Three-dimensional dynamic safety management and control method and system for power transmission and transformation hoisting operation

Abstract

The invention discloses a three-dimensional dynamic safety management and control method and system for power transmission and transformation hoisting operation, and relates to the technical field of construction safety, and the method comprises the steps: obtaining point cloud data of a target operation area through ground three-dimensional laser scanning; according to the point cloud data, a three-dimensional grid model is constructed through an oblique photography technology and a Beidou real-time dynamic carrier phase difference technology, and a semantic electronic fence is generated; according to a pure vision technology, through target detection, three-dimensional reconstruction and time sequence tracking, a hanging object 4D digital branch is constructed, and a hanging object prediction track is obtained; the shortest three-dimensional distance is calculated through an efficient collision detection algorithm according to the hanging object prediction track and the semantic electronic fence, multi-dimensional risk assessment is carried out, and graded early warning signals are generated; and carrying out comprehensive alarm prompting according to the graded early warning signal. According to the invention, a traditional passive protection mode depending on single threshold alarm is changed, and quantitative evaluation and predictive management and control of dynamic risks are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of construction safety, in particular to a three-dimensional dynamic safety control method and system for hoisting operation of power transmission and transformation. BACKGROUND

[0002] As a common and key link in power transmission and transformation project construction, hoisting operation is generally in a complex environment, and the operation site is often limited in space, with existing or under-construction equipment and facilities, and even adjacent high-voltage live equipment, forming a complex three-dimensional operation space. Under this environment, hoisting operation is of high risk and is the key object of safety control.

[0003] However, the existing safety guarantee highly depends on mechanical limiting, simple sensors and human lookouts, and these traditional methods have obvious limitations in complex operation environment. The core problem lies in the serious inconsistency between risk perception dimension and reality: the risk of hoisting operation is essentially a four-dimensional (4D) problem (space + time), while the traditional method only relies on one-dimensional scalar threshold for passive response, lacking dynamic understanding of three-dimensional spatial relationship and prediction ability of future risks. This "dimension reduction" processing leads to the loss of key safety information, and cannot perceive the three-dimensional relationship between the hoist arm and the environment, nor can it distinguish the types of obstacles to provide differentiated prevention and control measures, making it difficult to meet the needs of active and intelligent safety guarantee.

[0004] In order to solve the above problems, a three-dimensional dynamic safety control method and system for hoisting operation of power transmission and transformation are urgently needed. SUMMARY

[0005] The purpose of the present application is to provide a three-dimensional dynamic safety control method and system for hoisting operation of power transmission and transformation, which realizes quantitative evaluation and predictive control of dynamic risks by constructing a three-dimensional grid model and a 4D digital avatar of the hoisted object, and changes the passive protection mode of traditional single threshold alarm.

[0006] To achieve the above purpose, the present application provides a three-dimensional dynamic safety control method for hoisting operation of power transmission and transformation, comprising the following steps: S1, acquiring point cloud data of a target operation area by using ground three-dimensional laser scanning; S2, constructing a three-dimensional grid model and generating a semantic electronic fence by using oblique photography technology and Beidou real-time dynamic carrier phase difference technology according to the point cloud data; S3, constructing a 4D digital avatar of the hoisted object by using target detection, three-dimensional reconstruction and time sequence tracking according to pure vision technology, and obtaining a predicted trajectory of the hoisted object; S4, according to the prediction trajectory of the hanging object and the semantic electronic fence, the shortest three-dimensional distance between the prediction trajectory of the hanging object and the semantic electronic fence is calculated through an efficient collision detection algorithm, multi-dimensional risk assessment is performed, and a hierarchical early warning signal is generated; S5, according to the hierarchical early warning signal, comprehensive alarm prompts are performed through a sound and light device and a three-dimensional visual monitoring interface.

[0007] Preferably, the specific content of S2, according to the point cloud data, constructing a three-dimensional grid model through oblique photography technology and Beidou real-time dynamic carrier phase difference technology, includes: According to the Beidou real-time dynamic carrier phase difference technology, the point cloud data is high-precision geographically matched; The high-precision geographically matched point cloud data is sequentially subjected to splicing processing, denoising processing and fusion processing to obtain high-precision point cloud data; According to the high-precision point cloud data, a three-dimensional grid model is automatically generated through oblique photography technology; The three-dimensional grid model is subjected to semantic segmentation, static facilities are identified and labeled, and a semantic electronic fence is generated.

