Crane operating area out-of-bound early warning method and system based on three-dimensional fence

Abstract

The application discloses a crane operating area overrunning early warning method and system based on a three-dimensional fence, belongs to the technical field of crane safety monitoring, captures full attitude data of the crane through a distributed monitoring mechanism, and constructs a three-dimensional scene model to set an initial fence, which covers a dangerous area and an operating range, reduces a monitoring blind area, dynamically updates the fence according to engineering progress and environmental data, avoids the defect that a fixed boundary is disconnected from construction advancement and environmental changes, ensures that a safety threshold always matches actual risks under different working conditions, predicts a future trajectory of the crane, judges risks in combination with the dynamic fence, compared with a traditional trigger type alarm, the early warning speed is faster, sufficient time is reserved for risk disposal, so that the probability of equipment collision and crane motion overrun accidents is greatly reduced, and the safety of crane operation is improved.

Description

Technical Field

[0001] This invention relates to the field of crane safety monitoring technology, specifically to a method and system for early warning of crane operation area boundary crossing based on a three-dimensional fence. Background Technology

[0002] With the increasing scale and complexity of construction projects, crane operation areas are often mixed with human-machine interaction. Traditional physical barriers are easily damaged or ignored, leading to frequent collisions and accidents, such as crane booms hitting buildings or accidentally entering high-voltage power line areas. Furthermore, crane height control relies on manual judgment, posing a high risk of exceeding limits. As buildings become larger, crane collisions and height exceeding accidents are increasing year by year. To reduce the probability of crane accidents, many regions have mandated real-time crane monitoring and electronic fence collision prevention in recent years, and have also introduced new regulations to require heavy machinery to be equipped with positioning safety terminals. However, existing monitoring methods mostly rely on positioning systems to locate the position of various parts of the crane and judge whether the crane's movements exceed limits based on set safety thresholds. If limits are exceeded, an alarm is triggered. However, in actual operation, crane movements are affected by environmental factors and project progress. Static thresholds are difficult to adapt to environmental changes and construction progress, and alarms are mainly triggered by events, lacking a risk prediction mechanism. When a crane rapidly approaches a dangerous area, timely warnings are difficult to provide, potentially leading to the escalation of accidents in emergency situations.

[0003] Chinese Patent Publication No. CN116281682A discloses a method and system for safety monitoring of oilfield cranes. This method determines thermal faults in power cables during operation by using thermal infrared image detection. It acquires various parameters of the boom and hook, calculates the real-time operating area radius R of the boom, and determines the distance L1 between the boom and obstacles. It also determines whether the real-time rotation angle β of the hook is within the corresponding angle range ε ​​of the obstacle, and whether the real-time height h of the hook is higher than the height of the obstacle. If these factors affect crane operation, a collision warning is activated, enabling all-weather automatic monitoring of the crane and preventing collisions. The method also acquires real-time wind speed data; if the wind speed exceeds a preset limit, it determines that the wind speed will affect crane operation and activates a warning. While this method considers the impact of dynamic environmental changes on the warning effect to some extent, it still uses a trigger-based warning system, making it difficult to provide timely warnings when the crane rapidly approaches a danger zone. Summary of the Invention

[0004] This invention addresses the problem that most existing crane safety monitoring methods rely on trigger-based early warnings, which are slow to respond to changes in the environment and project progress. It provides a crane operation area boundary crossing early warning method and system based on a 3D fence. This system captures the crane's full-attitude data through a distributed monitoring mechanism and constructs a 3D scene model to set an initial fence, covering hazardous areas and the work area, reducing monitoring blind spots. Simultaneously, the fence is dynamically updated based on project progress and environmental data, avoiding the shortcomings of fixed boundaries being out of sync with construction progress and environmental changes. This ensures that safety thresholds always match actual risks under different working conditions and predicts the crane's future trajectory. Combined with dynamic fence risk assessment, this provides a faster early warning speed compared to traditional trigger-based alarms, allowing sufficient time for risk handling. This significantly reduces the probability of equipment collisions and crane movement exceeding limits, thereby improving the safety of crane operations.

[0005] In a first aspect, one technical solution provided in this embodiment of the invention is: a method for early warning of crane operation area boundary crossing based on a three-dimensional fence, comprising the following steps:

[0006] S1. Based on the distributed monitoring mechanism, monitor the crane's operating status in real time to obtain the crane's current posture data and collect the current environmental data; construct a three-dimensional scene model of the construction area and set initial three-dimensional fences based on different area types;

[0007] S2. Based on the dynamic adjustment mechanism of the three-dimensional fence and combined with the project progress and current environmental data, the initial three-dimensional fence is updated in real time to obtain a dynamic fence; a crane attitude prediction model is constructed, and the crane's current attitude data and current environmental data are used as model inputs to obtain the crane's predicted attitude;

[0008] S3. Determine the risk level based on the predicted attitude of the crane and the positional relationship of the target fence, and execute early warning actions based on the risk level.

