Optimization methods, devices and equipment for bird control on power transmission lines

Abstract

This invention discloses a method, apparatus, and equipment for optimizing bird control on power transmission lines, relating to the field of bird control technology for power transmission lines. The method includes the following steps: generating a bird-control pulse intervention signal containing initial control parameters by combining bird activity data and facility topology location data in the power transmission line area; alternately acquiring baseline perception data of the target bird and transient reflection characteristics generated by the intervention signal during the intervals and activation periods of the pulse intervention signal; performing time-series tracking and state determination on the baseline perception data and transient reflection characteristics to obtain stress behavior feedback data of the target bird, and generating a pulse adaptability index of the target bird based on the time-series evolution characteristics of the feedback data; when the adaptability index meets preset conditions, generating a parameter bias sequence superimposed on the initial control parameters, acquiring the spatial coordinates of the target bird, performing beamforming, and outputting the modulated pulse intervention signal. This invention addresses the problems of easily interrupted monitoring data and the tendency for target birds to undergo adaptive desensitization.

Description

Technical Field

[0001] This invention relates to the field of bird protection technology for power transmission lines, and more specifically, to a method, apparatus, and equipment for optimizing bird protection control on power transmission lines. Background Technology

[0002] Existing bird control solutions for power transmission lines mostly employ physical isolation or dynamic intervention methods based on multi-sensor fusion to address the dynamic defense needs in unattended and complex ecological environments.

[0003] For example, the invention patent with publication number CN121242012A discloses a multimodal bird habitat deterrence and status monitoring method, which integrates infrared, radar, and image sensors for multimodal perception and dynamically adjusts sound, light, and vibration intervention strategies based on Markov decision processes. Another example is the invention patent with publication number CN117981744B, which discloses a bird-damage protection system and method for power transmission lines. This system uses deep learning algorithms to extract bird species characteristics from different parts of the body, and automatically activates a deterrence device when a bird threat warning is detected.

[0004] However, the aforementioned existing technologies still have the following technical problems in practical applications: 1. During the implementation of pulsed intervention actions, bird deterrence equipment often fails to maintain high-definition capture of the behavioral characteristics of the target birds, resulting in blind spots or data interruptions in the dynamic tracking of the target during the intervention process. 2. The evaluation of bird control effectiveness often lags behind the intervention actions, making it difficult to respond within seconds based on the immediate stress state of the target birds. As a result, the adjustment of intervention strategies often lags behind changes in bird behavior. 3. The energy distribution and modulation patterns of existing intervention signals are relatively fixed, and it is difficult to balance precision strikes with environmental friendliness. This leads to target birds developing behavioral tolerance or physiological desensitization in the short term, resulting in a significant decrease in the success rate of repeated interventions over time.

[0005] Therefore, there is an urgent need for an optimized method for bird control of power transmission lines that combines anti-interference continuous monitoring, transient behavior analysis capabilities, and anti-desensitization directional intervention capabilities. Summary of the Invention

[0006] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a method, apparatus, and equipment for optimizing bird control on power transmission lines. By alternately acquiring baseline sensing data and transient stress characteristics to quantify the adaptability index, and superimposing an offset sequence for beamforming intervention, the present invention addresses the problems of easy interruption of monitoring data and easy adaptive desensitization of target birds in the prior art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: An optimized method for bird control on power transmission lines includes the following steps: By combining bird activity data in the power transmission line area with the topological location of the facilities, a bird-proof pulse intervention signal containing initial control parameters is generated; During the intervals and periods of action of the pulse intervention signal, the baseline perception data of the target birds and the transient reflection characteristics generated by the intervention signal were acquired alternately. The baseline sensing data and transient reflex features are subjected to time-series tracking and state determination to obtain the stress behavior feedback data of the target bird. Based on the time-series evolution characteristics of the feedback data, the pulse adaptability index of the target bird is quantified and generated. When the fitness index meets the preset critical conditions, the generated parameter bias sequence is superimposed on the initial control parameters, the spatial coordinates of the target bird are obtained and beamforming is performed, and the optimized modulated pulse intervention signal is output.

[0008] In a preferred embodiment, generating a bird-proof pulse intervention signal containing initial control parameters includes: extracting the spatial coordinates of intruding birds based on bird activity data in the power transmission line area and combining them with the facility topology location; calculating the relative distance between the intruding birds and the power transmission facility and screening target birds that meet preset risk triggering conditions; matching a basic bird-proofing strategy for the target birds; and generating a pulse intervention signal based on the initial control parameter values ​​in the strategy, wherein the initial control parameters are set based on the critical scintillation fusion frequency of the target birds.

[0009] In a preferred embodiment, acquiring the baseline perception data of the target birds includes: acquiring a time-synchronized sampled image sequence during the interval between pulse intervention signal transmissions; extracting the self-disturbance parameters generated when the bird-prevention intervention device transmits the pulse intervention signal; and performing reverse phase compensation on the sampled image sequence to obtain the baseline perception data after filtering out the device's self-disturbance.

[0010] In a preferred embodiment, the transient reflection features generated by the intervention signal include: during the intervention period of the pulse intervention signal, acquiring a transient exposure image sequence and extracting the non-steady pulse reflection spot in the exposure image sequence; performing multi-scale time-frequency decomposition on the optical envelope parameters of the pulse reflection spot, filtering out the preset jitter frequency, and taking the main frequency feature corresponding to the energy peak as the transient reflection feature.

[0011] In a preferred embodiment, obtaining the stress behavior feedback data of the target bird includes: performing time-series tracking of baseline perception data to obtain the dwell time of the target bird within the preset safety threshold of the power transmission facility after intervention; determining the cessation state of the target bird's approach to the power transmission facility based on the amplitude change of transient reflection characteristics; and generating stress behavior feedback data of the target bird by combining the dwell time, transient reflection characteristics, and cessation state.

[0012] In a preferred embodiment, the quantitative generation of the pulse adaptability index of the target birds includes: statistically analyzing the temporal evolution characteristics of dwell time and the intensity decay law of transient reflex characteristics in the feedback data of stress behavior within a continuous intervention period, and classifying the behavior cessation state within the feedback data; performing multi-dimensional feature fusion mapping on the temporal evolution characteristics, intensity decay law and classification results to obtain the pulse adaptability index characterizing the tolerance of the target birds to the current bird control strategy.

[0013] In a preferred embodiment, generating the parameter bias sequence includes: calculating a time-varying modulation parameter corresponding to the pulse fitness index based on a pre-established positive correlation mapping relationship, wherein the time-varying modulation parameter is used to simulate the spatial approximation effect; using the time-varying modulation parameter to generate a Doppler equivalent frequency shift bias sequence and a gain bias sequence with nonlinearly increasing amplitude; and combining the frequency shift bias sequence and the gain bias sequence to obtain the parameter bias sequence.

[0014] In a preferred embodiment, the output optimized modulation pulse intervention signal includes: extracting the three-dimensional spatial coordinates of the target bird based on reference sensing data and performing phased array beamforming to construct a directional intervention channel pointing towards the target bird; and outputting the pulse intervention signal after superimposing the bias sequence through the directional intervention channel to form a pulse intervention signal with spatial directivity and time-varying characteristics.

