Power transmission line bird activity prediction method and system fusing ecological data

Abstract

The invention relates to a power transmission line bird activity prediction method and system fused with ecological data, and belongs to the technical field of power grid safety protection, and the method comprises the steps: collecting ecological environment data and bird monitoring data of a power transmission line region, carrying out the preprocessing, feature extraction and space-time alignment of multi-source data, and forming a standardized data set; based on the data set, through a bird activity prediction model integrating a species ecological position sub-model, a multi-modal time sequence prediction sub-model and a migration prediction sub-model, prediction is carried out, and the probability, scale and species type of bird activity in a future time period are obtained; determining an early warning grade according to the prediction result, and generating a differential prevention and control strategy corresponding to the grade for execution; actual data after prevention and control are collected and compared with a prediction result, the model is optimized and updated, and a closed loop is formed. According to the method, the problems of prediction lag and strategy passivity in the prior art are solved, conversion from passive expelling to active early warning is realized, the accuracy and efficiency of bird damage prevention and control are improved, and ecological protection is considered.

Description

Technical Field

[0001] This application relates to the field of smart grid security protection technology, and more specifically, to a method and system for predicting bird activity on power transmission lines by integrating ecological data. Background Technology

[0002] Bird-related faults on power transmission lines are a significant threat to the safe and stable operation of the power grid. Traditional bird control methods mainly rely on manual patrols and fixed bird deterrent devices. These methods are characterized by significant lag and passivity, failing to provide effective intervention before birds gather and cause faults.

[0003] With technological advancements, existing solutions have begun to employ intelligent bird control systems based on video surveillance or sound recognition. However, most of these technologies remain limited to real-time identification and removal of birds already present, essentially operating on a "post-event response" basis. Their core flaw lies in the lack of forward-looking prediction capabilities regarding bird activity trends, resulting in a consistently passive bird control strategy. This not only wastes resources but also struggles to cope with sudden, large-scale bird gatherings. Existing technologies, such as the proposed method and platform for targeted bird control under scene adaptation (patent publication number CN119294869A), collect data by deploying a monitoring network in the target scene and utilizes predictive models for trend analysis, subsequently matching a strategy library for management and optimization. However, this approach still has certain shortcomings, such as the lack of integration of ecological data (e.g., meteorological, migration cycles), making it difficult to explain the environmental triggers for bird gatherings and resulting in insufficient strategy targeting; the lack of historical data accumulation and analysis, hindering the formation of long-term predictive models; and the absence of real-time image detection, making it impossible to predict bird activity trends. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention proposes a method and system for predicting bird activity on power transmission lines by integrating ecological data.

[0005] The technical solution of this invention is as follows: This invention proposes a method for predicting bird activity along power transmission lines by integrating ecological data, comprising the following steps: Ecological environment data and bird monitoring data of the power transmission line area were collected, and the collected data were preprocessed, feature extracted and spatiotemporally aligned to form a standardized dataset; Based on a standardized dataset, a bird activity prediction model is used to predict bird activity in the power transmission line area. The bird activity prediction results include the probability, scale, and species type of bird activity. The warning level is determined based on the bird activity forecast results, and corresponding prevention and control strategies are generated and implemented based on the warning level. The actual bird activity data after the implementation of the prevention and control strategy is collected, compared with the prediction results to calculate the error, and the bird activity prediction model is optimized and updated based on the error.

[0006] Preferably, the ecological environment data includes meteorological data, vegetation and food resource data, and bird migration data; the bird monitoring data includes multimodal data collected through images, voiceprints, lidar, and historical activity log data.

[0007] Preferably, the spatiotemporal alignment includes: Time alignment: Using a unified timestamp as a benchmark, bird monitoring data and ecological environment data with different sampling frequencies are unified to the same time granularity through interpolation methods to form time series data; Spatial alignment: The area of ​​the power transmission line is divided into grids, and the location of bird activity is spatially correlated with the ecological data within the grids.

