A method, system, device and medium for pre-warning of air fall of a construction worker

Abstract

The application provides a construction worker air fall early warning method, system, device and medium, the method comprises the following steps: constructing a construction site BIM model and importing it into the Unity3D platform, integrating the construction air fall body model and the dynamic update SWRL reasoning rule of the hazard source, forming a construction site hazard source digital twin base model; real-time collection of positioning data of intelligent wearable devices is stored in the personnel positioning database, the personnel positioning information is analyzed, the personnel movement style is described, and the personalized safety early warning rule is formulated; real-time image acquisition detects the protection condition of the target air area, combines the necessity of setting protection of the target air area with the ontology SWRL reasoning rule, and timely updates the Unity3D model to create a hazard source trigger area; the distance between the personnel and each hazard source in the model, the current hazard area approach speed and the current hazard area residence time are calculated in real time, whether the threshold value is reached to trigger the alarm is judged. The application can realize the precise management of the construction worker air fall early warning.

Description

Technical Field

[0001] This invention relates to the field of construction safety early warning technology, and in particular to a method, system, equipment and medium for early warning of falling workers from heights. Background Technology

[0002] Traditional safety management for operations at height relies primarily on manual inspections, passive protective equipment, and safety regulations. However, these measures suffer from low inspection frequency and insufficient real-time early warning capabilities. As the complexity of construction site environments increases, traditional methods are struggling to meet the demands for real-time monitoring of aerial protection measures and timely early warning of personnel fall risks.

[0003] With the continuous development of image processing technology, automated risk warning systems have been widely applied. These systems primarily track personnel movement trajectories using target recognition technology or obtain real-time personnel locations using positioning technology, thereby determining whether personnel have entered pre-set danger zones and achieving intelligent safety warnings. However, existing technologies mainly rely on fixed environmental parameters and real-time personnel locations for risk assessment, making them difficult to adapt to the complex and ever-changing construction site environment. In actual construction, workers often temporarily dismantle edge protection facilities due to construction needs and fail to restore them in a timely manner, resulting in a discrepancy between pre-set environmental parameters and actual site conditions. Existing risk assessment models struggle to make accurate judgments, leading to significant safety hazards. Furthermore, different workers have significantly different work habits and behavioral styles, and existing warning systems cannot dynamically adjust risk assessment strategies based on individual worker characteristics, limiting the accuracy and applicability of warnings.

[0004] To address the shortcomings and deficiencies of the existing technologies, this invention proposes a multi-dimensional parameter dynamically adjustable early warning scheme for construction workers' falls from heights, aiming to achieve refined and intelligent management of fall risks at construction sites. Summary of the Invention

[0005] The technical problem to be solved by this invention is to provide a method, system, equipment and medium for early warning of falls from heights by construction workers, so as to achieve precise management of the risk of falls from heights at construction sites.

[0006] In a first aspect, the present invention provides a method for early warning of falling from heights for construction workers. The method is based on a smart wearable device. The smart wearable device is equipped with a positioning tag, an image acquisition sensor, a transmission module, and an early warning feedback module. The smart wearable device is connected to a server via a network and transmits the location information and image data acquired by the positioning tag and the image acquisition sensor to the server through the transmission module. The method includes the following steps: Step S1: Construct a BIM model of the construction site, identify and define the information of the air hazard sources, set protective measures in all hazard source areas by default, and import the Unity3D model; integrate the construction air fall ontology model, access and parse the hazard source dynamic update reasoning rules based on SWRL, and form a construction site hazard source digital twin base model with logical reasoning capabilities. Step S2: Collect location data from smart wearable devices in real time and store it in the personnel location database; Step S3: By periodically calling the personnel location database information, obtain the personnel's historical location information, calculate three indicators for different personnel: approach speed to danger zone, stay time in danger zone, and approach distance to danger zone, characterize the personnel's movement style, and formulate personalized safety warning rules, including adjusting the warning risk area distance threshold, warning speed threshold, and stay time threshold for the corresponding personnel. Step S4: Pre-build and train an image recognition model for identifying protective facilities in airspace. Based on the airspace image data uploaded in real time by the smart wearable device, detect the actual status information of the hazard sources and corresponding protection status in the target airspace. Write the actual status information into the corresponding ontology instance and trigger the ontology inference engine to execute the inference rules. The inference engine combines the preset ontology model and inference rules to determine the target protection status that needs to be set for the target airspace. Compare the actual status information with the target protection status to determine whether there is a safety deficiency in the target airspace. If so, update the hazard source and update the risk status of the corresponding target airspace in the Unity3D model, and generate the corresponding risk warning trigger area simultaneously. Step S5: Acquire personnel location information in real time, render personnel movement trajectory, and calculate three data points: distance between the personnel and each hazard source in the model at the current moment, approach speed to the current hazard area, and stay time in the current hazard area. When any data reaches the corresponding threshold, trigger an alarm and send risk warning information to the warning feedback module on the corresponding smart wearable device.

[0007] Furthermore, the positioning tag on the smart wearable device is a UWB positioning tag, and the method also includes installing UWB positioning base stations at the construction site, calculating the distance by measuring the TOF of the UWB signal between the positioning tag and multiple positioning base stations, and thus determining the precise location of the target person in three-dimensional space.

[0008] Furthermore, step S3 specifically includes: By periodically accessing the personnel location database information, the location information of personnel within one week can be obtained. The personnel location database information includes the location tag ID, XYZ three-dimensional coordinates, acquisition time, distance between the tag and each base station, and the real-time speed of the personnel's movement. Three indicators were calculated for different personnel: approach speed to the danger zone, time spent in the danger zone, and distance of approach to the danger zone. The movement style of the personnel was constructed and classified into aggressive, conservative, or balanced to distinguish the daily behavior of different work positions. Personalized safety warning rules are formulated based on the individual's movement style. For a balanced movement style, the default distance threshold, speed threshold, and time threshold are used. For an aggressive movement style, the distance threshold and speed threshold are increased by a first percentage from the default threshold, and the time threshold is increased by a second percentage from the default threshold. For a conservative movement style, the distance threshold, speed threshold, and time threshold are decreased by a third percentage from the default threshold.