[0015] A power transmission line bird control optimization device for implementing any of the methods described above includes the following units: an initial drive unit, used to generate a bird-proofing pulse intervention signal containing initial control parameters by combining bird activity data in the power transmission line area with the topological location of the facility; a multi-mode sampling unit, used to alternately acquire the baseline perception data of the target bird and the transient reflection characteristics generated by the intervention signal during the interval and action period of the pulse intervention signal; a behavior quantization unit, used to perform time-series tracking and state determination on the baseline perception data and transient reflection characteristics respectively to obtain the stress behavior feedback data of the target bird, and quantify and generate the pulse adaptability index of the target bird based on the time-series evolution characteristics of the feedback data; and a directional optimization unit, used to generate a parameter bias sequence superimposed on the initial control parameters when the adaptability index meets the preset critical conditions, acquire the spatial coordinates of the target bird and perform beamforming, and output the optimized modulated pulse intervention signal.

[0016] An electronic device for implementing any of the above-described power transmission line bird control optimization methods includes: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of any of the above-described power transmission line bird control optimization methods.

[0017] The technical effects and advantages of the bird control optimization method, device, and equipment for power transmission lines of this invention are as follows: 1. This invention constructs a data acquisition mechanism that coordinates the timing of perception and intervention actions by alternately acquiring the baseline perception data of the target birds and the transient reflection characteristics generated by the intervention signal during the interval and action period of the pulse intervention signal. This effectively avoids the perception blind spot problem caused by strong intervention signal interference in the single-thread monitoring mode, and ensures the spatiotemporal continuity and data integrity of target dynamic monitoring throughout the bird control process.

[0018] 2. This invention obtains stress behavior feedback data by performing time-series tracking and state determination on the benchmark sensing data and transient reflection characteristics, respectively. Based on the time-series evolution characteristics of the feedback data, a pulse adaptability index is generated, realizing quantitative analysis of the behavioral response and stress state changes of target birds. This effectively overcomes the shortcomings of traditional macroscopic bird prevention assessment indicators, which are coarse and obviously lagging, and improves the timeliness and accuracy of the system's assessment of the target bird's state.

[0019] 3. This invention generates a parameter bias sequence and superimposes it onto the initial control parameters when the adaptability index meets the preset critical conditions. Based on the spatial coordinates of the target bird, beamforming is performed to output an optimized and modulated pulse intervention signal. While realizing spatially directional focusing of intervention energy, a dynamic variable of tolerance is introduced, which effectively suppresses the behavioral immunity and physiological desensitization phenomenon that target birds are prone to under single repeated intervention, and ensures the long-term effectiveness of the bird control strategy for power transmission lines. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the optimized bird control method for power transmission lines provided in an embodiment of the present invention.

[0021] Figure 2 This is a schematic diagram of the time-domain fluctuation signal sequence waveform of the pulse reflection spot provided in an embodiment of the present invention.

[0022] Figure 3 This is a schematic diagram of the low-frequency approximation coefficient waveform provided in an embodiment of the present invention.

[0023] Figure 4 This is a schematic diagram of the transient reflection feature waveform after denoising and inverse wavelet transform reconstruction, provided in an embodiment of the present invention.

[0024] Figure 5 This is a schematic diagram illustrating the trend of stress behavior and impulse adaptability index of target birds provided in an embodiment of the present invention.

[0025] Figure 6 This is a schematic diagram of a power transmission line bird control optimization device unit provided in an embodiment of the present invention.

[0026] Figure 7This is a schematic diagram of an exemplary storage medium that can be used to implement embodiments of the present disclosure, as provided in the embodiments of the present invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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.

[0028] Example 1, Figure 1 This invention provides an optimized method for bird control on power transmission lines, comprising the following steps: S1. Combining bird activity data in the power transmission line area with the topological location of facilities, generate a bird-proofing pulse intervention signal containing initial control parameters, including: S101. Based on bird activity data in the power transmission line area, extract the spatial coordinates of invasive birds and combine them with the facility topology location to calculate the relative distance between the invasive birds and the power transmission facilities, and screen target birds that meet the preset risk triggering conditions; specifically as follows: By using front-end sensing devices (such as a fusion node of a binocular vision camera and a microwave radar) installed on power transmission towers, the system collects continuous video streams and point cloud data in real time, including bird flight trajectories, resting postures, and species characteristics, i.e., bird activity data. The system achieves time synchronization between the camera and the radar through a hardware triggering mechanism, and performs spatiotemporal joint calibration and spatial registration on the continuous video streams and point cloud data based on the camera intrinsic parameters and radar-camera extrinsic parameter conversion matrix obtained by offline calibration using a joint calibration target (such as a checkerboard calibration board with built-in corner reflectors) before the equipment leaves the factory. Using a pre-trained deep learning object detection model (e.g., YOLOv8), object recognition and bounding box regression are performed on 2D images in the video stream. To ensure the model adapts to the complex environment of power transmission lines in the field, the specific training process of the deep learning object detection model is as follows: A dataset of images of birds along power transmission lines, including multiple species, lighting conditions, and weather conditions, is constructed. Species categories and the pixel coordinates of the actual 2D bounding boxes are labeled and uniformly scaled to a preset resolution (e.g., 640 x 640 pixels) as model input. A YOLOv8 architecture is used as the pre-trained network. The AdamW optimizer is used, with an initial learning rate set to 0.001 and a cosine annealing strategy for dynamic decay. The total loss function is composed of a weighted average of the class classification loss and the bounding box regression loss. After the model is trained in a preset batch (e.g., 300 times), the weight parameters with the highest average accuracy on the validation set are extracted and stored in the front-end sensing device as the final running parameters of the deep learning object detection model. During real-time detection, the system uses the deep learning object detection model for forward propagation and outputs the class confidence of the target in the image and the predicted two-dimensional bounding box. Subsequently, for each two-dimensional bounding box identified by the detection model, based on the offline calibrated radar-camera extrinsic parameter transformation matrix, the three-dimensional point cloud synchronously output by the microwave radar is projected onto the two-dimensional image plane, and the set of microwave radar three-dimensional point cloud coordinates whose spatial projection positions fall within the geometric range of the current two-dimensional bounding box is extracted as input to the density clustering algorithm. The density clustering process uses the DBSCAN algorithm. Specifically, the three-dimensional point cloud coordinate set corresponding to each two-dimensional bounding box is imported into the DBSCAN algorithm, and a search is performed in the spatial dimension according to the preset neighborhood search radius and minimum point count parameter. The DBSCAN algorithm automatically identifies and separates multiple discrete point cloud clusters contained in the current two-dimensional bounding box by identifying core points that meet the minimum point count requirement and performing cluster expansion. At the same time, outliers that do not meet the minimum point count requirement are judged as discrete background noise and removed. The neighborhood search radius is set based on the physical size of the trunk of common birds that endanger power transmission lines, with a value range of 0.1m to 0.3m. The minimum point count parameter has a value range of 5 to 10. Finally, the system further filters the multiple point cloud clusters corresponding to each two-dimensional bounding box output by the DBSCAN algorithm, extracts the point cloud cluster containing the most point clouds as the largest master point cloud cluster of the corresponding invasive bird, to represent the physical entity reflective surface of the invasive bird; extracts the geometric center coordinates of the largest master point cloud cluster as the real-time three-dimensional spatial coordinates of each invasive bird in the three-dimensional world coordinate system with the tower base as the origin, and the coordinates are represented as three-dimensional point vectors.

