Overhead line threat grading assessment method and system based on dynamic trajectory tracking

Abstract

The application discloses an overhead line threat grading evaluation method and system based on dynamic trajectory tracking and belongs to the field of power facility safety. The system acquires video data streams and spatial position data of an overhead line and its surrounding environment through a data acquisition module; a data processing and analysis module constructs a three-dimensional space model based on the data, detects and classifies potential threat objects, generates time sequence three-dimensional coordinate sequences thereof, predicts future spatial positions, and, in combination with dynamic threat parameters such as instantaneous distance, speed, acceleration, attitude parameters and expected intrusion time, calculates comprehensive threat scores to determine threat grades; and an early warning and visualization module generates and publishes graded early warning information according to the threat grades. Through three-dimensional reconstruction and trajectory prediction, the application solves the problems of traditional two-dimensional monitoring, such as fuzzy spatial relationship and delayed alarm, and realizes accurate evaluation and forward-looking early warning of potential threats to the overhead line.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of power facility safety, and particularly relates to an overhead line threat grading evaluation method and system based on dynamic trajectory tracking. BACKGROUND

[0002] As a key infrastructure of the power system, the safe and stable operation of overhead transmission lines is of great importance. However, they are exposed to complex environments for a long time and continuously face dynamic potential threats such as large construction machinery, ultra-high vehicles, and unmanned aerial vehicles. To deal with these threats, the existing technology mainly relies on manual regular inspection, fixed video monitoring, and unmanned aerial vehicle inspection. However, these methods have significant limitations. Manual inspection is inefficient and slow to respond, and cannot provide real-time early warning for sudden dynamic threats. Fixed video monitoring can achieve remote monitoring, but it is mostly based on two-dimensional image analysis and uses a two-dimensional warning area to make judgments, which makes it difficult to accurately perceive the real three-dimensional spatial distance and relative motion situation between the threat object and the line. It is prone to false alarms due to light changes, birds, and other irrelevant factors. More importantly, it lacks the ability to predict the development trend of the threat, and often triggers an alarm only when the danger is imminent.

[0003] The inspection using advanced technologies such as unmanned aerial vehicles or laser radars can obtain high-precision three-dimensional data, but its operation mode is usually periodic "snapshot" detection, which is mainly used to find static or slowly changing hidden dangers such as tree barriers and illegal buildings. It is powerless against dynamic threats that suddenly appear and move quickly during the inspection gap. In summary, the existing technology generally lacks the ability to continuously track the dynamic trajectory of potential threat objects in three-dimensional space, and cannot conduct forward-looking quantitative risk assessment by combining key kinematic parameters such as speed and acceleration. Its early warning mechanism also mostly stays at the level of simple binary (alarm / no alarm) judgment, making it difficult to achieve fine-grained graded early warning and meet the urgent need for early, accurate, and intelligent defense against dynamic threats. SUMMARY

[0004] To solve the above problems in the existing technology, the present application provides an overhead line threat grading evaluation method based on dynamic trajectory tracking, which includes the following steps:

[0005] Step S1, real-time acquisition of video data stream and spatial position data containing overhead lines and their surrounding environment; based on the video data stream and spatial position data, a three-dimensional space model of the overhead lines is constructed, and at least one three-dimensional safety area is defined around the three-dimensional space model;

[0006] Step S2, identifying a potential threat object in the video data stream; continuously tracking the motion process of the potential threat object, and generating a dynamic trajectory of the potential threat object in the three-dimensional space model; in this process, the category and size information of the potential threat object are obtained;

[0007] Step S3, predicting the future three-dimensional space position of the potential threat object based on the dynamic trajectory; analyzing the real-time working posture of the potential threat object in the video data stream, and quantifying the real-time working posture into one or more posture threat parameters;

[0008] Based on the posture threat parameter, the expected intrusion time is calculated, specifically, based on the dynamic trajectory, a motion trend model of the potential threat object is established; according to the motion trend model and the posture threat parameter, the expected intrusion time of the potential threat object into different three-dimensional safety regions is calculated;

[0009] Step S4, by comparing the future three-dimensional space position of the potential threat object with the three-dimensional space model of the overhead line, the expected intrusion time, the dynamic threat parameter, the posture threat parameter and the category and size information of the potential threat object are scored by a comprehensive threat scoring function to obtain a comprehensive threat score, and the threat level is determined based on the comprehensive threat score; according to the threat level, a graded early warning information is generated and published.