[0008] Preferably, the specific content of S3, according to the pure vision technology, constructing a 4D digital clone of the hanging object through target detection, three-dimensional reconstruction and time sequence tracking, includes: Multi-view video data is obtained, and the multi-view video data is subjected to de-distortion and brightness equalization processing to obtain multi-view video processing data; Based on a target detection neural network model, the hanging object in the multi-view video processing data is identified to obtain the spatial position of the hanging object, and a 4D digital clone of the hanging object is obtained; According to the spatial position of the hanging object, the motion state of the hanging object is estimated and predicted through an extended Kalman filter algorithm to obtain the prediction trajectory of the hanging object.

[0009] Preferably, the specific content of S4, according to the prediction trajectory of the hanging object and the semantic electronic fence, the shortest three-dimensional distance between the prediction trajectory of the hanging object and the semantic electronic fence is calculated through an efficient collision detection algorithm, multi-dimensional risk assessment is performed, and a hierarchical early warning signal is generated, includes: The semantic electronic fence and the prediction trajectory of the hanging object are uniformly converted into geometric bodies to obtain semantic electronic fence spatial data and hanging object prediction trajectory spatial data; According to the semantic electronic fence spatial data and the hanging object prediction trajectory spatial data, the shortest three-dimensional distance between the prediction trajectory of the hanging object and the semantic electronic fence is calculated through an efficient collision detection algorithm; According to the shortest three-dimensional distance and the semantic electronic fence, multi-dimensional risk assessment is performed to generate a hierarchical early warning signal.

[0010] Preferably, the multi-dimensional risk assessment is performed according to the shortest three-dimensional distance and the semantic electronic fence, and the specific content of the generated hierarchical early warning signal includes: The geometric risk degree assessment is performed according to the shortest three-dimensional distance, and a geometric risk degree assessment result is obtained. The semantic risk degree assessment is performed according to the semantic electronic fence, and a semantic risk degree assessment result is obtained. The risk decision matrix is constructed according to the geometric risk degree assessment result and the semantic risk degree assessment result, and a comprehensive risk index is evaluated and generated. The hierarchical early warning signal is generated according to the comprehensive risk index.

[0011] Preferably, the geometric risk degree assessment is performed according to the shortest three-dimensional distance, and the specific content of the obtained geometric risk degree assessment result includes: It is determined whether the shortest three-dimensional distance is lower than a preset alert threshold and a preset danger threshold, and a geometric risk degree value is determined. When the shortest three-dimensional distance is lower than the preset alert threshold, the geometric risk degree is level one. When the shortest three-dimensional distance is lower than the preset danger threshold, the geometric risk degree is level two.

[0012] Preferably, the semantic risk degree assessment is performed according to the semantic electronic fence, and the specific content of the obtained semantic risk degree assessment result includes: The semantic attribute of the collision object is identified according to the semantic electronic fence. The collision degree value is determined according to the semantic attribute.

[0013] Preferably, the risk decision matrix is constructed according to the geometric risk degree assessment result and the semantic risk degree assessment result, and the specific content of the evaluated and generated comprehensive risk index includes: The risk decision matrix is constructed according to the geometric risk degree value and the collision degree value. The risk comprehensive evaluation standard is defined. The current geometric risk degree value is obtained, the geometric risk degree value is configured with a geometric risk degree emphasis conversion coefficient according to the geometric risk degree emphasis requirement in the risk comprehensive evaluation, and a geometric risk degree evaluation result is obtained. The current collision degree value is obtained, the collision degree value is configured with a collision degree emphasis conversion coefficient according to the collision degree emphasis requirement in the risk comprehensive evaluation, and a collision degree evaluation result is obtained. The geometric risk degree evaluation result and the collision degree evaluation result are configured with a weight coefficient, and a comprehensive risk index is obtained.