[0008] Preferably, the bird activity prediction model includes: a species ecological niche sub-model, a multimodal time series prediction sub-model, and a migration prediction sub-model; A species niche model was developed, using the MaxEnt model to calculate the potential distribution probability of birds in the power transmission line area based on ecological factors. A multimodal time-series prediction sub-model integrates long short-term memory networks and attention mechanisms to predict the probability and scale of bird activity based on time-series data; The migration prediction sub-model uses a bidirectional long short-term memory network to predict the time window for migratory birds to arrive at the power transmission line area and their flock movement trajectory.

[0009] Preferably, determining the warning level based on bird activity prediction results specifically involves: When the predicted probability of activity is below the first threshold or the predicted bird species is a protected species, it is classified as a low-risk level. When the predicted probability of activity is between the first and second thresholds, or when there is a risk of protected species gathering, it is determined to be at a medium risk level. When the predicted probability of activity is higher than the second threshold, or when there is a risk of high-risk birds or common birds gathering, it is determined to be a high-risk level.

[0010] Preferably, the generation of differentiated prevention and control strategies based on early warning levels specifically includes: For low-risk levels, a control strategy of monitoring without expelling animals will be implemented, along with ecological observation and gentle guidance for protected species. For medium-risk levels, the first level of bird deterrence devices, including sound and light bird deterrence devices, should be activated in advance. For high-risk levels, the second level of bird deterrence devices should be activated in advance, including sound and light bird deterrence devices, laser bird deterrence devices, and ultrasonic bird deterrence devices.

[0011] Preferably, the method further includes a dynamic early warning threshold adaptive mechanism, which dynamically adjusts the trigger threshold of the early warning level based on historical activity log data and real-time ecological environment data. The historical activity log includes bird breeding seasons and migration seasons.

[0012] On the other hand, the present invention also provides a transmission line bird activity prediction system that integrates ecological data, comprising: The data acquisition module collects ecological environment data and bird monitoring data in the power transmission line area, and performs preprocessing, feature extraction and spatiotemporal alignment on the collected data to form a standardized dataset. The prediction module uses a bird activity prediction model based on a standardized dataset to obtain bird activity prediction results for the transmission line area. The prevention and control module determines the warning level based on the bird activity prediction results, and generates and executes the corresponding prevention and control strategy based on the warning level. The feedback optimization module collects actual bird activity data after the implementation of prevention and control strategies, compares it with the prediction results to calculate the error, and optimizes and updates the bird activity prediction model based on the error.

[0013] In another aspect, the present invention also provides an electronic device having a computer program stored thereon, which, when executed by a processor, implements a method for predicting bird activity on power transmission lines by integrating ecological data as described in any embodiment of the present invention.

[0014] In another aspect, the present invention also provides a computer-readable medium for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to implement a method for predicting bird activity along power transmission lines by integrating ecological data as described in any embodiment of the present invention.

[0015] The present invention has the following beneficial effects: 1. By integrating ecological data such as meteorology, vegetation, and migration, a causal relationship model between bird activity and the ecological environment was established, enabling predictions to move beyond surface phenomena and delve into underlying mechanisms. The multimodal time-series prediction model, combined with an attention mechanism, effectively utilizes the temporal patterns of historical and real-time data, achieving high-precision predictions of bird activity for the next 24 hours and providing ample time windows for proactive prevention and control.

[0016] 2. The risk-based grading and early warning mechanism, based on prediction results, enables differentiated and precise deployment of prevention and control strategies. Matching prevention and control measures are initiated for different risk levels and species types (such as common birds, high-risk birds, and protected species), avoiding the resource waste and equipment damage caused by the traditional "one-size-fits-all" approach. Simultaneously, it significantly reduces the risk of accidental injury to protected birds, achieving a win-win situation for both power safety and ecological protection.

[0017] 3. The dynamic early warning threshold adaptive mechanism enables the system to dynamically adjust its early warning sensitivity based on spatiotemporal changes such as seasons and phenology, enhancing the system's environmental adaptability. Through closed-loop optimization of "prediction-execution-feedback," the model can continuously perform incremental learning using actual performance data, allowing the prediction accuracy and decision-making intelligence of the entire system to continuously evolve over time, possessing powerful self-learning capabilities. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0019] It should be understood that the step numbers used in the text are for ease of description only and are not intended to limit the order in which the steps are performed.