[0029] During system initialization, the building information model of the tower is imported or the real-time dynamic differential positioning technology is used to mark and calibrate the points to obtain the static coordinate set of the power transmission facilities (such as insulator strings, high-voltage phase line hanging points, crossarm anti-pulling pins and other high-risk nodes that are prone to bird damage) in the same three-dimensional coordinate system, that is, the topological location of the facilities. Based on the three-dimensional spatial coordinates and the topological location of the facility, the shortest spatial relative distance between the invasive birds and the power transmission facility is calculated in real time using spatial analytical geometry. The specific calculation formula is as follows: , in, This refers to the set of all pre-defined high-risk nodes in power transmission facilities. The three-dimensional spatial coordinates of the current invasive bird species. For facility nodes The calibration coordinates.

[0030] Determine whether the shortest spatial relative distance meets the preset risk triggering condition. The specific risk triggering condition is: the shortest spatial relative distance is less than the preset bird warning zone radius, and the extension line of the intruding bird's movement vector trajectory intersects with the bird warning zone where the power transmission facility is located (a three-dimensional spherical space with the calibration coordinates of the power transmission facility node as the center and the radius of the bird warning zone as the radius). The radius of the bird warning zone is set based on the State Grid's safe physical clearance for live-line work and the wingspan of a large wading bird when fully extended, with a recommended value range of 3.0m to 5.0m. The invasive bird's motion vector trajectory is constructed and predicted based on real-time three-dimensional spatial coordinates across multiple consecutive frames using a multi-target tracking algorithm (such as the DeepSORT algorithm). The specific steps of the multi-target tracking algorithm include: (1) The two-dimensional bounding box image of each invasive bird detected in the current video frame is cropped and uniformly scaled to 128×64 resolution as input and fed into the lightweight feature extraction network built into the DeepSORT algorithm (e.g., a lightweight feature extraction model based on residual network). The network outputs a fixed-dimensional (e.g., 128-dimensional) appearance feature vector through convolution operation, which is used to characterize the global appearance information such as feather texture and color of the invasive bird. (2) For historical targets that have been assigned a globally unique tracking ID in the previous frame, the Kalman filter module inside the DeepSORT algorithm is used to predict the spatial motion state. The Kalman filter module takes the three-dimensional spatial coordinates and spatial motion speed of the historical tracking ID in the previous frame as input. It has a preset state transition matrix that represents the uniform motion law and a covariance matrix for tolerating sensor measurement noise. Based on the matrix, the algorithm calculates according to the uniform kinematics principle and outputs the predicted three-dimensional spatial position of the historical target in the current frame. (3) Pair the invasive birds actually detected in the current frame with the historical targets predicted in step (2); in this process, the system calculates the Mahalanobis distance of each pair of "actual detection coordinates-predicted position" combinations, as well as the cosine distance between the currently extracted apparent feature vector and the apparent feature vector of the historical target; the Mahalanobis distance threshold (set to 9.48 based on chi-square distribution) and the cosine distance threshold (recommended value range of 0.2 to 0.4) are used as mandatory rejection conditions to directly eliminate matching combinations whose spatial or appearance deviation exceeds the corresponding distance threshold; for candidate target combinations that pass the threshold filtering, the system assigns a preset weight ratio to the Mahalanobis distance and the cosine distance (recommended weight ratio of Mahalanobis distance is 0.3, cosine distance is 0.48). The weighted sum (with a weight of 0.7) is used to calculate the comprehensive deviation cost of the candidate target combination, and then a global two-dimensional scoring matrix containing all "current-historical" target combinations and their deviation cost values ​​is constructed. In this matrix, a combination optimization calculation that minimizes the global total cost is performed (for example, using the Hungarian algorithm to solve the minimum weighted matching of this two-dimensional matrix). The system finds the connection scheme that minimizes the sum of the total deviation values ​​and binds them one-to-one as the optimal matching result. Based on this, the historical tracking IDs of the successfully matched invasive birds are continued, and new tracking IDs are assigned to new targets that are not matched. In addition, the system sets a maximum loss tolerance frame number (recommended value is 30 frames) to maintain the tracking ID lifecycle of targets that are temporarily blocked by power transmission facilities. (4) After matching, the Kalman filter module uses the actual detected three-dimensional spatial coordinates of the current frame to perform error correction and smoothing on the motion state of the successfully matched tracking ID, and extracts the smoothed historical continuous three-dimensional spatial coordinate sequence for linear extrapolation, so as to calculate the future spatial position and motion direction of the invading bird, thereby generating the motion vector trajectory of the invading bird. When the spatial location and movement vector of an invading bird meet the risk triggering conditions, the invading bird is marked as a target bird and assigned a unique tracking ID; if multiple birds in the power transmission line area meet the conditions at the same time, the individual with the shortest spatial relative distance is locked as the primary target bird.

[0031] S102. Match the basic bird control strategy for the target bird, and generate a pulse intervention signal based on the initial control parameter values ​​in the strategy. The initial control parameters are set based on the critical scintillation fusion frequency of the target bird, as follows: The primary target bird species tags (such as magpies and sparrows) identified by the deep learning target detection algorithm are used as the query key to query a pre-set intervention strategy mapping expert database, resulting in a basic bird control strategy containing initial control parameter values. This intervention strategy mapping expert database is constructed based on extensive visual physiological experimental observation data of common birds that threaten power transmission lines (such as magpies, sparrows, and peregrine falcons), specifically a relational data mapping table pre-installed on the device's local storage medium or a cloud server. The initial control parameters specifically include the initial emission frequency, pulse duty cycle, and initial trigger level amplitude for driving the optical intervention module (such as a high-intensity LED strobe array) in the bird control intervention device for that specific species. The emission frequency is set based on the critical scintillation fusion frequency (CFF) range of the target bird. Specifically, the system is based on the visual physiological characteristics of birds... Based on engineering experience, the lower limit of the CFF characteristic value of the primary target bird species is obtained, and the initial emission frequency parameter of the optical intervention module is set to a value between the upper limit of human CFF and the lower limit of the target bird's CFF. Taking the magpie as an example, its lower limit of CFF value is about 80Hz, while the human CFF value is usually between 45Hz and 60Hz. The system generates the corresponding basic bird prevention strategy by querying the mapping expert database: set the initial emission frequency parameter of the optical intervention device to 70Hz, the pulse duty cycle to 50%, and the trigger level amplitude to 5V. Since the 70Hz strobe light is higher than the human visual fusion frequency, it appears as a soft, constant light source to the human eye. However, this frequency is lower than the lower limit of the magpie's CFF value, so the light source appears as a violently dazzling, unstable, high-frequency strobe light in the magpie's vision, thereby triggering visual deprivation and panic stress response in the bird. Based on the initial control parameter values ​​obtained from the query, the main control chip (such as an MCU hardware timer module with microsecond-level clock accuracy) generates a pulse width modulation square wave electrical signal of the corresponding frequency as a pulse intervention signal output to the target birds.

