Methods, systems, and equipment for early warning of transmission line risks based on bird migration.
By acquiring bird migration parameters and environmental constraints, candidate migration trajectories are predicted and a three-dimensional risk buffer is established. The overlap probability and centroid sequence are calculated, and graded early warnings are carried out. This solves the problems of missed and false alarms in bird damage early warning in existing technologies and achieves precise prevention and control of transmission line risks.
Patent Information
- Application Number
- CN202611262995.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-19
- Publication Date
- 2026-11-17
AI Technical Summary
Existing bird damage early warning technologies cannot accurately predict bird migration routes, leading to missed and false reports, and are unable to effectively prevent power transmission line accidents caused by bird damage.
By acquiring bird migration parameters and environmental constraints, candidate migration trajectories are predicted, a three-dimensional risk buffer zone is established, overlap probability and centroid sequence are calculated, and graded early warnings are implemented. By combining movement patterns and continuous overlap duration, the accuracy of early warnings is improved.
It significantly improves the accuracy of short-term bird damage early warning, reduces missed and false alarms, and ensures the effectiveness of operation and maintenance decisions.
Smart Images

Figure CN123117432A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of risk warning, and in particular to a method, system, and equipment for risk warning of power transmission lines based on bird migration. Background Technology
[0002] Large-scale migratory bird flocks often traverse mountainous areas, river valleys, and wetlands during their flights, areas that coincide significantly with power transmission line corridors. When flocks fly at low altitudes or cross power lines, their droppings contaminate insulators, their bodies short-circuit air gaps, and they collide with conductors in mid-air, all of which can trigger insulator flashovers, line trips, or even line breaks, causing serious consequences such as power outages, equipment damage, and even forest fires. With increased efforts in ecological protection, bird populations and their activity ranges continue to expand, leading to bird-related faults that are seasonally concentrated, sudden, and have a wide impact. Traditional "post-incident inspection" and "passive fault elimination" models are no longer sufficient to meet the requirements for safe power grid operation. Therefore, it is urgent to develop short-term early warning systems based on the real-time migration status of bird flocks, moving the prevention and control point forward, gaining maintenance response time before the flocks arrive at the lines, and transforming passive handling into proactive defense.
[0003] Existing bird damage early warning technologies mainly fall into two categories: one is based on historical bird-related fault records, topography, and bird habitats to create static bird damage distribution maps for long-term planning and deployment of bird control devices. However, its time resolution is based on annual or quarterly statistics, which cannot answer the short-term question of "when it happens," and its kilometer-level grid accuracy is far from meeting the positioning requirements of poles and towers. The other category uses radar, video, or audio equipment to identify and track bird targets in real time to determine whether birds have appeared or are approaching the line. However, its early warning logic usually extrapolates the observed single flight path into a predicted center line and superimposes a two-dimensional horizontal buffer zone for planar intersection judgment. This method has three inherent flaws: First, the actual migration paths of bird flocks are affected by multiple factors such as wind fields, topographic uplift, river valley guidance, and habitat attraction, resulting in significant branching and diffusion. A single centerline is prone to missing bird flocks that will actually pass through the route. Second, simply enlarging the two-dimensional buffer radius to compensate for missed reports will cause a large number of bird flocks flying at altitudes far above the guide line to be included in the warning, resulting in serious false alarms. Third, completely ignoring the altitude dimension makes it impossible to determine whether bird flocks have actually entered the risk altitude of the guide line sag, leading to a large number of invalid warning pushes, which not only wastes operation and maintenance resources but also weakens the attention of operation and maintenance personnel to real danger signals. Summary of the Invention
[0004] This invention provides a method, system, and device for early warning of transmission line risks based on bird migration, which can improve the accuracy of risk warning.
[0005] This invention provides a method for early warning of transmission line risks based on bird migration, comprising: To obtain real-time migration parameters of the target flock of birds, environmental constraints during flight, and spatial structural parameters of the power transmission line; Based on the real-time migration parameters and environmental constraints, several candidate migration trajectories of the target bird flock within the future time window are predicted, and the candidate migration trajectories within each time slice are mapped to spatial units to calculate the flight probability envelope. Based on the aforementioned spatial structure parameters, a three-dimensional risk buffer zone for the transmission line is established. Within each time slice, the overlapping probability sequence and the probability centroid sequence are calculated based on the flight probability envelope and the three-dimensional risk buffer. The duration of continuous overlap between the target bird flock and the transmission line is determined based on the overlap probability sequence, and the movement pattern of the target bird flock relative to the transmission line is determined based on the probability centroid sequence. A graded early warning system is implemented based on the motion pattern and the duration of continuous overlap.
[0006] This invention provides a data foundation reflecting the true movement patterns and environmental response characteristics of bird flocks by fusing real-time migration parameters and environmental constraints. This avoids prediction bias caused by relying solely on historical statistics or simple extrapolation, ensuring data quality for accurate early warnings from the source. By generating multiple candidate migration trajectories and mapping them to spatial units to form a flight probability envelope, a probability distribution replaces a single deterministic path. This fully characterizes the bifurcation and diffusion characteristics of bird migration paths influenced by multiple factors such as wind field, terrain, and habitat, effectively solving the technical problem of missed reports under path bifurcation conditions when extrapolating a single path. It also avoids blindly expanding the early warning range to compensate for missed reports. By establishing a three-dimensional risk buffer zone for transmission lines that includes a height dimension, the early warning judgment is elevated from a two-dimensional plane to a three-dimensional space. This accurately reflects the true risk boundaries of conductor sag, protection gaps, and tower structures, effectively excluding irrelevant bird flocks flying at altitudes far above the conductors. This significantly reduces false alarms caused by traditional two-dimensional horizontal buffering methods from a height dimension. By calculating the flight probability envelope within each time slice... The overlapping probability sequence and probability centroid sequence of the network area and the three-dimensional risk buffer zone incorporate trajectory uncertainty, spatial three-dimensional intersection relationship, degree of high matching and arrival probability into a unified temporal quantification framework. This enables a precise characterization of the spatiotemporal coupling relationship between bird flocks and line risks, avoiding random errors caused by qualitative judgment or single-point threshold comparison. By extracting the continuous overlap duration from the overlapping probability sequence, the continuous exposure time of bird flocks in the line risk space is incorporated into the evaluation system as an independent risk dimension. This solves the problem that relying solely on instantaneous probability cannot distinguish between brief brushes and long-term continuous threats, making the early warning triggering conditions more consistent with the actual risk evolution law. By determining the movement pattern based on the probability centroid sequence and combining it with the continuous overlap duration for graded early warning, it is possible to distinguish between two risk mechanisms: long-term accompanying flight along the line direction and rapid crossing in the vertical direction. This avoids the distortion of early warning levels caused by different risk modes using the same evaluation scale, making the final output early warning level more consistent with the actual risk level. As a result, the accuracy of short-term bird damage early warning is significantly improved overall, and the interference of missed and false alarms on operation and maintenance decisions is reduced.
[0007] Furthermore, the prediction of several candidate migration trajectories of the target bird flock within a future time window based on the real-time migration parameters and the environmental constraints includes: Based on the environmental constraints, determine the state transition probability of the current state relative to the previous state; By constraining the deflection direction and spatial reachability of the trajectory through the state transition probability, several candidate migration trajectories within a future time window are generated.
[0008] By introducing environmental constraints to calculate state transition probabilities, the generated candidate migration trajectories are not simply extrapolated, but fully consider the guiding and restrictive effects of environmental factors such as wind field and terrain on the actual flight direction of bird flocks. This avoids the single extrapolation model from deviating significantly from the real path in complex environments, making the trajectory prediction closer to the actual movement patterns of bird flocks. The coverage accuracy of the candidate trajectory set is improved, providing a reliable data foundation for the subsequent construction of the probability envelope.
[0009] Furthermore, the flight probability envelope region includes a core probability envelope region and an extended probability envelope region, and the step of mapping the candidate migration trajectories within each time slice to spatial cells to calculate the flight probability envelope region includes: For each time slice, the three-dimensional position of each candidate migration trajectory is mapped to a spatial unit containing the height dimension; The probability of a flock arriving in the current time slice of a spatial unit is obtained by summing the normalized weights of each candidate migration trajectory falling into the same spatial unit in the current time slice, wherein the normalized weights are calculated based on the real-time migration parameters and the environmental constraints. Within each time slice, the spatial units are sorted from largest to smallest according to the arrival probability of the flock of birds, and the sorting results are obtained. The arrival probabilities of the flock of birds in the sorting results are accumulated according to the sorting direction to obtain the cumulative arrival probability. When the cumulative arrival probability reaches a first preset threshold, the corresponding spatial unit constitutes the core probability envelope region. Continue selecting along the sorting direction until the cumulative arrival probability reaches the second preset threshold. All selected spatial units constitute an extended probability envelope region, wherein the second preset threshold is greater than the first preset threshold.
