Artificial intelligence-based bim system information processing method and system

Abstract

The application relates to the field of building technology, in particular to a BIM system information processing method and system based on artificial intelligence, which comprises a mobile monitoring terminal, a construction scheme analysis module, a process risk mapping module, a personnel presence monitoring module, a gap monitoring and judging module, a monitoring time sequence planning module, a BIM data fusion unit and an analysis and processing unit. The personnel presence monitoring module is used for collecting the number of personnel present in a construction area in real time, and the gap monitoring and judging module is used for comparing the number of personnel present with a safety monitoring personnel number threshold in real time, so that a monitoring permission instruction is generated only when the gap monitoring condition is met, and the mobile monitoring terminal is driven to enter the construction area. The mechanism effectively solves the conflict between the mobile monitoring equipment and the construction operation, guarantees the completeness of monitoring data collection, maximally reduces the interference of monitoring activities on normal construction, and realizes intelligent scheduling of construction priority and gap monitoring.

Description

Technical Field

[0001] This invention relates to the field of building technology, and more specifically to an information processing method and system for BIM systems based on artificial intelligence. Background Technology

[0002] With the rapid development of Building Information Modeling (BIM) technology, BIM has evolved from a simple 3D visualization tool into an information management platform that spans the entire lifecycle of a building project. In the field of construction site management, BIM technology enables 3D simulation of the construction process, progress tracking, and resource allocation, providing crucial technical support for construction organization. However, traditional BIM systems primarily focus on geometric information management during the design phase and progress simulation during the construction phase. Their information processing capabilities for real-time environmental parameter acquisition, dynamic safety risk identification, and emergency response handling at the construction site remain relatively weak, making it difficult to meet the real-time, accurate, and intelligent management needs of modern smart construction sites.

[0003] Construction sites present complex and variable environments with various safety risks, including falls from heights, falling objects, dust explosions, and structural instability. Currently, environmental monitoring at construction sites relies primarily on two methods: first, regular manual inspections by safety management personnel using portable monitoring equipment to inspect key areas; and second, the deployment of fixed environmental monitoring sensors in critical areas to continuously monitor parameters such as temperature, humidity, and harmful gas concentrations. Manual inspections are limited by the experience and physical capabilities of the inspectors, resulting in low detection frequency, limited coverage, difficulty in reaching high-risk areas, and significant subjective judgment errors. While fixed sensors enable continuous monitoring, their fixed installation locations make it difficult to dynamically adjust monitoring points according to construction progress, easily creating blind spots and failing to provide a comprehensive understanding of the three-dimensional environment of the construction site.

[0004] In recent years, artificial intelligence (AI) technology, especially deep learning algorithms, has made significant progress in image recognition, data mining, and predictive analytics, providing new technical means for safety management at construction sites. Some studies have attempted to introduce AI algorithms into construction sites, using video surveillance images to identify personnel violations or dangerous situations, or using data analysis to predict safety risk trends. However, in existing technologies, AI analysis results often exist independently of the BIM model, lacking deep integration with the three-dimensional spatial information of the construction site. This results in less intuitive risk location, making it difficult for managers to quickly determine the specific spatial location of risk sources and their impact range.

[0005] On the other hand, there are often temporal and spatial conflicts between monitoring operations and normal construction work at construction sites. Existing mobile monitoring equipment typically uses preset fixed cycles or simple path planning algorithms when planning inspection routes and monitoring times, without fully considering real-time work schedules, personnel density, and high-risk work periods. This can lead to monitoring equipment entering densely populated areas or overlapping work zones during inspections, disrupting normal construction order, reducing efficiency, and potentially posing new safety hazards to on-site personnel due to the equipment's operation. Summary of the Invention

[0006] The purpose of this invention is to provide an information processing method and system for BIM system based on artificial intelligence, which solves the problem that there are often time and space conflicts between monitoring operations and normal construction operations at the construction site. Existing mobile monitoring equipment usually adopts a preset fixed cycle or simple path planning algorithm when planning inspection paths and monitoring time, without fully considering the real-time work schedule, personnel distribution density and high-risk operation periods on site.

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

[0008] An AI-based BIM system information processing system includes: a mobile monitoring terminal for real-time collection of building environment data; a construction plan parsing module for acquiring the construction plan text and associated BIM component semantic information for the Nth construction process, and parsing and extracting process type and temporary facility change information; a process risk mapping module for matching a preset process risk feature library based on process type and temporary facility change information, identifying dynamic high-risk areas formed by the implementation of the construction plan in the Nth construction process, and generating a dynamic high-risk area map; a personnel presence monitoring module for collecting the number of personnel present in the construction area, including RFID identification devices deployed at the construction site entrance and sub-entrances of the construction area, as well as RFID wearable tags worn by workers; and a gap monitoring judgment module for comparing the number of personnel present with a preset safety monitoring personnel number threshold in real time, and judging when the number of personnel present is lower than the safety monitoring personnel number threshold. The system is designed to meet gap monitoring conditions and generate monitoring permit instructions. If the number of safety monitoring personnel reaches or exceeds a threshold, it determines that the gap monitoring conditions are not met and generates a monitoring termination instruction or a monitoring prohibition instruction. A monitoring timing planning module responds to monitoring permit instructions by determining the regional priority and data collection density of monitoring tasks based on a dynamic high-risk area map, and plans the collection time and path for mobile monitoring terminals to enter the corresponding construction areas. A BIM data fusion unit receives real-time environmental data uploaded by mobile monitoring terminals and performs spatial coordinate matching and semantic alignment with a preset BIM 3D model to generate a 3D dynamic field model with environmental attribute labels. An analysis and processing unit, equipped with a pre-trained safety risk identification model, performs real-time analysis of the 3D dynamic field model, identifies abnormal environmental conditions, potential safety risks, and missing temporary protective measures at the construction site, and generates corresponding risk level judgment results.