[0032] This step eliminates interference calculations from invalid targets outside the safe zone by integrating three-dimensional spatial positioning and tracking with a frequency modulation strategy based on species CFF characteristics. At the same time, by setting differentiated frequency parameters, it applies initial physical sensory suppression with specific effects to the target birds, laying the foundation for subsequent implementation of transient stress feature analysis and dynamic optimization of bird control parameters.

[0033] S2. During the intervals and periods of action of the pulse intervention signal, the baseline perception data of the target bird and the transient reflection characteristics generated by the intervention signal are acquired alternately.

[0034] In this embodiment, obtaining the baseline perception data of the target bird in step S2 includes: S201. During the interval between pulse intervention signal transmissions, acquire a time-synchronized sampled image sequence; specifically as follows: The intervention state is determined by the high and low level logic output by the hardware timer of the main control chip (MCU). When the pulse width modulation signal (PWM) sent by the main control chip to the drive circuit is in the low duty cycle stage, the system determines it to be a transmission interval. At this time, the control clock of the intervention pulse is used as an external trigger source to generate a timing synchronization signal. The camera is triggered to sample at the moment the pulse is extinguished to obtain a timing-synchronized digital sampled image sequence. Specifically, the MCU outputs a synchronization trigger signal to the camera's external trigger pin (such as the TriggerIn interface) through the GPIO port and sets the trigger mode to edge trigger mode. At the same time, a trigger delay time with a value range of 10μs to 200μs is set to avoid electromagnetic spike interference at the moment the PWM signal is turned off. At this time, the continuous digital image sequence acquired avoids the peak power output time of the intervention signal and avoids the sensor saturation blinding phenomenon.

[0035] S202. Extract the self-disturbance parameters generated when the bird-prevention intervention device emits a pulse intervention signal, and perform reverse phase compensation on the sampled image sequence to obtain the baseline sensing data after filtering out the device's self-disturbance; specifically as follows: By analyzing the sub-pixel offset of static background anchor points in the sampled image, or by combining the self-disturbance parameters with the onboard high-precision inertial measurement unit (IMU), specifically: the system synchronously acquires the three-axis acceleration data of the IMU at a set high-frequency sampling rate (recommended range of 1000Hz to 4000Hz), calculates the three-axis dynamic acceleration magnitude after removing the static gravitational acceleration component through vector synthesis, and extracts the maximum value of the acceleration magnitude within the current pulse exposure time window to obtain the real-time acceleration peak value; when the real-time acceleration peak value is lower than the preset vibration tolerance threshold (set based on the transient exposure time of the camera and the physical size of the image sensor pixels, recommended range of 2.0m / s), the system can obtain the real-time acceleration peak value. 2 Up to 5.0 m / s 2When a high-intensity pulse triggers a real-time acceleration peak exceeding the vibration tolerance threshold, the system switches to IMU mode. A Kalman filter algorithm is introduced to perform attitude calculation and zero-bias error correction on the IMU output angular velocity and acceleration data, avoiding the cumulative drift error caused by direct quadratic integration. Subsequently, the image plane displacement is calculated using the camera focal length parameters and used as the self-perturbation parameter. This self-perturbation parameter characterizes the instantaneous mechanical micro-vibration of the equipment caused by the high-intensity pulse emission, or the displacement vector of the imaging sensor caused by severe power fluctuations; both are two-dimensional spatial displacement vectors. ; The self-perturbation parameter is used to perform reverse phase compensation on the sampled image sequence to obtain reference sensing data; the reverse phase compensation is achieved by performing a two-dimensional spatial displacement vector on the sampled image in the spatial domain. Resampling and shifting in the opposite direction physically cancels out image jitter. The calculation formula is as follows: , in, To filter out self-disturbances, the regional baseline sensing data for the target birds and local background of power transmission facilities is presented as a stable grayscale or color digital image. For the sampled image sequence, For self-disturbance displacement parameters; when When the pixel coordinates are non-integer pixels, bilinear interpolation or cubic convolution interpolation is used to resample the grayscale data before recalculating the baseline sensing data to avoid jagged edges caused by resampling.

[0036] In this embodiment, the transient reflection characteristics generated by the intervention signal in S2 include: S203. During the intervention period of the pulse intervention signal, acquire a transient exposure image sequence and extract the non-steady-state pulse reflection spot from the exposure image sequence; specifically as follows: When the pulse width modulation (PWM) signal sent by the main control chip to the drive circuit is in the high duty cycle stage, the system determines that it is the intervention period; specifically, during the peak period of the pulse signal illumination, a transient exposure image sequence is acquired using a preset shutter exposure time (usually set between 1ms and 5ms), and the transient exposure image sequence is represented as a single frame or multiple frames of digital images. The transient exposure image is compared with the reference sensing data using inter-frame difference method combined with background subtraction technology. After removing the static background, the abnormal brightness areas belonging only to the bird's body surface, namely the non-steady-state pulse reflection spots, are extracted. The specific steps are as follows: (1) After aligning the transient exposure image with the corresponding reference sensing image space, perform grayscale value subtraction operation on each pixel and take its absolute value to obtain the grayscale difference image; (2) Statistical analysis is performed on the two-dimensional positioning region of the target bird in the differential image. The two-dimensional positioning region is derived from the two-dimensional bounding box of the bird in step S1. The two-dimensional positioning region is updated in real time in each frame of the image. In order to avoid interference from reflection artifacts caused by pulse irradiation of static facilities such as power transmission towers, a dynamic motion mask is generated in combination with the bird's movement trajectory in step S1. (3) In the two-dimensional positioning area covered by the dynamic motion mask, the mean and standard deviation of the pixel values ​​of the differential image are statistically analyzed, and the standard deviation is multiplied by a preset empirical coefficient and added to the mean pixel value. The result is used as the brightness judgment threshold. The empirical coefficient ranges from 2.5 to 3.5. (4) Mark the pixels whose pixel values ​​exceed the brightness determination threshold as candidate bright pixels, and perform 8-neighborhood connected component analysis on the candidate bright pixels to merge spatially connected pixels into connected regions; (5) All connected regions are screened for consistency in area and brightness. Connected regions with an area greater than a preset pixel threshold (e.g., 50 pixels) and an average pixel value of the connected regions higher than the brightness determination threshold are identified as non-steady-state pulse reflection spots.