[0010] By sorting by arrival probability and using a two-layer cumulative threshold strategy to construct a core and an extended probability envelope, high-probability regions and low-probability peripheral regions can be distinguished and expressed. The core envelope ensures high-confidence judgment, while the extended envelope covers peripheral risks caused by path bifurcation and observation errors. This expands the early warning coverage while avoiding excessive expansion that could lead to false alarms, accurately balancing the contradiction between missed and false alarms.
[0011] Furthermore, the normalized weights are calculated based on the real-time migration parameters and the environmental constraints, including: Within the current time slice, obtain the historical normalized weights of each candidate migration trajectory from the previous time slice; Calculate the observation matching likelihood based on the received bird flock observation data; Based on the historical normalized weights, the state transition probabilities, and the observation matching likelihood, the unnormalized weights of the current time slice are calculated. The unnormalized weights of all the candidate migration trajectories are summed to obtain the summation result, and the normalized weight of each candidate migration trajectory in the current time slice is calculated based on the summation result.
[0012] This approach integrates the historical weights from the previous moment, the probability of environmental transition, and the matching likelihood of new observation data, enabling rolling updates of normalized weights. This ensures that the weights of each candidate trajectory are continuously adjusted as new observations arrive, preventing significant weight jumps or lags behind the actual movement changes of the flock. It guarantees that trajectory probability allocation always keeps pace with the real-time dynamics of the flock, providing accurate support for arrival probability calculations.
[0013] Furthermore, the spatial structure parameters include tower coordinates, span, conductor suspension point position, conductor sag parameters, and line corridor width. Establishing a three-dimensional risk buffer zone for the transmission line based on these spatial structure parameters includes: Based on the tower coordinates, the span, and the conductor suspension point positions, determine the suspension points at both ends of each phase conductor; Based on the coordinates of the two hanging points and the sag parameters of the conductor, a spatial curve of the conductor is constructed; The horizontal buffer radius is determined based on the line channel width and the preset horizontal error margin, and the vertical buffer radius is determined based on the protection gap and the preset height error margin. Based on the conductor space curve, the spatial ranges corresponding to the horizontal buffer radius and the vertical buffer radius are superimposed to form a three-dimensional risk buffer zone with corresponding span. A tower risk body is constructed with the tower coordinates as the center. The three-dimensional risk buffer zone of adjacent spans is then combined with the tower risk body to obtain the complete three-dimensional risk buffer zone of the transmission line.
[0014] By using the guideline space curve as a reference and superimposing horizontal and vertical buffer radii to construct a three-dimensional buffer zone, the risk assessment of bird flocks is elevated from a two-dimensional plane to a three-dimensional space. At the same time, the error margin corresponding to the accuracy of the data source is introduced, so that the buffer zone boundary not only fits the actual spatial shape of the guideline but also has reasonable engineering redundancy. This effectively avoids the two-dimensional judgment from misjudging bird flocks flying at altitudes far above the guideline as risk, and significantly reduces the number of invalid warnings.
[0015] Further, the calculation of the overlapping probability sequence and the probability centroid sequence within each time slice, based on the flight probability envelope and the three-dimensional risk buffer, includes: For each time slice, all spatial units within the flight probability envelope are selected, and it is determined whether each spatial unit intersects with the three-dimensional risk buffer zone of the transmission line, thus identifying several intersecting spatial units. Obtain the flock flight height corresponding to each of the intersecting spatial units and the guide risk height corresponding to the position of the three-dimensional risk buffer, calculate the height difference between the flock flight height and the guide risk height, and obtain the height matching weight; The probability of the flock of birds arriving in each of the intersecting spatial units is multiplied by the corresponding height matching weight and then summed to obtain the overlap probability under that time slice. The overlapping probabilities are arranged in chronological order of the time slices to form an overlapping probability sequence; Based on the spatial position of each intersecting spatial unit, a weighted average is calculated according to the matching weight of the bird flock arrival probability and the altitude to obtain the probability centroid of the time slice, and the probability centroids are arranged in chronological order to form a probability centroid sequence.
[0016] By introducing height matching weights to weight the arrival probabilities of intersecting spatial units, only flocks of birds flying at altitudes close to the risk height of the guide line can contribute a higher overlap probability. Even if flocks of birds flying at altitudes far above or far below the guide line overlap with the line in horizontal position, their overlap probability will be significantly reduced. This achieves precise filtering in the height dimension at the probability source, significantly improves the spatial accuracy of early warnings, and reduces false alarms caused by height mismatch.
[0017] Further, the spatial structure parameters include the line alignment; determining the continuous overlap duration of the target bird flock with respect to the transmission line based on the overlap probability sequence, and determining the movement pattern of the target bird flock relative to the transmission line based on the probability centroid sequence, includes: Traverse the overlapping probabilities in the overlapping probability sequence to identify a set of continuous time periods where the overlapping probability is not lower than a preset effective threshold. Select the time period with the longest duration from the set of continuous time periods as the target influence time period and use the duration of the target influence time period as the continuous overlapping duration. Within each time slice covered by the target's influence period, the positions of the probability centroids of adjacent time slices are differentially calculated and normalized to obtain the local movement direction of the bird flock's trajectory. Calculate the angle between the local movement direction and the line direction, and determine the movement pattern of the target bird flock relative to the transmission line based on the numerical range of the angle.
[0018] By extracting the longest continuous period not lower than the effective threshold from the overlapping probability sequence as the continuous overlapping duration, and combining it with the probability centroid difference to obtain the angle between the local movement direction and the route direction, it is possible to distinguish whether the flock of birds is flying along the route direction for a long time or crossing it quickly laterally. This avoids confusing the two different risk mechanisms for evaluation, making the subsequent early warning judgment more in line with the actual risk causes, and further improving the accuracy of the early warning.
[0019] Furthermore, the motion mode includes a continuous exposure mode and a momentary crossing mode, and the graded early warning based on the motion mode and the duration of continuous overlap includes: If it is the continuous exposure mode, then the continuous exposure weight is applied to the continuous overlap duration to correct it, and the corrected duration is obtained; If it is the instantaneous crossing mode, then the maximum overlap probability in the overlap probability sequence is corrected by applying an instantaneous overlap weight to obtain the corrected probability; The warning result of the corresponding level is output by comprehensively judging the correction duration, correction probability, bird flock size and route importance parameters.
[0020] By applying differentiated directional weights to the overlap duration and maximum overlap probability for the continuous exposure mode and the instantaneous crossing mode respectively, the early warning basis for the two risk modes is matched with their actual disaster-causing mechanisms. This avoids underestimating or overestimating the risks caused by using a uniform evaluation standard for all overlap situations, and achieves accurate graded early warning.
[0021] Another embodiment of the present invention provides a transmission line risk early warning system based on bird migration, comprising: the system being used to perform the steps of the transmission line risk early warning method based on bird migration as described in the present invention.
[0022] Another embodiment of the present invention provides a terminal device, including: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the steps of the transmission line risk warning method based on bird migration as described in the present invention. Attached Figure Description
[0023] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.
[0024] Figure 1 This is a flowchart illustrating an embodiment of the power transmission line risk warning method based on bird migration provided in this application. Figure 2 This is a flowchart illustrating one embodiment of steps S201 to S204 provided in this application; Figure 3 This is a flowchart illustrating one embodiment of steps S301 to S305 provided in this application; Figure 4This is a flowchart illustrating one embodiment of steps S401 to S405 provided in this application; Figure 5 This is a flowchart illustrating one embodiment of steps S501 to S503 provided in this application. Detailed Implementation
[0025] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0026] 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 this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.
[0027] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.
[0028] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.
[0029] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.
[0030] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).
[0031] In the description of the embodiments of this application, unless otherwise expressly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. For those skilled in the art, the specific meaning of the above terms in the embodiments of this application can be understood according to the specific circumstances.
[0032] See Figure 1 To improve the accuracy of risk warning, an embodiment of the present invention provides a method, system, and device for risk warning of power transmission lines based on bird migration, including steps S101 to S106: Step S101: Obtain the real-time migration parameters of the target flock of birds, the environmental constraints of their flight, and the spatial structure parameters of the power transmission line. In some embodiments, for the real-time migration parameters of the target bird flock, the system relies on weather radar, low-altitude monitoring radar, video or audio-visual terminals to collect raw data, and combines the results of manual verification to establish a target bird flock status record. The recorded parameters include at least latitude and longitude, flight altitude, flight speed, migration direction, target density, flock size and identification fusion confidence, and can also be associated with the corresponding radar trajectory number. All parameters are uniformly used as the input basis for subsequent candidate trajectory generation and prediction error estimation.
