional distribution. During the evaluation process, the Matplotlib tool is used to generate a quality evaluation map to display the spatial distribution results of the safety quality of the engineering area. The engineering safety quality is finally output in a structured JSON file format. The threshold comparison analysis of the engineering safety quality is performed. When the engineering safety quality is lower than the predicted safety quality threshold, the engineering repair plan is mapped based on the preferred exclusion path. By analyzing the matching degree between the structural characteristics of the preferred exclusion path and the actual situation of the project, the path correction algorithm is used to dynamically adjust the path data and generate a preliminary design drawing of the repair plan. The design drawing is subjected to multi-dimensional weight analysis, and the final engineering repair strategy is generated by combining the risk indicator and the path feature optimization. The repair strategy includes the resource allocation, time arrangement and operation instructions required for the repair. The engineering repair strategy is used to manage the engineering repair of the construction project. The BIM (Building Information Model) based The intelligent management system of Building Information Modeling (BIM) decomposes the repair strategy, combines the repair resource database with the task allocation model to generate a detailed repair task list, and assigns the tasks to the construction team and equipment management system. During the construction process, intelligent sensors and IoT devices are used to monitor the repair progress and project quality in real time. After the repair is completed, the BIM system is used to conduct a secondary assessment of the overall quality of the construction project to ensure that the quality of the project after repair meets safety requirements. All construction data and repair records are synchronized to the cloud platform for long-term safety supervision.

[0152] The present invention also provides a construction project quality and safety supervision system based on multi-source data, which is used to execute the construction project quality and safety supervision method based on multi-source data as described above. The construction project quality and safety supervision system based on multi-source data includes:

[0153] The safety monitoring module is used to obtain the positioning data of engineering personnel and engineering equipment; identify the motion trajectory of the engineering personnel positioning data to obtain the personnel motion trajectory; analyze the positioning changes of the engineering equipment positioning data to generate the equipment motion trajectory;

[0154] The abnormal behavior analysis module is used to simulate the dynamic engineering site based on the movement trajectory of personnel and equipment, and generate a dynamic simulation building engineering model; analyze the abnormal construction behavior of the dynamic simulation building engineering model and generate abnormal construction behavior data;

[0155] The development prediction module is used to conduct a cause-tracing analysis on the abnormal construction behavior data to generate abnormal behavior inducements; based on the abnormal behavior inducements, the abnormal development of the abnormal construction behavior data is predicted to obtain predicted abnormal development data;

[0156] A supervision focus module is used to cut independent regions of the dynamic simulation building engineering model according to the predicted abnormal development data to generate focused supervision regions; based on the focused supervision regions, an exclusion path analysis is performed on the predicted abnormal development data to obtain candidate exclusion paths;

[0157] The abnormal avoidance analysis module is used to map the candidate rejection paths in path space based on the abnormal construction behavior data to generate abnormal behavior paths; perform speed optimization rejection matching on the candidate rejection paths according to the abnormal behavior paths to obtain the optimal rejection paths;

[0158] The engineering repair management module is used to conduct safety quality assessment based on the predicted abnormal development data to obtain the engineering safety quality; when the engineering safety quality is lower than the predicted safety quality threshold, engineering repair management is performed based on the preferred exclusion path to implement quality and safety supervision of intelligent building projects.

[0159] The present invention obtains the positioning data of engineering personnel and equipment in real time through the safety monitoring module, realizes the accurate identification of motion trajectory, the dynamic simulation construction engineering model generated by the abnormal behavior analysis module improves the monitoring ability of construction behavior, the development prediction module reveals the root cause of abnormal behavior through inducement backtracking analysis, and provides a basis for subsequent decision-making, the independent area cutting of the supervision focus module realizes the accurate supervision of key areas, the preferred rejection path generated by the abnormal avoidance analysis module improves the response ability to potential risks, and the safety quality assessment of the engineering repair management module provides a quantitative standard for construction quality. The overall system significantly enhances the safety management efficiency and intelligence level of construction projects.