[0037] S204. Perform multi-scale time-frequency decomposition on the optical envelope parameters of the pulsed reflection spot, and after filtering out the preset jitter frequency, use the dominant frequency feature corresponding to the energy peak as the transient reflection feature; specifically as follows: The optical envelope parameters of the pulse reflection spot within multiple consecutive intervention cycles are extracted. The recommended range for the number of consecutive intervention cycles is 64 to 256 (e.g., 128). The specific value of the optical envelope parameter is the average gray value of all pixels within the connected region of the pulse reflection spot. The extracted optical envelope parameters of the consecutive intervention cycles are discretely spliced ​​in chronological order to construct a one-dimensional time-domain fluctuation signal sequence. The discrete sampling rate of the time-domain fluctuation signal sequence is synchronously equivalent to the initial transmission frequency of the pulse intervention device at a 1:1 ratio. Considering that bird stress responses are non-stationary transient signals, the system employs Discrete Wavelet Transform (DWT) to perform multi-scale time-frequency decomposition on the time-domain fluctuating signal sequence: Daubechies4 (db4) or Symlets5 (sym5) are selected as wavelet basis functions, and the number of wavelet decomposition levels is calculated based on the discrete sampling rate of the time-domain fluctuating signal sequence and the preset upper limit of low-frequency mechanical noise, as shown in the following formula: , in, The wavelet decomposition level is denoted as . The rounding up symbol, The discrete sampling rate of the time-domain fluctuating signal sequence. The preset upper limit for low-frequency mechanical noise (e.g., 1.5Hz); Based on the wavelet decomposition level, the signal is decoupled into the first wavelet decomposition layer, representing low-frequency background fluctuations. Layer approximation coefficients (corresponding) (low frequency band) and detail coefficients characterizing high frequency transient stress (corresponding to) (the above high-frequency bands); The approximation coefficients are thresholded to zero to filter out low-frequency mechanical interference noise at a preset jitter frequency (0.1Hz to 1.5Hz). After filtering out the low-frequency mechanical noise, a soft thresholding function is used to denoise the detail coefficients, smoothly shrinking white noise components below the threshold while retaining effective detail coefficients above the threshold. The threshold used is the VisuShrink universal threshold. ,in The total number of data points in the time-domain fluctuation signal sequence. To calculate the standard deviation of background noise, the median absolute deviation of the first layer detail coefficients (corresponding to the highest frequency band) is calculated and divided by a constant of 0.6745. Finally, the denoised detail coefficients are subjected to inverse wavelet transform to reconstruct a clean waveform that characterizes the transient panic action and muscle tremor response of birds. The dominant frequency feature corresponding to the envelope energy peak and the time node of the abrupt change in the reconstructed waveform are extracted as transient reflection features.

[0038] To demonstrate the effects of the above multi-scale time-frequency decomposition and denoising reconstruction, such as Figure 2 , 3 As shown in Figure 4, the figure illustrates the simulation signal decomposition results of the system using discrete wavelet transform to process the optical envelope parameters of the pulse reflection spot; among which, Figure 2 The waveform diagram is the original time-domain fluctuation signal sequence, which is interspersed with low-frequency swaying of the tower, high-frequency stress of birds, and high-frequency white noise from the environment. Figure 3 The waveform diagram of the low-frequency approximation coefficients from 0.1Hz to 1.5Hz, which were decoupled and isolated from the system, reflects the background noise of environmental mechanical interference; Figure 4 The image shows the clean transient waveform after soft thresholding to remove white noise and inverse wavelet transform reconstruction. The image clearly highlights the peak envelope energy and abrupt change time points generated when the bird is stimulated by the pulse, effectively filtering out interference from non-target environmental background. This step, through microsecond-level temporal coordination of sensing and intervention actions, achieves alternating detection of steady-state background during the interval period and transient stress during the action period. By using reverse phase compensation and frequency domain filtering techniques, the influence of equipment self-disturbance and environmental clutter is reduced, providing a high-confidence data source for subsequent quantitative analysis of birds' microscopic physiological adaptation to specific pulse stimuli.

[0039] S3. The baseline sensing data and transient reflex features are subjected to time-series tracking and state determination to obtain the stress behavior feedback data of the target bird. Based on the time-series evolution characteristics of the feedback data, the pulse adaptability index of the target bird is quantified and generated.

[0040] In this embodiment, the stress behavior feedback data of the target bird obtained in step S3 includes: S301. Perform time-series tracking on the baseline sensing data to obtain the dwell time of the target birds within the preset safety threshold of the power transmission facility after intervention, as detailed below: Based on the unique tracking ID locked for the target bird in step S101, the multi-target tracking algorithm is used to continuously track the continuous reference perception image sequence and extract the continuous motion trajectory of the target bird in three-dimensional space. The continuous motion trajectory is formed by connecting the center coordinates of the three-dimensional bounding box of the target bird in the time frame corresponding to the tracking ID. For the continuous movement trajectory, the bird-proof warning area (recommended radius 3.0m to 5.0m) set in step S1 is used as the preset safety threshold for the power transmission facility to determine the dwell time. Specifically: when the multi-target tracking algorithm first assigns the tracking ID to the target bird entering the bird-proof warning area, the system records the system timestamp carried by the corresponding image frame as the initial entry time; subsequently, the system continuously extracts the bounding box center coordinates corresponding to the tracking ID in subsequent consecutive frames along the continuous movement trajectory, and determines whether it remains within the three-dimensional geometric boundary of the bird-proof warning area; when the bounding box center coordinates of the tracking ID move out of the bird-proof warning area, or when the tracking ID is continuously lost in the image due to the target flying away for more than the preset target loss tolerance frame number (e.g., 15 frames), the system determines that the spatial tracking of the current intervention cycle has ended, and records the system timestamp of the current end frame as the departure time; finally, the difference between the initial entry time and the departure time is calculated to obtain the dwell time of the target bird.

[0041] S302. Based on the amplitude changes of transient reflection characteristics, determine the cessation state of the target bird's behavior towards the power transmission facility; specifically as follows: The transient reflection feature is a one-dimensional transient high-frequency signal waveform after wavelet reconstruction. The one-dimensional transient high-frequency signal waveform is analyzed using Hilbert transform, and the upper envelope formed by the analytical signal magnitude is extracted. The maximum amplitude of the upper envelope is extracted within the current intervention period as the peak value of the transient amplitude. The peak value of the transient amplitude characterizes the intensity of physiological panic behaviors such as eyelid closure, severe muscle spasm, or stress wing flapping when birds are subjected to specific stroboscopic stimuli. The instantaneous velocity vector of the target bird is obtained by performing a difference operation on the three-dimensional coordinate points of two adjacent frames on the continuous motion trajectory obtained in step S301; the real-time three-dimensional spatial coordinates of the target bird are subtracted from the static coordinate set of the power transmission facility nodes pre-calibrated in step S1 to obtain the spatial position vector from the target bird to the center of the power transmission facility; the spatial angle between the velocity vector and the spatial position vector, i.e., the bird's flight yaw angle, is calculated using the dot product and inverse cosine operation formulas of spatial analytic geometry. When the peak value of the transient amplitude exceeds the preset physiological stress threshold (based on long-term wading bird behavior experiment observations and expert experience, the recommended value range is 45 to 75), and the bird's flight yaw angle is greater than... When the instantaneous velocity vector deviates from the direction of the power transmission facility, indicating that the bird has turned around or taken an emergency avoidance action, the bird's approach to the power transmission facility within the current intervention period is determined to be successfully terminated and assigned a Boolean value of 1; otherwise, it is determined to be not terminated and assigned a Boolean value of 0. The bird's approach to the power transmission facility includes the bird's intrusive action of flying straight towards the tower, the bird's approaching action such as circling and probing at the edge of the critical area or slowing down to prepare to land on the tower and perch.