[0009] Furthermore, step S4 specifically includes: Pre-build and train an image recognition model for identifying protective facilities in the airspace; The system acquires real-time image data of the airspace area collected and uploaded by smart wearable devices, inputs it into the image recognition model, and outputs the actual status information of the hazard sources and corresponding protective facilities within the target airspace area. The target protection status required for the target airspace is determined according to the preset ontology SWRL reasoning rules; The target protection status is compared with the actual status information; When the comparison result indicates that the target airspace should be equipped with protective facilities but has not been equipped with corresponding protective facilities, it is determined that the target airspace has a safety deficiency and the area is marked as a new hazard source. The ontology is updated based on the newly added hazard source, and the risk status of the corresponding target airspace in the Unity3D model is updated to generate a corresponding risk warning trigger zone and issue warning information to safety management personnel.

[0010] Secondly, the present invention provides a fall warning system for construction workers, the system being based on a smart wearable device, the smart wearable device being equipped with a positioning tag, an image acquisition sensor, a transmission module and a warning feedback module, the smart wearable device being connected to a server via a network, and transmitting the location information and image data acquired by the positioning tag and the image acquisition sensor to the server via the transmission module; The system includes: The 3D model building module is used to build a BIM model of the construction site, identify and define the information of the air hazard sources, set protective measures in all hazard source areas by default, and import the Unity3D model; integrate the construction air fall ontology model, access and parse the dynamic update reasoning rules of hazard sources based on SWRL, and form a digital twin base model of construction site hazard sources with logical reasoning capabilities. The location database module is used to collect location data from smart wearable devices in real time and store it in the personnel location database; The safety risk preference assessment module is used to periodically call the personnel location database to obtain historical location information of personnel, calculate three indicators for different personnel: approach speed to danger zone, stay time in danger zone, and approach distance to danger zone, characterize personnel movement style, and formulate personalized safety warning rules, including adjusting the corresponding warning risk area distance threshold, warning speed threshold, and stay time threshold for personnel. The real-time environment update module is used to pre-build and train an image recognition model for identifying protective facilities in airspace. Based on the airspace image data uploaded in real time by the smart wearable device, it detects the actual status information of the hazard sources and corresponding protection status in the target airspace. The actual status information is written into the corresponding ontology instance, and the ontology inference engine is triggered to execute inference rules. The inference engine combines the preset ontology model and inference rules to determine the target protection status that needs to be set for the target airspace. The actual status information is compared with the target protection status to determine whether there is a safety deficiency in the target airspace. If so, the hazard source is updated, and the risk status of the corresponding target airspace in the Unity3D model is updated. The corresponding risk warning trigger area is generated simultaneously. The risk warning module is used to acquire personnel location information in real time, render personnel movement trajectory, and calculate three data points: the distance between the personnel and each hazard source in the model at the current moment, the approach speed to the current hazard area, and the stay time in the current hazard area. When any data reaches the corresponding threshold, an alarm is triggered and risk warning information is sent to the warning feedback module on the corresponding smart wearable device.

[0011] Furthermore, the positioning tag on the smart wearable device is a UWB positioning tag, and the method also includes installing UWB positioning base stations at the construction site, calculating the distance by measuring the TOF of the UWB signal between the positioning tag and multiple positioning base stations, and thus determining the precise location of the target person in three-dimensional space.

[0012] Furthermore, the security risk preference assessment module specifically includes: By periodically accessing the personnel location database information, the location information of personnel within one week can be obtained. The personnel location database information includes the location tag ID, XYZ three-dimensional coordinates, acquisition time, distance between the tag and each base station, and the real-time speed of the personnel's movement. Three indicators were calculated for different personnel: approach speed to the danger zone, time spent in the danger zone, and distance of approach to the danger zone. The movement style of the personnel was constructed and classified into aggressive, conservative, or balanced to distinguish the daily behavior of different work positions. Personalized safety warning rules are formulated based on the individual's movement style. For a balanced movement style, the default distance threshold, speed threshold, and time threshold are used. For an aggressive movement style, the distance threshold and speed threshold are increased by a first percentage from the default threshold, and the time threshold is increased by a second percentage from the default threshold. For a conservative movement style, the distance threshold, speed threshold, and time threshold are decreased by a third percentage from the default threshold.

[0013] Furthermore, the real-time environment update module specifically includes: Pre-build and train an image recognition model for identifying protective facilities in the airspace; The system acquires real-time image data of the airspace area collected and uploaded by smart wearable devices, inputs it into the image recognition model, and outputs the actual status information of the hazard sources and corresponding protective facilities within the target airspace area. The target protection status required for the target airspace is determined according to the preset ontology SWRL reasoning rules; The target protection status is compared with the actual status information; When the comparison result indicates that the target airspace should be equipped with protective facilities but has not been equipped with corresponding protective facilities, it is determined that the target airspace has a safety deficiency and the area is marked as a new hazard source. The ontology is updated based on the newly added hazard source, and the risk status of the corresponding target airspace in the Unity3D model is updated to generate a corresponding risk warning trigger zone and issue warning information to safety management personnel.

[0014] Thirdly, the present invention provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described in the first aspect.

[0015] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in the first aspect.

[0016] One or more technical solutions provided by this invention have at least the following technical effects or advantages: By transforming fall protection strategies from passive to proactive and from fixed to dynamically adjusted, managers can issue early warnings before falls occur, effectively reducing the occurrence of falls from heights during construction. Real-time tracking of the actual protection status of building openings and edges near the located personnel, timely updating of environmental parameters, ensuring that the dangerous areas marked by the Unity3D model match the actual situation on site, avoiding safety misjudgments by the system when the status of edge protection facilities changes, and reducing accidents of falling from heights due to personnel negligence. The combination of positioning technology and Unity3D model enables the early warning system to present and record the interaction between people and the environment in real time, providing data support for early prediction of people falling from buildings and post-accident investigation, which is conducive to improving subsequent management measures. Fall risk indicators comprehensively characterize people's movement styles. Personalized safety early warning rules developed based on these indicators help to provide accurate early warnings of fall risks to construction workers, reduce the number of false alarms, and improve the overall safety level of projects under construction. Attached Figure Description

[0017] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0018] Figure 1 This is a flowchart illustrating the execution of a method for early warning of falls by construction workers in an airspace, as described in Embodiment 1 of the present invention. Figure 2 This is a schematic diagram of the deployment of a UWB positioning system in a specific embodiment of the present invention; Figure 3 This is a schematic diagram of the interaction interface between UWB positioning data and the environment Unity3D model in a specific embodiment of the present invention; Figure 4 This is a UI interface for warning of a person falling from a height in a Unity3D model, as described in a specific embodiment of the present invention. Figure 5 This is a schematic diagram of the danger warning logic judgment in a specific embodiment of the present invention; Figure 6 This is a schematic diagram of a construction worker fall warning system according to Embodiment 2 of the present invention; Figure 7 This is a schematic diagram of the electronic device in Embodiment 3 of the present invention; Figure 8 This is a schematic diagram of the structure of the medium in Embodiment 4 of the present invention. Detailed Implementation

[0019] This application provides a method, system, device, and medium for early warning of falls from heights by construction workers, in order to improve the accuracy of early warning of falls from heights by construction workers.