[0010] Further, step S2 is specifically:

[0011] Step S201, detecting the potential threat object from the video data stream, and identifying the category and size information of the potential threat object;

[0012] Step S202, combining the position of the potential threat object in the video data stream and the spatial position data of the inspection equipment, the three-dimensional space coordinates of the potential threat object at each time are calculated;

[0013] Step S203, arranging the three-dimensional space coordinates in time sequence into a time sequence three-dimensional coordinate sequence as a dynamic trajectory in the three-dimensional space.

[0014] Further, the process of determining the threat level by evaluation in step S4 is specifically:

[0015] Step S401, by a comprehensive threat scoring function with multiple input variables, at least one parameter selected from the following group is mapped into a comprehensive threat score;

[0016] The parameters include: dynamic threat parameter, posture threat parameter, expected intrusion time, category and size information of the potential threat object;

[0017] The dynamic threat parameter includes:

[0018] Instantaneous spatial distance, the instantaneous spatial distance is the shortest distance between the three-dimensional spatial coordinates of the potential threat object at a certain moment and the three-dimensional spatial model of the overhead line;

[0019] Instantaneous velocity vector, the instantaneous velocity vector is the first-order derivative of the three-dimensional spatial coordinates of the potential threat object with respect to time;

[0020] Instantaneous radial velocity, the instantaneous radial velocity is the projection component of the instantaneous velocity vector in the direction of the shortest distance connecting the potential threat object and the overhead line;

[0021] Instantaneous acceleration vector, the instantaneous velocity vector is the second-order derivative of the three-dimensional spatial coordinates of the potential threat object with respect to time; with the three-dimensional spatial model of the overhead line as the reference, the instantaneous acceleration vector is decomposed into mutually orthogonal radial acceleration component and tangential acceleration component; the radial acceleration component represents the acceleration of approaching or moving away from the overhead line; the tangential acceleration component represents the acceleration perpendicular to the radial direction;

[0022] Step S402, compare the comprehensive threat score with a set of preset grade division thresholds to determine the corresponding threat grade; according to the threat grade, generate and publish the graded early warning information.

[0023] The overhead line threat grading evaluation system based on dynamic trajectory tracking, for the overhead line threat grading evaluation method based on dynamic trajectory tracking described above, comprising:

[0024] A data acquisition module is configured to acquire video data stream and spatial position data containing the overhead line and its surrounding environment in real time;

[0025] A data processing and analysis module is connected with the data acquisition module, and is configured to construct a three-dimensional spatial model of the overhead line based on the video data stream and the spatial position data; identify the potential threat object in the video data stream, and track the motion process of the potential threat object, generate a dynamic trajectory of the potential threat object in the three-dimensional space based on the motion process; predict the future three-dimensional spatial position of the potential threat object based on the dynamic trajectory; and evaluate the threat grade by comparing the future three-dimensional spatial position of the potential threat object with the three-dimensional spatial model of the overhead line;

[0026] An early warning and visualization module is connected with the data processing and analysis module, and is configured to generate and publish the graded early warning information according to the threat grade.

[0027] Further, the data processing and analysis module comprises:

[0028] A line three-dimensional spatial positioning submodule is configured to process the video data stream and the spatial position data, construct a three-dimensional spatial model of the overhead line, and define at least one three-dimensional safety area around the three-dimensional spatial model;

[0029] a dynamic target detection and classification submodule, configured to detect and identify potential threat objects from the video data stream and determine the category and size information of the potential threat objects;

[0030] a multi-target dynamic tracking submodule, configured to generate a time-series three-dimensional coordinate sequence as a dynamic trajectory of each identified potential threat object; wherein the time-series three-dimensional coordinate sequence is , wherein is an index of the potential threat object, is a three-dimensional space coordinate of the th potential threat object at time ;

[0031] a dynamic threat assessment submodule, configured to predict whether the potential threat object will enter the three-dimensional safe area based on the time-series three-dimensional coordinate sequence and assess the threat level based thereon.

[0032] Further, the dynamic threat assessment submodule further comprises a dynamic threat reasoning engine, and a processing procedure of the dynamic threat reasoning engine is specifically as follows:

[0033] receiving a set of input data, the input data comprising: dynamic threat parameters, at least one of which is extracted from the category and size information of the potential threat object and external scene information; the external scene information comprising meteorological data, traffic flow data and construction activity data;

[0034] based on the input data, reasoning through a built-in probabilistic graph model to output a comprehensive threat score as a basis for the threat level; the probabilistic graph model adopts a hidden Markov model or a Bayesian network;

[0035] the probabilistic graph model comprises at least one hidden state node, and the hidden state node is used to represent the threat intention of the potential threat object which cannot be directly observed.