[0009] A further technical solution is that the workflow of the construction scheme analysis module and the process risk mapping module includes: Step S101, obtaining the construction scheme text of the Nth construction process and extracting process keywords; Step S102, parsing the description of temporary facility changes in the construction scheme text, identifying the edge protection removal instructions, hole opening instructions, and temporary support removal instructions, and mapping the instructions to specific spatial coordinates in combination with the semantic information of BIM components to generate a temporary facility change spatial list; Step S103, matching the preset process risk feature library according to the process type and retrieving the typical risk mode corresponding to the process; Step S104, spatially overlaying the temporary facility change spatial list with the typical risk mode and assigning risk level values ​​to generate a dynamic high-risk area map; Step S105, overlaying the dynamic high-risk area map onto the BIM 3D model in real time, driving the monitoring timing planning module to perform priority monitoring and scheduling for high-risk areas.

[0010] A further technical solution is that the monitoring time sequence planning module works as follows: Step S201, receive the dynamic high-risk area map, and determine the regional priority ranking of the monitoring tasks based on the risk level of each high-risk area, the expected existence period, and the proximity to the current monitoring time; Step S202, obtain the monitoring permission instruction output by the gap monitoring judgment module and the number of personnel present in each construction area; Step S203, construct the acquisition path based on the regional priority ranking and the number of personnel present.

[0011] A further technical solution involves the following workflow for the personnel presence monitoring module and the gap monitoring and determination module: Step S301: Using RFID identification devices deployed at the entrances and exits of the construction site and sub-entrances of the construction area, the entry and exit records of RFID tags worn by workers are collected, and the number of personnel present in each construction area and the entire site is counted in real time; Step S302: Based on the type of the Nth construction process, the risk level of the construction area, and the risk distribution density in the dynamic high-risk area map, the threshold for the number of safety monitoring personnel in each area is calculated; Step S303: The number of personnel present is compared with the threshold in real time: when it is lower than the threshold, a monitoring permission instruction is generated; when it reaches or exceeds the threshold, a monitoring termination instruction or a monitoring prohibition instruction is generated.

[0012] A further technical solution also includes a processing solution generation module and an early warning push unit. The processing solution generation module is used to automatically match and output the corresponding safety processing solution based on the risk level determination result. The early warning push unit is used to push the risk level determination result, safety processing solution, dynamic high-risk area map and planned monitoring time information to the on-site management personnel terminal and mobile monitoring terminal.

[0013] A further technical solution is that the analysis and processing unit's verification process for temporary protective measures in dynamic high-risk areas includes: Step S401, controlling the mobile monitoring terminal to enter the high-risk area marked on the dynamic high-risk area map and collecting environmental status data of the area; Step S402, comparing the environmental status data with the temporary facility change information extracted by the construction plan analysis module to identify whether the temporary protective measures required by the construction plan are in place; Step S403, when the absence or failure of temporary protective measures is identified, a high-risk warning is generated, triggering the processing plan generation module to output a special rectification plan.

[0014] A further technical solution includes a pre-construction monitoring module. The pre-construction monitoring module operates as follows: Step S501, based on the BIM 3D model, determines the actual completion time of the previous construction process and the planned start time of the Nth construction process, generating a dynamic gap availability window; Step S502, confirms via RFID identification equipment that there are no personnel remaining in the construction area or that the number of personnel present is below a preset threshold; Step S503, controls a mobile monitoring terminal to enter the construction area within the dynamic gap availability window, and prioritizes environmental data collection for high-risk areas based on the dynamic high-risk area map output by the process risk mapping module; Step S504, compares the collected environmental data with the designed completed environment at the end of the previous process in multiple dimensions to identify environmental differences; Step S505, matches the environmental differences with the safety risk feature library of the Nth construction process to identify potential pre-construction risks; Step S506, when the potential pre-construction risk is higher than or equal to a preset threshold, pushes a prohibition on construction instruction to the construction team leader and generates a safety handling plan; when it is lower than the preset threshold, pushes a construction readiness permit.