[0020] The overall approach of the technical solution in this application is as follows: The construction environment is modeled using a BIM model and the Unity3D engine, and logical reasoning is implemented through ontology technology. Smart wearable devices worn by construction workers are used to collect on-site image information and personnel location information. The real-time image data uploaded by the smart wearable devices is transmitted to a target detection backend service. The backend service calls a pre-trained target detection model to detect surrounding hazards and protection status, maps the detection results to the ontology model to update the attribute information of the corresponding instance, and calls an inference engine combined with SWRL rules to perform reasoning, obtaining the hazard information judgment result for the target's airspace. The reasoning result is then updated in real-time to the corresponding position in the Unity3D model. Subsequently, based on the current hazard identification, real-time risk assessment is performed on on-site construction workers. Using pre-acquired historical location data for each person, the movement style of the personnel is analyzed and characterized to formulate personalized risk assessment strategies, improving the accuracy of early warnings and reducing false alarms and missed alarms, thereby achieving precise management of the risk of falls from heights at construction sites.

[0021] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods. Example 1

[0022] like Figures 1 to 5 As shown, this embodiment provides a method for early warning of falls from heights for construction workers. The method is based on a smart wearable device, which is equipped with a positioning tag, an image acquisition sensor, a transmission module, and an early warning feedback module. The smart wearable device is connected to a server via a network, and the transmission module transmits the location information and image data collected by the positioning tag and the image acquisition sensor to the server respectively. The positioning tag is used to collect the location information of construction workers in real time, the image acquisition sensor is used to visually capture the status of on-site protective facilities, the transmission module uses at least one of 5G / Wi-Fi / Bluetooth transmission methods, and the early warning feedback module uses at least one of audible and visual alarms and vibration feedback to output early warning information to relevant personnel on-site. The method includes the following steps: Step S1: Construct a BIM model of the construction site, identify and define the information of the air hazard sources, set protective measures in all hazard source areas by default according to the safety construction design requirements, import the Unity3D model, integrate the construction air fall ontology model, access and parse the hazard source dynamic update reasoning rules based on SWRL, and form a digital twin base model of the construction site hazard sources with logical reasoning capabilities. Step S2: Real-time collection of location data from smart wearable devices and storage in the personnel location database; the location information database stores the personnel location coordinates received by the positioning base station in real time, so that the Unity3D model can read the personnel location coordinates in the database in real time through C# scripts, and monitor the updates of each tag's location data, such as... Figure 3 As shown, it renders the movement trajectory of people in real time; Step S3: By periodically calling the personnel location database information, obtain the personnel's historical location information, calculate three indicators for different personnel: approach speed to danger zone, stay time in danger zone, and approach distance to danger zone, characterize the personnel's movement style, and formulate personalized safety warning rules, including adjusting the warning risk area distance threshold, warning speed threshold, and stay time threshold for the corresponding personnel. Step S4: Pre-build and train an image recognition model for identifying protective facilities in airspace. Based on the airspace image data uploaded in real time by the smart wearable device, detect the actual status information of the hazard sources and corresponding protection status in the target airspace. Write the actual status information of the protective facilities obtained by image recognition into the corresponding airspace hazard source ontology instance and trigger the ontology inference engine to execute SWRL inference rules. The inference engine combines the preset ontology model and SWRL inference rules to determine the target protection status that needs to be set in the target airspace. Compare the actual status information with the target protection status to determine whether there is a safety deficiency (including protection deficiency and protection damage) in the target airspace. If so, update the hazard source and update the risk status of the corresponding target airspace in the Unity3D model, and generate the corresponding risk warning trigger area simultaneously. Step S5: Acquire personnel location information in real time, render personnel movement trajectory, and calculate three data points: distance between the personnel and each hazard source in the model at the current moment, approach speed to the current hazard area, and stay time in the current hazard area. When any data reaches the corresponding threshold, trigger an alarm and send risk warning information to the warning feedback module on the corresponding smart wearable device.

[0023] Preferably, the positioning tag on the smart wearable device is a UWB (Ultra-Wideband) positioning tag, and the method further includes installing UWB positioning base stations at the construction site. Distance is calculated by measuring the Time of Flight (TOF) of the UWB signal between the positioning tag and multiple positioning base stations. After obtaining the distance between the positioning tag and multiple base stations at a certain moment, the precise relative position of the positioning tag is determined using triangulation, thereby determining the precise position of the target person in three-dimensional space. The UWB positioning tag includes a UWB positioning chip and a wireless communication chip, moving with the person. The positioning base station includes a UWB positioning chip and a microcontroller unit responsible for logic control and data processing, fixedly installed at the construction site. The positioning data is input into a Unity3D model to achieve real-time and accurate mapping of the person's position between physical dimensions and digital dimensions.