[0020] It should be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0021] The terms “comprising” and “including” indicate the presence of the described feature, whole, step, operation, element and / or component, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or collections thereof.

[0022] The term “and / or” refers to any combination of one or more of the associated listed items, as well as all possible combinations, and includes these combinations.

[0023] Example 1: To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with specific embodiments of the present application.

[0024] To address the problems in existing technologies, this embodiment provides a method for predicting bird activity along power transmission lines by integrating ecological data, comprising the following steps: Ecological environment data and bird monitoring data of the power transmission line area were collected, and the collected data were preprocessed, feature extracted and spatiotemporally aligned to form a standardized dataset.

[0025] In a preferred embodiment of this invention, the ecological environment data includes meteorological data, vegetation and food resource data, and bird migration data; in this embodiment: Meteorological data: By deploying high-precision meteorological stations (sampling frequency 1 time / 10 minutes), data such as temperature (accuracy ±0.2℃), humidity (±2%RH), wind speed (0-60m / s), precipitation (0.1mm resolution), and air pressure (±0.1hPa) are collected. The focus is on recording the impact of seasonal climate characteristics (such as the passage of cold fronts during migration and continuous sunny weather during the breeding season) on bird activities.

[0026] Vegetation and food resource data: Using hyperspectral remote sensing images (10m×10m resolution) and ground vegetation monitoring instruments, vegetation type (coniferous forest / broadleaf forest / farmland), vegetation phenology (germination / flowering / fruiting), and insect density (counted by insect-attracting lamps) within a 500-meter radius around the power transmission line were obtained to establish a baseline linking food resource abundance and bird activity.

[0027] Bird migration data: By accessing historical migration trajectory data from global bird migration databases (such as eBird and GBIF) and combining them with local banding records, a spatiotemporal distribution model of bird migration in the target area is constructed, and key migration corridors (such as the essential routes for migratory birds to migrate north and south) and stopover sites are marked.

[0028] The bird monitoring data includes multimodal data collected through images, voiceprints, and lidar, as well as historical activity log data. In this embodiment, Image data: An 8-megapixel black light-level camera (supporting day and night mode switching) is used to capture one frame of bird roosting and flight images every 5 seconds, extracting features such as feather color, body size, and behavioral posture; Voiceprint data: Bird calls were recorded using an omnidirectional microphone (frequency response 20Hz-20kHz), and voiceprint feature parameters including fundamental frequency, spectral entropy, and Mel-frequency cepstral coefficients were extracted. Behavioral data: Bird activity trajectories were recorded using lidar (sampling rate 10Hz), and behavioral indicators such as dwell time, flight speed, and flock size were analyzed.

[0029] Historical activity data: Establish a bird activity log database to store data on bird species, numbers, appearance times, and activity areas for the past 3 years. Perform statistical analysis by season, month, and time period to uncover periodic activity patterns (e.g., the probability of magpies nesting on poles increases by 60% during the breeding season from April to June).

[0030] In a preferred embodiment of this invention, the preprocessing includes: Ecological data standardization: Meteorological data: Z-score standardization was used to eliminate the influence of dimensions, the formula is (x'=(x μ) / σ) (where (μ) is the mean and (σ) is the standard deviation); Migration data: Latitude and longitude coordinates are converted into relative positions with respect to transmission lines (e.g., within 50 meters of the tower), and spatial smoothing is performed using a Gaussian kernel function; Vegetation data: Vegetation cover is quantified using the Normalized Difference Vegetation Index (NDVI), calculated using the formula (NDVI = (NIR)). R) / (NIR+R)) (NIR is the reflectance in the near-infrared band, and R is the reflectance in the red band).