[0015] Compared with existing technologies, the beneficial effects of this invention are as follows: 1. The construction plan text of the Nth construction process is parsed by the construction plan analysis module to extract the process type and temporary facility change information. The process risk mapping module then matches the preset process risk feature library to identify dynamic high-risk areas formed by the implementation of the construction plan, generating a dynamic high-risk area map. This technical solution breaks through the limitations of static and experience-based risk areas in traditional safety management, achieving real-time dynamic identification and precise spatial positioning of risk areas during construction progress, shifting safety management from post-event handling to pre-event warning and in-event monitoring. 2. The number of personnel on-site in the construction area is collected in real time by the personnel presence monitoring module. The gap monitoring judgment module compares the number of personnel on-site with the safety monitoring personnel number threshold in real time. A monitoring permission instruction is generated only when the gap monitoring conditions are met (the number of personnel on-site is below the threshold), driving the mobile monitoring terminal into the construction area. This mechanism effectively solves the conflict between mobile monitoring equipment and construction operations, ensuring the integrity of monitoring data collection while minimizing interference with normal construction activities, achieving intelligent scheduling of construction priority and gap monitoring. 3. The monitoring sequence planning module responds to monitoring permit instructions, determines the regional priority and data collection density of monitoring tasks based on the dynamic high-risk area map, and plans the collection time points and paths for mobile monitoring terminals to enter the corresponding construction areas. This solution enables limited monitoring resources to prioritize coverage of dynamic high-risk areas and ensures continuous tracking of risk evolution through reasonable data collection density settings, significantly improving monitoring efficiency and the timeliness of risk identification. 4. The BIM data fusion unit performs spatial coordinate matching and semantic alignment between the real-time environmental data uploaded by the mobile monitoring terminal and the preset BIM 3D model, generating a 3D dynamic field model with environmental attribute labels. The analysis and processing unit has a built-in trained safety risk identification model that performs real-time analysis of the model, automatically identifying abnormal environmental states, potential safety risks, and missing temporary protective measures at the construction site, and generating corresponding risk level judgments. This solution achieves a real-time data closed loop between the physical construction site and the digital twin model, upgrading safety risk identification from subjective judgment by manual inspection to objective quantitative analysis based on artificial intelligence, providing a precise data foundation for subsequent automated handling. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0017] Example:

[0018] An AI-based BIM system information processing system includes: a mobile monitoring terminal for real-time collection of building environment data; a construction plan parsing module for acquiring the construction plan text and associated BIM component semantic information of the Nth construction process, and parsing and extracting process type and temporary facility change information; a process risk mapping module for matching a preset process risk feature library according to the process type and temporary facility change information, identifying dynamic high-risk areas formed by the implementation of the construction plan in the Nth construction process, and generating a dynamic high-risk area map; a personnel presence monitoring module for collecting the number of personnel present in the construction area, including RFID identification devices deployed at the construction site entrance and sub-entrances of the construction area, as well as RFID wearable tags worn by workers; and a gap monitoring judgment module for comparing the number of personnel present with a preset safety monitoring personnel number threshold in real time, and judging when the number of personnel present is lower than the safety monitoring personnel number threshold. The system is designed to meet gap monitoring conditions and generate monitoring permit instructions. If the number of safety monitoring personnel reaches or exceeds a threshold, the system determines that gap monitoring conditions are not met and generates a monitoring termination instruction or a monitoring prohibition instruction. A monitoring timing planning module responds to monitoring permit instructions by determining the regional priority and data collection density of monitoring tasks based on a dynamic high-risk area map, and plans the collection time and path for mobile monitoring terminals to enter the corresponding construction areas. A BIM data fusion unit receives real-time environmental data uploaded by mobile monitoring terminals and performs spatial coordinate matching and semantic alignment with a preset BIM 3D model to generate a 3D dynamic field model with environmental attribute labels. An analysis and processing unit, equipped with a pre-trained safety risk identification model, performs real-time analysis of the 3D dynamic field model, identifies abnormal environmental conditions, potential safety risks, and missing temporary protective measures at the construction site, and generates corresponding risk level judgments. The Nth construction process is any construction process within the entire construction phase. Mobile monitoring terminals include, but are not limited to: autonomous mobile robots (AMRs) equipped with LiDAR, unmanned aerial vehicles (UAVs) equipped with visible light cameras and infrared thermal imagers, or handheld 3D scanning devices. The terminal integrates a 5G / 4G communication module, an RTK-GNSS positioning module, and an inertial measurement unit (IMU), enabling autonomous navigation and obstacle avoidance with a positioning accuracy of at least centimeter level. RFID identification devices are deployed at the main entrances and exits of the construction site, sub-entrances and exits of construction areas, and floor passageways. They utilize ultra-high frequency (UHF) RFID technology, operating at a frequency of 860-960MHz, with an adjustable reading distance of 0-15 meters, and support simultaneous identification of multiple tags. Wearable RFID tags are integrated into the safety helmets or reflective vests of workers and possess a unique identification code.