[0024] Preferably, step S1 specifically includes: constructing a BIM model of the construction site, wherein the data source of the BIM model includes CAD drawings of the construction site, point cloud data of the site, and construction-specific plans; performing component analysis and spatial semantic annotation on the BIM model; identifying and defining the information of the hazard sources in the hazard source according to the preset criteria for judging the hazard source in the air (such as edge height ≥1.2m, opening size ≥0.2m×0.2m, etc.); the hazard source information in the air includes the location coordinates of the hazard source, spatial range, and hazard source type (opening, edge); and automatically generating and configuring standardized protective measures (including guardrails, safety nets, and warning signs) in all the hazard source areas by default. At least one of the following (specifically determined according to actual needs) is considered; the protective measures form a spatial association and constraint relationship with the corresponding hazard source area; the BIM model with the hazard source and protective measures is converted to a format (converted to FBX format recognizable by Unity3D), retaining component IDs, semantic attributes, and spatial coordinates, and imported into the Unity3D engine to form a three-dimensional visualization scene model; a pre-built construction fall-from-height ontology model is integrated, the ontology model adopts a class hierarchy structure, classifying knowledge related to fall-from-height hazards into three categories: building elements, construction activities, and construction resources, with standard measures as the control class, and hasActivity is set. The four object attributes, `hasResource`, `needMeasure`, `hasMeasure`, and `hasResource`, respectively describe the dependency relationship between building elements and construction activities, the relationship between construction activities and the use of construction resources, the relationship between building elements and the protection requirements of regulatory measures, and the relationship between construction resources and the allocation of regulatory measures. This establishes a mapping relationship between BIM components, hazardous sources in open areas, protective measures, and entities, attributes, and relationships in the ontology model. Through four top-level concepts in the ontology model—building elements, construction activities, construction resources, and regulatory measures—this invention establishes clear object attribute relationships to describe the construction scenario from work objects and work behaviors to resource allocation and protection. The system establishes the basic semantic link for protection requirements; it accesses and loads dynamic update inference rules for hazard sources written in the SWRL rule language. This inference rule set includes protection deficiency inference rules and hazard source area intrusion inference rules. The SWRL inference rules are parsed and executed using an ontology inference engine (Pellet inference engine or HermiT inference engine), ultimately forming a digital twin base model of the construction site hazard sources with real-time logical inference and dynamic update capabilities. For example, protection deficiency inference rules include: if there is an opening in the elevator shaft requiring a protective door, but the door is missing, then protection is deemed lacking; if the elevator shaft protective door exists, but its height is less than 1.5 m, then protection is deemed non-standard; if the elevator shaft protective door exists, but its bottom is at least 0.05 m above the ground, then protection is deemed non-standard; if the elevator shaft protective door exists, but the toe board is missing, then protection is deemed incomplete, etc. Specific settings are adjusted according to construction specifications and actual requirements.For the intrusion intrusion reasoning rules for hazardous source areas, trigger zones are created for hazardous sources in Unity 3D. These trigger zones define the hazardous and risk areas of the hazardous source. Upon detecting construction personnel entering these hazardous or risk areas, a UI pop-up notification is automatically triggered, and a risk warning is sent to a smart wearable device. The hazardous area is the core area of ​​the hazardous source, and the risk area is a certain area extending outward from the hazardous area. For example, the red area within 1 meter of the edge is the hazardous area, and the yellow area within 1-2 meters of the edge is the risk area. The system detects personnel entering the risk or hazardous area and automatically displays a UI warning notification upon detection, as shown in the attached image. Figure 4 As shown.

[0025] In a preferred embodiment, a worker movement risk index is introduced to assess the risk of workers falling from near edges and to develop personalized early warning rules. Step S3 specifically includes: By periodically accessing the personnel location database information, the location information of personnel within one week can be obtained. The personnel location database information includes attributes such as location tag ID, XYZ three-dimensional coordinates, acquisition time, distance between the tag and each base station, and real-time speed of personnel movement. Three indicators were calculated for different personnel: approach speed to the danger zone, time spent in the danger zone, and distance of approach to the danger zone. The movement style of the personnel was constructed and classified into aggressive, conservative, or balanced to distinguish the daily behavior of different work positions. Personalized safety warning rules are formulated based on the individual's movement style. For a balanced movement style, the default distance threshold, speed threshold, and time threshold are used. For an aggressive movement style, the distance threshold and speed threshold are increased by a first percentage (e.g., 20%) from the default threshold, and the time threshold is increased by a second percentage (e.g., 50%) from the default threshold. For a conservative movement style, the distance threshold, speed threshold, and time threshold are decreased by a third percentage (e.g., 20%) from the default threshold.

[0026] This invention utilizes historical location data of each construction worker stored in a database, introducing three indicators: approach speed to the danger zone, dwell time in the danger zone, and approach distance to the danger zone, to characterize the worker's movement style. Specifically, it calculates each worker's speed when approaching the danger zone by measuring the distance traveled per second, records their dwell time in the danger zone, and measures their minimum distance from the danger zone. This identifies each worker's movement style and establishes personalized danger zone warning rules, specifically expanding or shrinking the danger zone range to match individual movement styles and safety needs, preventing falls from heights. In practice, this is achieved by assigning corresponding risk and danger zones to each worker in Unity3D and automatically adding colliders to these zones to trigger actions.

[0027] The personnel positioning database is built using a lightweight SQLite database. By embedding C# scripts, it possesses the ability to autonomously monitor and write personnel positioning information transmitted from UWB positioning base stations to a host computer in real time. Positioning tags on smart wearable devices simultaneously send signals to multiple base stations. Each base station calculates the distance to the tag based on the signal transmission time-of-flight. Through a multi-point ranging positioning algorithm combined with a geometric model, the system can accurately determine the tag's location. Finally, the main base station (base station 1) transmits this location information to a computer in real time for processing via wireless communication protocols or wired serial communication protocols, ensuring rapid data updates and accurate analysis. Figure 2 As shown, four positioning base stations, installed in spherically or approximately spherically symmetric positions, were configured at the construction site. Construction workers wearing smart safety helmets with UWB positioning tags were able to achieve 3D positioning. The main base station of the positioning system continuously collected signals and automatically created data tables in an SQLite database using C# scripts. The system assigned a unique ID to each person and generated a corresponding database table based on that ID. The table contained multiple fields, including time, UWB positioning tag ID, the person's X, Y, and Z coordinates, the distance between each tag and each base station, and the person's real-time movement speed. Appropriate data types were defined for each field's attributes: the time field was set to TEXT type, the tag ID and distance fields to INT type, and the coordinate and speed fields to DOUBLE type, ensuring data storage accuracy and efficient querying.