[0036] The present application has the following beneficial effects:

[0037] The present application combines the three-dimensional space model of the overhead line with the dynamic trajectory tracking of the potential threat object, provides a precise spatial reference carrier for the category and size information, and completely solves the pain points of "fuzzy spatial relationship and inaccurate positioning" in traditional two-dimensional monitoring - for example, the trajectory offset of a large-size threat object (such as a crane) can be quickly judged in combination with the three-dimensional model to determine the vertical / horizontal safe distance from the line, while the size parameter of a small-size threat (such as a drone) can assist in excluding the interference of "misjudging birds", realizing the dual precise correlation of "spatial position + physical property".

[0038] More importantly, the present application constructs a dynamic risk prediction closed loop through the linkage calculation of the posture threat parameter, the motion trend model and the expected intrusion time; specifically:​

[0039] The posture threat parameter (such as the angle of the crane arm lifting, the flight posture of the unmanned aerial vehicle) can quantify the real-time operation state of the threat object, and directly correct the motion trend model; for example, when the threat object presents an operation posture of continuously approaching the line (such as the crane arm continuously lifting), the posture threat parameter will increase the invasion probability weight of the motion trend model, and then accurately shorten the calculation result of the expected invasion time; on the contrary, if the posture of the threat object tends to be stationary (such as a temporarily parked engineering vehicle), the posture parameter will reduce the threat of its motion trend, avoiding the problem of misjudging the emergency degree caused by relying on trajectory prediction alone.

[0040] Finally, the comprehensive threat score function is not simply the superposition of each parameter, but the risk quantification is realized through the weight linkage between parameters; for example, the combination of "large size + high posture threat parameter + short expected invasion time" will form a high comprehensive threat score, and the combination of "small size + low posture threat parameter + long expected invasion time" will correspond to a low score, and the evaluation logic of "different parameters confirming and correcting each other" is truly realized. This parameter coordination mechanism promotes the early warning mode from passive waiting for threshold triggering (such as relying only on distance threshold alarm) to forward-looking prediction based on multi-parameter linkage, which not only greatly improves the effective advance of early warning, but also reduces the false alarm rate through the mutual constraint between parameters, ensuring the balance between early warning accuracy and timeliness. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1 The block diagram of the system modules of the present application is shown. DETAILED DESCRIPTION

[0042] The present application will be further described below in conjunction with specific embodiments.

[0043] In order to further illustrate the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the following describes a new energy vehicle collision safety crush alloy crash performance detection method according to the present application, its specific implementation, structure, features and effects in detail in conjunction with the preferred embodiments and the drawings. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0044] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0045] An overhead line threat grading evaluation method based on dynamic trajectory tracking, comprising the following steps:

[0046] Step S1, real-time acquisition of video data stream and spatial position data containing overhead line and its surrounding environment; based on the video data stream and the spatial position data, a three-dimensional spatial model of the overhead line is constructed, and at least one three-dimensional safety area is defined around the three-dimensional spatial model;

[0047] Step S2, identifying a potential threat object in the video data stream; continuously tracking the motion process of the potential threat object to generate a dynamic trajectory of the potential threat object in the three-dimensional spatial model; in this process, the category and size information of the potential threat object are acquired; the detection method can adopt the YOLOv5 algorithm based on deep learning, which can detect unmanned aerial vehicles, birds, kites, balloons and other objects in real time under different lighting conditions; the extraction of size information can be realized by combining the known camera internal parameter and spatial depth data with triangulation; for example, in an embodiment, the system detects that the target is an unmanned aerial vehicle with a wingspan of 1.5m, and classifies it as a "large unmanned aerial vehicle" category;

[0048] Step S3, predicting the future three-dimensional spatial position of the potential threat object based on the dynamic trajectory; analyzing the real-time operation posture of the potential threat object in the video data stream, and quantifying the real-time operation posture into one or more posture threat parameters;

[0049] Based on the posture threat parameter, the expected intrusion time is calculated, specifically, based on the dynamic trajectory, a motion trend model of the potential threat object is established; according to the motion trend model and the posture threat parameter, the expected intrusion time of the potential threat object into different three-dimensional safety areas is calculated;

[0050] Step S4, by comparing the future three-dimensional spatial position of the potential threat object with the three-dimensional spatial model of the overhead line, the expected intrusion time, the dynamic threat parameter, the posture threat parameter and the category and size information of the potential threat object are scored by a comprehensive threat scoring function to obtain a comprehensive threat score, and the threat level is determined based on the comprehensive threat score; according to the threat level, graded warning information is generated and published.