[0019] The workflow of the construction scheme analysis module and the process risk mapping module includes: Step S101, obtaining the construction scheme text of the Nth construction process and extracting process keywords; Step S102, parsing the description of temporary facility changes in the construction scheme text, identifying the edge protection removal instructions, hole opening instructions, and temporary support removal instructions, and mapping the instructions to specific spatial coordinates by combining BIM component semantic information to generate a temporary facility change spatial list; Step S103, matching the preset process risk feature library according to the process type and retrieving the typical risk mode corresponding to the process; Step S104, spatially overlaying the temporary facility change spatial list with the typical risk mode and assigning risk level values ​​to generate a dynamic high-risk area map; Step S105, overlaying the dynamic high-risk area map onto the BIM 3D model in real time, driving the monitoring timing planning module to perform priority monitoring and scheduling for high-risk areas. By extracting process keywords from construction plan texts, analyzing descriptions of temporary facility changes, and identifying instructions for removing edge protection, opening holes, and removing temporary supports, and mapping these instructions to specific spatial coordinates using BIM component semantic information, a spatial list of temporary facility changes is generated. This list is then spatially overlaid with typical risk patterns and assigned risk levels, generating a dynamic high-risk area map which is overlaid onto the BIM 3D model in real time. This process automates the conversion from unstructured construction plan texts to structured, spatial risk maps, eliminating subjective biases in traditional safety management, ensuring the comprehensiveness and accuracy of risk identification, and directly driving the monitoring sequence planning module to prioritize monitoring and scheduling of high-risk areas.

[0020] The monitoring time-series planning module operates as follows: Step S201, receives a dynamic high-risk area map and determines the regional priority ranking of monitoring tasks based on the risk level, expected duration, and proximity to the current monitoring time of each high-risk area; Step S202, obtains the monitoring permit instructions output by the gap monitoring judgment module and the number of personnel present in each construction area; Step S203, constructs the data acquisition path based on the regional priority ranking and the number of personnel present. The monitoring time-series planning module determines the regional priority ranking of monitoring tasks based on the risk level, expected duration, and proximity to the current monitoring time of each high-risk area, and constructs the data acquisition path based on the regional priority ranking and the number of personnel present. This scheme comprehensively considers the urgency of risks and the degree of construction interference, avoiding resource waste or monitoring blind spots caused by experience-based scheduling in traditional monitoring dispatching, enabling mobile monitoring terminals to efficiently complete data acquisition tasks within a limited gap window.

[0021] The workflow of the personnel presence monitoring module and the gap monitoring and judgment module includes: Step S301, collecting RFID tag entry and exit records of workers using RFID identification devices deployed at the construction site entrances and exits and sub-entrances of construction areas, and calculating the number of personnel present in each construction area and the entire site in real time; Step S302, calculating the threshold number of safety monitoring personnel for each area based on the type of the Nth construction process, the risk level of the construction area, and the risk distribution density in the dynamic high-risk area map; Step S303, comparing the number of personnel present with the threshold in real time: generating a monitoring permission instruction when the number is below the threshold; generating a monitoring termination instruction or a monitoring prohibition instruction when the number reaches or exceeds the threshold. By collecting RFID tag entry and exit records of workers using RFID identification devices deployed at the construction site entrances and exits and sub-entrances of construction areas, the number of personnel present in each construction area and the entire site in real time is calculated. Simultaneously, the threshold number is calculated based on the type of the Nth construction process, the risk level of the construction area, and the risk distribution density in the dynamic high-risk area map, transforming the threshold setting from a fixed value into a variable dynamically adapted to the construction scenario. This solution compares the number of personnel on site with dynamically calculated thresholds in real time to accurately determine gap monitoring conditions, which better meets the complex and ever-changing safety management needs of construction sites.

[0022] It also includes a treatment plan generation module and an early warning push unit. The treatment plan generation module automatically matches and outputs the corresponding safety treatment plan based on the risk level assessment results. The early warning push unit pushes the risk level assessment results, safety treatment plans, dynamic high-risk area maps, and planned monitoring time information to the on-site management personnel terminals and mobile monitoring terminals. The treatment plan generation module automatically matches and outputs the corresponding safety treatment plan based on the risk level assessment results output by the analysis and processing unit, while the early warning push unit simultaneously pushes the risk level assessment results, safety treatment plans, dynamic high-risk area maps, and planned monitoring time information to the on-site management personnel terminals and mobile monitoring terminals. This closed-loop mechanism significantly shortens the time window from risk identification to response, realizes the automation and intelligence of construction site safety management, and ensures the synchronous delivery of key safety information between the management and execution levels.

[0023] The analysis and processing unit's verification process for temporary protective measures in dynamic high-risk areas includes: Step S401, controlling a mobile monitoring terminal to enter the high-risk area marked on the dynamic high-risk area map and collecting environmental status data for that area; Step S402, comparing the environmental status data with the temporary facility change information extracted by the construction plan analysis module to identify whether the temporary protective measures required by the construction plan are in place; Step S403, when a missing or ineffective temporary protective measure is detected, a high-risk warning is generated, triggering the processing plan generation module to output a special rectification plan. By controlling a mobile monitoring terminal to enter the high-risk area marked on the dynamic high-risk area map to collect environmental status data and comparing it with the temporary facility change information extracted by the construction plan analysis module, the system automatically identifies whether the temporary protective measures required by the construction plan are in place. This solution upgrades the traditional manual inspection of protective measures to automated verification, enabling timely detection of potential hazards such as the failure to simultaneously install temporary railings after the removal of edge protection and the failure to cover protective covers after opening holes. When temporary protective measures are found to be missing or ineffective, a high-risk warning is generated and the handling solution generation module is triggered to output a special rectification plan, which significantly improves the supervision efficiency and response speed of the implementation of protective measures.