[0028] In a preferred embodiment, machine vision technology is used to track the protection status of the edges at the construction site and update the hazardous areas in the BIM model. Step S4 specifically includes: Pre-build and train an image recognition model for identifying protective facilities in the airspace; The system acquires real-time image data of the airspace area collected and uploaded by smart wearable devices, inputs it into the image recognition model, and outputs the actual status information of the hazard sources and corresponding protective facilities within the target airspace area. The target protection status required for the target airspace is determined according to the preset ontology SWRL reasoning rules; The target protection status is compared with the actual status information; When the comparison result indicates that the target airspace should be equipped with protective facilities but has not been equipped with corresponding protective facilities, it is determined that the target airspace has a safety deficiency and the area is marked as a new hazard source. The ontology is updated based on the newly added hazard source, and the risk status of the corresponding target airspace in the Unity3D model is updated to generate the corresponding risk warning trigger zone and issue warning information to safety management personnel. The image recognition model is built using the YOLOv5 model. The Unity3D model can map the coordinates of the hazard source to the corresponding coordinate structure in the BIM model to determine its coordinates in the Unity3D model, so as to accurately update the hazard triggering area of ​​the Unity3D model. The hazard source recognition model is built using the YOLOv5 model and trained using pre-collected images marked with hazard sources and protective facilities. This enables it to quickly identify the status of protective facilities for hazard sources in the target area and obtain the status of protection settings for edges and openings. Specifically, a sufficient number of edge protection-related images are collected from construction sites in different environments. The images are preprocessed and labeled to build training and validation sets. During training and testing, the network parameters, such as the weights of localization loss, confidence loss, and classification loss, are fine-tuned based on the model's performance. Finally, an efficient image recognition model is trained. This image recognition model is only used to identify whether a specific hazard source exists in the image.

[0029] Before performing recognition based on real-time uploaded image data, the aforementioned image recognition model also includes image position alignment. This is achieved by mapping the corresponding structural attributes in the BIM model to obtain the actual coordinates of the hazard source in the image. The Unity3D model can then map the hazard source coordinates to the corresponding coordinate structure in the BIM model to determine its coordinates within the Unity3D model, thereby accurately updating the hazard trigger zone in the Unity3D model.

[0030] Preferably, step S5 specifically includes: the server acquiring the location information of construction workers uploaded by the smart wearable device in real time, synchronizing the location information to the digital twin base model, and rendering the real-time movement trajectory of the construction workers; simultaneously, based on the current location information, calculating three data points: the actual distance between the worker and each hazard source in the model at the current moment, the approach speed to the current hazard area, and the dwell time in the current hazard area; comparing the three data points with the thresholds corresponding to the personalized safety warning rules defined in step S3, and when any data point reaches or exceeds the corresponding threshold, the server immediately triggers an alarm mechanism, generates risk warning information, and sends it to the warning feedback module on the corresponding smart wearable device through the transmission module. The warning feedback module outputs the warning information in a preset manner to remind the construction workers to avoid risks. In this step, by embedding a data interface written in C# into the Unity3D model, the personnel location database is read in real time. Combining the advantages of Unity3D's real-time rendering and interaction, the information of personnel who have entered the risk area is displayed on the UI interface, simulating the scene of personnel interacting with the environment in real time and generating feedback information.

[0031] Based on the same inventive concept, this application also provides a system corresponding to the method in Embodiment 1, as detailed in Embodiment 2. Example 2

[0032] like Figure 6 As shown, this embodiment provides a construction worker fall warning system. The system is based on a smart wearable device. The smart wearable device is equipped with a positioning tag, an image acquisition sensor, a transmission module, and a warning feedback module. The smart wearable device is connected to the server via a network, and the transmission module transmits the location information and image data collected by the positioning tag and the image acquisition sensor to the server respectively. The system includes: The 3D model building module is used to build a BIM model of the construction site, identify and define the information of the air hazard sources, set protective measures in all hazard source areas by default according to the safety construction design requirements, import the Unity3D model, integrate the construction air fall ontology model, access and parse the dynamic update reasoning rules of the hazard sources based on SWRL, and form a digital twin base model of the construction site hazard sources with logical reasoning capabilities. The location database module is used to collect location data from smart wearable devices in real time and store it in a personnel location database. It stores the location coordinates of personnel received from location base stations in real time through the location information database, so that Unity3D models can read the personnel location coordinates from the database in real time via C# scripts, and monitor the updates of each tag's location data, such as... Figure 3 As shown, it renders the movement trajectory of people in real time; The safety risk preference assessment module is used to periodically call the personnel location database to obtain historical location information of personnel, calculate three indicators for different personnel: approach speed to danger zone, stay time in danger zone, and approach distance to danger zone, characterize personnel movement style, and formulate personalized safety warning rules, including adjusting the corresponding warning risk area distance threshold, warning speed threshold, and stay time threshold for personnel. The real-time environment update module is used to pre-build and train an image recognition model for identifying protective facilities in airspace. Based on the airspace image data uploaded in real time by the smart wearable device, it detects the actual status information of the hazard sources and corresponding protection status in the target airspace. The actual status information is written into the corresponding ontology instance, and the ontology inference engine is triggered to execute inference rules. The inference engine combines the preset ontology model and inference rules to determine the target protection status that needs to be set for the target airspace. The actual status information is compared with the target protection status to determine whether there is a safety deficiency (including protection deficiency and protection damage) in the target airspace. If so, the hazard source is updated, and the risk status of the corresponding target airspace in the Unity3D model is updated. The corresponding risk warning trigger zone is generated simultaneously. The risk warning module is used to acquire personnel location information in real time, render personnel movement trajectory, and calculate three data points: the distance between the personnel and each hazard source in the model at the current moment, the approach speed to the current hazard area, and the stay time in the current hazard area. When any data reaches the corresponding threshold, an alarm is triggered and risk warning information is sent to the warning feedback module on the corresponding smart wearable device.

[0033] At the same time, the UI interface displays information about personnel who have entered the risk area.

[0034] Preferably, the positioning tag on the smart wearable device is a UWB (Ultra-Wideband) positioning tag, and the method further includes installing UWB positioning base stations at the construction site. Distance is calculated by measuring the Time of Flight (TOF) of the UWB signal between the positioning tag and multiple positioning base stations. After obtaining the distance between the positioning tag and multiple base stations at a certain moment, the precise relative position of the positioning tag is determined using triangulation, thereby determining the precise position of the target person in three-dimensional space. The UWB positioning tag includes a UWB positioning chip and a wireless communication chip, moving with the person. The positioning base station includes a UWB positioning chip and a microcontroller unit responsible for logic control and data processing, fixedly installed at the construction site. The positioning data is input into a Unity3D model to achieve real-time and accurate mapping of the person's position between physical dimensions and digital dimensions.