[0033] To address the environmental constraints of bird flocks' flight, the system simultaneously collects information on wind speed, wind direction, terrain elevation, slope, valley orientation, historical migration routes, candidate habitats, no-fly or obstruction areas, time periods, and seasons. Among these, wind field data is used to correct for the difference between ground speed and air speed of bird flocks, terrain and valley data are used to constrain the accessibility and deflection direction of candidate routes, and historical migration route and candidate habitat data are used to reflect the bird flocks' route selection preferences.
[0034] For the spatial structural parameters of transmission lines, the system retrieves them from the transmission line ledger, GIS platform, 3D line model, inspection system or operation and maintenance management system. The parameters include at least the line name, voltage level, line importance, tower coordinates, tower type, span range, conductor hanging point height, conductor sag, line direction, ground elevation, protection gap, passage width, crossing area attributes, equipment sensitivity level and corresponding operation and maintenance personnel information; among them, the line importance and equipment sensitivity level can be marked by combining the original fault ledger of the project.
[0035] After the data collection is completed, all bird flock data and route data are uniformly converted to the same spatial coordinate system, time reference, and altitude reference to eliminate the impact of reference differences on subsequent spatial coupling calculations.
[0036] It should be noted that for target transmission line data, the system prioritizes using measured 3D line models, inspection point clouds, or UAV inspection results. When some parameters are missing, the system sets a safety margin based on the accuracy of the data source. For spans lacking precise sag parameters, linear interpolation can be performed based on the conductor suspension point height of adjacent towers, or catenary or parabolic approximations can be used, superimposed with a vertical safety margin determined by voltage level, span length, terrain undulation, and data source accuracy. Bird flock data and line data are synchronized in spatial coordinate system, time reference, and height reference before overlapping calculations.
[0037] Step S102: Based on the real-time migration parameters and the environmental constraints, predict several candidate migration trajectories of the target bird flock within the future time window, and map the candidate migration trajectories in each time slice to spatial units to calculate the flight probability envelope. In some embodiments, predicting several candidate migration trajectories for a target bird flock within a future time window based on the real-time migration parameters and environmental constraints includes: determining the state transition probability of the current state relative to the previous state based on the environmental constraints; and constraining the deflection direction and spatial accessibility of the trajectory through the state transition probability to generate several candidate migration trajectories within the future time window. Specifically, the state transition probability considers five types of constraint factors: migration direction consistency, wind field accessibility, terrain and altitude change costs, proximity to historical migration routes, and attractiveness of candidate habitats. Among them, wind field accessibility corrects the difference between the bird flock's ground speed and air speed based on real-time wind speed and direction, with downwind paths corresponding to higher transition probabilities; terrain and altitude change costs set cost thresholds for path deflection and altitude gain based on terrain elevation and slope, with terrain barriers and steep elevation gain areas corresponding to lower transition probabilities; valley orientation, historical migration routes, and candidate habitats correspond to path preference weights, allocating higher transition probabilities to directions and spatial locations that conform to the bird flock's migratory habits, thereby constraining the deflection direction and spatial accessibility of candidate trajectories and eliminating invalid paths that do not meet environmental constraints. Subsequently, using a three-dimensional voxel grid and a transition probability matrix as a unified data foundation, a number of candidate migration trajectories within the future time window are generated through path search based on the three-dimensional grid transition probability matrix. Each trajectory consists of multiple three-dimensional location points sorted by time, and each location point stores the horizontal coordinates, flight altitude, corresponding time, and trajectory weight. The prediction time window can be dynamically set according to the radar refresh cycle, bird flock speed, route importance, and maintenance response time.
[0038] By introducing environmental constraints to calculate state transition probabilities, the generated candidate migration trajectories are not simply extrapolated, but fully consider the guiding and restrictive effects of environmental factors such as wind field and terrain on the actual flight direction of bird flocks. This avoids the single extrapolation model from deviating significantly from the real path in complex environments, making the trajectory prediction closer to the actual movement patterns of bird flocks. The coverage accuracy of the candidate trajectory set is improved, providing a reliable data foundation for the subsequent construction of the probability envelope.
[0039] Please refer to Figure 2 In some embodiments, the flight probability envelope region includes a core probability envelope region and an extended probability envelope region. The step of mapping the candidate migration trajectories within each time slice to spatial cells to calculate the flight probability envelope region includes steps S201 to S204: Step S201: For each time slice, map the three-dimensional position of each candidate migration trajectory to a spatial unit containing the height dimension. In some embodiments, the union of the predicted coverage area and the target transmission line corridor is used as the computational region. This region is discretized into several three-dimensional voxels in the horizontal and vertical directions, with each voxel being a spatial unit. The dimensions of the voxels in the horizontal direction are determined based on the radar horizontal positioning error, the width of the line corridor, and the required positioning accuracy of the towers or spans. The dimensions of the voxels in the vertical direction are determined based on the radar height measurement error, the protection gap, and the vertical safety margin. The prediction time window is divided into sequentially arranged time slices according to a fixed time step, with the time step determined based on the radar refresh cycle and the early warning response time. In a non-limiting embodiment, the side length of the voxels in the horizontal direction can be 50 to 200 meters, and the height of the voxels in the vertical direction can be 10 to 25 meters. The specific values are calibrated according to the actual engineering scenario and do not constitute a necessary limitation of the present invention. For each candidate migration trajectory within each time slice, its three-dimensional position coordinates are assigned to a unique corresponding spatial unit through coordinate mapping, that is, each voxel can accommodate the landing points of multiple candidate trajectories within each time slice. After completing the above division, the system also establishes a spatial index based on the voxel center coordinates and height range, and pre-records the candidate association relationships between each voxel and the tower risk body and the span conductor buffer body, so as to reduce the spatial retrieval calculation amount during subsequent time-by-time slice coupling.
[0040] Step S202: Accumulate the normalized weights of each candidate migration trajectory falling into the same spatial unit in the current time slice to obtain the bird flock arrival probability of the spatial unit in the current time slice, wherein the normalized weights are calculated based on the real-time migration parameters and the environmental constraints. In some embodiments, the normalized weights are calculated based on the real-time migration parameters and the environmental constraints, including: obtaining the historical normalized weights of each candidate migration trajectory in the previous time slice within the current time slice; calculating the observation matching likelihood based on the received bird flock observation data; calculating the unnormalized weights of the current time slice based on the historical normalized weights, the state transition probabilities, and the observation matching likelihood; summing the unnormalized weights of all the candidate migration trajectories to obtain a summation result, and calculating the normalized weights of each candidate migration trajectory in the current time slice based on the summation result. Specifically, the system updates the weights of each candidate migration trajectory sequentially according to time slices. Within the current time slice, the historical normalized weights of each candidate migration trajectory calculated in the previous time slice are retrieved first. This serves as the basis for the calculation of this weight iteration. When new bird flock observation data is received from weather radar, low-altitude monitoring radar, video, or audio-visual terminals... At the same time, the degree of matching between the current predicted state and the measured observation data of each candidate migration trajectory is calculated one by one to obtain the observation matching likelihood. The higher the degree of agreement between the predicted trajectory position, flight altitude, and velocity and the actual observation results, the greater the corresponding observation matching likelihood value. Subsequently, the historical normalized weights of each candidate trajectory are applied. Corresponding state transition probability Multiplying the observation-matching likelihood by the three factors yields the unnormalized weight of the trajectory in the current time slice. After all M candidate trajectories have completed the calculation of their unnormalized weights, the system uses the sum of the unnormalized weights of all trajectories as the normalization denominator. The unnormalized weights of each trajectory Divide by the normalized denominator That is, the normalized weights of each candidate trajectory in the current time slice are obtained. This ensures that the sum of the normalized weights of all candidate trajectories within the same time slice is 1.
[0041] In some embodiments, the relevant calculation formula is as follows: ; In the formula, Environmental constraints The next state vector is derived from the previous state vector. Transition to the current state vector The state transition probability is given by the following formula: where the state vector contains at least horizontal position, flight altitude, and velocity; m is the candidate trajectory index, m=1,2,…,M; k is the time slice index; n is the candidate trajectory index used for normalization summation; and M is the total number of candidate trajectories. For bird flock observation Current state of the trajectory The observation matching likelihood; The unnormalized weights for the current time slice; The historical normalized weights for the current time slice satisfy... .