[0024] It also includes a pre-construction monitoring module. The pre-construction monitoring module operates as follows: Step S501, based on the BIM 3D model, determine the actual completion time of the previous construction process and the planned start time of the Nth construction process, and generate a dynamic gap availability window; Step S502, confirm through RFID identification equipment that there are no personnel remaining in the construction area or that the number of personnel present is lower than the preset threshold; Step S503, control the mobile monitoring terminal to enter the construction area within the dynamic gap availability window, and based on the dynamic high-risk area map output by the process risk mapping module, prioritize the collection of environmental data for high-risk areas; Step S504, compare the collected environmental data with the design completion environment at the end of the previous process in multiple dimensions to identify environmental differences; Step S505, match the environmental differences with the safety risk feature library of the Nth construction process to identify potential pre-construction risks; Step S506, when the potential pre-construction risk is higher than or equal to the preset threshold, push a prohibition on construction instruction to the construction team leader and generate a safety handling plan; when it is lower than the preset threshold, push a construction readiness permit. The pre-construction monitoring module uses a BIM 3D model to determine the actual completion time of the previous construction process and the planned start time of the Nth construction process, generating a dynamic gap availability window. After confirming that there are no personnel remaining in the construction area or that the number of personnel on site is below a preset threshold, the module controls a mobile monitoring terminal to enter the construction area and prioritizes environmental data collection in high-risk areas. The collected data is compared with the designed completed environment at the end of the previous process in multiple dimensions to identify environmental differences and match them with the safety risk feature library of the Nth construction process to identify potential pre-construction risks. This solution achieves automatic verification of the environmental baseline and intelligent identification of pre-construction risks during the critical window period of process transition, enabling timely detection and handling of residual risks before the start of the Nth construction process. When a potential pre-construction risk is higher than or equal to a preset threshold, a construction prohibition order is pushed and a safety handling plan is generated; when it is lower than the preset threshold, a construction readiness permit is pushed, effectively avoiding the phenomenon of starting construction with defects and improving the safety assurance capability during the process transition phase.

[0025] The process keyword extraction submodule employs a pre-trained language model based on the Transformer architecture (such as BERT or its variants), fine-tuned using corpora from the construction field, to perform named entity recognition (NER) on the construction plan text. Entity types include: process type (such as concrete pouring, steel structure hoisting, formwork removal, etc.), work surface floor, and area number.

[0026] Temporary Facility Change Instruction Recognition Submodule: Based on regular expressions and contextual syntax rules, it extracts structured temporary facility change instructions from the construction plan text. For example:

[0027] Edge protection removal command rule pattern: Remove + [Location Description] + Edge Protection + [Scope Description]

[0028] Hole opening instruction rule pattern: Open + [Size] + Hole + [Location] + Used for + [Purpose]

[0029] Temporary support removal instruction rule pattern: Remove + [Location] + Temporary Support + [Number / Range]

[0030] BIM component semantic information matching: Component entities matching process keywords are queried and extracted using BIM model APIs (such as the IFC standard interface or Revit API) to obtain their globally unique identifiers (GUIDs) and spatial boundary coordinates. Temporary facility change instructions are mapped to the spatial coordinates of specific BIM components, generating a temporary facility change spatial list containing component GUIDs, change types, and spatial coordinates.

[0031] The process risk characteristic database is pre-stored in the system database, with process type as the primary key. Each record contains:

[0032] List of typical risk patterns (falling from edges, falling through holes, being struck by objects, and risks from blind spots, etc.)

[0033] Spatial impact range parameters for each risk model (e.g., the impact range after the removal of edge protection extends outward 1.5 meters along the removal edge).

[0034] Risk level assignment rules (Level 1 / Level 2 / Level 3)

[0035] The formula for calculating the expected time period (based on the planned start and end times of the construction scheme and the statistical average duration of historical processes).

[0036] Dynamic high-risk area map generation algorithm:

[0037] Spatial overlay: Perform a Boolean union operation on each change area in the temporary facility change space list and the spatial impact range of typical risk models to generate a risk impact space volume;

[0038] Risk level assignment: If multiple risk patterns are superimposed on the same spatial area, the highest risk level shall be taken as the level of the area;

[0039] Calculation of the expected duration of existence: Based on the planned start and end times in the construction plan and combined with the average duration of this type of process according to historical data, the expected duration of existence (T_start, T_end) of each high-risk area is calculated.

[0040] Map output: Output dynamic high-risk area map in JSON format. Each area record includes: area ID, BIM spatial coordinates (polygon vertex sequence or bounding box), risk type, risk level, expected period of existence, and associated process ID.