[0042] S303. Combining the dwell time, transient reflex characteristics, and termination state, generate stress behavior feedback data for the target bird; specifically as follows: The macroscopic behavioral parameters and microscopic physiological parameters extracted in each single intervention cycle are structurally spliced ​​together to obtain a low-dimensional feature column vector of "dwell time, transient amplitude peak value, and behavior termination state Boolean value", which is the stress behavior feedback data. The stress behavior feedback data characterizes the target bird's comprehensive stress performance under the current round of pulse action, including the tolerance time, the intensity of the physiological response, and whether it ultimately abandons the approach.

[0043] In this embodiment, the quantification of the pulse fitness index of the target bird in step S3 includes: S304. Analyze the temporal evolution characteristics of dwell time and the intensity decay pattern of transient reflex characteristics in the feedback data of stress behavior within a continuous intervention period, and classify the behavioral cessation states within the feedback data; specifically as follows: In order to capture the habitual desensitization phenomenon in birds, the system needs to extract the temporal trend of stress behavior feedback data within the historical continuous intervention cycle. The recommended number of continuous intervention cycles is 5 to 10 cycles to balance the number of behavioral change samples and the statistical duration. First, based on the natural chronological order of continuous intervention cycles, the dwell time values ​​within historical continuous intervention cycles are extracted, and a positive dwell time series with the current intervention cycle value as the tail is constructed. The least squares method is used to perform linear regression fitting on the positive dwell time series, and the slope of the fitted line is calculated as the temporal evolution characteristic of dwell time. If the slope is greater than 0, it indicates that as the number of intervention cycles increases, the dwell time of birds in the danger zone gradually becomes longer. Secondly, following the natural temporal sequence of continuous intervention cycles, a positive sequence of transient amplitude peaks is constructed, with the current intervention cycle value at the end. The difference between the transient amplitude peak value at the beginning of the sequence (i.e., the oldest cycle) and the transient amplitude peak value at the latest cycle is calculated and divided by the transient amplitude peak value at the beginning of the sequence to obtain the beginning-end attenuation rate of the transient reflection characteristics, i.e., the intensity attenuation law. To avoid negative attenuation caused by increased stress in birds, which could lead to the subsequent pulse adaptability index exceeding the limit, if the calculated beginning-end attenuation rate is less than 0, the beginning-end attenuation rate value is forcibly set to zero. The larger the beginning-end attenuation rate, the faster the bird's muscle spasms and panic response subside. Finally, the sequence of behavior termination states in continuous intervention cycles was statistically classified, and the classification results were mapped to discrete state parameters based on the frequency of successful termination (Boolean value of 1). Specifically: if the frequency is M times the number of continuous intervention cycles, the discrete state parameter is 0 (indicating high sensitivity); if the frequency is between 1 and M-1, the discrete state parameter is 0.5 (indicating that the target birds are in the adaptation transition period); if the frequency is 0, the discrete state parameter is 1 (indicating that the target birds are completely immune and desensitized).

[0044] S305. Perform multi-dimensional feature fusion mapping on the temporal evolution characteristics, intensity decay patterns, and classification results to obtain the pulse adaptability index characterizing the tolerance of target birds to the current bird control strategy; specifically as follows: Based on a pre-established normalized weighted evaluation mathematical function, a multi-dimensional feature fusion mapping is performed on the temporal evolution characteristics, intensity decay patterns, and classification results to obtain the pulse fitness index; the pulse fitness index The calculation formula is as follows: , in, The recommended values ​​for the preset weighting coefficients are 0.3, 0.4, and 0.3, respectively. By assigning the highest weight to the attenuation of microscopic physiological reflexes, advanced early warning can be achieved. The slope of the fitted line. The first and last decay rates, These are discrete state parameters; For standard An activation function is used to smoothly and nonlinearly map a slope parameter with infinite positive and negative spans to... interval; The pulse adaptability index is a value that takes on a range of... The closer the value of the scalar between 1 and 1, the faster the physiological fear response of the target bird disappears, the longer it stays, and the more it ignores the driving action. In other words, the target bird has become physiologically and psychologically desensitized to the current frequency and light intensity.

[0045] To intuitively present the dynamic evolution of the above-mentioned multi-dimensional feature fusion, such as Figure 5 As shown in the figure, the evolution of stress behavior and the trend of the impulse fitness index of a target bird species over 10 consecutive intervention cycles are illustrated. The left-hand main Y-axis corresponds to a bar chart representing the trend of gradually increasing dwell time within the intervention cycle (the slope of the fitted line is greater than 0). The first secondary Y-axis (inner side) on the right side corresponds to a dashed line with dots, representing the nonlinear decay trend of the transient amplitude peak. The second secondary Y-axis (outer side) on the right side corresponds to a dark solid line with squares, representing the impulse fitness index derived from a normalized weighted model. Figure 5 As indicated by the label, during the 7th consecutive intervention cycle, the curve of the pulse fitness index rose sharply and exceeded the preset desensitization threshold of 0.75. The system determined that the individual bird had been immune desensitized and then output a command to trigger the subsequent deep intervention mechanism.

[0046] This step, by integrating multi-target continuous tracking and cross-dimensional statistics of time-frequency features, transforms the abstract bird tolerance metric into an adaptability index derived from continuous periodic data. This improves upon the technical problem of traditional bird control equipment relying on high-delay post-event judgment and provides a unique and proactive triggering basis for subsequent early changes to intervention strategies.

[0047] S4. When the adaptability index meets the preset critical conditions, the parameter bias sequence is generated and superimposed on the initial control parameters, the spatial coordinates of the target bird are obtained and beamforming is performed, and the optimized modulated pulse intervention signal is output.

[0048] The system monitors the pulse adaptability index in real time. When the adaptability index is greater than or equal to the preset desensitization threshold (recommended value is 0.75), it is determined that the target bird has developed habitual desensitization. At this time, the system triggers the deep intervention mechanism, that is, it starts to execute the optimized modulation and directional output process of steps S401 to S405. In this embodiment, generating the parameter bias sequence in step S4 includes: S401. Based on a pre-established positive correlation mapping relationship, calculate the time-varying modulation parameter corresponding to the pulse fitness index, wherein the time-varying modulation parameter is used to simulate the spatial approximation effect; specifically as follows: After triggering the deep intervention mechanism, time-varying modulation parameters are calculated based on a pre-established exponential positive correlation mapping relationship. These time-varying modulation parameters are represented by a dimensionless scaling factor that grows non-linearly with the intervention time. Under the premise that the physical intervention device's position is fixed, non-linear frequency and amplitude distortion is used to simulate the physical characteristics of a predator approaching at high speed in the bird's vision and hearing—that is, the spatial approximation effect. The calculation formula for the time-varying modulation parameters is as follows: , in, The duration (in seconds) after triggering deep intervention is obtained by extracting the real-time timestamp of the system's current clock and subtracting the timestamp of the moment when the system determines that the adaptability index is greater than or equal to the desensitization critical threshold. For a moment The time-varying modulation parameter is expressed as a dimensionless, continuously increasing floating-point number. The pulse adaptability index; The recommended value for the basic aggression coefficient is 0.5. To approximate the rate gain constant, a value of 1.2 is recommended.