[0031] Feature extraction from bird monitoring data: Image features: 2048-dimensional depth features were extracted using the ResNet50 network, and combined with morphological features (such as wingspan and body length ratio) to distinguish similar species; Voiceprint features: The EfficientNet-B0 network is used to extract features from the Mel spectrogram, and the dynamic time warping (DTW) algorithm is used to achieve voiceprint matching; Behavioral characteristics: Convert trajectory data into kinematic parameters such as velocity, acceleration, and steering angle to identify behavioral patterns such as clustering and solitary behavior.

[0032] In a preferred embodiment of this invention, the spatiotemporal alignment includes: Time alignment: Using a unified timestamp as a benchmark, bird monitoring data and ecological environment data with different sampling frequencies are unified to the same time granularity through interpolation methods to form time series data. Specifically, using the UTC timestamp as a benchmark, high-frequency monitoring data (once every 5 seconds) is interpolated and matched with low-frequency ecological data (such as daily weather and weekly vegetation data) to construct a time series dataset in hourly units. For example, the daily average temperature is interpolated using a sine function to obtain hourly temperature data, maintaining consistency in time granularity.

[0033] Spatial alignment: The power transmission line area is divided into grids, and the locations of bird activities are spatially correlated with the ecological data within the grids. Specifically, the power transmission line area is discretized using a grid division method (10m×10m grid), and the locations of bird activities (latitude and longitude) are spatially correlated with ecological data (such as vegetation type within the grid and distance from water sources), forming a three-dimensional data cube of "spatial unit-time-multi-feature".

[0034] Based on a standardized dataset, a bird activity prediction model is used to predict bird activity in the power transmission line area. The bird activity prediction results include the probability, scale, and species type of bird activity.

[0035] As a preferred embodiment of this example, the bird activity prediction model includes: a species ecological niche sub-model, a multimodal time series prediction sub-model, and a migration prediction sub-model.

[0036] The species niche model uses the MaxEnt model to calculate the potential distribution probability of birds in the power transmission line area based on ecological factors. In this embodiment, the input of the species niche model includes 15 key ecological factors as prediction variables, including: meteorological factors: annual average temperature, extreme high temperature, precipitation seasonality, wind speed; topographic factors: altitude, slope, aspect; resource factors: distance from water source, vegetation cover, food resource abundance; and anthropogenic factors: distance from road, intensity of human activities.

[0037] Feature engineering: Non-collinear factors with VIF < 10 are screened out by factor expansion factor (VIF), and hinge feature transformation is used to enhance the ability to capture nonlinear relationships (such as converting temperature features into binary features of "inside and outside the suitable temperature range").

[0038] Model training: Five-fold cross-validation was used, with AUC (area under the curve) as the evaluation metric (target AUC > 0.85). Model parameters (such as regularization coefficient and feature weights) were adjusted through Bayesian optimization to generate a potential distribution probability map of species and identify highly suitable habitats within the transmission line area.

[0039] A multimodal time-series prediction sub-model, integrating a long short-term memory network and an attention mechanism, predicts the probability and scale of bird activity based on time-series data. In this embodiment, the model architecture of the multimodal time-series prediction sub-model includes: Input layer: Receives multimodal feature sequences (image feature sequences, voiceprint feature sequences, and ecological factor sequences); Encoding layer: Parallel LSTM networks are used to process temporal features of different modalities, with 128 LSTM units in the image branch, 64 in the speaker branch, and 32 in the ecological factor branch; Fusion layer: Introduces a spatiotemporal attention mechanism to dynamically allocate the weights of features of each modality (e.g., giving higher weights to image features during the breeding season and increasing the weights of ecological factors during the migration season), and captures cross-modal associations through multi-head attention; Output layer: Hybrid density network (MDN) is used to predict the probability of bird activity (0-1), aggregation size (0-10 birds / 100 birds / 1000 birds) and dominant species type per hour in the next 24 hours.

[0040] Model optimization: The AdamW optimizer (learning rate 0.001, weight decay 0.01) is used, focal loss is introduced to solve the sample imbalance problem (the weight of rare bird samples is increased by 5 times), and early stopping is used to prevent overfitting.