[0160] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0161] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A construction project quality and safety supervision method based on multi-source data, characterized in that: The following steps are involved: Step S1: Acquire engineering personnel positioning data and engineering equipment positioning data; Perform motion trajectory recognition on the engineering personnel positioning data to obtain the personnel motion trajectory; perform positioning change analysis on the engineering equipment positioning data to generate the equipment motion trajectory; Step S2: Perform dynamic engineering site simulation based on the movement trajectory of personnel and equipment to generate a dynamic simulation building engineering model; perform abnormal construction behavior analysis on the dynamic simulation building engineering model to generate abnormal construction behavior data; Step S3: Perform cause retrospective analysis on the abnormal construction behavior data to generate abnormal behavior causes; perform abnormal development prediction on the abnormal construction behavior data based on the abnormal behavior causes to obtain predicted abnormal development data; Step S4: cutting the dynamic simulation building engineering model into independent regions according to the predicted abnormal development data to generate a focused supervision area; Perform rejection pathway analysis on the predicted abnormal development data based on the focused regulatory region to obtain candidate rejection pathways; Step S5: Based on the abnormal construction behavior data, path space mapping is performed on the candidate rejection paths to generate abnormal behavior paths; speed optimization rejection matching is performed on the candidate rejection paths according to the abnormal behavior paths to obtain the preferred rejection paths; Step S6: Perform safety quality assessment based on the predicted abnormal development data to obtain the project safety quality; when the project safety quality is lower than the predicted safety quality threshold, perform project repair management based on the preferred exclusion path to implement quality and safety supervision of intelligent building projects.

2. The construction engineering quality and safety supervision method based on multi-source data according to claim 1 is characterized in that: Step S1 includes the following steps: Step S11: Acquire engineering personnel positioning data and engineering equipment positioning data; Step S12: marking the personnel positioning data and the engineering equipment positioning data to generate a personnel spatiotemporal point set and an equipment spatiotemporal point set; Step S13: Continuing the trajectory of the personnel space-time point set to obtain personnel trajectory segments; performing feature fitting processing on the personnel trajectory segments to generate the personnel movement trajectory; Step S14: performing noise filtering on the instrument spatiotemporal point set to obtain a filtered instrument point set; performing motion vector calculation on the filtered instrument point set to generate an instrument motion trajectory.

3. The construction engineering quality and safety supervision method based on multi-source data according to claim 1 is characterized in that: Step S2 includes the following steps: Step S21: Acquire engineering terrain data; perform time-series alignment on the movement trajectory of personnel and the movement trajectory of equipment to obtain synchronous trajectory data; Step S22: performing terrain matching processing on the synchronous trajectory data based on the engineering terrain data to obtain terrain constraint data; performing three-dimensional spatial reconstruction on the synchronous trajectory data according to the terrain constraint data to obtain a construction site model; Step S23: dynamically evolving the construction site model based on the synchronous trajectory data to generate a dynamic simulation construction engineering model; Step S24: performing trajectory behavior pattern recognition on the dynamic simulation building engineering model to obtain fused construction behavior features; performing abnormal construction behavior analysis on the fused construction behavior features to generate construction behavior abnormality data.

4. The construction engineering quality and safety supervision method based on multi-source data according to claim 3 is characterized in that: Step S24 includes the following steps: Performing spatiotemporal slicing on the dynamic simulation building engineering model to obtain model slicing data; performing trajectory projection on the model slicing data to generate trajectory mapping data; Conduct behavioral pattern recognition on trajectory mapping data to obtain fused construction behavior features; Perform trajectory collision identification on the fused construction behavior features to obtain trajectory collision data; perform trajectory development simulation based on the fused construction behavior features to obtain trajectory development simulation data; Performing trajectory collision identification on trajectory development simulation data to generate trajectory development collision data; Based on the trajectory collision data and trajectory development collision data, the fused construction behavior features are used to screen abnormal construction behaviors and generate abnormal construction behavior data.

5. The construction engineering quality and safety supervision method based on multi-source data according to claim 1 is characterized in that: Step S3 includes the following steps: Step S31: Perform time series analysis on abnormal construction behavior data to obtain abnormal association chain data; Step S32: tracing the abnormal cause of the abnormal association chain data based on a preset standard association network to generate abnormal behavior causes; Step S33: performing numerical simulation processing on the abnormal behavior inducement to obtain development simulation data; performing trend extrapolation calculation on the development simulation data to obtain abnormal development trend data; Step S34: performing abnormal development prediction based on the abnormal development trend data to obtain predicted abnormal development data.

6. The construction engineering quality and safety supervision method based on multi-source data according to claim 5 is characterized in that: Step S32 includes the following steps: Perform chain deconstruction processing on the abnormal association chain data to obtain the abnormal chain node distribution; Project the abnormal chain node distribution onto the preset standard association network to obtain node matching data; Calculate the path interaction degree based on the node matching data to generate an abnormal interaction path; extract the interaction inducement node features based on the abnormal interaction path to obtain the characteristic inducement node; According to the preset standard association network, an abnormal interaction network is constructed for the characteristic inducement nodes to obtain an abnormal interaction network; Based on the abnormal interaction network, the path tracing and inversion of the characteristic inducement nodes are performed to generate the inducements of abnormal behavior.