[0035] Preferably, the 3D model construction module specifically includes: constructing a BIM model of the construction site, wherein the data source of the BIM model includes CAD drawings of the construction site, point cloud data of the site, and construction-specific plans; performing component analysis and spatial semantic annotation on the BIM model; identifying and defining the information of the hazard sources in the airspace according to preset criteria for judging hazard sources in the airspace (such as edge height ≥ 1.2m, opening size ≥ 0.2m × 0.2m, etc.); the hazard source information in the airspace includes the location coordinates of the hazard source, spatial range, and hazard source type (opening, edge); and automatically generating and configuring standardized protective measures (including guardrails, safety nets, etc.) in all hazard source areas by default. At least one of the warning signs (specifically determined according to actual needs), the protective measures form a spatial association and constraint relationship with the corresponding hazard source area; the BIM model with the hazard source and protective measures is converted to a format (converted to FBX format recognizable by Unity3D), retaining the component ID, semantic attributes and spatial coordinates, and imported into the Unity3D engine to form a three-dimensional visualization scene model; a pre-built construction fall-from-height ontology model is integrated, the ontology model adopts a class hierarchy structure, classifying the knowledge related to fall-from-height hazards into three categories: building elements, construction activities, and construction resources, with standard measures as the control class, and hasActivit is set. The four object attributes, y, hasResource, needMeasure, and hasMeasure, respectively describe the dependency relationship between building elements and construction activities, the relationship between construction activities and the use of construction resources, the relationship between building elements and the protection requirements of regulatory measures, and the relationship between construction resources and the allocation of regulatory measures. This establishes a mapping relationship between BIM components, hazardous sources in open areas, protective measures, and entities, attributes, and relationships in the ontology model. This invention establishes clear object attribute relationships between four top-level concepts in the ontology model—building elements, construction activities, construction resources, and regulatory measures—to describe the work objects, work behaviors, resource allocation, and... The basic semantic link of protection requirements; access and load dynamic update inference rules for hazard sources written in the SWRL rule language. The inference rule set includes protection deficiency inference rules and hazard source area intrusion inference rules. The SWRL inference rules are parsed and executed by the ontology inference engine (Pellet inference engine or HermiT inference engine) to finally form a digital twin base model of construction site hazard sources with real-time logical inference and dynamic update capabilities. For example, protection deficiency inference rules include: if there is an opening in the elevator shaft and a protective door is required, but the protective door is not present, then protection is deemed lacking; if the elevator shaft protective door exists, but its height is less than 1.5 m, then protection is deemed non-standard; if the elevator shaft protective door exists, but its bottom is greater than or equal to 0.05 m above the ground, then protection is deemed non-standard; if the elevator shaft protective door exists, but the toe board is missing, then protection is deemed incomplete, etc. The specific settings are adjusted according to construction specifications and actual requirements.For the intrusion intrusion reasoning rules for hazardous source areas, trigger zones are created for hazardous sources in Unity3D. These trigger zones define the hazardous and risk areas of the hazardous source. Upon detecting construction personnel entering these hazardous or risk areas, a UI pop-up notification is automatically triggered, and a risk warning is sent to a smart wearable device. The hazardous area is the core area of ​​the hazardous source, and the risk area is a certain area extending outward from the hazardous area. For example, the red area within 1 meter of the edge is the hazardous area, and the yellow area within 1-2 meters of the edge is the risk area. The system detects personnel entering the risk or hazardous area and automatically displays a UI warning notification upon detection, as shown in the attached diagram. Figure 4 As shown.

[0036] Preferably, the security risk preference assessment module specifically includes: By periodically accessing the personnel location database information, the location information of personnel within one week can be obtained. The personnel location database information includes attributes such as location tag ID, XYZ three-dimensional coordinates, acquisition time, distance between the tag and each base station, and real-time speed of personnel movement. Three indicators were calculated for different personnel: approach speed to the danger zone, time spent in the danger zone, and distance of approach to the danger zone. The movement style of the personnel was constructed and classified into aggressive, conservative, or balanced to distinguish the daily behavior of different work positions. Personalized safety warning rules are formulated based on the individual's movement style. For a balanced movement style, the default distance threshold, speed threshold, and time threshold are used. For an aggressive movement style, the distance threshold and speed threshold are increased by a first percentage (e.g., 20%) from the default threshold, and the time threshold is increased by a second percentage (e.g., 50%) from the default threshold. For a conservative movement style, the distance threshold, speed threshold, and time threshold are decreased by a third percentage (e.g., 20%) from the default threshold.

[0037] Preferably, machine vision technology is used to track the protection status of the edges at the construction site and update the hazardous areas in the BIM model. The real-time environment update module specifically includes: Pre-build and train an image recognition model for identifying protective facilities in the airspace; The system acquires real-time image data of the airspace area collected and uploaded by smart wearable devices, inputs it into the image recognition model, and outputs the actual status information of the hazard sources and corresponding protective facilities within the target airspace area. The target protection status required for the target airspace is determined according to the preset ontology SWRL reasoning rules; The target protection status is compared with the actual status information; When the comparison result indicates that the target airspace should be equipped with protective facilities but has not been equipped with corresponding protective facilities, it is determined that the target airspace has a safety deficiency and the area is marked as a new hazard source. The ontology is updated based on the newly added hazard source, and the risk status of the corresponding target airspace in the Unity3D model is updated to generate the corresponding risk warning trigger zone and issue warning information to safety management personnel. The image recognition model is built using the YOLOv5 model. The Unity3D model can map the coordinates of the hazard source to the corresponding coordinate structure in the BIM model to determine its coordinates in the Unity3D model, so as to accurately update the hazard triggering area of ​​the Unity3D model. The hazard source recognition model is built using the YOLOv5 model and trained using pre-collected images marked with hazard sources and protective facilities. This enables it to quickly identify the status of protective facilities for hazard sources in the target area and obtain the status of protective settings for edges and openings. Specifically, a sufficient number of edge protection-related images are collected from construction sites in different environments. The images are preprocessed and labeled to build training and validation sets. During training and testing, the network parameters, such as the weights of localization loss, confidence loss, and classification loss, are fine-tuned based on the model performance to finally train an efficient image recognition model. This image recognition model is only used to identify whether a specific hazard source exists in the image. A simplified working interface for the detection module developed based on PySide6 is shown below. Figure 4 As shown.