[0014] Preferably, the specific content of the generated hierarchical early warning signal according to the comprehensive risk index includes: If the comprehensive risk index is lower than the risk comprehensive evaluation standard, level one early warning is performed. If the comprehensive risk index is equal to the risk comprehensive evaluation standard, a secondary early warning is carried out. If the comprehensive risk index is higher than the risk comprehensive evaluation standard, a tertiary early warning is carried out.

[0015] The application further provides a three-dimensional dynamic safety management and control system for power transmission and transformation hoisting operation, comprising: A data acquisition module is configured to acquire point cloud data of a target operation area by using ground three-dimensional laser scanning. A model construction module is configured to construct a three-dimensional grid model and generate a semantic electronic fence by using oblique photography technology and high-precision satellite positioning technology according to the point cloud data. A digital double construction module is configured to construct a 4D digital double of a hoisted object by using target detection, three-dimensional reconstruction and time sequence tracking according to pure visual technology, and obtain a hoisted object predicted trajectory. A risk assessment module is configured to calculate the shortest three-dimensional distance between the hoisted object predicted trajectory and the semantic electronic fence by using an efficient collision detection algorithm according to the hoisted object predicted trajectory and the semantic electronic fence, and perform multi-dimensional risk assessment to generate a graded early warning signal. An alarm prompt module is configured to perform comprehensive alarm prompting by using a sound and light device and a three-dimensional visual monitoring interface according to the graded early warning signal.

[0016] Compared with the traditional technology, the three-dimensional dynamic safety management and control method and system for power transmission and transformation hoisting operation provided by the application have the following beneficial effects: the three-dimensional grid model is constructed by using oblique photography technology and high-precision satellite positioning technology, the semantic electronic fence is generated, and a static knowledge base containing various entity safety attributes is constructed; the 4D digital double of the hoisted object is constructed by using pure visual technology, and the hoisted object motion trajectory is predicted; the complex scene is converted into an efficient “dynamic-static” geometric verification problem by using the efficient collision detection algorithm, and the graded early warning is realized; the three-dimensional grid model and the 4D digital double of the hoisted object are constructed, the problem that the risk perception dimension is seriously inconsistent with the reality is solved, the quantitative assessment and predictive control of the dynamic risk are realized, and the traditional passive protection mode depending on a single threshold alarm is changed.

[0017] The technical method of the application will be further described in detail below by using the accompanying drawings and embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 The flowchart of the three-dimensional dynamic safety management and control method for power transmission and transformation hoisting operation of the application is shown in FIG. 1. Figure 2 The module framework diagram of the three-dimensional dynamic safety management and control system for power transmission and transformation hoisting operation of the application is shown in FIG. 2. DETAILED DESCRIPTION

[0019] The technical method of the present invention will be further described below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps described in these embodiments do not limit the scope of this application.

[0020] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the scope of this application and its application or use.

[0021] Techniques, systems, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, they should be considered part of the instruction manual.

[0022] In all the examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0023] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0024] This invention provides a three-dimensional dynamic safety management and control method for power transmission and transformation hoisting operations, such as... Figure 1 As shown, it includes the following steps: S1. Use terrestrial 3D laser scanning to acquire point cloud data of the target work area. Acquiring point cloud data requires multi-angle, full-coverage data collection of the target work area.

[0025] S2. Based on point cloud data, a three-dimensional mesh model is constructed using oblique photography and BeiDou Real-Time Kinematic (RTK) technology to generate a semantic electronic fence.

[0026] The three-dimensional mesh model constructed by this invention is a high-fidelity three-dimensional digital twin model that is highly consistent with the physical world in terms of geometry, space and semantics, and has absolute geographic coordinates. It can be used as a background for three-dimensional visualization and a geometric reference for real-time collision detection, and is a "static digital map" on which safety functions depend.

[0027] Furthermore, step S2 includes steps S201 to S204, as detailed below: S201. High-precision georegistration of point cloud data is performed based on BeiDou real-time dynamic carrier phase differential technology. This can be understood as using BeiDou RTK technology to accurately measure the coordinates of ground control points, which serve as a global reference for high-precision georegistration of point cloud data, ensuring the overall accuracy of the model.