[0051] Further, step S2 is specifically:

[0052] Step S201, detecting the potential threat object from the video data stream, and identifying the category and size information of the potential threat object;

[0053] Step S202, combining the position of the potential threat object in the video data stream and the spatial position data of the inspection equipment, the three-dimensional spatial coordinates of the potential threat object at each time are calculated; the calculation process adopts PnP (Perspective-n-Point) algorithm combined with EKF (Extended Kalman Filter) for optimization, to ensure that the target can still obtain smooth and continuous three-dimensional coordinates when moving quickly;

[0054] Step S203, organize the three-dimensional space coordinates arranged in time sequence into a time sequence three-dimensional coordinate sequence as a dynamic trajectory in three-dimensional space.

[0055] Further, the process of determining the threat level by evaluation in step S4 is specifically:

[0056] Step S401, map at least one parameter selected from the following group into a comprehensive threat score by a comprehensive threat scoring function of a multivariate input:

[0057] The parameters include: dynamic threat parameters, posture threat parameters, expected intrusion time, and category and size information of potential threat objects.

[0058] The dynamic threat parameters include:

[0059] Instantaneous spatial distance, which is the shortest distance between the three-dimensional space coordinates of the potential threat object at a certain moment and the three-dimensional space model of the overhead line;

[0060] Instantaneous velocity vector, which is the first derivative of the three-dimensional space coordinates of the potential threat object with respect to time;

[0061] Instantaneous radial velocity, which is the projection component of the instantaneous velocity vector in the direction of the shortest distance between the potential threat object and the overhead line;

[0062] Instantaneous acceleration vector, which is the second derivative of the three-dimensional space coordinates of the potential threat object with respect to time; with the three-dimensional space model of the overhead line as the reference, the instantaneous acceleration vector is decomposed into mutually orthogonal radial acceleration component and tangential acceleration component; the radial acceleration component represents the acceleration of approaching or moving away from the overhead line; the tangential acceleration component represents the acceleration perpendicular to the radial direction;

[0063] Step S402, compare the comprehensive threat score with a group of preset level division threshold values to determine the corresponding threat level; according to the threat level, generate and issue a graded warning information.

[0064] Reference Figure 1 , the overhead line threat grading evaluation system based on dynamic trajectory tracking, for the overhead line threat grading evaluation method based on dynamic trajectory tracking described above, comprising:

[0065] The data acquisition module is configured to acquire video data stream and spatial position data of the overhead line and its surrounding environment in real time. In an embodiment, the video data stream is preferably acquired by a high-definition industrial camera installed on a UAV, a tower monitoring camera or a patrol robot, and the camera resolution is preferably 1920x1080 or above to ensure the accuracy of threat object identification. The spatial position data can be acquired by fusing GPS, Beidou navigation system and inertial measurement unit (IMU) to ensure that the positioning accuracy is still centimeter level in complex terrain. Preferably, the industrial camera and the IMU are installed on the gimbal platform of the UAV, the gimbal stabilizes the picture through a three-axis stabilizing mechanism, and the GPS antenna is fixed on the top of the UAV body to acquire high-precision position data.

[0066] The data processing and analysis module is connected with the data acquisition module, and is configured to construct a three-dimensional space model of the overhead line based on the video data stream and the spatial position data, identify potential threat objects in the video data stream, track the motion process of the potential threat objects, generate dynamic trajectories of the potential threat objects in the three-dimensional space based on the motion process, predict the future three-dimensional space positions of the potential threat objects based on the dynamic trajectories, and evaluate the threat level by comparing the future three-dimensional space positions of the potential threat objects with the three-dimensional space model of the overhead line.

[0067] The early warning and visualization module is connected with the data processing and analysis module, and is configured to generate and issue graded early warning information according to the threat level.