[0038] Before performing recognition based on real-time uploaded image data, the aforementioned image recognition model also includes image position alignment. This is achieved by mapping the corresponding structural attributes in the BIM model to obtain the actual coordinates of the hazard source in the image. The Unity3D model can then map the hazard source coordinates to the corresponding coordinate structure in the BIM model to determine its coordinates within the Unity3D model, thereby accurately updating the hazard trigger zone in the Unity3D model.

[0039] Preferably, the risk warning module specifically includes: the server acquiring the location information of construction workers uploaded by the smart wearable device in real time, synchronizing the location information to the digital twin base model, and rendering the real-time movement trajectory of the construction workers; simultaneously, based on the current location information, calculating three data points: the actual distance between the person and each hazard source in the model at the current moment, the approach speed to the current hazard area, and the stay time in the current hazard area; comparing the three data points with the thresholds corresponding to the personalized safety warning rules defined in step S3, and when any data point reaches or exceeds the corresponding threshold, the server immediately triggers an alarm mechanism, generates risk warning information, and sends it to the warning feedback module on the corresponding smart wearable device through the transmission module, whereby the warning feedback module outputs the warning information in a preset manner to remind the construction workers to avoid risks. In this step, by embedding a data interface written in C# into the Unity3D model, the personnel location database is read in real time. Combining the advantages of Unity3D's real-time rendering and interaction, the information of personnel who have entered the risk area is displayed on the UI interface, simulating the scene of personnel interacting with the environment in real time and generating feedback information.

[0040] Since the apparatus described in Embodiment 2 of the present invention is an apparatus used to implement the method of Embodiment 1 of the present invention, those skilled in the art can understand the specific structure and variations of the apparatus based on the method described in Embodiment 1 of the present invention, and therefore will not be described again here. All apparatuses used in the method of Embodiment 1 of the present invention fall within the scope of protection of the present invention.

[0041] Based on the same inventive concept, this application provides an electronic device embodiment corresponding to Embodiment 1, as detailed in Embodiment 3. Example 3

[0042] This embodiment provides an electronic device, such as... Figure 7 As shown, it includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it can implement any of the embodiments in Example 1.

[0043] Since the electronic device described in this embodiment is the device used to implement the method in Embodiment 1 of this application, those skilled in the art can understand the specific implementation method and various variations of the electronic device in this embodiment based on the method described in Embodiment 1 of this application. Therefore, how the electronic device implements the method in the embodiment of this application will not be described in detail here. Any device used by those skilled in the art to implement the method in the embodiment of this application falls within the scope of protection of this application.

[0044] Based on the same inventive concept, this application provides a storage medium corresponding to Embodiment 1, as detailed in Embodiment 4. Example 4

[0045] This embodiment provides a computer-readable storage medium, such as... Figure 8 As shown, a computer program is stored thereon, which, when executed by a processor, can implement any of the embodiments in Example 1.

[0046] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0047] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, systems, and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0048] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0049] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0050] The technical solutions provided in this application embodiment have at least the following technical effects or advantages: The present invention combines technologies such as BIM, real-time animation interaction, UWB positioning and machine vision, which can accurately acquire and display the three-dimensional position information of personnel in real time, update the edge setting status, ensure the environment is real and effective, and formulate personalized early warning rules in combination with personnel movement risk indicators, effectively reduce the risk of personnel falling from heights and improve the safety management level of construction sites.

[0051] While specific embodiments of the present invention have been described above, those skilled in the art should understand that the specific embodiments described are merely illustrative and not intended to limit the scope of the invention. Equivalent modifications and variations made by those skilled in the art in accordance with the spirit of the invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for early warning of falls from heights for construction workers, characterized in that: The method is implemented based on a smart wearable device, which is equipped with a positioning tag, an image acquisition sensor, a transmission module, and an early warning feedback module. The smart wearable device is connected to the server via a network, and the transmission module transmits the location information and image data collected by the positioning tag and the image acquisition sensor to the server respectively. The method includes the following steps: Step S1: Construct a BIM model of the construction site, identify and define the information of the air hazard sources, set protective measures in all hazard source areas by default, import the Unity3D model, integrate the construction air fall ontology model, access and parse the dynamic update reasoning rules of the hazard sources based on SWRL, and form a digital twin base model of the construction site hazard sources with logical reasoning capabilities. Step S2: Collect location data from smart wearable devices in real time and store it in the personnel location database; Step S3: By periodically calling the personnel location database information, obtain the personnel's historical location information, calculate three indicators for different personnel: approach speed to danger zone, stay time in danger zone, and approach distance to danger zone, characterize the personnel's movement style, and formulate personalized safety warning rules, including adjusting the warning risk area distance threshold, warning speed threshold, and stay time threshold for the corresponding personnel. Step S4: Pre-build and train an image recognition model for identifying protective facilities in airspace. Based on the airspace image data uploaded in real time by the smart wearable device, detect the actual status information of the hazard sources and corresponding protection status in the target airspace. Write the actual status information into the corresponding ontology instance and trigger the ontology inference engine to execute the inference rules. The inference engine combines the preset ontology model and inference rules to determine the target protection status that needs to be set for the target airspace. Compare the actual status information with the target protection status to determine whether there is a safety deficiency in the target airspace. If so, update the hazard source and update the risk status of the corresponding target airspace in the Unity3D model, and generate the corresponding risk warning trigger area simultaneously. Step S5: Acquire personnel location information in real time, render personnel movement trajectory, and calculate three data points: distance between the personnel and each hazard source in the model at the current moment, approach speed to the current hazard area, and stay time in the current hazard area. When any data reaches the corresponding threshold, trigger an alarm and send risk warning information to the warning feedback module on the corresponding smart wearable device.

2. The method for early warning of falls from heights for construction workers according to claim 1, characterized in that: The positioning tag on the smart wearable device is a UWB positioning tag, and the method also includes installing UWB positioning base stations at the construction site, calculating the distance by measuring the TOF of the UWB signal between the positioning tag and multiple positioning base stations, and thus determining the precise location of the target person in three-dimensional space.