[0041] Specific implementation of personnel presence monitoring and gap monitoring determination

[0042] RFID personnel counting: The RFID identification device reads in and out tags in real time at the entrance and exit, and updates the number N of present personnel in each area through a direction determination algorithm (which judges the movement direction based on the phase difference of a dual-antenna array or RSSI intensity change).

[0043] Th=w1×N_base+w2×R_level+w3×D_density+C

[0044] Th is the threshold for the number of safety monitoring personnel;

[0045] N_base is the base threshold (e.g., 20% of the rated number of operating personnel in the area, or determined according to construction quotas);

[0046] R_level is the current risk level coefficient of the construction area (1.5 for first-level risk, 1.0 for second-level risk, 0.5 for third-level risk);

[0047] D_density is the risk distribution density of the area in the dynamic high-risk area map (the number of high-risk areas per unit area, units / 100m²);

[0048] w1, w2, and w3 are weight coefficients, which are determined by training through multiple regression analysis or random forest algorithm based on historical monitoring data.

[0049] C is a correction constant, set according to on-site management requirements.

[0050] Determination logic:

[0051] If N<Th: a monitoring permission instruction is generated to allow the mobile monitoring terminal to enter the area;

[0052] If N≥Th and there is no ongoing monitoring task: a monitoring prohibition instruction is generated;

[0053] If N≥Th and the current monitoring task is in progress: a monitoring termination instruction is generated to control the mobile monitoring terminal to evacuate immediately.

[0054] P_j=α×(4-L_j)+β×(1 / |T_j-T_now|)+γ×E_j

[0055] Where:

[0056] P_j is the priority score of area j, and a higher score indicates a higher priority;

[0057] L_j is the risk level of area j (first level=3, second level=2, third level=1);

[0058] T_j is the end time of the estimated existence period of area j;

[0059] T_now represents the current monitoring time;

[0060] E_j represents the proximity (reciprocal of distance) between region j and the current monitoring terminal location.

[0061] α, β, and γ are weighting coefficients that satisfy α+β+γ=1 and are set according to on-site management preferences.

[0062] Data collection path construction algorithm: Based on the regional priority ranking results and the number of personnel present, an improved ant colony algorithm (ACO) or genetic algorithm (GA) is used to construct the data collection path of the mobile monitoring terminal. The path nodes include the entry points of each high-risk area and key monitoring station locations. The optimization objective is to minimize the total path time and avoid areas where the number of personnel present is greater than or equal to a threshold.

[0063] BIM Data Fusion Unit and Analysis Processing Unit

[0064] Spatial coordinate matching: Real-time environmental data (point cloud, imagery) collected by mobile monitoring terminals is converted to the unified coordinate system of the BIM model using SLAM (Simultaneous Localization and Mapping) algorithm or RTK-GNSS / IMU combined positioning data. ICP (Iterative Closest Point) algorithm or feature point-based registration method is then used to perform coarse and fine registration between the point cloud data and the BIM 3D model.

[0065] Semantic alignment: Utilizing the semantic information of components in the BIM model (such as walls, slabs, columns, edge protection components, and hole components), semantic segmentation and annotation are performed on the registered environmental data. For example, the area in the point cloud corresponding to the edge protection component in the BIM model is marked as edge protection status to be inspected, generating a 3D dynamic field model with environmental attribute labels.

[0066] Security risk identification model: A 3D convolutional neural network (3DCNN) or PointNet++ network built using a deep learning framework (such as PyTorch or TensorFlow). The input is voxelized data or point cloud data of a 3D dynamic field model, and the output is:

[0067] Environmental anomaly classification results (such as excessive noise, abnormal dust concentration, insufficient lighting, etc.)

[0068] Potential safety risk types and levels (e.g., fall risk level 1, object strike risk level 2, etc.)

[0069] Results of identification of missing temporary protective measures (such as missing edge protection, uncovered holes, etc.)

[0070] The model training dataset includes: historical construction site point cloud / image data, labeled risk areas, risk types, and levels. The model is deployed on an edge computing gateway or a cloud GPU server, with an inference latency of no more than 2 seconds.

[0071] Processing solution generation module and early warning push unit

[0072] Safety Handling Solution Matching: The solution generation module has a built-in safety risk management knowledge base. Using risk type + risk level as the search key, it matches corresponding safety handling solution templates. Solution templates include: description of corrective measures, required materials and equipment, responsible personnel, completion deadline, and acceptance criteria. When multiple risks are identified (multiple risk types exist in the same area), a rule engine is used to merge solutions and resolve conflicts.

[0073] Early warning push mechanism: Asynchronous message distribution is implemented using message queues (such as RabbitMQ or Kafka).

[0074] Push the following information to the on-site management personnel's terminals: risk level assessment results, dynamic high-risk area map (displayed in the form of a lightweight BIM model), safety handling plan, and post-planning monitoring time information;

[0075] Push task instructions and path planning results to mobile monitoring terminals;

[0076] Key instructions (such as prohibiting construction or emergency evacuation) are verified using digital signatures to ensure information integrity and non-repudiation.