[0028] S202. The point cloud data after high-precision georegistration is sequentially processed by stitching, denoising and fusion to obtain high-precision point cloud data.

[0029] S203. Based on high-precision point cloud data, automatically generate a three-dimensional mesh model using oblique photogrammetry technology.

[0030] S204. Perform semantic segmentation on the 3D mesh model, identify and label static facilities, and generate a semantic electronic fence.

[0031] When identifying and labeling static facilities, key entities such as "main transformers," "tower foundations," and "live power lines" can be labeled using manual or intelligent identification methods. The three-dimensional geometry of each identified static object is defined as a "static electronic fence" or "untouchable zone" with centimeter-level precision. The process of generating semantic electronic fences endows the three-dimensional mesh model with security semantic information that can be understood by computers, forming a database of static obstacles for dynamic collision detection.

[0032] S3. Based on pure vision technology, through target detection, 3D reconstruction and temporal tracking, a 4D digital clone of the suspended object is constructed to obtain the predicted trajectory of the suspended object.

[0033] Furthermore, step S3 can be replaced by the following steps S301 to S303: S301. Acquire multi-view video data and perform distortion correction and brightness equalization processing on the multi-view video data to obtain multi-view processed video data. The acquisition of multi-view video data adopts a pure vision-based approach, which can be achieved by using cameras deployed in the target working area.

[0034] S302. Based on the target detection neural network model (YOLOv8 model), identify suspended objects in multi-view video processing data, obtain the spatial position of the suspended objects, and obtain a 4D digital clone of the suspended objects. Specifically, when identifying suspended objects in multi-view video processing data, the three-dimensional geometric structure of the scene and the motion vector field of dynamic targets such as suspended objects can be directly and synchronously inferred in the world coordinate system.

[0035] The 4D digital avatar of the suspended object can generate the complete motion form of the suspended object in a time series at once, accurately depicting the position, posture and motion trajectory of the suspended object with dense point clouds.

[0036] S303. Based on the spatial position of the suspended object, the motion state of the suspended object is estimated and predicted by the extended Kalman filter algorithm to obtain the predicted trajectory of the suspended object.

[0037] The trajectory prediction process integrates visual observation and kinematic laws, which can not only estimate the current state, but also predict the probabilistic range of motion of the suspended object within the next 100-200 milliseconds, realizing predictive safety warnings and serving as the core basis for operators to gain reaction time.

[0038] This invention reconstructs the four-dimensional (3D space + time) motion state of the hook and the hoisted object in real time and accurately using a 4D digital clone of the hoisted object, and predicts its short-term future trajectory, realizing a leap from passive perception to active prediction, and providing a dynamic perception core for system safety decision-making.

[0039] S4. Based on the predicted trajectory of the suspended object and the semantic electronic fence, the shortest three-dimensional distance between the predicted trajectory of the suspended object and the semantic electronic fence is calculated through an efficient collision detection algorithm. Multi-dimensional risk assessment is then performed to generate graded early warning signals.

[0040] Furthermore, step S4 specifically includes the following steps: S401. Convert the semantic electronic fence and the predicted trajectory of the suspended object into a unified geometry (convex hull or directed bounding box) to obtain the spatial data of the semantic electronic fence and the spatial data of the predicted trajectory of the suspended object.

[0041] S402. Based on the semantic electronic fence spatial data and the predicted trajectory spatial data of the suspended object, calculate the shortest three-dimensional distance between the predicted trajectory of the suspended object and the semantic electronic fence using an efficient collision detection algorithm. The shortest three-dimensional distance can provide a geometric basis for risk assessment.

[0042] S403. Based on the shortest three-dimensional distance and semantic electronic fence, conduct multi-dimensional risk assessment and generate graded early warning signals. The multi-dimensional risk assessment involves integrating multiple pieces of information for dynamic risk assessment.

[0043] Furthermore, step S403 specifically includes the following: Based on the shortest three-dimensional distance, a geometric hazard assessment is performed to obtain the geometric hazard assessment result. Specifically, it involves determining whether the shortest three-dimensional distance is lower than a preset warning threshold and a preset danger threshold to determine the geometric hazard level: when the shortest three-dimensional distance is lower than the preset warning threshold, the geometric hazard level is Level 1; when the shortest three-dimensional distance is lower than the preset danger threshold, the geometric hazard level is Level 2.