[0049] S402. Using the time-varying modulation parameters, generate a Doppler equivalent frequency shift bias sequence and a gain bias sequence whose amplitude increases non-linearly; specifically as follows: Based on a physical model of the Doppler effect approximating natural sound sources, a linear mapping is used to generate a Doppler equivalent frequency shift bias sequence to characterize the blue shift phenomenon of the transmission frequency. Simultaneously, based on the inverse square law of energy attenuation, a quadratic mapping is used to generate a gain bias sequence with non-linearly increasing amplitude to characterize the sharp energy jump caused by distance reduction. The calculation formulas are as follows: , , in, It is the Doppler equivalent frequency shift offset sequence, which is expressed as a time-varying frequency compensation value (unit: Hz). This is the frequency offset conversion factor (e.g., 2.0 Hz / s). This is a gain bias sequence, expressed as a time-varying voltage compensation value (in V). This is the gain conversion factor (e.g., 0.5V / s²).

[0050] S403. Combine the frequency shift offset sequence and the gain offset sequence to obtain the parameter offset sequence; specifically as follows: The frequency shift bias sequence and the gain bias sequence are combined in the digital domain in a matrix form to form a timing compensation matrix containing both frequency and amplitude control dimensions, i.e., the parameter bias sequence, which is represented as a two-dimensional time-varying vector array. .

[0051] In this embodiment, the output of the optimized modulated pulse intervention signal in step S4 includes: S404. Based on baseline sensing data, extract the three-dimensional spatial coordinates of the target bird and perform phased array beamforming to construct a directional intervention channel pointing towards the target bird; specifically as follows: Through the underlying system bus, the three-dimensional spatial coordinates of the target bird in the current video frame are extracted from the continuous movement trajectory of the target bird in step S3, in a three-dimensional world coordinate system with the tower base as the origin; the factory calibration coordinates of the center of the bird-proof intervention device array surface are extracted and used as the new coordinate origin; by performing vector subtraction between the three-dimensional spatial coordinates of the target bird and the factory calibration coordinates, the absolute spatial coordinates of the target bird with the center of the device array surface as the coordinate origin are obtained. Based on the absolute spatial coordinates, phased array beamforming is performed on the acoustic intervention module (such as the ultrasonic transducer) in the bird control intervention device. Specifically, to clarify the physical reference for spatial interference calculation, the ultrasonic transducer device preferentially adopts a uniform rectangular array architecture with rows and columns, and the physical distance between adjacent independent transmitting units is preset to half the wavelength corresponding to the center operating frequency. Under this array topology, a spatial analytical geometry algorithm is used to calculate the path geometric distance of the target spatial position of each transmitting unit in the device array by combining the fixed internal coordinates of each transmitting unit in the array plane with the absolute spatial coordinates of the target bird, and to obtain the spatial path difference between adjacent units. Then, through the field programmable gate array (FPGA) main control chip, the microsecond-level phase delay of the driving signal of each transmitting unit is controlled based on the spatial path difference. The phase delay causes the ultrasonic beams emitted by each unit to produce constructive interference (peaks superimposed, energy maximized) at the spatial coordinates of the target bird, while producing destructive interference in other directions. The directional energy main lobe synthesized in three-dimensional space through phased array beamforming, which has highly concentrated main energy and always points to the coordinates of the target bird, is the directional intervention channel.

[0052] S405. The pulse intervention signal after superimposed bias sequence is output through the directional intervention channel to form a pulse intervention signal with spatial directivity and time-varying characteristics; specifically as follows: Based on the initial transmission frequency and trigger level amplitude generated in step S1, the frequency shift offset sequence and gain offset sequence are superimposed to obtain optimized control parameters. To avoid the signal frequency exceeding the lower limit of the target bird's CFF due to excessive blue shift increment, the superimposed transmission frequency is protected by upper limit amplitude based on a preset safety margin (recommended value range is 2Hz to 5Hz). Taking the magpie as an example, assuming the initial transmission frequency parameter for the magpie is 70Hz and the trigger level amplitude is set to 5V; since the lower limit of the magpie's CFF is approximately 80Hz, and the preset safety margin is 2Hz, the upper limit of the transmission frequency is set to 78Hz; the parameter bias sequence calculated at the current moment is frequency bias "+15Hz" and gain bias "+2.5V", and the superimposed transmission frequency parameter is 85Hz and the trigger level amplitude is 7.5V. Since 85Hz is greater than the set upper limit of the transmission frequency, the upper limit amplitude protection is triggered; therefore, the system forces the final optimized transmission frequency to 78Hz and the trigger level amplitude to 7.5V. At the same time, the FPGA chip, through the pre-loaded built-in duty cycle reload register, synchronously calculates and reloads the duration of the absolute high level within the pulse period at each clock interrupt trigger edge of updating the transmission frequency, so as to keep the pulse duty cycle unchanged; The optimized optical pulse emission frequency and level amplitude gain are synchronously mapped to the low-frequency amplitude modulation envelope and transmission power gain of the ultrasonic high-frequency carrier signal in the acoustic intervention module to optimize the control parameters of the acoustic intervention module. For example, if the optimized optical pulse emission frequency is 78Hz and the trigger level amplitude is increased from the initial 5V to 7.5V (i.e., the amplitude is amplified by 1.5 times), the system directly uses 78Hz as the low-frequency amplitude modulation envelope frequency of the ultrasonic signal (e.g., a carrier frequency of 40kHz). That is, the 40kHz high-frequency continuous ultrasonic wave emitted by the acoustic module will be amplitude modulated by the 78Hz envelope signal, thus being regularly divided into 78 high-frequency ultrasonic pulses per second, thereby achieving absolute synchronization with the 78Hz strobe of the LED in terms of excitation rhythm. At the same time, the system equivalently transfers the 1.5 times gain ratio of the optical end to the acoustic high-voltage drive end. For example, if the initial high-voltage drive level of the ultrasonic transducer is 50V, the mapped optimized ultrasonic drive level is synchronously amplified by 1.5 times, increasing to 75V. The drive signal containing optimized control parameters is fed in parallel into the LED strobe array and the ultrasonic transducer array via the FPGA chip, and the forced pulse intervention signal is output to the target bird through the directional energy main lobe (i.e., the directional intervention channel). Under the spatial filtering cover of directional acoustic beamforming, the target bird will be simultaneously subjected to intense high-frequency light splatter that approaches its visual CFF limit, as well as spatially directional high sound pressure ultrasonic pulse stimulation at the same frequency as the light splatter.