7. The construction engineering quality and safety supervision method based on multi-source data according to claim 1 is characterized in that: Step S4 includes the following steps: Step S41: extracting spatial features from the predicted abnormal development data to obtain spatial feature data; performing coordinate transformation processing on the spatial feature data to obtain transformed spatial data; Step S42: spatially map the transformed spatial data to obtain abnormal impact domain data; perform regional segmentation calculation on the abnormal impact domain data to obtain supervision sub-region data; Step S43: performing risk focusing processing on the supervision sub-areas according to the predicted abnormal development data to generate a focused supervision area; Step S44: projecting the predicted abnormal development data onto an abnormal trajectory based on the focused supervision area to generate a supervision area trajectory; extending the supervision area trajectory onto an avoidance trajectory based on the abnormal behavior inducement to generate an abnormal avoidance trajectory; Step S45: Perform path planning on the abnormal avoidance trajectory to generate a candidate rejection path.

8. The construction engineering quality and safety supervision method based on multi-source data according to claim 1 is characterized in that: Step S5 includes the following steps: Step S51: performing spatial feature recognition on abnormal construction behavior data to obtain abnormal spatial feature data; Step S52: performing spatial feature mapping on the candidate rejection paths based on the abnormal spatial feature data to generate abnormal behavior paths; Step S53: performing similarity matching on the abnormal behavior path and the candidate rejection path to obtain path similarity data; Step S54: performing path transformation simulation on the abnormal behavior path and the candidate rejection path according to the path similarity data to obtain a simulated path transformation set; Step S55: Perform transformation speed analysis on the simulated path transformation set to obtain path transformation speed parameters; perform speed optimization matching on the candidate rejection paths according to the path transformation speed parameters to obtain the optimized rejection paths.

9. The construction engineering quality and safety supervision method based on multi-source data according to claim 1 is characterized in that: Step S6 includes the following steps: Step S61: Perform risk quantification processing on the predicted abnormal development data to obtain risk indicator data; Step S62: Perform engineering safety quality assessment based on risk indicator data to obtain engineering safety quality; Step S63: Perform threshold comparison analysis on the engineering safety quality. When the engineering safety quality is lower than the predicted safety quality threshold, perform engineering repair scheme mapping based on the preferred exclusion path to generate an engineering repair strategy. Step S64: Use the engineering repair strategy to manage the construction project to implement quality and safety supervision of the intelligent building project.

10. A construction engineering quality and safety supervision system based on multi-source data, characterized in that: Used to execute the construction engineering quality and safety supervision method based on multi-source data as claimed in claim 1, the construction engineering quality and safety supervision system based on multi-source data includes: The safety monitoring module is used to obtain the positioning data of engineering personnel and engineering equipment; identify the motion trajectory of the engineering personnel positioning data to obtain the personnel motion trajectory; analyze the positioning changes of the engineering equipment positioning data to generate the equipment motion trajectory; The abnormal behavior analysis module is used to simulate the dynamic engineering site based on the movement trajectory of personnel and equipment, and generate a dynamic simulation building engineering model; analyze the abnormal construction behavior of the dynamic simulation building engineering model and generate abnormal construction behavior data; The development prediction module is used to conduct a cause-tracing analysis on the abnormal construction behavior data to generate abnormal behavior inducements; based on the abnormal behavior inducements, the abnormal development of the abnormal construction behavior data is predicted to obtain predicted abnormal development data; A supervision focus module is used to cut independent regions of the dynamic simulation building engineering model according to the predicted abnormal development data to generate focused supervision regions; based on the focused supervision regions, an exclusion path analysis is performed on the predicted abnormal development data to obtain candidate exclusion paths; The abnormal avoidance analysis module is used to map the candidate rejection paths in path space based on the abnormal construction behavior data to generate abnormal behavior paths; perform speed optimization rejection matching on the candidate rejection paths according to the abnormal behavior paths to obtain the optimal rejection paths; The engineering repair management module is used to conduct safety quality assessment based on the predicted abnormal development data to obtain the engineering safety quality; when the engineering safety quality is lower than the predicted safety quality threshold, engineering repair management is performed based on the preferred exclusion path to implement quality and safety supervision of intelligent building projects.

Images