[0044] Based on semantic geofences, semantic hazard assessments are performed to obtain the results. Specifically, this involves identifying the semantic attributes of collision objects using the semantic geofence, and then determining the collision severity value based on these attributes. In essence, this involves identifying the semantic attributes of collision objects (such as "main transformer" or "temporary scaffolding") and dynamically adjusting the warning level and sensitivity based on their importance.

[0045] Based on the geometric hazard assessment results and semantic hazard assessment results, a risk decision matrix is ​​constructed, and a comprehensive risk index is generated. Specifically, the process involves first constructing the risk decision matrix based on the geometric hazard level and collision severity values; then defining the comprehensive risk assessment criteria; next, obtaining the current geometric hazard level value; and then, according to the geometric hazard emphasis requirements in the comprehensive risk assessment, configuring a geometric hazard emphasis conversion coefficient on the geometric hazard level value to obtain the geometric hazard assessment result. Similarly, obtaining the current collision severity value; and then, according to the collision severity emphasis requirements in the comprehensive risk assessment, configuring a collision severity emphasis conversion coefficient on the collision severity value to obtain the collision severity assessment result. Finally, weighting coefficients are assigned to both the geometric hazard assessment result and the collision severity assessment result to obtain the comprehensive risk index.

[0046] Based on the comprehensive risk index, a tiered early warning signal is generated. Specifically, if the comprehensive risk index is lower than the comprehensive risk assessment standard, a Level 1 early warning is issued; if the comprehensive risk index is equal to the comprehensive risk assessment standard, a Level 2 early warning is issued; and if the comprehensive risk index is higher than the comprehensive risk assessment standard, a Level 3 early warning is issued.

[0047] The warning signals include brightness warnings and sound intensity warnings. Level 3 warnings have stronger brightness and sound intensity warnings than Level 2 warnings, and Level 2 warnings have stronger brightness and sound intensity warnings than Level 1 warnings.

[0048] Multidimensional risk assessment also includes behavioral hazard assessment, which can detect the wearing status of personnel's safety helmets and intrusion behavior in dangerous areas through real-time visual analysis, so as to achieve proactive control of personnel safety.

[0049] In this invention, multi-source information is incorporated into the evaluation and decision engine, and a comprehensive risk index is generated through a configurable risk matrix, which achieves accurate and intelligent composite alarms and can effectively avoid false alarms and alarm fatigue.

[0050] S5. Based on the graded early warning signals, comprehensive alarm prompts are issued through audio-visual devices and a three-dimensional visualization monitoring interface.

[0051] Therefore, the core steps of the three-dimensional dynamic safety management method for power transmission and transformation hoisting operations of the present invention are: (I) Digital Twin Modeling and Semanticization. In the target operation scenario, high-precision 3D reality models are automatically generated by acquiring overlapping images and precise coordinate points using oblique photogrammetry. Subsequently, the model is semantically processed through human interaction to delineate and label key "static electronic fence" areas, providing a foundation for subsequent collision detection.

[0052] (II) Core Algorithm Development and Verification. For key targets such as suspended objects and personnel, datasets were prepared and the target detection model was fine-tuned and optimized to achieve accurate identification. Simultaneously, a pure vision-based 4D reconstruction and tracking algorithm was developed to perceive the movement trajectory of suspended objects in real time. Based on this, collision detection and hierarchical alarm logic were developed to provide algorithmic support for core functions.

[0053] This invention provides a three-dimensional dynamic safety management and control system for power transmission and transformation hoisting operations, such as... Figure 2 As shown, it includes: The data acquisition module is used to acquire point cloud data of the target work area using terrestrial 3D laser scanning.

[0054] The model building module is used to construct a 3D mesh model based on point cloud data using oblique photogrammetry and high-precision satellite positioning technology, and generate semantic electronic fences.

[0055] The clone construction module is used to construct a 4D digital clone of the suspended object based on pure vision technology, through object detection, 3D reconstruction and temporal tracking, and obtain the predicted trajectory of the suspended object.