[0009] In this solution, a distributed monitoring mechanism captures the attitude data of various parts of the crane, thereby capturing the overall attitude of the crane. Simultaneously, a 3D scene model is constructed and an initial fence is set, covering hazardous areas and the work area, thus eliminating blind spots in traditional manual monitoring. By linking project progress with real-time environmental data such as wind speed and visibility, the 3D fence is dynamically updated, solving the problems of fixed boundaries and disconnection from scene changes in traditional monitoring, ensuring that safety thresholds adapt to working conditions in real time. A crane attitude prediction model is constructed to predict the crane's future trajectory, enabling proactive risk assessment and providing ample response time compared to traditional trigger-based alarms. By executing corresponding risk actions according to risk levels, excessive alarms can be avoided from interfering with actual operations, ensuring construction stability.

[0010] Preferably, in S1, the crane's current attitude data is obtained by real-time monitoring of the crane's operating status based on a distributed monitoring mechanism, including the following steps:

[0011] Monitoring nodes are installed at the top, middle and hook of the crane boom, and all monitoring nodes are located on the same longitudinal axis of the boom.

[0012] The spatial coordinates of the monitoring node are determined based on satellite positioning signals. The current vibration acceleration and swing angular velocity of the corresponding part of the boom are determined based on the changes in the spatial coordinates as the current attitude data of the crane.

[0013] In this solution, by deploying monitoring nodes on the top, middle and hook of the boom along the same axis, the motion status of each key part of the crane can be comprehensively captured, avoiding the one-sidedness of single-point monitoring. By accurately acquiring the spatial coordinates of the nodes and the derived vibration acceleration and swing angular velocity through satellite positioning, centimeter-level attitude perception can be achieved. This allows for the output of comprehensive and high-precision attitude data, providing reliable input for dynamic fence updates and attitude prediction, and providing a data foundation for subsequent active prevention and control.

[0014] Preferably, in S1, a three-dimensional scene model of the construction area is constructed, and initial three-dimensional fences are set based on different area types, including the following steps:

[0015] The construction site 3D model is obtained by collecting 3D point cloud information of the construction area using LiDAR and performing 3D modeling of the construction area according to a fixed ratio.

[0016] Fixed hazardous objects in the 3D model of the construction site are marked with different colors according to their degree of danger, and corresponding initial safety thresholds are set for fixed hazardous objects of different colors to generate initial 3D fences.

[0017] The fixed hazardous objects include high-voltage lines, buildings, foundation pits, work areas, and personnel access routes.

[0018] In this solution, three-dimensional point cloud information is collected by LiDAR and modeled at a fixed scale to accurately restore the terrain and distribution of hazardous objects in the construction area, thereby generating a 1:1 three-dimensional model of the construction site. This eliminates the boundary errors and blind spots of traditional manual calibration. By color-coding fixed hazardous objects such as high-voltage lines and foundation pits according to their degree of danger, the risk levels are intuitively distinguished. At the same time, corresponding initial safety thresholds are set for hazardous objects of different colors, so that the initial three-dimensional fence fits the actual risk differences. This lays a precise foundation for subsequent dynamic adjustments based on the environment and progress, meeting the core requirements of comprehensive safety management.

[0019] As a preferred embodiment, in S2, a dynamic fence is obtained by updating the initial three-dimensional fence in real time based on the dynamic adjustment mechanism of the three-dimensional fence and in combination with the project progress and current environmental data, including the following steps:

[0020] Real-time collection of construction data from the construction area;

[0021] The environmental data includes wind speed and visibility, and the construction data includes the completed height of the floor pouring, the installation location of components, and the relocation information of the material storage area.

[0022] The current construction scenario is determined based on environmental and construction data. The corresponding dynamic threshold calculation rules are matched based on the current construction scenario to obtain the threshold adjustment value. The initial three-dimensional fence is updated in real time based on the threshold adjustment value to obtain the dynamic fence.

[0023] This solution comprehensively captures dynamic factors affecting fence control by collecting real-time environmental data such as wind speed and visibility, as well as construction data such as floor pouring height and material stacking area relocation. This avoids adjustment deviations caused by relying on a single data source. By accurately determining the current construction scenario based on multi-dimensional data and matching corresponding dynamic threshold calculation rules, fence adjustments are made scientifically based rather than blindly changed. At the same time, the initial three-dimensional fence is updated in real time to obtain a dynamic fence, which completely solves the pain point of traditional fixed fences being out of touch with construction progress and environmental changes. This ensures that safety thresholds always fit the actual working conditions and provides accurate boundary references for subsequent crane attitude prediction and risk level determination.