[0041] The migration prediction sub-model uses a bidirectional long short-term memory network to predict the time window for migratory birds to arrive at the power transmission line area and their flock movement trajectory. In this embodiment, the migration prediction sub-model uses a bidirectional LSTM network to learn the migration time sequence features and combines meteorological data such as air pressure field and wind direction to predict the time window for migratory birds to arrive at the power transmission line area (with the error controlled within ±12 hours).

[0042] Cluster risk prediction: For flocking birds (such as starlings), a swarm intelligence algorithm (particle swarm optimization) is introduced to simulate the movement trajectory of the cluster and predict whether the cluster will enter the 50-meter warning zone of the power transmission line within the next 3 hours, and output the predicted values ​​of cluster size and stay time.

[0043] A fusion layer is set at the output end of the bird activity prediction model. After normalization of the output results of the above three sub-models, the comprehensive activity prediction probability is obtained by weighted fusion. The historical prediction accuracy or real-time ecological environment characteristics are dynamically adjusted.

[0044] The comprehensive prediction results output by the fusion layer reflect the probability, scale, and species distribution of bird activities at specific times and locations. After visualization, they form a spatiotemporal dynamic heat map, which serves as the input basis for subsequent early warning level determination and prevention and control strategy generation.

[0045] The warning level is determined based on the bird activity forecast results, and corresponding prevention and control strategies are generated and implemented based on the warning level.

[0046] In a preferred embodiment of this invention, determining the warning level based on bird activity prediction results specifically involves: When the predicted activity probability is lower than the first threshold or the predicted bird species is a protected species, it is determined to be a low-risk level. In this embodiment, the low-risk level is divided into a first-level warning, which is triggered when the predicted bird activity probability is less than 30% within 24 hours, or the active birds are protected species and there are no high-risk birds.

[0047] When the predicted activity probability is between the first threshold and the second threshold, or when there is a risk of protected species gathering, it is determined to be a medium-risk level. In this embodiment, the medium-risk level is divided into a level two warning, and the triggering condition is that the predicted bird activity probability is 30%-70% within 24 hours, or there is a risk of a small number of common birds gathering.

[0048] Prediction basis: The niche model shows that the current area is a habitat with low species suitability, and the time series model predicts that the activity intensity is stable and at a historical low.

[0049] Prediction is based on suitable weather conditions (such as temperature 15-25℃ and light wind) and the peak activity period of local resident birds (such as morning feeding time from 6 to 8 am).

[0050] When the predicted activity probability is higher than the second threshold, or there is a risk of high-risk birds or common birds gathering, it is determined to be a high-risk level. In this embodiment, the high-risk level is divided into three levels of warning. The triggering conditions are that the predicted bird activity probability is >70% within 24 hours, or there are high-risk birds (such as crows and birds of prey).

[0051] As a preferred embodiment of this practice, the generation of differentiated prevention and control strategies based on early warning levels specifically includes: For low-risk levels, a control strategy of monitoring without driving away is initiated, and ecological observation and gentle guidance are implemented for protected species; the gentle guidance involves using gentle sound waves (3-5kHz low-frequency guiding sound) to guide the protected species.

[0052] For medium-risk levels, the first-level bird deterrent device, including sound and light bird deterrent device, is activated in advance; specifically, strobe lights and bionic bird calls (similar warning sounds) are activated 2 hours in advance, triggered once every 30 minutes to drive away ordinary birds.

[0053] For high-risk areas, the second level of bird deterrence devices will be activated in advance, including sound and light bird deterrence devices, laser bird deterrence devices, and ultrasonic bird deterrence devices. Specifically, enhanced prevention and control measures will be deployed 4 hours in advance. In addition to activating the first level of bird deterrence devices, high-risk birds will be targeted by combined laser scanning (532nm green laser, scanning frequency 5Hz) and ultrasonic waves (25-30kHz).

[0054] The actual bird activity data after the implementation of the prevention and control strategy is collected, compared with the prediction results to calculate the error, and the bird activity prediction model is optimized and updated based on the error.