[0056] The risk assessment module is used to calculate the shortest three-dimensional distance between the predicted trajectory of the suspended object and the semantic electronic fence using an efficient collision detection algorithm, and to perform multi-dimensional risk assessment and generate graded early warning signals based on the predicted trajectory of the suspended object and the semantic electronic fence.

[0057] The alarm notification module is used to provide comprehensive alarm notifications based on graded early warning signals, through audio-visual devices and a three-dimensional visualization monitoring interface.

[0058] Finally, it should be noted that the above embodiments are only used to illustrate the technical methods of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical methods of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical methods to deviate from the spirit and scope of the technical methods of the present invention.

Claims

1. A three-dimensional dynamic safety management and control method for power transmission and transformation hoisting operations, characterized in that, The three-dimensional dynamic safety management and control method for power transmission and transformation hoisting operations includes the following steps: S1. Use ground-based three-dimensional laser scanning to obtain point cloud data of the target work area; S2. Based on point cloud data, a three-dimensional mesh model is constructed using oblique photography technology and BeiDou real-time dynamic carrier phase differential technology to generate a semantic electronic fence. S3. Based on pure vision technology, a 4D digital clone of the suspended object is constructed through target detection, 3D reconstruction and temporal tracking to obtain the predicted trajectory of the suspended object; S4. Based on the predicted trajectory of the suspended object and the semantic electronic fence, the shortest three-dimensional distance between the predicted trajectory of the suspended object and the semantic electronic fence is calculated through an efficient collision detection algorithm, and a multi-dimensional risk assessment is performed to generate a graded early warning signal. S5. Based on the graded early warning signals, comprehensive alarm prompts are issued through audio-visual devices and a three-dimensional visualization monitoring interface.

2. The three-dimensional dynamic safety control method for power transmission and transformation hoisting operations according to claim 1, characterized in that, Based on point cloud data, S2 constructs a 3D mesh model using oblique photography and BeiDou real-time dynamic carrier phase differential technology, generating the following specific content for a semantic electronic fence: High-precision georegistration of point cloud data is performed using BeiDou real-time dynamic carrier phase differential technology. The point cloud data after high-precision georegistration is sequentially processed by stitching, denoising and fusion to obtain high-precision point cloud data. Based on high-precision point cloud data, a 3D mesh model is automatically generated using oblique photogrammetry technology. Semantic segmentation is performed on the 3D mesh model to identify and label static facilities, generating a semantic electronic fence.

3. The three-dimensional dynamic safety control method for power transmission and transformation hoisting operations according to claim 1, characterized in that, In S3, based on pure vision technology, a 4D digital avatar of the suspended object is constructed through object detection, 3D reconstruction, and temporal tracking. The specific content of the predicted trajectory of the suspended object includes: Acquire multi-view video data, and perform distortion correction and brightness equalization on the multi-view video data to obtain multi-view video processed data; Based on the target detection neural network model, the suspended object in multi-view video processing data is identified, the spatial position of the suspended object is obtained, and a 4D digital clone of the suspended object is obtained. Based on the spatial position of the suspended object, the motion state of the suspended object is estimated and predicted by the extended Kalman filter algorithm to obtain the predicted trajectory of the suspended object.

4. The three-dimensional dynamic safety control method for power transmission and transformation hoisting operations according to claim 1, characterized in that, In S4, based on the predicted trajectory of the suspended object and the semantic electronic fence, an efficient collision detection algorithm is used to calculate the shortest three-dimensional distance between the predicted trajectory of the suspended object and the semantic electronic fence, and a multi-dimensional risk assessment is performed to generate a graded early warning signal. The specific content includes: The semantic electronic fence and the predicted trajectory of the suspended object are uniformly converted into geometric objects to obtain the spatial data of the semantic electronic fence and the spatial data of the predicted trajectory of the suspended object. Based on the semantic electronic fence spatial data and the spatial data of the predicted trajectory of the suspended object, the shortest three-dimensional distance between the predicted trajectory of the suspended object and the semantic electronic fence is calculated using an efficient collision detection algorithm. Based on the shortest three-dimensional distance and semantic electronic fence, a multi-dimensional risk assessment is conducted to generate graded early warning signals.