[0024] Preferably, the construction scenarios include conventional operation scenarios, harsh environment scenarios, schedule change scenarios, human-machine intensive scenarios, and mixed scenarios;

[0025] The harsh environment scenario matches the dynamic threshold calculation rules, generates wind speed influence coefficient and visibility influence coefficient based on changes in wind speed and visibility, and updates the initial safety threshold in the form of linear correlation.

[0026] The progress change scenario matches the progress dynamic threshold calculation rule, and the initial safety threshold is directly updated based on the construction data.

[0027] The dynamic threshold calculation rule for matching density in human-machine dense scenes generates a safety buffer with a radius of r centered on the boundary of the human-machine dense area, and extends the initial safety threshold to the boundary of the safety buffer.

[0028] The mixed scenario calculates a comprehensive adjustment value by weighting and summing the initial security threshold adjustment values ​​of each single-type scenario, and then adjusts the initial security threshold based on the comprehensive adjustment value.

[0029] This solution categorizes construction scenarios into various types, such as routine, harsh environments, and schedule changes, and matches specific dynamic threshold calculation rules accordingly. This avoids a one-size-fits-all approach to fence adjustments, thereby improving the accuracy of scenario adaptation and ensuring that threshold adjustments in different scenarios are scientifically aligned with actual risks. By combining classification adaptation with comprehensive calculation, the solution achieves precise iteration of dynamic fence thresholds, providing a reliable boundary reference for risk prediction and hierarchical linkage, and enhancing the pertinence and effectiveness of comprehensive safety management.

[0030] Preferably, in step S2, a crane attitude prediction model is constructed, using the current crane attitude data and current environmental data as model inputs to obtain the predicted crane attitude, including the following steps:

[0031] The historical operation data of n cranes of the same type and the environmental data of the corresponding time period are correlated as experimental samples;

[0032] The model aims to obtain the coordinate changes of the top of the boom within a future time period T. A neural network is used to train the experimental samples to obtain a crane attitude prediction model.

[0033] Using the current attitude data of the crane and the current environmental data as model inputs, the predicted sequence of three-dimensional coordinates of the top of the crane boom within the next T time period is obtained as the predicted attitude of the crane.

[0034] This solution integrates historical operational data from n similar cranes and corresponding environmental data as experimental samples, covering diverse working conditions and risk scenarios. This provides comprehensive data support for subsequent model training, thereby improving prediction generalization ability and reliability. By using a long short-term memory neural network to train the model, the temporal dependencies of the crane's posture (such as boom swing inertia) can be accurately captured, enabling high-precision prediction of the three-dimensional coordinate sequence of the boom top at time T in the future. By visualizing the predicted crane posture using coordinate sequences, this provides advance warning and accurate data for subsequent risk level determination, shifting safety control from alarms after boundary crossings to advance prediction, and allowing operators sufficient time to respond.

[0035] Preferably, in S3, the risk level is determined based on the predicted attitude of the crane and the positional relationship of the target fence, and an early warning action is executed based on the risk level, including the following steps:

[0036] If, in the 3D coordinate prediction sequence, there exists a coordinate point whose distance to the boundary of the dynamic fence is between the first and second distance thresholds, it is considered a minor risk; if it is between the second and third distance thresholds, it is considered a moderate risk; and if it is less than the third distance threshold, it is considered a severe risk.

[0037] The warning advance time is determined based on the risk level, and the warning advance time decreases as the risk level increases. The alarm light is activated based on the warning advance time.

[0038] This solution clearly categorizes risks into light, medium, and heavy levels based on distance thresholds, providing quantifiable standards for risk assessment, avoiding vague management, and improving the accuracy of risk identification. By setting a warning lead time that dynamically adjusts with the risk level, it ensures rapid response to heavy risks and allows sufficient handling windows for light risks, balancing emergency efficiency with operational convenience. This achieves targeted risk warnings, avoids excessive alarm interference, and strengthens the last line of defense for proactive prevention and control, meeting the needs of comprehensive safety management.

[0039] Secondly, one technical solution provided in this embodiment of the invention is: a crane operation area boundary crossing early warning system based on a three-dimensional fence, including a data acquisition module, a scene construction module, an attitude prediction module, an early warning module, and an alarm device;

[0040] The data acquisition module monitors the crane's operating status in real time based on a distributed monitoring mechanism to obtain the crane's current posture data, and also collects current environmental data;

[0041] The scene construction module constructs a three-dimensional scene model of the construction area, sets initial three-dimensional fences based on different area types, and updates the initial three-dimensional fences in real time based on the three-dimensional fence dynamic adjustment mechanism and combined with the project progress and current environmental data to obtain dynamic fences.