[0042] It should be noted that the historical normalization weights of the current time slice vary depending on the candidate trajectory generation method. The determination method corresponding to the generation mechanism can be adopted: when extrapolating the direction, velocity and height perturbations, the initial weights are determined by the joint probability density of each perturbation; when using the three-dimensional voxel transition matrix for path search, the initial weights are determined by the product of the transition probabilities of each time step of the path; when there is a lack of reliable priors or when equal weight benchmark calculation is performed, each candidate trajectory takes the same initial weight, which is the reciprocal of the total number of candidate trajectories.
[0043] It should be noted that if the current iteration is in the pure prediction stage where no new observations have been obtained, the observation matching likelihood is fixed at 1, and the weight evolution is driven only by the state transition probability.
[0044] In some embodiments, within each time slice, the system counts candidate migration trajectories falling into each spatial unit and assigns normalized weights to these candidate migration trajectories in the current time slice. The probabilities are accumulated and used as the arrival probability of the bird flock in that spatial unit. Specifically, when using the probability accumulation method, if the normalized weight of the m-th candidate trajectory in the current time slice is... Its three-dimensional position Falling into the spatial unit If the weight is not specified, it will be included in the arrival probability of that unit; the sum of the normalized weights of all candidate trajectories within the same time slice is 1, therefore the sum of the arrival probabilities of all spatial units is also 1. The above calculation relationship is expressed by the following formula: ;in, spatial unit In time slices The probability of a flock of birds arriving, M is the total number of candidate trajectories, and the indicator function is... Used to determine whether the three-dimensional position of the trajectory falls within a spatial unit. The value is set to 1 if the trajectory falls into the space, and 0 otherwise. Through the above accumulation method, the weight of each candidate trajectory is assigned to the corresponding voxel according to its spatial landing point, forming the spatial probability distribution of the bird flock within that time slice.
[0045] It should be noted that if kernel density estimation is used as a replacement, it is not necessary to strictly assign the trajectory to a unique voxel. Instead, the weights are assigned in the form of a kernel function based on the spatial distance between the trajectory position and the voxel center. However, the probability accumulation method is more direct in engineering implementation and the amount of computation is more controllable, so it is preferred.
[0046] Step S203: In each time slice, the spatial units are sorted from largest to smallest according to the arrival probability of the flock of birds to obtain a sorting result. The arrival probability of the flock of birds in the sorting result is accumulated according to the sorting direction to obtain a cumulative arrival probability. When the cumulative arrival probability reaches a first preset threshold, the corresponding spatial unit constitutes a core probability envelope region. In some embodiments, after calculating the arrival probability of all spatial units within each time slice, the system sorts all spatial units in descending order of arrival probability to obtain a descending order of arrival probability within that time slice. Then, spatial units are selected sequentially from front to back along this sorting direction, while simultaneously accumulating the sum of the arrival probabilities of the selected spatial units to obtain the cumulative arrival probability. When the cumulative arrival probability first reaches a first preset threshold, selection stops, and at this point, all the selected spatial units collectively constitute the core probability envelope region of that time slice.
[0047] As a non-limiting example, the first preset threshold can be set to 60%, indicating that the core probability envelope covers the spatial units with the highest probability of bird flocks arriving within that time slice, and its cumulative probability quality reaches 60%. This threshold can also be calibrated based on the historical costs of missed and false alarms. For example, during peak bird damage seasons, the first threshold can be appropriately lowered to expand the coverage of the core envelope and reduce the risk of missed alarms, while during low-risk seasons, the first threshold can be increased to make the core envelope more concentrated and reduce unnecessary attention areas.
[0048] It should be noted that the spatial units within the core probability envelope represent high-probability areas with concentrated candidate trajectories and the highest confidence, serving as the primary source of evidence for subsequent high-confidence warning confirmation.
[0049] Step S204: Continue selecting along the sorting direction until the cumulative arrival probability reaches the second preset threshold. All selected spatial units constitute an extended probability envelope region, wherein the second preset threshold is greater than the first preset threshold.
[0050] In some embodiments, based on the determination of the core probability envelope region, the system continues to select subsequent spatial units sequentially along the same sorting direction and continuously updates the cumulative arrival probability. When the cumulative arrival probability reaches a second preset threshold higher than the first preset threshold, the selection stops. At this time, all the selected spatial units (including all spatial units that previously constituted the core probability envelope region and the newly selected parts) together constitute the extended probability envelope region of the time slice.
[0051] As a non-limiting example, the second preset threshold can be 95%, meaning that the extended probability envelope covers all high arrival probability spatial units within the time slice where the cumulative probability quality reaches 95%.
[0052] It should be noted that since the construction of the extended probability envelope region is based on the selected spatial units of the core probability envelope region and is extended along the sorting direction, the core probability envelope region is always contained within the extended probability envelope region.
[0053] As an optional preferred implementation, after determining the extended probability envelope region, the system can further extract spatial units outside the core probability envelope region and define them as the core peripheral incremental layer. (A certain time slice is recorded.) The set of spatial units of the lower core probability envelope region is The set of spatial units of the extended probability envelope region is Then the spatial unit set of the outer incremental layer can be represented as This peripheral incremental layer represents spatial units whose arrival probability is below the core threshold but still falls within the coverage area of the extended threshold. In subsequent coupled calculations, the system can use this layer alone to calculate the peripheral incremental overlap probability, enabling early identification of situations where peripheral paths may approach the line. In actual operational handling, the results generated by the peripheral incremental layer serve as evidence for early attention and do not directly trigger higher-level warning confirmation actions. Instead, warning escalation and confirmation judgment are triggered only when the core probability envelope overlaps with the line's three-dimensional risk buffer zone, thereby achieving progressive early warning management from peripheral attention to core confirmation.
[0054] By sorting by arrival probability and using a two-layer cumulative threshold strategy to construct a core and an extended probability envelope, high-probability regions and low-probability peripheral regions can be distinguished and expressed. The core envelope ensures high-confidence judgment, while the extended envelope covers peripheral risks caused by path bifurcation and observation errors. This expands the early warning coverage while avoiding excessive expansion that could lead to false alarms, accurately balancing the contradiction between missed and false alarms.
[0055] Step S103: Based on the spatial structure parameters, establish a three-dimensional risk buffer zone for the transmission line; Please refer to Figure 3 In some embodiments, the spatial structure parameters include tower coordinates, span, conductor suspension point position, conductor sag parameters, and line channel width. Step S103 includes steps S301 to S305: Step S301: Determine the two ends of each phase conductor based on the tower coordinates, the span, and the conductor suspension point positions; In some embodiments, when constructing a three-dimensional risk buffer zone for a transmission line, the system uses tower coordinates, span, and line alignment as basic data, and reads the coordinates of two adjacent towers and their conductor suspension point elevations from the line ledger or GIS platform. For each phase conductor of the target transmission line, the system determines the start and end positions of the span based on the coordinates of adjacent towers and the span range, and then, in conjunction with the suspension point design parameters of each phase conductor on the tower, determines the three-dimensional coordinates of the suspension points at both ends of the span for that phase conductor.
[0056] It should be noted that if the line contains multi-phase conductors, the phase-by-phase extraction should be repeated. For tension towers or angle towers, the suspension point coordinates need to be corrected based on the tower head dimensions, crossarm orientation, and suspension point offset; for straight-line towers, the suspension points for each phase are determined directly using the tower center coordinates combined with the crossarm length and suspension point height. The above operations provide accurate endpoint constraints for the subsequent generation of conductor space curves.
[0057] Step S302: Based on the coordinates of the two hanging points and the conductor sag parameters, construct the conductor space curve; In some embodiments, after the coordinates of the two end points are determined, the system constructs a conductor space curve based on the coordinates of the two end points and the conductor sag parameter. For a span between adjacent towers, the parameter u is allowed to vary between 0 and 1. The horizontal coordinate of the conductor curve is obtained by linear interpolation of the two end points, and the vertical coordinate is obtained by subtracting the sag from the height of the line connecting the two end points. Taking a parabolic approximation as an example, the coordinate components of the conductor space curve are calculated according to the following relationships: ;in, Number the line objects; This is a dimensionless parameter along the span direction. , and These correspond to the attachment points at both ends of the gear length; and The coordinates of the two hanging points in a unified three-dimensional coordinate system; For and The three-dimensional coordinates of the j-th phase conductor at parameter u; f is the maximum sag of the line connecting the two end suspension points of the conductor in this span, and the sag term is... exist The maximum value f is obtained at the specified location. The parametric equations of the above three coordinate components together constitute the conductor space curve for this span. When u takes continuous values from 0 to 1, the coordinate points are connected sequentially to form a continuous spatial trajectory, which is the actual spatial shape of the conductor within this span. If the line ledger provides the weight per unit length q and the horizontal tension H but does not directly give the maximum sag ff, then estimate it using the following formula and substitute it into the curve equation: In the formula, f is the maximum sag obtained by approximating the parabola; q is the gravity load per unit length of the conductor; L is the horizontal span between the two suspension points; H is the horizontal component of the conductor tension; when q, L and H use the same unit system, f is the length quantity.