[0068] Further, the data processing and analysis module comprises:

[0069] The line three-dimensional space positioning submodule is configured to process the video data stream and the spatial position data, construct a three-dimensional space model of the overhead line, and define at least one three-dimensional safety area around the three-dimensional space model. In this embodiment, the three-dimensional space model can be realized by a structured light reconstruction algorithm, stereo vision calculation or laser radar scanning. The laser radar is installed below the UAV and can output point cloud data in real time, which is fused with the video frames to form a complete three-dimensional line and environment model.

[0070] The dynamic target detection and classification submodule is configured to detect and identify potential threat objects from the video data stream, and determine the category and size information of the potential threat objects.

[0071] The multi-target dynamic tracking submodule is configured to generate a time sequence three-dimensional coordinate sequence as a dynamic trajectory of each identified potential threat object. The time sequence three-dimensional coordinate sequence is as follows: wherein, is the index of the potential threat object, is the time sequence three-dimensional coordinate sequence of the i-th potential threat object at time t, and is the time sequence three-dimensional coordinate sequence of the i-th potential threat object at time t+1.​ three-dimensional spatial coordinates; in this embodiment, the early warning information is sent in real time to the ground monitoring center through a wireless link and is displayed in a graphical manner in the monitoring terminal interface. The visualization interface can be implemented using a three-dimensional GIS platform, which can display overhead lines, potential threat objects, and their dynamic trajectories on a three-dimensional map, and color-code different levels of threats;

[0072] a dynamic threat assessment submodule that predicts whether a potential threat object will enter a three-dimensional safety region based on the time-series three-dimensional coordinate sequence, and assesses the threat level accordingly. The three-dimensional safety region can be defined as a buffer zone that is expanded outward by 5m to 20m from the outer diameter of the conductor, so as to provide a high-risk early warning for objects that intrude into the buffer zone. In actual applications, the safety zone range can be dynamically adjusted according to the voltage level of the power transmission line, for example, a 10m safety zone for a 220kV line and a 20m safety zone for a 500kV line.

[0073] Further, the dynamic threat assessment submodule further includes a dynamic threat reasoning engine, and the processing flow of the dynamic threat reasoning engine is specifically as follows:

[0074] receiving a set of input data, the input data including: dynamic threat parameters, at least one of which is extracted from the category and size information of the potential threat object, and external scene information; the external scene information includes meteorological data, traffic flow data, and construction activity data;

[0075] based on the input data, reasoning through an embedded probabilistic graph model to output a comprehensive threat score as the basis for the threat level; the probabilistic graph model uses a hidden Markov model or a Bayesian network;

[0076] the probabilistic graph model includes at least one hidden state node, which is used to represent the threat intention of the potential threat object that cannot be directly observed. For example, when the target is a "large unmanned aerial vehicle" and its size information indicates that the wingspan exceeds 2m, and the external scene information is "close to a high-voltage power transmission corridor", the probabilistic graph model can infer that the threat intention node probability is relatively high, and the final comprehensive threat score is greater than that of an ordinary small bird.

[0077] The above embodiments are only used to illustrate the technical solutions of the present application, but not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A method for overhead line threat ranking assessment based on dynamic trajectory tracking, characterized in that, The method comprises the following steps: Step S1, real-time acquisition of video data stream and spatial position data containing overhead line and its surrounding environment; based on the video data stream and the spatial position data, a three-dimensional space model of the overhead line is constructed, and at least one three-dimensional safety area is defined around the three-dimensional space model; Step S2, identifying a potential threat object in the video data stream; continuously tracking the movement process of the potential threat object to generate a dynamic trajectory of the potential threat object in the three-dimensional space model; in this process, the category and size information of the potential threat object are obtained; Step S3, predicting the future three-dimensional space position of the potential threat object based on the dynamic trajectory; analyzing the real-time working posture of the potential threat object in the video data stream, and quantifying the real-time working posture into one or more posture threat parameters; based on the posture threat parameter, calculating the expected intrusion time, specifically, based on the dynamic trajectory, establishing a motion trend model of the potential threat object; according to the motion trend model and the posture threat parameter, the expected intrusion time of the potential threat object into different three-dimensional safety areas is calculated; Step S4, by comparing the future three-dimensional space position of the potential threat object with the three-dimensional space model of the overhead line, scoring the expected intrusion time, the dynamic threat parameter, the posture threat parameter and the category and size information of the potential threat object through a comprehensive threat scoring function to obtain a comprehensive threat score, and determining a threat level based on the comprehensive threat score; according to the threat level, generating and publishing a graded warning information; The process of determining the threat level by evaluation in step S4 is specifically: Step S401, through a comprehensive threat scoring function with multiple variable inputs, at least one parameter selected from the following group is mapped to a comprehensive threat score; The parameters include: dynamic threat parameter, posture threat parameter, expected intrusion time, category and size information of the potential threat object; The dynamic threat parameter includes: instantaneous spatial distance, which is the shortest distance between the three-dimensional space coordinates of the potential threat object at a certain moment and the three-dimensional space model of the overhead line; instantaneous velocity vector, which is the first derivative of the three-dimensional space coordinates of the potential threat object with respect to time; instantaneous radial velocity, which is the projection component of the instantaneous velocity vector in the direction of the shortest distance between the potential threat object and the overhead line; instantaneous acceleration vector, which is the second derivative of the three-dimensional space coordinates of the potential threat object with respect to time; taking the three-dimensional space model of the overhead line as a reference, the instantaneous acceleration vector is decomposed into a radial acceleration component and a tangential acceleration component which are orthogonal to each other; the radial acceleration component represents the acceleration of approaching or moving away from the overhead line; the tangential acceleration component represents the acceleration perpendicular to the radial direction; Step S402, comparing the comprehensive threat score with a group of preset level division thresholds to determine the corresponding threat level; according to the threat level, generating and publishing a graded warning information.