[0042] The attitude prediction module is equipped with a crane attitude prediction model, which uses the current crane attitude data and current environment data as model inputs to obtain the crane's predicted attitude.

[0043] The early warning module determines the risk level based on the predicted attitude of the crane and the positional relationship of the target fence, and controls the alarm device to perform early warning actions based on the risk level.

[0044] In this solution, a corresponding system is built to integrate the work area boundary crossing early warning method, realizing human-computer interaction and improving the user experience.

[0045] Preferably, the system also includes a positioning module, which comprises a local differential network and a fixed RTK base station, for locating the coordinates of various parts of the crane boom based on satellite differential signals and reference position data from the fixed RTK base station.

[0046] In this solution, by deploying fixed RTK base stations and a local differential network, network RTK costs are avoided, thereby reducing management and control costs.

[0047] Preferably, the system also includes a communication module, which uses LoRa communication to enable information exchange between the modules.

[0048] This solution employs local communication methods such as 2.4G LoRa, which significantly shortens the alarm link and improves the early warning response speed.

[0049] The beneficial effects of this invention are as follows: This invention captures the full attitude data of the crane through a distributed monitoring mechanism and constructs a three-dimensional scene model to set an initial fence, covering the dangerous area and the working range, reducing monitoring blind spots. At the same time, the fence is dynamically updated according to the project progress and environmental data, avoiding the defects of fixed boundaries being out of sync with construction progress and environmental changes. This ensures that the safety threshold always matches the actual risk under different working conditions, and the future trajectory of the crane is predicted. Combined with the dynamic fence to predict risks, the early warning speed is faster than the traditional trigger-based alarm, leaving sufficient time for risk handling. This significantly reduces the probability of equipment collisions and crane movement exceeding limits, and improves the safety of crane operation.

[0050] The above description of the invention is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0051] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings. The drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings.

[0052] Figure 1 This is a flowchart of the crane operation area boundary crossing early warning method based on three-dimensional fence according to the present invention;

[0053] Figure 2 This is a schematic diagram of the crane operation area boundary crossing early warning system based on three-dimensional fence according to the present invention. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only one preferred embodiment of this invention and are only used to explain this invention. They do not limit the scope of protection of this invention. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0055] Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe the operations (or steps) as sequential processes, many of the operations (or steps) can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but it may also have additional steps not included in the figures; the process may correspond to a method, function, procedure, subroutine, subroutine, etc.

[0056] Example 1: To address the problem that most existing crane safety monitoring methods rely on trigger-based early warnings, which are slow to respond to changes in the environment and project progress, this example provides a crane operation area boundary crossing early warning method based on a three-dimensional fence. Figure 1 As shown, it includes the following steps:

[0057] S1: Based on the distributed monitoring mechanism, monitor the crane's operating status in real time to obtain the crane's current posture data and collect the current environmental data; construct a three-dimensional scene model of the construction area and set initial three-dimensional fences based on different area types.

[0058] In this embodiment, the crane's current attitude data is obtained by real-time monitoring of the crane's operating status based on a distributed monitoring mechanism, including the following steps:

[0059] Monitoring nodes are installed at the top, middle and hook of the crane boom, and all monitoring nodes are located on the same longitudinal axis of the boom.

[0060] The spatial coordinates of the monitoring node are determined based on satellite positioning signals. The current vibration acceleration and swing angular velocity of the corresponding part of the boom are determined based on the changes in the spatial coordinates as the current attitude data of the crane.

[0061] Specifically, to avoid a decrease in positioning accuracy due to interference with a single positioning system, this embodiment employs multi-source positioning. RTK positioning is used to receive differential positioning signals to initially calculate the three-dimensional coordinates of the nodes. UWB positioning is used to measure distances to calculate the relative distances and relative positions between nodes. IMU modules are installed on the nodes to collect the acceleration and angular velocity generated by the boom operation, which are used to correct dynamic errors in the positioning process. The three methods work together to obtain the spatial coordinates of the monitoring nodes, preventing positioning errors caused by interference.

[0062] This embodiment, by deploying monitoring nodes on the coaxial axis of the boom top, middle and hook, can comprehensively capture the motion status of key parts of the crane, avoiding the one-sidedness of single-point monitoring; by accurately acquiring the spatial coordinates of the nodes and the derived vibration acceleration and swing angular velocity through satellite positioning, centimeter-level attitude perception can be achieved, thereby outputting comprehensive and high-precision attitude data, providing reliable input for dynamic fence updates and attitude prediction, and providing a data foundation for subsequent active prevention and control.

[0063] In this embodiment, a three-dimensional scene model of the construction area is constructed, and initial three-dimensional fences are set based on different area types, including the following steps:

[0064] The construction site 3D model is obtained by collecting 3D point cloud information of the construction area using LiDAR and performing 3D modeling of the construction area according to a fixed ratio.