[0058] It should be noted that if the line ledger has both horizontal tension and unit length mass parameters, the catenary model can be used to obtain the conductor space curve. This model has higher fitting accuracy for long spans or large sag scenarios. When the heights of the two suspension points are different, the two suspension points are used as boundary conditions to solve for the catenary curve parameters. For spans with missing parameters, linear interpolation can be performed based on the heights of the conductor suspension points of adjacent towers, or an empirical sag approximation can be used. This is then superimposed with a vertical safety margin determined by the voltage level, span length, terrain undulation, and data source accuracy to form an approximate conductor height band.
[0059] Step S303: Determine the horizontal buffer radius based on the line channel width and the preset horizontal error margin, and determine the vertical buffer radius based on the protection gap and the preset height error margin. In some embodiments, after the conductor space curve is determined, the system further sets buffer ranges in the horizontal and vertical directions. The horizontal buffer radius is determined by the sum of half the width of the line channel and the horizontal error margin, wherein the horizontal error margin is determined based on a combination of the line model plane error and the radar horizontal positioning error; the vertical buffer radius is determined by the sum of the protection gap and the height error margin, wherein the height error margin is determined based on a combination of the sag estimation error, the terrain elevation error, and the radar height measurement error. The specific calculation relationships are as follows: ;in, and These are the horizontal and vertical buffer radii of the line object j, respectively. For the width of the line channel, The reference protection gap is determined based on voltage level and safety requirements. For horizontal error margin, This represents the height error margin. The vertical buffer radius can be set upwards and downwards to correspond to different vertical risk ranges above and below the conductor.
[0060] Step S304: Using the conductor space curve as a reference, superimpose the spatial range corresponding to the horizontal buffer radius and the vertical buffer radius to form a three-dimensional risk buffer zone with corresponding span. In some embodiments, after the horizontal and vertical buffer radii are determined, the system uses the conductor space curve as a reference and expands the two buffer radii outward in the horizontal and vertical directions respectively to form a three-dimensional buffer body for a single-phase conductor. Specifically, for each point on the space curve of a certain phase conductor within the target span, all spatial points that satisfy the condition that the horizontal projection distance to that point does not exceed the horizontal buffer radius and the vertical distance does not exceed the vertical buffer radius together constitute the three-dimensional risk buffer zone for that phase conductor. When determining whether a spatial unit falls into the buffer body, firstly, the horizontal center of the spatial unit is projected onto the conductor space curve, and the curve height at the projection position is the conductor risk height corresponding to that unit; then, the horizontal projection distance is compared with the horizontal buffer radius, and the height difference is compared with the vertical buffer radius. Intersection is considered only when both conditions are met simultaneously. The buffer bodies of each phase conductor within the same span are constructed and then the union is taken to form the three-dimensional risk buffer zone for that span.
[0061] Step S305: Construct a tower risk body centered on the tower coordinates, and combine the three-dimensional risk buffer zones of adjacent spans with the tower risk body to obtain a complete three-dimensional risk buffer zone for the transmission line.
[0062] In some embodiments, for the tower area, the system constructs a tower risk body centered on the tower coordinates. Its spatial extent includes the outer envelope surrounding the tower body, crossarms, insulator strings, jumpers, and conductor suspension points. A safety margin is then added to this outer envelope to cover the electrical clearance requirements and structural space occupancy of the tower area. After constructing the three-dimensional risk buffer zone for each span and the tower risk body for each tower, the system aggregates all span buffer zones and all tower risk bodies in sequence according to the line route, forming a complete three-dimensional risk buffer set covering the entire transmission line.
[0063] It should be noted that the risk body size can be set differently for different types of towers: tension towers and angle towers, due to their complex tower head structure and numerous jumpers, can have a larger horizontal threshold set; crossing towers that cross railways, highways, rivers or important passages can also have their tower risk body range expanded; while ordinary straight towers with low historical risk sections can use the standard threshold.
[0064] It should be noted that the three-dimensional risk buffer set is indexed by tower, span, or continuous line segment. In addition to spatial boundary information, each buffer object also stores the voltage level, line importance, equipment sensitivity level, and corresponding maintenance personnel information of the line object to which it belongs. For crossing areas, tension towers, angle towers, important transmission channels, or historically high-incidence bird-related sections, their differentiated horizontal distance threshold and vertical height threshold settings are retained during aggregation, so that critical sections can obtain more stringent risk assessment conditions in subsequent overlap probability calculations.
[0065] By using the guideline space curve as a reference and superimposing horizontal and vertical buffer radii to construct a three-dimensional buffer zone, the risk assessment of bird flocks is elevated from a two-dimensional plane to a three-dimensional space. At the same time, the error margin corresponding to the accuracy of the data source is introduced, so that the buffer zone boundary not only fits the actual spatial shape of the guideline but also has reasonable engineering redundancy. This effectively avoids the two-dimensional judgment from misjudging bird flocks flying at altitudes far above the guideline as risk, and significantly reduces the number of invalid warnings.
[0066] Step S104: Within each time slice, the overlapping probability sequence and the probability centroid sequence are calculated based on the flight probability envelope and the three-dimensional risk buffer. Please refer to Figure 4 In some embodiments, step S104 includes steps S401 to S405: Step S401: For each time slice, select all spatial units within the flight probability envelope region, determine whether each spatial unit intersects with the three-dimensional risk buffer zone of the transmission line, and identify several intersecting spatial units. In some embodiments, for a certain line object j in a time slice To calculate the overlap probability, the system first selects the set of spatial units to participate in the calculation based on the calculation objective. This set can be selected as the core probability envelope region as needed. Extended probability envelope region or peripheral incremental layer Once selected, the system sequentially determines the three-dimensional risk buffer of each spatial unit g within the set and the line object j. The existence of spatial intersection is determined by checking whether the voxel range represented by the spatial unit overlaps with the spatial envelope of the buffer. If they intersect, the spatial unit is marked as an intersecting spatial unit; otherwise, it is not included in the subsequent accumulation of the overlap probability for that line object. The core probability envelope and the extended probability envelope participate in the above judgment process, respectively, and output their respective sets of intersecting spatial units.
[0067] It should be noted that when it is necessary to characterize the contribution of the peripheral path separately, the peripheral increment is used. Perform intersection checks and probability accumulation.
[0068] Step S402: Obtain the flock flight height corresponding to each of the intersecting spatial units and the guide risk height corresponding to the position of the three-dimensional risk buffer, calculate the height difference between the flock flight height and the guide risk height, and obtain the height matching weight; In some embodiments, for each spatial cell g that has been determined to intersect with the three-dimensional risk buffer of the line object j, the system further obtains the representative altitude of bird flock flight for that spatial cell. (Preferably, the probability-weighted average of the voxel center height or the bird flock height within the unit is taken), and the traverse risk height corresponding to the horizontal projection position of the spatial unit is also obtained. The traverse risk height is determined as follows: the horizontal center of the spatial unit is projected onto the traverse space curve; the curve height at the projection location is the traverse risk height corresponding to that spatial unit. Based on this, the system calculates the signed height difference between the two. The system then uses the square of the absolute value of the height difference, or the absolute value itself, as a measure of the degree of vertical deviation. Subsequently, the system calculates the height matching weight based on this height difference. The calculation formula is: ,in, The value is determined by the error of the protection gap or radar height measurement, and is usually in the range of 10 to 25 meters, representing the attenuation scale of the height matching weight; Let be the difference between the flock height at spatial unit g and the conductor risk height at line object j. If the conductor risk height is given as an interval value, then the minimum distance from the flock height to that interval is taken as the height difference. When there are multiple phase conductors in the same span, the system can calculate the height difference between the spatial unit and each phase conductor, and select the conductor with the smallest height difference as the height matching object; alternatively, it can calculate the height matching weight of each phase conductor, and take the maximum value as the final height matching weight of the spatial unit.
[0069] It should be noted that the high matching weight The weight is used to quantify the degree of matching between the flight altitude of the flock of birds and the risk altitude of the guide line: the smaller the altitude difference, the closer the weight is to 1, indicating that the altitude of the flock of birds in this spatial unit matches the height of the guide line; the larger the altitude difference, the weight gradually decays and approaches 0, indicating that even if the flock of birds intersects with the line buffer zone in a horizontal position, its flight altitude is far from the risk altitude of the guide line, and the actual risk is low.