[0053] This step, through nonlinear superposition of parameter bias and spatial filtering techniques of beamforming, endows the pulse signal with spatial directivity for dynamic tracking of the target trajectory, as well as the time-varying characteristics of exponentially increasing stimulus intensity. This improves the technical problems of easy energy dissipation and habitual desensitization in traditional non-directional defense equipment, effectively enhancing the spatial efficiency of intervention signals and the long-term effectiveness of bird deterrence equipment.

[0054] Example 2, Figure 6 A bird control optimization device for implementing the method is provided, comprising the following units: The initial drive unit is used to combine bird activity data in the power transmission line area with the topological location of the facilities to generate a bird-proof pulse intervention signal containing initial control parameters; The multi-mode sampling unit is used to alternately acquire the baseline perception data of the target bird and the transient reflection characteristics generated by the intervention signal during the interval and action period of the pulse intervention signal; The behavior quantification unit is used to perform time-series tracking and state determination on the baseline perception data and transient reflex features respectively, to obtain the stress behavior feedback data of the target bird, and to quantify and generate the pulse adaptability index of the target bird based on the time-series evolution features of the feedback data. The orientation optimization unit is used to generate a parameter bias sequence superimposed on the initial control parameters when the fitness index meets the preset critical conditions, obtain the spatial coordinates of the target bird, perform beamforming, and output an optimized and modulated pulse intervention signal.

[0055] Example 3, Figure 7 An electronic device for implementing the aforementioned bird control optimization method for power transmission lines is provided, comprising: Memory, used to store computer programs; A processor, configured to implement the steps of any of the above-described power transmission line bird control optimization methods when executing the computer program.

[0056] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0057] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0058] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. 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.

[0059] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0060] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0061] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An optimized method for bird control on power transmission lines, characterized in that, Includes the following steps: By combining bird activity data in the power transmission line area with the topological location of the facilities, a bird-proof pulse intervention signal containing initial control parameters is generated; During the intervals and periods of action of the pulse intervention signal, the baseline perception data of the target birds and the transient reflection characteristics generated by the intervention signal were acquired alternately. The baseline sensing data and transient reflex features are subjected to time-series tracking and state determination to obtain the stress behavior feedback data of the target bird. Based on the time-series evolution characteristics of the feedback data, the pulse adaptability index of the target bird is quantified and generated. When the fitness index meets the preset critical conditions, the generated parameter bias sequence is superimposed on the initial control parameters, the spatial coordinates of the target bird are obtained and beamforming is performed, and the optimized modulated pulse intervention signal is output.

2. The optimized method for bird control on power transmission lines according to claim 1, characterized in that, The generation of the bird-proof pulse intervention signal containing initial control parameters includes: Based on bird activity data in power transmission line areas, the spatial coordinates of invasive birds are extracted and combined with the topological location of the facilities to calculate the relative distance between the invasive birds and the power transmission facilities and screen target birds that meet the preset risk triggering conditions. A basic bird control strategy targeting the target birds is matched, and a pulse intervention signal is generated based on the initial control parameter values ​​in the strategy. The initial control parameters are set based on the critical scintillation fusion frequency of the target birds.

3. The optimized method for bird control on power transmission lines according to claim 1, characterized in that, The acquisition of baseline sensory data for the target bird includes: During the interval between pulse intervention signal transmissions, a time-synchronized sequence of sampled images is acquired. The self-disturbance parameters generated when the bird-proof intervention device emits a pulse intervention signal are extracted, and the sampled image sequence is subjected to reverse phase compensation to obtain the baseline sensing data after filtering out the device's self-disturbance.

4. The optimized method for bird control on power transmission lines according to claim 1, characterized in that, The transient reflection characteristics generated by the intervention signal include: During the intervention period of the pulse intervention signal, a transient exposure image sequence is acquired and the non-steady pulse reflection spot in the exposure image sequence is extracted; The optical envelope parameters of the pulse reflection spot are decomposed into multiple scales and time frequencies. After filtering out the preset jitter frequency, the main frequency feature corresponding to the energy peak is taken as the transient reflection feature.

5. The optimized method for bird control on power transmission lines according to claim 1, characterized in that, The obtained stress behavior feedback data of the target birds includes: By performing time-series tracking on baseline sensing data, the dwell time of target birds within the preset safety threshold of the power transmission facility after intervention was obtained; Based on the amplitude changes of transient reflection characteristics, the cessation state of the target bird's behavior of approaching the power transmission facility is determined; By combining the dwell time, transient reflex characteristics, and termination state, stress behavior feedback data of the target bird is generated.

6. The optimized method for bird control on power transmission lines according to claim 1, characterized in that, The quantitative generation of the pulse fitness index for the target bird includes: The study analyzed the temporal evolution characteristics of dwell time and the intensity decay patterns of transient reflex characteristics in the feedback data of stress behavior within a continuous intervention cycle, and classified the behavioral cessation states within the feedback data. The temporal evolution characteristics, intensity decay patterns, and classification results are subjected to multi-dimensional feature fusion mapping to obtain the pulse adaptability index, which characterizes the tolerance of target birds to the current bird control strategy.

7. The optimized method for bird control on power transmission lines according to claim 1, characterized in that, The generated parameter bias sequence includes: Based on a pre-established positive correlation mapping relationship, the time-varying modulation parameter corresponding to the pulse fitness index is calculated, and the time-varying modulation parameter is used to simulate the spatial approximation effect; Using the time-varying modulation parameters, a Doppler equivalent frequency shift bias sequence and a gain bias sequence with non-linearly increasing amplitude are generated; The frequency shift bias sequence and the gain bias sequence are combined to obtain the parameter bias sequence.

8. The optimized method for bird control on power transmission lines according to claim 1, characterized in that, The output optimized modulated pulse intervention signal includes: Based on baseline sensing data, the three-dimensional spatial coordinates of the target bird are extracted and phased array beamforming is performed to construct a directional intervention channel pointing towards the target bird. The pulse intervention signal after superimposing the bias sequence is output through the directional intervention channel to form a pulse intervention signal with spatial directivity and time-varying characteristics.

9. A power transmission line bird control optimization device for implementing the method of any one of claims 1 to 8, characterized in that, Includes the following units: The initial drive unit is used to combine bird activity data in the power transmission line area with the topological location of the facilities to generate a bird-proof pulse intervention signal containing initial control parameters; The multi-mode sampling unit is used to alternately acquire the baseline perception data of the target bird and the transient reflection characteristics generated by the intervention signal during the interval and action period of the pulse intervention signal; The behavior quantification unit is used to perform time-series tracking and state determination on the baseline perception data and transient reflex features respectively, to obtain the stress behavior feedback data of the target bird, and to quantify and generate the pulse adaptability index of the target bird based on the time-series evolution features of the feedback data. The orientation optimization unit is used to generate a parameter bias sequence superimposed on the initial control parameters when the fitness index meets the preset critical conditions, obtain the spatial coordinates of the target bird, perform beamforming, and output an optimized and modulated pulse intervention signal.

10. An electronic device for implementing the power transmission line bird control optimization method according to any one of claims 1 to 8, characterized in that, include: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the steps of the power transmission line bird control optimization method as described in any one of claims 1 to 8.

Images