[0065] Fixed hazardous objects in the 3D model of the construction site are marked with different colors according to their degree of danger, and corresponding initial safety thresholds are set for fixed hazardous objects of different colors to generate initial 3D fences.

[0066] The fixed hazardous objects include high-voltage lines, buildings, foundation pits, work areas, and personnel access routes.

[0067] Specifically, a drone equipped with a lidar can be used to perform a full-coverage scan of the construction area to obtain three-dimensional point cloud information. The three-dimensional point cloud information is then imported into professional three-dimensional modeling software such as CloudCompare and Geomagic. Fixed hazardous objects in the point cloud information are identified and classified through AI algorithms or manual recognition. For example, high-voltage lines are marked in red and buildings are marked in blue. Based on the classification results, corresponding initial safety thresholds are set for these fixed hazardous objects, and closed three-dimensional polygonal boundaries are generated as initial three-dimensional fences.

[0068] This embodiment uses LiDAR to collect 3D point cloud information and model it at a fixed scale, accurately restoring the terrain and distribution of hazardous objects in the construction area, thereby generating a 1:1 3D model of the construction site. This eliminates the boundary errors and blind spots of traditional manual calibration. By color-coding fixed hazardous objects such as high-voltage lines and foundation pits according to their degree of danger, the risk levels are intuitively distinguished. At the same time, corresponding initial safety thresholds are set for hazardous objects of different colors, so that the initial 3D fence fits the actual risk differences. This lays a precise foundation for subsequent dynamic adjustments based on the environment and progress, meeting the core requirements of comprehensive safety management.

[0069] S2: Based on the dynamic adjustment mechanism of the three-dimensional fence and combined with the project progress and current environmental data, the initial three-dimensional fence is updated in real time to obtain a dynamic fence; a crane attitude prediction model is constructed, and the crane's current attitude data and current environmental data are used as model inputs to obtain the crane's predicted attitude.

[0070] In this embodiment, a dynamic fence is obtained by updating the initial three-dimensional fence in real time based on the three-dimensional fence dynamic adjustment mechanism and combined with the project progress and current environmental data, including the following steps:

[0071] Real-time collection of construction data from the construction area;

[0072] The environmental data includes wind speed and visibility, and the construction data includes the completed height of the floor pouring, the installation location of components, and the relocation information of the material storage area.

[0073] The current construction scenario is determined based on environmental and construction data. The corresponding dynamic threshold calculation rules are matched based on the current construction scenario to obtain the threshold adjustment value. The initial three-dimensional fence is updated in real time based on the threshold adjustment value to obtain the dynamic fence.

[0074] This embodiment comprehensively captures dynamic factors affecting fence control by collecting environmental data such as wind speed and visibility in real time, as well as construction data such as floor pouring height and material stacking area relocation. This avoids adjustment deviations caused by relying on a single data source. By accurately determining the current construction scenario based on multi-dimensional data and matching corresponding dynamic threshold calculation rules, fence adjustments are made scientifically based rather than blindly changed. At the same time, the initial three-dimensional fence is updated in real time to obtain a dynamic fence, which completely solves the pain point of traditional fixed fences being out of touch with construction progress and environmental changes. This ensures that safety thresholds always fit the actual working conditions and provides accurate boundary references for subsequent crane attitude prediction and risk level determination.

[0075] In this embodiment, the construction scenarios include conventional operation scenarios, harsh environment scenarios, schedule change scenarios, human-machine intensive scenarios, and mixed scenarios;

[0076] The harsh environment scenario matches the dynamic threshold calculation rules, generates wind speed influence coefficient and visibility influence coefficient based on changes in wind speed and visibility, and updates the initial safety threshold in the form of linear correlation.

[0077] The progress change scenario matches the progress dynamic threshold calculation rule, and the initial safety threshold is directly updated based on the construction data.

[0078] The dynamic threshold calculation rule for matching density in human-machine dense scenes generates a safety buffer with a radius of r centered on the boundary of the human-machine dense area, and extends the initial safety threshold to the boundary of the safety buffer.

[0079] The mixed scenario calculates a comprehensive adjustment value by weighting and summing the initial security threshold adjustment values ​​of each single-type scenario, and then adjusts the initial security threshold based on the comprehensive adjustment value.

[0080] Specifically, for example, a normal operation scenario is defined as follows: wind speed less than 8 m / s, visibility greater than 200 m, no change in project progress, and no dense movement of personnel or equipment. A severe environment scenario is defined as follows: wind speed exceeding 8 m / s or visibility less than or equal to 200 m. A progress change scenario is defined as follows: if the project progress is updated, such as by increasing the building height. A scenario is defined as follows: if the population density in a certain area is greater than 3 people / 10m². 3 If mobile devices enter the work area, it is defined as a human-machine intensive scene; if multiple of the above judgment conditions are triggered, it is defined as a mixed scene.