2. The overhead line threat grading evaluation method based on dynamic trajectory tracking according to claim 1, wherein: Step S2 is specifically: Step S201, detecting a potential threat object from the video data stream, and identifying the category and size information of the potential threat object; Step S202, combining the position of the potential threat object in the video data stream and the spatial position data of the inspection equipment, calculating the three-dimensional spatial coordinates of the potential threat object at each time; Step S203, organizing the three-dimensional spatial coordinates arranged in time sequence into a time sequence three-dimensional coordinate sequence as a dynamic trajectory in three-dimensional space.

3. An overhead line threat classification and assessment system based on dynamic trajectory tracking for the overhead line threat classification and assessment method based on dynamic trajectory tracking as claimed in claim 1 or 2, characterized in that: Comprise: A data acquisition module for acquiring video data stream and spatial position data containing overhead lines and their surrounding environment in real time; A data processing and analysis module connected with the data acquisition module, based on the video data stream and the spatial position data, constructing a three-dimensional spatial model of the overhead line; identifying potential threat objects in the video data stream, and tracking the motion process of the potential threat objects, generating a dynamic trajectory of the potential threat objects in three-dimensional space based on the motion process; Based on the dynamic trajectory, predicting the future three-dimensional spatial position of the potential threat object; By comparing the future three-dimensional spatial position of the potential threat object with the three-dimensional spatial model of the overhead line, the threat level is evaluated; An early warning and visualization module connected with the data processing and analysis module, for generating and publishing graded warning information according to the threat level.

4. The overhead line threat classification and assessment system based on dynamic trajectory tracking of claim 3, wherein: The data processing and analysis module comprises: A line three-dimensional spatial positioning submodule for processing video data stream and spatial position data, constructing a three-dimensional spatial model of the overhead line, and defining at least one three-dimensional safety area around the three-dimensional spatial model; A dynamic target detection and classification submodule for detecting and identifying potential threat objects from the video data stream, and determining the category and size information of the potential threat objects; The multi-target dynamic tracking sub-module is configured to generate a time-series three-dimensional coordinate sequence as a dynamic trajectory of each identified potential threat object; wherein the time-series three-dimensional coordinate sequence is for: , wherein, is an index of the potential threat object, is a three-dimensional space coordinate of the th potential threat object at time . A dynamic threat assessment submodule for predicting whether the potential threat object will enter the three-dimensional safety area based on the time sequence three-dimensional coordinate sequence, and evaluating the threat level accordingly.

5. The overhead line threat classification and assessment system based on dynamic trajectory tracking as claimed in claim 4 wherein: The dynamic threat assessment submodule further comprises a dynamic threat reasoning engine, and the processing flow of the dynamic threat reasoning engine is specifically: Receive a set of input data, including: dynamic threat parameters, at least one of which is extracted from the category and size information of the potential threat object, external scene information including meteorological data, traffic flow data and construction activity data; Based on the input data, reasoning through the built-in probabilistic graph model to output a comprehensive threat score as the basis for threat level; the probabilistic graph model uses hidden Markov model or Bayesian network; The probabilistic graph model contains at least one hidden state node, which is used to represent the threat intention of the potential threat object that cannot be directly observed.

Images