3. The method for early warning of falls from heights for construction workers according to claim 1, characterized in that: Step S3 specifically involves: By periodically accessing the personnel location database information, the location information of personnel within one week can be obtained. The personnel location database information includes the location tag ID, XYZ three-dimensional coordinates, acquisition time, distance between the tag and each base station, and the real-time speed of the personnel's movement. Three indicators were calculated for different personnel: approach speed to the danger zone, time spent in the danger zone, and distance of approach to the danger zone. The movement style of the personnel was constructed and classified into aggressive, conservative, or balanced to distinguish the daily behavior of different work positions. Personalized safety warning rules are formulated based on the individual's movement style. For a balanced movement style, the default distance threshold, speed threshold, and time threshold are used. For an aggressive movement style, the distance threshold and speed threshold are increased by a first percentage from the default threshold, and the time threshold is increased by a second percentage from the default threshold. For a conservative movement style, the distance threshold, speed threshold, and time threshold are decreased by a third percentage from the default threshold.

4. The method for early warning of falls from heights for construction workers according to claim 1, characterized in that: Step S4 specifically includes: Pre-build and train an image recognition model for identifying protective facilities in the airspace; The system acquires real-time image data of the airspace area collected and uploaded by smart wearable devices, inputs it into the image recognition model, and outputs the actual status information of the hazard sources and corresponding protective facilities within the target airspace area. The target protection status required for the target airspace is determined according to the preset ontology SWRL reasoning rules; The target protection status is compared with the actual status information; When the comparison result indicates that the target airspace should be equipped with protective facilities but has not been equipped with corresponding protective facilities, it is determined that the target airspace has a safety deficiency and the area is marked as a new hazard source. The ontology is updated based on the newly added hazard source, and the risk status of the corresponding target airspace in the Unity3D model is updated to generate a corresponding risk warning trigger zone and issue warning information to safety management personnel.

5. A fall warning system for construction workers, characterized in that: The system is based on a smart wearable device, which is equipped with a positioning tag, an image acquisition sensor, a transmission module, and an early warning feedback module. The smart wearable device is connected to the server via a network and transmits the location information and image data collected by the positioning tag and the image acquisition sensor to the server through the transmission module. The system includes: The 3D model building module is used to build a BIM model of the construction site, identify and define the information of the air hazard sources, set protective measures in all hazard source areas by default, and import the Unity3D model; integrate the construction air fall ontology model, access and parse the dynamic update reasoning rules of hazard sources based on SWRL, and form a digital twin base model of construction site hazard sources with logical reasoning capabilities. The location database module is used to collect location data from smart wearable devices in real time and store it in the personnel location database; The safety risk preference assessment module is used to periodically call the personnel location database to obtain historical location information of personnel, calculate three indicators for different personnel: approach speed to danger zone, stay time in danger zone, and approach distance to danger zone, characterize personnel movement style, and formulate personalized safety warning rules, including adjusting the corresponding warning risk area distance threshold, warning speed threshold, and stay time threshold for personnel. The real-time environment update module is used to pre-build and train an image recognition model for identifying protective facilities in airspace. Based on the airspace image data uploaded in real time by the smart wearable device, it detects the actual status information of the hazard sources and corresponding protection status in the target airspace. The actual status information is written into the corresponding ontology instance, and the ontology inference engine is triggered to execute inference rules. The inference engine combines the preset ontology model and inference rules to determine the target protection status that needs to be set for the target airspace. The actual status information is compared with the target protection status to determine whether there is a safety deficiency in the target airspace. If so, the hazard source is updated, and the risk status of the corresponding target airspace in the Unity3D model is updated. The corresponding risk warning trigger area is generated simultaneously. The risk warning module is used to acquire personnel location information in real time, render personnel movement trajectory, and calculate three data points: the distance between the personnel and each hazard source in the model at the current moment, the approach speed to the current hazard area, and the stay time in the current hazard area. When any data reaches the corresponding threshold, an alarm is triggered and risk warning information is sent to the warning feedback module on the corresponding smart wearable device.

6. A fall warning system for construction workers according to claim 5, characterized in that: The positioning tag on the smart wearable device is a UWB positioning tag, and the system also includes installing UWB positioning base stations at the construction site. The distance is calculated by measuring the Time-of-Flight (TOF) of the UWB signal between the positioning tag and multiple positioning base stations, thereby determining the precise location of the target person in three-dimensional space.

7. A fall warning system for construction workers according to claim 5, characterized in that: The security risk preference assessment module specifically includes: By periodically accessing the personnel location database information, the location information of personnel within one week can be obtained. The personnel location database information includes the location tag ID, XYZ three-dimensional coordinates, acquisition time, distance between the tag and each base station, and the real-time speed of the personnel's movement. Three indicators were calculated for different personnel: approach speed to the danger zone, time spent in the danger zone, and distance of approach to the danger zone. The movement style of the personnel was constructed and classified into aggressive, conservative, or balanced to distinguish the daily behavior of different work positions. Personalized safety warning rules are formulated based on the individual's movement style. For a balanced movement style, the default distance threshold, speed threshold, and time threshold are used. For an aggressive movement style, the distance threshold and speed threshold are increased by a first percentage from the default threshold, and the time threshold is increased by a second percentage from the default threshold. For a conservative movement style, the distance threshold, speed threshold, and time threshold are decreased by a third percentage from the default threshold.

8. A fall warning system for construction workers according to claim 5, characterized in that: The real-time environment update module specifically includes: Pre-build and train an image recognition model for identifying protective facilities in the airspace; The system acquires real-time image data of the airspace area collected and uploaded by smart wearable devices, inputs it into the image recognition model, and outputs the actual status information of the hazard sources and corresponding protective facilities within the target airspace area. The target protection status required for the target airspace is determined according to the preset ontology SWRL reasoning rules; The target protection status is compared with the actual status information; When the comparison result indicates that the target airspace should be equipped with protective facilities but has not been equipped with corresponding protective facilities, it is determined that the target airspace has a safety deficiency and the area is marked as a new hazard source. The ontology is updated based on the newly added hazard source, and the risk status of the corresponding target airspace in the Unity3D model is updated to generate a corresponding risk warning trigger zone and issue warning information to safety management personnel.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method as described in any one of claims 1 to 4.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1 to 4.

Images