[0081] Specifically, the calculation rules for environmental dynamic thresholds are expressed by the following formula:

[0082] ;

[0083] Where S is the updated security threshold. As the initial security threshold, The wind speed influence coefficient, The visibility impact coefficient. and The weighting coefficients can be adjusted according to actual needs; for wind speed influence coefficients, when the wind speed is greater than 8 m / s, the coefficient increases by 0.05 for every 1 m / s increase (e.g., the coefficient is 0.1 for a wind speed of 10 m / s, and the safety threshold is 1.0 × (1 + 0.1) = 1.1 meters); for visibility influence coefficients, when the visibility is less than 200 m, the coefficient increases by 0.08 for every 50 m decrease.

[0084] The calculation rule for the dynamic progress threshold, taking the completed floor pouring height as an example, is expressed by the following formula:

[0085] ;

[0086] Where H represents the height of the floor after pouring.

[0087] In the dynamic threshold calculation rules, the radius r can be selected as 1.5m.

[0088] This embodiment subdivides various construction scenarios, such as conventional, harsh environments, and schedule changes, and matches specific dynamic threshold calculation rules accordingly. This avoids a one-size-fits-all approach to fence adjustments, thereby improving the accuracy of scenario adaptation and ensuring that threshold adjustments in different scenarios are scientifically aligned with actual risks. By combining classification adaptation with comprehensive calculation, it achieves precise iteration of dynamic fence thresholds, providing a reliable boundary reference for risk prediction and hierarchical linkage, and enhancing the pertinence and effectiveness of comprehensive safety management.

[0089] In this embodiment, a crane attitude prediction model is constructed, using the current crane attitude data and current environment data as model inputs to obtain the predicted crane attitude, including the following steps:

[0090] The historical operation data of n cranes of the same type and the environmental data of the corresponding time period are correlated as experimental samples;

[0091] The model aims to obtain the coordinate changes of the top of the boom within a future time period T. A neural network is used to train the experimental samples to obtain a crane attitude prediction model.

[0092] Using the current attitude data of the crane and the current environmental data as model inputs, the predicted sequence of three-dimensional coordinates of the top of the crane boom within the next T time period is obtained as the predicted attitude of the crane.

[0093] This embodiment integrates historical operation data and corresponding environmental data from n similar cranes as experimental samples, covering diverse working conditions and risk scenarios. This provides comprehensive data support for subsequent model training, thereby improving prediction generalization ability and reliability. By using a long short-term memory neural network to train the model, it can accurately capture the temporal dependencies of the crane's posture (such as the boom swing inertia), thus enabling high-precision prediction of the three-dimensional coordinate sequence of the boom top at time T in the future. By visualizing the predicted crane posture using coordinate sequences, it provides advance warning and accurate data basis for subsequent risk level determination, promoting safety control from alarms after boundary crossings to advance prediction, and reserving sufficient time for operators to handle the situation.

[0094] S3: Determine the risk level based on the predicted attitude of the crane and the positional relationship of the target fence, and execute early warning actions based on the risk level.

[0095] In this embodiment, the risk level is determined based on the predicted attitude of the crane and the positional relationship of the target fence, and an early warning action is executed based on the risk level, including the following steps:

[0096] If, in the 3D coordinate prediction sequence, there exists a coordinate point whose distance to the boundary of the dynamic fence is between the first and second distance thresholds, it is considered a minor risk; if it is between the second and third distance thresholds, it is considered a moderate risk; and if it is less than the third distance threshold, it is considered a severe risk.

[0097] The warning advance time is determined based on the risk level, and the warning advance time decreases as the risk level increases. The alarm light is activated based on the warning advance time.

[0098] Specifically, when the risk is mild, the crane is slowly approaching the danger zone without any immediate collision risk. Adjustments can be made promptly upon the issuance of a warning, and a warning signal can be issued within 3-5 seconds. When the risk is moderate, the crane is moving at a faster speed or is affected by environmental factors (such as strong winds), posing a greater risk of collision. In this case, a warning signal needs to be issued within a shorter timeframe, such as 2-3 seconds, so that operators can recognize the risk and make adjustments. When the risk is severe, the predicted trajectory clearly indicates it will touch the fence boundary, or the crane has entered the edge of the danger zone and is about to collide. Immediate warning is required, using a clear system of flashing red warning lights and high-decibel voice alarms within 1-2 seconds. Otherwise, it may touch the danger zone.