5. The three-dimensional dynamic safety control method for power transmission and transformation hoisting operations according to claim 4, characterized in that, Based on the shortest three-dimensional distance and semantic electronic fence, a multi-dimensional risk assessment is conducted to generate tiered early warning signals. The specific content includes: Based on the shortest three-dimensional distance, a geometric hazard assessment is performed to obtain the geometric hazard assessment result; Based on the semantic electronic fence, a semantic risk assessment is performed to obtain the semantic risk assessment results; Based on the geometric risk assessment results and the semantic risk assessment results, a risk decision matrix is ​​constructed, and a comprehensive risk index is generated. Based on the comprehensive risk index, a tiered early warning signal is generated.

6. The three-dimensional dynamic safety control method for power transmission and transformation hoisting operations according to claim 5, characterized in that, Based on the shortest three-dimensional distance, a geometric hazard assessment is performed. The specific content of the geometric hazard assessment results includes: Determine whether the shortest three-dimensional distance is lower than the preset warning threshold and the preset danger threshold, and determine the geometric danger level value; When the shortest three-dimensional distance is lower than the preset warning threshold, the geometric danger level is level one; When the shortest three-dimensional distance is lower than the preset danger threshold, the geometric danger level is level two.

7. The three-dimensional dynamic safety control method for power transmission and transformation hoisting operations according to claim 6, characterized in that, Based on the semantic electronic fence, a semantic risk assessment is conducted, and the specific content of the semantic risk assessment results includes: Based on semantic electronic fences, identify the semantic attributes of colliding objects; The collision level is determined based on semantic attributes.

8. The three-dimensional dynamic safety control method for power transmission and transformation hoisting operations according to claim 7, characterized in that, Based on the geometric risk assessment results and semantic risk assessment results, a risk decision matrix is ​​constructed, and the specific content of the comprehensive risk index generated includes: Construct a risk decision matrix based on geometric hazard level and collision severity level; Define comprehensive risk assessment criteria; Obtain the current geometric hazard level value, and according to the geometric hazard level emphasis requirements in the comprehensive risk assessment, configure the geometric hazard level emphasis conversion coefficient for the geometric hazard level value to obtain the geometric hazard level assessment result; Obtain the current collision severity value, and configure a collision severity emphasis conversion coefficient for the collision severity value according to the collision severity emphasis requirements in the comprehensive risk assessment to obtain the collision severity assessment result; Weighting coefficients are assigned to the geometric hazard assessment results and the collision severity assessment results to obtain a comprehensive risk index.

9. The three-dimensional dynamic safety control method for power transmission and transformation hoisting operations according to claim 8, characterized in that, Based on the comprehensive risk index, the specific content for generating tiered early warning signals includes: If the comprehensive risk index is lower than the comprehensive risk assessment standard, a Level 1 warning will be issued; If the comprehensive risk index equals the comprehensive risk assessment standard, then a level-two early warning will be issued; If the comprehensive risk index is higher than the comprehensive risk assessment standard, a level three warning will be issued.

10. A three-dimensional dynamic safety control system for power transmission and transformation hoisting operations, characterized in that, include: The data acquisition module is used to acquire point cloud data of the target work area using ground-based 3D laser scanning; The model building module is used to construct a 3D mesh model based on point cloud data using oblique photogrammetry and high-precision satellite positioning technology, and generate a semantic electronic fence. The clone construction module is used to construct a 4D digital clone of the suspended object based on pure vision technology through target detection, 3D reconstruction and temporal tracking, and obtain the predicted trajectory of the suspended object; The risk assessment module is used to calculate the shortest three-dimensional distance between the predicted trajectory of the suspended object and the semantic electronic fence through an efficient collision detection algorithm, based on the predicted trajectory of the suspended object and the semantic electronic fence, and to perform multi-dimensional risk assessment and generate graded early warning signals. The alarm notification module is used to provide comprehensive alarm notifications based on graded early warning signals, through audio-visual devices and a three-dimensional visualization monitoring interface.

Images