[0070] For a given spatial unit, its horizontal center is projected onto the traverse curve corresponding to the span. The height of the curve at the projected location is the traverse risk height for that spatial unit. This represents the height of the spatial unit. The difference between the probability-weighted average of the voxel center height or the height of bird flocks within a unit and the guide risk height. Used to measure the height of the guide line risk between the flight altitude of birds and the guide line.
[0071] Step S403: Multiply the arrival probability of the bird flock in each of the intersecting spatial units by the corresponding height matching weight and then sum them up to obtain the overlap probability under this time slice. In some embodiments, after completing the screening of intersecting spatial cells and the calculation of height matching weights, the system multiplies the bird flock arrival probability of each intersecting spatial cell by its corresponding height matching weight, and sums all the product results to obtain the overlap probability of the line object in that time slice. The calculation formula is as follows: ,in, For line object j in time slice The overlap probability, where G is the set of spatial units participating in this calculation. Let g be the arrival probability of spatial cell g in this time slice. This is a spatial intersection indicator function; it takes the value 1 when a spatial cell intersects with the buffer, and 0 otherwise. For highly matched weights.
[0072] It should be noted that, in the specific calculation, the core probability envelope and the extended probability envelope participate in the above accumulation process respectively to obtain the core overlap probability and the extended overlap probability of the time slice. Since the core probability envelope is contained within the extended probability envelope, when using the same intersection conditions and high-level matching rules, the value of the extended overlap probability is not less than the core overlap probability. The system saves the calculation results of these two sets of overlap probabilities separately: the core overlap probability is used for high-confidence confirmation and triggering of higher-level warnings, and the extended overlap probability is used to detect the risks that may be caused by low-probability paths in the periphery in advance. The above two sets of overlap probabilities form independent sequences and do not overlap with each other.
[0073] It should be noted that the core overlap probability and the extended overlap probability are calculated using the following formulas: ; In the formula, and These are the time slices of line object j. The core overlap probability and extended overlap probability; g is a spatial unit. Its probability of arrival; A three-dimensional risk buffer for object j; This is a spatial intersection indicator function, which takes the value 1 when there is an intersection and 0 otherwise. The height matching weight between the flock height of the spatial unit and the risk height of the conductor of object j.
[0074] It should be noted that, since the core probability envelope includes the extended probability envelope, the core overlap probability and the extended overlap probability are stored as high-confidence evidence and high-recall evidence, respectively, and are not directly added together. If the contribution of the outer path needs to be calculated separately, the outer incremental layer is used for calculation. ; In the formula, For line object j in time slice The probability of overlapping peripheral increments; It is a collection of spatial units in the outer incremental layer.
[0075] Step S404: Arrange the overlapping probabilities according to the order of the time slices to form an overlapping probability sequence; In some embodiments, for each route object, after calculating the overlap probability in each time slice, the system arranges the obtained overlap probabilities into a sequence according to the chronological order of each time slice, forming an overlap probability sequence for that route object as it changes with the prediction time window. The core probability envelope and the extended probability envelope form independent sequences, namely the core overlap probability sequence and the extended overlap probability sequence. The extended overlap probability sequence is used to identify peripheral paths that may be approaching the route earlier and determine the early attention time; the core overlap probability sequence is used to confirm high-confidence overlaps, calculate continuous exposure, and trigger higher-level warnings.
[0076] Step S405: Based on the spatial position of each of the intersecting spatial units, a weighted average is calculated according to the matching weight of the bird flock arrival probability and the height to obtain the probability centroid of the time slice, and the probability centroids are arranged in chronological order to form a probability centroid sequence.
[0077] In some embodiments, within each effective overlapping time slice, the system further calculates the probability centroid of that time slice, i.e., the weighted center position of the bird flock overlap probability in three-dimensional space, based on the spatial positions of all spatial units intersecting with the three-dimensional risk buffer of the line object, according to the bird flock arrival probability and height matching weight of each spatial unit. The formula for calculating the probability centroid is as follows: Where r∈{c,e} are the probability envelope level identifiers, representing the core layer and the extension layer, respectively; For line object j in The probability centroid of overlapping spatial units at time and level r. Let g be the three-dimensional position vector of the center point of spatial unit g. For the corresponding level of spatial unit set, Let be the arrival probability of a spatial unit. For high-level matching weights, centroid calculation is performed only within consecutive effective overlapping time periods. The centroid positions of adjacent effective time slices are differentiated and normalized to obtain the local movement direction of the bird flock trajectory. The probability centroids calculated for each time slice are arranged in chronological order to form a probability centroid sequence corresponding to the route object, which serves as the basis for subsequent judgment of the bird flock's positional changes relative to the route movement pattern (flying along the route or lateral crossing).
[0078] By introducing height matching weights to weight the arrival probabilities of intersecting spatial units, only flocks of birds flying at altitudes close to the risk height of the guide line can contribute a higher overlap probability. Even if flocks of birds flying at altitudes far above or far below the guide line overlap with the line in horizontal position, their overlap probability will be significantly reduced. This achieves precise filtering in the height dimension at the probability source, significantly improves the spatial accuracy of early warnings, and reduces false alarms caused by height mismatch.
[0079] Step S105: Determine the duration of continuous overlap of the target bird flock with the transmission line based on the overlap probability sequence, and determine the movement pattern of the target bird flock relative to the transmission line based on the probability centroid sequence. Further, the spatial structure parameters include the route alignment. Step S105 includes steps S501 to S503: Step S501: Traverse the overlapping probabilities in the overlapping probability sequence to identify a set of continuous time periods where the overlapping probability is not lower than a preset effective threshold. Select the time period with the longest duration from the set of continuous time periods as the target influence time period and use the duration of the target influence time period as the continuous overlapping duration. In some embodiments, after the core overlap probability sequence and extended overlap probability sequence of each line object are formed, the system independently performs the extraction operation of continuous overlap duration for each sequence. For an overlap probability sequence of a certain level, the system traverses each overlap probability value in the sequence from beginning to end, and determines whether it is not lower than the preset effective threshold corresponding to that level. When the overlap probability of multiple consecutive time slices is not lower than the effective threshold, these consecutive time slices constitute a continuous effective time period. The system summarizes all the continuous effective time periods identified during the traversal into a continuous time period set, and selects the time period with the longest duration from the set as the target influence time period, and determines the duration of this time period as the continuous overlap duration of the line object at that level. If there are multiple continuous effective overlaps in the overlap probability sequence, only the segment with the most time slices is selected as the longest continuous time period. Under the condition of equally spaced time slices, the continuous overlap duration is calculated by the following formula: In the formula, r is the probability envelope level identifier, r∈{c,e}, representing the core layer and the extension layer, respectively; The longest consecutive overlap duration of line object j at level r; The longest consecutive time slice with a corresponding overlap probability that is not lower than the effective threshold. The time step is the time step of adjacent equally spaced time slices; the superscript max indicates that the time slice with the most times is selected from multiple consecutive effective overlaps, and the minus 1 in the formula is used to convert N sampling times into N-1 time intervals between the first and last times.
[0080] Step S502: Within each time slice covered by the target influence period, perform a difference operation and normalize the position of the probability centroid of adjacent time slices to obtain the local movement direction of the bird flock trajectory. In some embodiments, within each time slice covered by the target impact period, the system uses spatial units that simultaneously belong to the corresponding probability envelope and intersect with the line's three-dimensional risk buffer zone to calculate the probability centroid position of overlapping spatial units within each effective time slice according to the probability centroid calculation formula in step S405. After obtaining the probability centroids of each effective time slice within the continuous period, the system performs a difference operation on the probability centroid positions of adjacent time slices to obtain the position change vector between them. This vector is then normalized by dividing it by its own Euclidean norm, thereby obtaining the unit vector of the local movement direction of the bird flock trajectory within the adjacent time interval. The above calculation relationship can be expressed by the following formula: ;in, The local migration direction unit vector is obtained by normalizing the difference between adjacent centroids. and These are slices of line object j in two adjacent valid time intervals. and The probability centroid; the ∥⋅∥ in the denominator represents the Euclidean norm.
[0081] If the probability centroids of two adjacent time slices coincide, the difference vector is zero. In this case, no division operation is performed; the most recent valid direction can be used, or the direction of the current time slice can be marked as undefined. Through the above time-slice differential normalization calculation, the system can obtain the sequence of local bird flock movement directions for each adjacent time interval within the target's influence period, which can be used for subsequent angle comparison with the route direction.
[0082] Step S503: Calculate the angle between the local movement direction and the line direction, and determine the movement pattern of the target bird flock relative to the power transmission line based on the numerical range of the angle.