[0099] This embodiment clearly classifies risks into light, medium, and heavy levels based on distance thresholds, providing a quantifiable standard for risk assessment, avoiding vague control, and improving the accuracy of risk identification. By setting a warning lead time that dynamically adjusts with the risk level, it ensures rapid response to heavy risks and allows sufficient handling window for light risks, balancing emergency efficiency and operational convenience. This achieves targeted risk warning, avoids excessive alarm interference, and strengthens the last line of defense for proactive prevention and control, meeting the needs of comprehensive safety management.

[0100] Example 2: This example also provides a crane operation area boundary crossing early warning system based on a three-dimensional fence, such as... Figure 2 As shown, it includes a data acquisition module, a scene construction module, an attitude prediction module, an early warning module, and an alarm device;

[0101] The data acquisition module monitors the crane's operating status in real time based on a distributed monitoring mechanism to obtain the crane's current posture data, and also collects current environmental data;

[0102] The scene construction module constructs a three-dimensional scene model of the construction area, sets initial three-dimensional fences based on different area types, and updates the initial three-dimensional fences in real time based on the three-dimensional fence dynamic adjustment mechanism and combined with the project progress and current environmental data to obtain dynamic fences.

[0103] The attitude prediction module is equipped with a crane attitude prediction model, which uses the current crane attitude data and current environment data as model inputs to obtain the crane's predicted attitude.

[0104] The early warning module determines the risk level based on the predicted attitude of the crane and the positional relationship of the target fence, and controls the alarm device to perform early warning actions based on the risk level.

[0105] This embodiment integrates the work area boundary crossing warning method in this solution by constructing a corresponding system, realizing human-computer interaction and improving the user experience.

[0106] In this embodiment, the early warning system also includes a positioning module, which includes a local differential network and a fixed RTK base station, and is used to locate the coordinates of various parts of the crane boom based on the satellite differential signal and the reference position data of the fixed RTK base station.

[0107] This embodiment avoids network RTK costs by deploying fixed RTK base stations and a local differential network, thereby reducing management and control costs.

[0108] In this embodiment, the early warning system also includes a communication module, which uses LoRa communication to enable information exchange between modules.

[0109] This solution employs local communication methods such as 2.4G LoRa, which significantly shortens the alarm link and improves the early warning response speed.

[0110] As can be seen from the above embodiments, it has at least the following substantial effects:

[0111] (1) This invention captures the attitude data of various parts of the crane through a distributed monitoring mechanism, thereby capturing the overall attitude of the crane. At the same time, a three-dimensional scene model is constructed and an initial fence is set to cover the dangerous area and the work area, thereby eliminating the blind spots of traditional manual monitoring.

[0112] (2) This invention dynamically updates the three-dimensional fence by linking the project progress with real-time environmental data such as wind speed and visibility, which solves the problem of traditional monitoring using fixed boundaries and being disconnected from scene changes, and ensures that the safety threshold can be adapted to the working conditions in real time.

[0113] (3) This invention predicts the future trajectory of the crane by constructing a crane attitude prediction model, thereby realizing the proactive prediction of risks and reserving sufficient time for handling compared with traditional trigger alarms;

[0114] (4) By performing corresponding risk actions according to the risk level, the present invention can avoid excessive alarms interfering with actual operations and ensure construction stability.

[0115] The specific embodiments described above are preferred embodiments of the present invention and are not intended to limit the specific scope of the present invention. The scope of the present invention includes, but is not limited to, these specific embodiments. All equivalent changes made in accordance with the shape and structure of the present invention are within the protection scope of the present invention.

Claims

1. A method for early warning of crane operation area boundary crossing based on three-dimensional fence, characterized in that: Includes the following steps: S1. Based on a distributed monitoring mechanism, the crane's operating status is monitored in real time to obtain the crane's current attitude data, specifically: Monitoring nodes are installed at the top, middle and hook of the crane boom, and all monitoring nodes are located on the same longitudinal axis of the boom. The spatial coordinates of the monitoring node are determined based on satellite positioning signals, and the current vibration acceleration and swing angular velocity of the corresponding part of the boom are determined based on the changes in the spatial coordinates as the current attitude data of the crane. Collect current environmental data; construct a 3D scene model of the construction area, and set initial 3D fences based on different area types; S2. Based on the dynamic adjustment mechanism of the three-dimensional fence and combined with the project progress and current environmental data, the initial three-dimensional fence is updated in real time to obtain a dynamic fence. Specifically, construction data of the construction area is collected in real time. The environmental data includes wind speed and visibility, and the construction data includes the completed height of the floor pouring, the installation location of components, and the relocation information of the material storage area. The current construction scenario is determined based on environmental and construction data. The corresponding dynamic threshold calculation rules are matched based on the current construction scenario to obtain the threshold adjustment value. The initial three-dimensional fence is updated in real time based on the threshold adjustment value to obtain the dynamic fence. The construction scenarios include conventional operation scenarios, harsh environment scenarios, schedule change scenarios, human-machine intensive scenarios, and mixed scenarios; The harsh environment scenario matches the dynamic threshold calculation rules, generates wind speed influence coefficient and visibility influence coefficient based on changes in wind speed and visibility, and updates the initial safety threshold in the form of linear correlation. The progress change scenario matches the progress dynamic threshold calculation rule, and the initial safety threshold is directly updated based on the construction data. The dynamic threshold calculation rule for matching density in human-machine dense scenes generates a safety buffer with a radius of r centered on the boundary of the human-machine dense area, and extends the initial safety threshold to the boundary of the safety buffer. The mixed scenario calculates a comprehensive adjustment value by weighting and summing the initial security threshold adjustment values ​​of each single-type scenario, and then adjusts the initial security threshold based on the comprehensive adjustment value. Construct a crane attitude prediction model, and use the current crane attitude data and current environment data as model inputs to obtain the crane's predicted attitude; S3. Determine the risk level based on the predicted attitude of the crane and the positional relationship of the target fence, and execute early warning actions based on the risk level.