[0077] Temporary protective measures verification process

[0078] Data Acquisition: Control the mobile monitoring terminal to enter high-risk areas and collect data.

[0079] Edge protection status image: High-definition image of the edge area captured by a visible light camera (resolution not less than 1920×1080).

[0080] Point cloud data of hole coverage status: LiDAR scan of the area around the hole, with a point cloud density of not less than 100 points / m²;

[0081] Temporary barrier setup status: panoramic camera or multi-angle camera array shooting.

[0082] Alignment algorithm:

[0083] Image comparison: Target detection models (such as YOLOv8 or Faster R-CNN) are used to identify targets such as temporary railings, protective covers, and safety nets in the images and match them with the types of protective measures required by the construction plan;

[0084] Point cloud comparison: The collected point cloud of the hole area is compared with the design boundary of the hole in the BIM model to calculate the opening area of ​​the hole. If the actual opening area is greater than the design value or the point cloud features of the covering are missing, it is determined that the protective measures are missing.

[0085] Rectification plan output: When the absence or failure of temporary protective measures is detected, a high-risk warning (level 1 risk) is generated, triggering the handling plan generation module to output a special rectification plan (such as immediately installing a 1.2-meter-high temporary railing in the edge area of ​​axis 1-3 on the 3rd floor of area A, with the railing posts spaced no more than 2 meters apart).

[0086] The specific implementation of the pre-construction monitoring module.

[0087] Dynamic gaps can be generated using a window: Based on the BIM 3D model and the construction schedule (such as XML / CSV files exported from Microsoft Project or Primavera P6, or 4D schedule information built into the BIM model), the actual completion time T_finish of the previous construction process and the planned start time T_start of the Nth construction process are determined, and the dynamic gap window ΔT = T_start - T_finish - T_buffer is calculated, where T_buffer is the safety buffer time (such as 30 minutes).

[0088] Personnel Detention Confirmation: Confirm via RFID identification device that there are no personnel detained in the construction area (N=0) or the number of personnel on site is lower than the preset retention threshold (e.g., N≤2 people, and their identities are on-site management personnel or cleaning personnel).

[0089] Environmental data acquisition: Within the available window ΔT of the dynamic gap, control the mobile monitoring terminal to enter the construction area and prioritize the acquisition of high-risk area data marked by the dynamic high-risk area map, including: 3D point cloud snapshot (for geometric shape comparison), visible light panoramic image (for surface condition recording), edge protection status and hole coverage status.

[0090] Multi-dimensional comparison algorithm:

[0091] Geometric comparison: The current point cloud snapshot is registered and compared with the design completion environment point cloud (theoretical point cloud extracted from the BIM model) at the end of the previous process to identify differences that exceed the tolerance (e.g. ±50mm).

[0092] Semantic comparison: The status of the components identified in the current image is semantically matched with the design completion status of the process in the BIM model to identify missing or damaged components;

[0093] Risk matching for differences: The identified environmental differences are matched with the safety risk feature library of the Nth construction process by keyword matching and spatial correlation analysis to identify potential pre-existing risks that may lead to safety accidents in the Nth construction process.

[0094] Decision-making logic:

[0095] If the potential pre-existing risk level is greater than or equal to the preset threshold (e.g., level 2 or above): send a prohibition order to the construction team leader and trigger the handling solution generation module to generate a safety handling solution.

[0096] If the potential pre-existing risk level is less than the preset threshold or the risk has been eliminated (rectification has been completed and verified by monitoring): push out the construction readiness permit.

[0097] Although the invention has been described herein with reference to several illustrative embodiments, it should be understood that many other modifications and implementations can be devised by those skilled in the art, which will fall within the scope and spirit of the principles disclosed herein. More specifically, various variations and modifications can be made to the components and / or layout of the subject matter combination within the scope of the disclosure and claims. Besides variations and modifications to the components and / or layout, other uses will be apparent to those skilled in the art.

Claims

1. An information processing system for BIM systems based on artificial intelligence, characterized in that, Includes: mobile monitoring terminals for real-time collection of building environment data; The construction plan parsing module is used to obtain the construction plan text and associated BIM component semantic information of the Nth construction process, and to parse and extract the process type and temporary facility change information. The process risk mapping module is used to match the process type and temporary facility change information with a preset process risk feature library, identify the dynamic high-risk areas formed by the implementation of the construction plan in the Nth construction process, and generate a dynamic high-risk area map. The personnel presence monitoring module is used to collect the number of personnel present in the construction area, including RFID identification devices deployed at the entrance and exit of the construction site and the sub-entrances and exits of the construction area, as well as RFID wearable tags worn by workers. The gap monitoring judgment module is used to compare the number of personnel present with a preset safety monitoring personnel number threshold in real time. When the number of personnel present is lower than the safety monitoring personnel number threshold, it is determined that the gap monitoring conditions are met and a monitoring permission instruction is generated. When the number of personnel present reaches or exceeds the safety monitoring personnel number threshold, it is determined that the gap monitoring conditions are not met and a monitoring termination instruction or a monitoring prohibition instruction is generated. The monitoring timing planning module is used to respond to the monitoring permission instruction, determine the regional priority and data collection density of the monitoring tasks based on the dynamic high-risk area map, and plan the collection time and path for the mobile monitoring terminal to enter the corresponding construction area. The BIM data fusion unit is used to receive the real-time environmental data uploaded by the mobile monitoring terminal, and perform spatial coordinate matching and semantic alignment of the real-time environmental data with the preset BIM 3D model to generate a 3D dynamic field model with environmental attribute labels. The analysis and processing unit has a built-in trained safety risk identification model, which is used to perform real-time analysis on the 3D dynamic field model, identify the abnormal environmental state, potential safety risks and lack of temporary protective measures at the construction site, and generate corresponding risk level judgment results.