[0055] In a preferred embodiment of this invention, the method further includes a dynamic early warning threshold adaptive mechanism. This mechanism dynamically adjusts the trigger threshold of the early warning level based on historical activity log data and real-time ecological environment data. The historical activity log includes bird breeding seasons and migration seasons. Specifically, it includes: Adjustment basis: Historical data identified a significant seasonal fluctuation pattern in bird activity intensity. For example: During the spring and autumn migration seasons (such as March-May and September-November): large numbers of migrating birds are active and highly random, increasing the overall risk background value.

[0056] Summer breeding season (e.g., April to July): Resident birds increase nesting and raising behavior, and their activity intensity and stay time on poles and towers are significantly prolonged.

[0057] Winter: Most bird species reduce their activity, resulting in a lower overall risk background value.

[0058] Specific adjustment rules: During migration and breeding seasons, the system automatically lowers the first threshold (low risk turning into medium risk) and the second threshold (medium risk turning into high risk). For example, the baseline threshold is temporarily adjusted from (30%, 70%) to (20%, 60%). This makes the system more sensitive during high-risk seasons, enabling it to issue early warnings and seize the initiative in prevention and control.

[0059] During periods of low activity, such as winter, the system may raise the thresholds to (40%, 80%) to avoid triggering unnecessary warnings due to sporadic activity and to conserve prevention and control resources.

[0060] Dynamic fine-tuning and immediate response based on real-time ecological and environmental data include: temporarily fine-tuning thresholds when real-time meteorological data indicates extreme weather (such as the passage of a cold front or severe convective weather) or extremely suitable weather (such as temperatures of 15-25°C and light winds). The threshold is lowered in the former case due to the risk of birds scrambling to land, and appropriately lowered in the latter case due to increased activity.

[0061] Phenological and resource signals: When real-time phenological data (such as farmland maturity period and fruit maturity period) obtained through hyperspectral remote sensing and vegetation monitoring instruments indicate significant changes in food resource abundance, the warning thresholds for relevant areas will be adjusted accordingly to address the risk of birds gathering to forage.

[0062] Example 2: This embodiment provides a power transmission line bird activity prediction system that integrates ecological data, including: The data acquisition module collects ecological environment data and bird monitoring data in the power transmission line area, and performs preprocessing, feature extraction and spatiotemporal alignment on the collected data to form a standardized dataset. The prediction module uses a bird activity prediction model based on a standardized dataset to obtain bird activity prediction results for the transmission line area. The prevention and control module determines the warning level based on the bird activity prediction results, and generates and executes the corresponding prevention and control strategy based on the warning level. The feedback optimization module collects actual bird activity data after the implementation of prevention and control strategies, compares it with the prediction results to calculate the error, and optimizes and updates the bird activity prediction model based on the error.

[0063] Example 3: This embodiment provides an electronic device that stores a computer program. When the computer program is executed by a processor, it implements a method for predicting bird activity on power transmission lines that integrates ecological data, as described in any embodiment of the present invention.

[0064] Example 4: This embodiment provides a computer-readable medium for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to implement a method for predicting bird activity along power transmission lines by integrating ecological data, as described in any embodiment of the present invention.

[0065] In this application embodiment, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent the existence of A alone, A and B simultaneously, or B alone. A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of singular or plural items. For example, at least one of a, b, and c can represent: a, b, c, a and b, a and c, b and c, or a and b and c, where a, b, and c can be single or multiple.

[0066] Those skilled in the art will recognize that the units and algorithm steps described in the embodiments disclosed herein can be implemented using electronic hardware, computer software, or a combination of electronic hardware and software. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0067] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0068] In the several embodiments provided in this application, any function, if implemented as a software functional unit and sold or used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0069] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention specification, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for predicting bird activity along power transmission lines by integrating ecological data, characterized in that, Includes the following steps: Ecological environment data and bird monitoring data of the power transmission line area were collected, and the collected data were preprocessed, feature extracted and spatiotemporally aligned to form a standardized dataset; Based on a standardized dataset, a bird activity prediction model is used to predict bird activity in the power transmission line area. The bird activity prediction results include the probability, scale, and species type of bird activity. The warning level is determined based on the bird activity forecast results, and corresponding prevention and control strategies are generated and implemented based on the warning level. The actual bird activity data after the implementation of the prevention and control strategy is collected, compared with the prediction results to calculate the error, and the bird activity prediction model is optimized and updated based on the error.