[0083] In some embodiments, the unit vectors of the local movement directions of the bird flock trajectory at each adjacent time interval are obtained. Subsequently, the system simultaneously obtains the line direction unit vector of the line object j. (Determined by tower coordinates and span direction). For each effective time interval, the system calculates the angle between the local movement direction and the line direction, specifically using the absolute value of the vector dot product combined with the inverse cosine function, and constrains the result to the range of zero to a right angle to eliminate the difference between the positive and negative directions along the line: ;in, The local tangential unit vector of the bird flock's trajectory; Let j be the unit vector representing the direction of the line object j; The angle between the two is denoted as π / 2, and its value ranges from 0 to π / 2.
[0084] It should be noted that the system classifies the movement patterns of the target bird flock relative to the transmission line based on the numerical range of the included angle: when the included angle is close to 0, it means that the bird flock's trajectory is basically parallel to the direction of the line, and the bird flock flies or moves parallel to the line for a long time, which is a continuous exposure mode; when the included angle is close to π / 2, it means that the bird flock's trajectory is basically perpendicular to the direction of the line, and the bird flock crosses the line laterally in a short time, which is an instantaneous crossing mode.
[0085] By extracting the longest continuous period not lower than the effective threshold from the overlapping probability sequence as the continuous overlapping duration, and combining it with the probability centroid difference to obtain the angle between the local movement direction and the route direction, it is possible to distinguish whether the flock of birds is flying along the route direction for a long time or crossing it quickly laterally. This avoids confusing the two different risk mechanisms for evaluation, making the subsequent early warning judgment more in line with the actual risk causes, and further improving the accuracy of the early warning.
[0086] Step S106: Perform graded early warning based on the motion pattern and the duration of continuous overlap.
[0087] In some embodiments, step S106 includes: if it is the continuous exposure mode, then applying a continuous exposure weight to the continuous overlap duration to correct it, obtaining a corrected duration; if it is the instantaneous crossing mode, then applying an instantaneous overlap weight to the maximum overlap probability in the overlap probability sequence to correct it, obtaining a corrected probability; combining the corrected duration, the corrected probability, the bird flock size, and the route importance parameters to comprehensively determine and output a warning result of the corresponding level. Specifically, when the included angle... When the value approaches 0, it is determined to be a continuous exposure mode. At this time, the system applies a continuous exposure weight to correct the continuous overlap duration. according to Calculation, where A directional correction coefficient of not less than 0 is used to improve the contribution of overlap duration to the early warning score in long-duration escort scenarios along the route; when the included angle When the value approaches π / 2, it is determined to be an instantaneous crossing mode. At this time, the system applies an instantaneous overlap weight to the maximum overlap probability in the overlap probability sequence for correction. according to Calculation, where A directional correction coefficient, not less than 0, is used to improve the contribution of peak overlap probability to the warning score in lateral crossing scenarios. After directional correction, the system combines the corrected overlap duration and maximum overlap probability with the flock size N and the degree of height matching. Importance of the route and equipment sensitivity level A comprehensive early warning scoring formula is used to determine the multi-factor fusion: ; in, The comprehensive early warning score for line object j is truncated to [0,1]. To predict the maximum probability of overlap between time slices within a time window, The duration of continuous overlap; It is a monotonically normalized function used to map the probability, duration, flock size and altitude matching degree to the interval between 0 and 1, respectively. All scores are non-negative and can be assigned based on historical samples or operational preferences; the importance of the line. Related to equipment sensitivity level Values have been normalized to 0 to 1 and directly participate in the weighting; the outer layer min{1,⋅} ensures that the overall score does not exceed 1.
[0088] It should be noted that the above weight settings ensure that the continuous exposure weight corresponding to the accompanying flight along the line mainly applies to the overlap duration term, while the instantaneous overlap weight corresponding to the lateral crossing mainly applies to the maximum overlap probability term. The two weights apply to two different basic quantities and are not confused with each other.
[0089] As an alternative implementation, the system can also use a combination of rules and scoring for graded judgment: a yellow warning is generated when the overlap probability reaches the first threshold and the bird flock size reaches the scale threshold; an orange warning is generated when the overlap probability reaches the second threshold, the overlap duration reaches the duration threshold, or important channels are affected; and a red warning is generated when the overlap probability reaches the third threshold and important lines or sensitive equipment are affected. The thresholds for each level are configured by the operation and maintenance unit based on regional experience and voltage levels. Regardless of whether a scoring formula or a rule-based approach is used, the core and extended probability envelope areas are not counted repeatedly: when the extended overlap probability reaches the extended effective threshold but the core overlap probability has not yet reached the core effective threshold, the system generates early warning information for peripheral paths; when the core overlap probability reaches the core effective threshold, the warning is confirmed and its level is upgraded based on the maximum core overlap probability, the core continuous overlap duration, and the corresponding directional weight. The final warning result includes the trigger envelope level, overlap probability category and value, expected impact time window, affected line objects and tower span numbers, and suggested actions, and is pushed to the GIS platform and the operation and maintenance personnel, and is continuously upgraded, downgraded, or lifted with the next cycle of observation data.
[0090] By applying differentiated directional weights to the overlap duration and maximum overlap probability for the continuous exposure mode and the instantaneous crossing mode respectively, the early warning basis for the two risk modes is matched with their actual disaster-causing mechanisms. This avoids underestimating or overestimating the risks caused by using a uniform evaluation standard for all overlap situations, and achieves accurate graded early warning.
[0091] It should be noted that the system adopts a rolling update mechanism. After receiving new bird flock status in each new monitoring update cycle, it re-executes the entire process of candidate trajectory generation, probability envelope construction, and spatiotemporal overlap calculation, and dynamically corrects issued warnings: when new observations show that the bird flock has deviated from the route or the overlap probability remains below the effective threshold, the system automatically downgrades or cancels the warning; when new observations show that the bird flock has entered a higher probability overlap area, the system upgrades the warning level and expands the push range. The warning results are sent to the GIS platform and the maintenance personnel. The GIS platform displays the probability migration corridor, the three-dimensional buffer projection of the route, the affected towers and spans, the warning level, the estimated arrival time, and the duration within the future time window. The push information includes the route name, tower number or span number, warning level, estimated impact time, main risk causes, and suggested actions, facilitating maintenance personnel to directly perform on-site inspections or handling. In terms of differentiated early warning strategies, the system can set larger horizontal distance thresholds or vertical height thresholds for tension towers, angle towers, crossing areas across railways / highways / rivers, important power transmission channels, important load power supply lines, or areas with a high incidence of historical bird damage. This allows key sections to obtain stricter risk judgment conditions in the overlap probability calculation, while standard thresholds are used for areas with ordinary straight spans and low historical risks. This ensures the safety of key sections while avoiding excessive early warnings.
[0092] This invention provides a data foundation reflecting the true movement patterns and environmental response characteristics of bird flocks by fusing real-time migration parameters and environmental constraints. This avoids prediction bias caused by relying solely on historical statistics or simple extrapolation, ensuring data quality for accurate early warnings from the source. By generating multiple candidate migration trajectories and mapping them to spatial units to form a flight probability envelope, a probability distribution replaces a single deterministic path. This fully characterizes the bifurcation and diffusion characteristics of bird migration paths influenced by multiple factors such as wind field, terrain, and habitat, effectively solving the technical problem of missed reports under path bifurcation conditions when extrapolating a single path. It also avoids blindly expanding the early warning range to compensate for missed reports. By establishing a three-dimensional risk buffer zone for transmission lines that includes a height dimension, the early warning judgment is elevated from a two-dimensional plane to a three-dimensional space. This accurately reflects the true risk boundaries of conductor sag, protection gaps, and tower structures, effectively excluding irrelevant bird flocks flying at altitudes far above the conductors. This significantly reduces false alarms caused by traditional two-dimensional horizontal buffering methods from a height dimension. By calculating the flight probability envelope within each time slice... The overlapping probability sequence and probability centroid sequence of the network area and the three-dimensional risk buffer zone incorporate trajectory uncertainty, spatial three-dimensional intersection relationship, degree of high matching and arrival probability into a unified temporal quantification framework. This enables a precise characterization of the spatiotemporal coupling relationship between bird flocks and line risks, avoiding random errors caused by qualitative judgment or single-point threshold comparison. By extracting the continuous overlap duration from the overlapping probability sequence, the continuous exposure time of bird flocks in the line risk space is incorporated into the evaluation system as an independent risk dimension. This solves the problem that relying solely on instantaneous probability cannot distinguish between brief brushes and long-term continuous threats, making the early warning triggering conditions more consistent with the actual risk evolution law. By determining the movement pattern based on the probability centroid sequence and combining it with the continuous overlap duration for graded early warning, it is possible to distinguish between two risk mechanisms: long-term accompanying flight along the line direction and rapid crossing in the vertical direction. This avoids the distortion of early warning levels caused by different risk modes using the same evaluation scale, making the final output early warning level more consistent with the actual risk level. As a result, the accuracy of short-term bird damage early warning is significantly improved overall, and the interference of missed and false alarms on operation and maintenance decisions is reduced.