2. The BIM system information processing system based on artificial intelligence according to claim 1, characterized in that: The workflow of the construction scheme analysis module and the process risk mapping module includes: Step S101, obtaining the construction scheme text of the Nth construction process and extracting process keywords; Step S102, parsing the description of temporary facility changes in the construction scheme text, identifying the edge protection removal instruction, hole opening instruction, and temporary support removal instruction, and mapping the instructions to specific spatial coordinates in combination with BIM component semantic information to generate a temporary facility change spatial list; Step S103, matching the preset process risk feature library according to the process type and retrieving the typical risk mode corresponding to the process; Step S104, spatially overlaying the temporary facility change spatial list with the typical risk mode and assigning risk level values ​​to generate a dynamic high-risk area map; Step S105, overlaying the dynamic high-risk area map onto the BIM 3D model in real time, driving the monitoring timing planning module to perform priority monitoring and scheduling for the high-risk area.

3. The BIM system information processing system based on artificial intelligence according to claim 1, characterized in that: The working steps of the monitoring time sequence planning module are as follows: Step S201, receive the dynamic high-risk area map, and determine the regional priority order of the monitoring tasks according to the risk level, expected existence period and proximity to the current monitoring time of each high-risk area. Step S202: Obtain the monitoring permission instruction output by the gap monitoring and determination module and the number of personnel present in each construction area; Step S203: Construct a data acquisition path based on the area priority ranking and the number of personnel present.

4. The BIM system information processing system based on artificial intelligence according to claim 1, characterized in that: The workflow of the personnel presence monitoring module and the gap monitoring and determination module includes: Step S301, collecting RFID tag entry and exit records of workers by using RFID identification devices deployed at the entrance and exit of the construction site and the sub-entrances and exits of the construction area, and counting the number of personnel present in each construction area and the entire site in real time; Step S302, calculating the threshold number of safety monitoring personnel in each area based on the type of the Nth construction process, the risk level of the construction area, and the risk distribution density in the dynamic high-risk area map; Step S303, comparing the number of personnel present with the threshold in real time: generating a monitoring permission instruction when the number is lower than the threshold; generating a monitoring termination instruction or a monitoring prohibition instruction when the number reaches or exceeds the threshold.

5. The BIM system information processing system based on artificial intelligence according to claim 1, characterized in that: It also includes a processing solution generation module and an early warning push unit. The processing solution generation module is used to automatically match and output the corresponding safety processing solution based on the risk level determination result. The early warning push unit is used to push the risk level determination result, safety handling plan, dynamic high-risk area map and planned monitoring time information to the on-site management personnel terminal and the mobile monitoring terminal.

6. The BIM system information processing system based on artificial intelligence according to claim 5, characterized in that: The analysis and processing unit's verification process for temporary protective measures in the dynamic high-risk area includes: Step S401, controlling the mobile monitoring terminal to enter the high-risk area marked on the dynamic high-risk area map and collecting environmental status data of the area; Step S402, comparing the environmental status data with the temporary facility change information extracted by the construction plan analysis module to identify whether the temporary protective measures required by the construction plan are in place; Step S403, when the temporary protective measures are found to be missing or ineffective, a high-risk warning is generated, triggering the processing plan generation module to output a special rectification plan.

7. The BIM system information processing system based on artificial intelligence according to claim 1, characterized in that: It also includes a pre-construction monitoring module, whose working steps are as follows: Step S501, based on the BIM 3D model, determine the actual completion time of the previous construction process and the planned start time of the Nth construction process, and generate a dynamic gap availability window; Step S502, confirm through the RFID identification device that there are no personnel remaining in the construction area or that the number of personnel present is lower than a preset threshold; Step S503, control the mobile monitoring terminal to enter the construction area within the dynamic gap availability window, and based on the dynamic high-risk area map output by the process risk mapping module, prioritize the collection of environmental data for the high-risk areas; Step S504, compare the collected environmental data with the design completion environment at the end of the previous process in multiple dimensions to identify environmental differences; Step S505, match the environmental differences with the safety risk feature library of the Nth construction process to identify potential pre-construction risks; Step S506, when the potential pre-construction risk is higher than or equal to a preset threshold, push a prohibition on construction instruction to the construction team leader and generate a safety handling plan; when it is lower than the preset threshold, push a construction readiness permit.