2. The method for predicting bird activity on power transmission lines by integrating ecological data according to claim 1, characterized in that: The ecological environment data includes meteorological data, vegetation and food resource data, and bird migration data; the bird monitoring data includes multimodal data collected through images, voiceprints, and lidar, as well as historical activity log data.

3. The method for predicting bird activity on power transmission lines by integrating ecological data according to claim 2, characterized in that: The spatiotemporal alignment includes: Time alignment: Using a unified timestamp as a benchmark, bird monitoring data and ecological environment data with different sampling frequencies are unified to the same time granularity through interpolation methods to form time series data; Spatial alignment: The area of ​​the power transmission line is divided into grids, and the location of bird activity is spatially correlated with the ecological data within the grids.

4. The method for predicting bird activity on power transmission lines by integrating ecological data according to claim 3, characterized in that: The bird activity prediction model includes: a species ecological niche sub-model, a multimodal time-series prediction sub-model, and a migration prediction sub-model; A species niche model was developed, using the MaxEnt model to calculate the potential distribution probability of birds in the power transmission line area based on ecological factors. A multimodal time-series prediction sub-model integrates long short-term memory networks and attention mechanisms to predict the probability and scale of bird activity based on time-series data; The migration prediction sub-model uses a bidirectional long short-term memory network to predict the time window for migratory birds to arrive at the power transmission line area and their flock movement trajectory.

5. The method for predicting bird activity on power transmission lines by integrating ecological data according to claim 1, characterized in that: The determination of the warning level based on bird activity prediction results is as follows: When the predicted probability of activity is below the first threshold or the predicted bird species is a protected species, it is classified as a low-risk level. When the predicted probability of activity is between the first and second thresholds, or when there is a risk of protected species gathering, it is determined to be at a medium risk level. When the predicted probability of activity is higher than the second threshold, or when there is a risk of high-risk birds or common birds gathering, it is determined to be a high-risk level.

6. The method for predicting bird activity on power transmission lines by integrating ecological data according to claim 1, characterized in that: The differentiated prevention and control strategy generated based on the early warning level specifically includes: For low-risk levels, a control strategy of monitoring without expelling animals will be implemented, along with ecological observation and gentle guidance for protected species. For medium-risk levels, the first level of bird deterrence devices, including sound and light bird deterrence devices, should be activated in advance. For high-risk levels, the second level of bird deterrence devices should be activated in advance, including sound and light bird deterrence devices, laser bird deterrence devices, and ultrasonic bird deterrence devices.

7. The method for predicting bird activity on power transmission lines by integrating ecological data according to claim 2, characterized in that: The method also includes a dynamic early warning threshold adaptive mechanism, which dynamically adjusts the trigger threshold of the early warning level based on historical activity log data and real-time ecological environment data. The historical activity log includes bird breeding seasons and migration seasons.

8. A transmission line bird activity prediction system integrating ecological data, characterized in that, include: The data acquisition module collects ecological environment data and bird monitoring data in the power transmission line area, and performs preprocessing, feature extraction and spatiotemporal alignment on the collected data to form a standardized dataset. The prediction module uses a bird activity prediction model based on a standardized dataset to obtain bird activity prediction results for the transmission line area. The prevention and control module determines the warning level based on the bird activity prediction results, and generates and executes the corresponding prevention and control strategy based on the warning level. The feedback optimization module collects actual bird activity data after the implementation of prevention and control strategies, compares it with the prediction results to calculate the error, and optimizes and updates the bird activity prediction model based on the error.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method for predicting bird activity on power transmission lines that integrates ecological data as described in claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method for predicting bird activity on power transmission lines that integrates ecological data as described in claims 1-7.