[0093] Based on the above method embodiments, corresponding apparatus embodiments are provided; An embodiment of the present invention provides a transmission line risk early warning system based on bird migration, comprising: the system being used to perform the steps of the transmission line risk early warning method based on bird migration as described in the present invention.
[0094] It is understood that the above-described device embodiments correspond to the method embodiments of the present invention, and can implement the transmission line risk warning method based on bird migration provided by any of the above-described method embodiments of the present invention.
[0095] It should be noted that the device embodiments described above are merely illustrative, and some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can specifically be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.
[0096] Based on the above embodiments of the transmission line risk warning method based on bird migration, another embodiment of the present invention provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the transmission line risk warning method based on bird migration of any embodiment of the present invention.
[0097] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the terminal device.
[0098] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.
[0099] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.
[0100] Based on the above-described method embodiments, another embodiment of the present invention provides a computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to execute the transmission line risk early warning method based on bird migration as described in any of the above-described method embodiments of the present invention.
[0101] The modules / units integrated in the device / terminal equipment, if implemented as software functional units and sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.
[0102] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.
Claims
1. A method for early warning of transmission line risks based on bird migration, characterized in that, include: To obtain real-time migration parameters of the target flock of birds, environmental constraints during flight, and spatial structural parameters of the power transmission line; Based on the real-time migration parameters and environmental constraints, several candidate migration trajectories of the target bird flock within the future time window are predicted, and the candidate migration trajectories within each time slice are mapped to spatial units to calculate the flight probability envelope. Based on the aforementioned spatial structure parameters, a three-dimensional risk buffer zone for the transmission line is established. Within each time slice, the overlapping probability sequence and the probability centroid sequence are calculated based on the flight probability envelope and the three-dimensional risk buffer. The duration of continuous overlap between the target bird flock and the transmission line is determined based on the overlap probability sequence, and the movement pattern of the target bird flock relative to the transmission line is determined based on the probability centroid sequence. A graded early warning system is implemented based on the motion pattern and the duration of continuous overlap.
2. The method for early warning of transmission line risks based on bird migration according to claim 1, characterized in that, The prediction of several candidate migration trajectories for the target bird flock within a future time window based on the real-time migration parameters and environmental constraints includes: Based on the environmental constraints, determine the state transition probability of the current state relative to the previous state; By constraining the deflection direction and spatial reachability of the trajectory through the state transition probability, several candidate migration trajectories within a future time window are generated.
3. The method for early warning of transmission line risks based on bird migration according to claim 2, characterized in that, The flight probability envelope region includes a core probability envelope region and an extended probability envelope region. Mapping the candidate migration trajectories within each time slice to spatial cells to calculate the flight probability envelope region includes: For each time slice, the three-dimensional position of each candidate migration trajectory is mapped to a spatial unit containing the height dimension; The probability of a flock arriving in the current time slice of a spatial unit is obtained by summing the normalized weights of each candidate migration trajectory falling into the same spatial unit in the current time slice, wherein the normalized weights are calculated based on the real-time migration parameters and the environmental constraints. Within each time slice, the spatial units are sorted from largest to smallest according to the arrival probability of the flock of birds, and the sorting results are obtained. The arrival probabilities of the flock of birds in the sorting results are accumulated according to the sorting direction to obtain the cumulative arrival probability. When the cumulative arrival probability reaches a first preset threshold, the corresponding spatial unit constitutes the core probability envelope region. Continue selecting along the sorting direction until the cumulative arrival probability reaches the second preset threshold. All selected spatial units constitute an extended probability envelope region, wherein the second preset threshold is greater than the first preset threshold.
4. The method for early warning of transmission line risks based on bird migration according to claim 3, characterized in that, The normalized weights are calculated based on the real-time migration parameters and the environmental constraints, including: Within the current time slice, obtain the historical normalized weights of each candidate migration trajectory from the previous time slice; Calculate the observation matching likelihood based on the received bird flock observation data; Based on the historical normalized weights, the state transition probabilities, and the observation matching likelihood, the unnormalized weights of the current time slice are calculated. The unnormalized weights of all the candidate migration trajectories are summed to obtain the summation result, and the normalized weight of each candidate migration trajectory in the current time slice is calculated based on the summation result.
5. The method for early warning of transmission line risks based on bird migration according to claim 1, characterized in that, The spatial structure parameters include tower coordinates, span, conductor suspension point positions, conductor sag parameters, and line corridor width. The establishment of a three-dimensional risk buffer zone for the transmission line based on these spatial structure parameters includes: Based on the tower coordinates, the span, and the conductor suspension point positions, determine the suspension points at both ends of each phase conductor; Based on the coordinates of the two hanging points and the sag parameters of the conductor, a spatial curve of the conductor is constructed; The horizontal buffer radius is determined based on the line channel width and the preset horizontal error margin, and the vertical buffer radius is determined based on the protection gap and the preset height error margin. Based on the conductor space curve, the spatial ranges corresponding to the horizontal buffer radius and the vertical buffer radius are superimposed to form a three-dimensional risk buffer zone with corresponding span. A tower risk body is constructed with the tower coordinates as the center. The three-dimensional risk buffer zone of adjacent spans is then combined with the tower risk body to obtain the complete three-dimensional risk buffer zone of the transmission line.
6. The method for early warning of transmission line risks based on bird migration according to claim 3, characterized in that, Within each time slice, based on the flight probability envelope and the three-dimensional risk buffer, the overlapping probability sequence and the probability centroid sequence are calculated, including: For each time slice, all spatial units within the flight probability envelope are selected, and it is determined whether each spatial unit intersects with the three-dimensional risk buffer zone of the transmission line, thus identifying several intersecting spatial units. Obtain the flock flight height corresponding to each of the intersecting spatial units and the guide risk height corresponding to the position of the three-dimensional risk buffer, calculate the height difference between the flock flight height and the guide risk height, and obtain the height matching weight; The probability of the flock of birds arriving in each of the intersecting spatial units is multiplied by the corresponding height matching weight and then summed to obtain the overlap probability under that time slice. The overlapping probabilities are arranged in chronological order of the time slices to form an overlapping probability sequence; Based on the spatial position of each intersecting spatial unit, a weighted average is calculated according to the matching weight of the bird flock arrival probability and the altitude to obtain the probability centroid of the time slice, and the probability centroids are arranged in chronological order to form a probability centroid sequence.
7. The method for early warning of transmission line risks based on bird migration according to claim 6, characterized in that, The spatial structure parameters include the line alignment. Determining the duration of continuous overlap of the target bird flock with respect to the transmission line based on the overlap probability sequence, and determining the movement pattern of the target bird flock relative to the transmission line based on the probability centroid sequence, includes: Traverse the overlapping probabilities in the overlapping probability sequence to identify a set of continuous time periods where the overlapping probability is not lower than a preset effective threshold. Select the time period with the longest duration from the set of continuous time periods as the target influence time period and use the duration of the target influence time period as the continuous overlapping duration. Within each time slice covered by the target's influence period, the positions of the probability centroids of adjacent time slices are differentially calculated and normalized to obtain the local movement direction of the bird flock's trajectory. Calculate the angle between the local movement direction and the line direction, and determine the movement pattern of the target bird flock relative to the transmission line based on the numerical range of the angle.
8. The method for early warning of transmission line risks based on bird migration according to claim 1, characterized in that, The motion modes include a continuous exposure mode and a transient crossing mode. The tiered early warning system based on the motion mode and the duration of continuous overlap includes: If it is the continuous exposure mode, then the continuous exposure weight is applied to the continuous overlap duration to correct it, and the corrected duration is obtained; If it is the instantaneous crossing mode, then the maximum overlap probability in the overlap probability sequence is corrected by applying an instantaneous overlap weight to obtain the corrected probability; The warning result of the corresponding level is output by comprehensively judging the correction duration, correction probability, bird flock size and route importance parameters.
9. A power transmission line risk early warning system based on bird migration, characterized in that, include: The system is used to execute the transmission line risk early warning method based on bird migration as described in any one of claims 1-8.
10. A terminal device, characterized in that, include: One or more processors; A memory, coupled to the processor, for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the steps of the transmission line risk warning method based on bird migration as described in any one of claims 1-8.