2. The method for early warning of crane operation area boundary crossing based on three-dimensional fence according to claim 1, characterized in that: In S1, a 3D scene model of the construction area is constructed, and initial 3D fences are set based on different area types, including the following steps: The construction site 3D model is obtained by collecting 3D point cloud information of the construction area using LiDAR and performing 3D modeling of the construction area according to a fixed ratio. Fixed hazardous objects in the 3D model of the construction site are marked with different colors according to their degree of danger, and corresponding initial safety thresholds are set for fixed hazardous objects of different colors to generate initial 3D fences. The fixed hazardous objects include high-voltage lines, buildings, foundation pits, work areas, and personnel access routes.

3. The method for early warning of crane operation area boundary crossing based on three-dimensional fence according to claim 1, characterized in that: In S2, a crane attitude prediction model is constructed. The crane's current attitude data and current environmental data are used as model inputs to obtain the crane's predicted attitude, including the following steps: The historical operation data of n cranes of the same type and the environmental data of the corresponding time period are correlated as experimental samples; The model aims to obtain the coordinate changes of the top of the boom within a future time period T. A neural network is used to train the experimental samples to obtain a crane attitude prediction model. Using the current attitude data of the crane and the current environmental data as model inputs, the predicted sequence of three-dimensional coordinates of the top of the crane boom within the next T time period is obtained as the predicted attitude of the crane.

4. The method for early warning of crane operation area boundary crossing based on three-dimensional fence according to claim 3, characterized in that: In S3, the risk level is determined based on the predicted attitude of the crane and the positional relationship of the target fence. Based on the risk level, early warning actions are executed, including the following steps: If, in the 3D coordinate prediction sequence, there exists a coordinate point whose distance to the boundary of the dynamic fence is between the first and second distance thresholds, it is considered a minor risk; if it is between the second and third distance thresholds, it is considered a moderate risk; and if it is less than the third distance threshold, it is considered a severe risk. The warning advance time is determined based on the risk level, and the warning advance time decreases as the risk level increases. The alarm light is activated based on the warning advance time.

5. A crane operation area boundary crossing early warning system based on a three-dimensional fence, applicable to the crane operation area boundary crossing early warning method based on a three-dimensional fence as described in any one of claims 1-4, characterized in that: It includes a data acquisition module, a scene construction module, an attitude prediction module, an early warning module, and an alarm device; The data acquisition module monitors the crane's operating status in real time based on a distributed monitoring mechanism to obtain the crane's current posture data, and also collects current environmental data; The scene construction module constructs a three-dimensional scene model of the construction area, sets initial three-dimensional fences based on different area types, and updates the initial three-dimensional fences in real time based on the three-dimensional fence dynamic adjustment mechanism and combined with the project progress and current environmental data to obtain dynamic fences. The attitude prediction module is equipped with a crane attitude prediction model, which uses the current crane attitude data and current environment data as model inputs to obtain the crane's predicted attitude. The early warning module determines the risk level based on the predicted attitude of the crane and the positional relationship of the target fence, and controls the alarm device to perform early warning actions based on the risk level.

6. The crane operation area boundary crossing early warning system based on three-dimensional fence according to claim 5, characterized in that: It also includes a positioning module, which comprises a local differential network and a fixed RTK base station, used to locate the coordinates of various parts of the crane boom based on satellite differential signals and reference position data from the fixed RTK base station.

7. The crane operation area boundary crossing early warning system based on three-dimensional fence according to claim 5, characterized in that: It also includes a communication module, which uses LoRa communication to enable information exchange between the modules.

Images