An AI vision-based comprehensive intelligent laser bird repelling device, bird repelling method and electronic equipment

Abstract

This invention relates to the field of power equipment protection technology, specifically to a comprehensive intelligent laser bird deterrent device and method based on AI vision. The device includes an image acquisition module for acquiring environmental images and transmitting them to an AI vision analysis module. This AI vision analysis module, based on a lightweight target detection model and multi-target tracking algorithm, identifies, tracks, and predicts the flight trajectories of birds in the images, generating analytical data including location, speed, number, and behavior type. A laser emission module projects green laser dots at the target birds based on the analytical data. A high-intensity light module emits strong light towards the target birds when the laser emission module's deterrent efficiency is insufficient. This bird deterrent device achieves eco-friendly bird deterrence by integrating artificial intelligence visual recognition, a multimodal bird deterrent execution unit, and an adaptive collaborative control strategy. It eliminates physical harm to birds, establishes an active identification and response mechanism, improves the long-term effectiveness of bird deterrence, reduces reliance on manual labor, and lowers operation and maintenance costs.

Description

Technical Field

[0001] This invention relates to the field of power equipment protection technology, specifically to a comprehensive intelligent laser bird deterrent device and method based on AI vision. Background Technology

[0002] In the existing power system operating environment, the behavior of birds perching, resting, and nesting on transmission towers has long threatened the safe and stable operation of the power grid. To address this issue, physical barriers such as bird spikes and bird-proof pinboards are currently widely used as the main means of prevention and control.

[0003] These physical bird deterrent devices are typically fixed to the angle steel structure of towers using stainless steel cable ties, bolts, or welding. They create physical barriers by placing sharp metal components in areas where birds frequently land, thus inhibiting birds from landing and nesting. However, these physical methods have significant ecological and ethical flaws. Their mechanism of action relies on causing potential physical harm to birds, potentially even resulting in actual stings. This pain-based deterrence violates the principles of non-lethal and non-harmful intervention advocated by modern wildlife conservation. Especially in migratory bird flyways, nature reserves, or other ecologically sensitive areas, they may inadvertently injure rare or protected birds, sparking ecological controversies and public concern.

[0004] Furthermore, the effectiveness of physical devices is limited to a fixed geometric space, making it impossible to dynamically adjust protective strategies based on bird behavior. Once birds adapt to the deployment pattern or choose detours, the protective effect will rapidly diminish. The devices themselves lack target recognition capabilities, cannot distinguish between high-risk birds and harmless species, and cannot determine whether birds actually intend to stay, resulting in a lack of specificity and flexibility in control measures.

[0005] Existing technologies also suffer from low operation and maintenance efficiency. Relying entirely on manual on-site installation and periodic inspections makes remote status monitoring and fault early warning difficult, and it's impossible to detect issues like corrosion, loosening, or detachment of equipment in a timely manner. In widely distributed transmission line scenarios, maintenance costs are high, response times are long, and aging equipment may become a falling hazard, posing secondary safety risks.

[0006] More significantly, traditional methods lack multimodal coordination mechanisms and cannot adaptively adjust based on dynamic factors such as bird species, lighting conditions, and flight paths. Single bird deterrence methods are easily adapted to by birds, leading to a decline in overall control efficiency over time. Especially at night, in low light conditions, or in rainy or foggy weather, physical barriers lack sensing capabilities and cannot actively trigger enhancement measures, resulting in frequent protection gaps and severely weakening the system's sustained effectiveness. Summary of the Invention

[0007] The purpose of this invention is to overcome the shortcomings of the prior art and provide an integrated intelligent laser bird deterrence device and method based on AI vision. This bird deterrence device achieves eco-friendly bird deterrence by integrating artificial intelligence visual recognition, multimodal bird deterrence execution unit and adaptive collaborative control strategy, eliminates physical harm to birds, establishes an active recognition and response mechanism, improves the long-term effectiveness of bird deterrence, reduces reliance on manual labor and lowers operation and maintenance costs.

[0008] To achieve one of the above objectives, the present invention provides the following technical solution: Provides a comprehensive intelligent laser bird deterrent device based on AI vision, including: The image acquisition module is used to acquire environmental images and transmit them to the AI ​​visual analysis module. The AI ​​visual analysis module, based on a lightweight target detection model and multi-target tracking algorithm, identifies, tracks, and predicts the flight trajectories of birds in images, generating analytical data including location, speed, number, and behavior type. The laser emitting module projects a green laser dot at the target bird based on the analyzed data. The high-intensity light module emits a strong light towards the target bird when the laser emission module's repelling efficiency is insufficient.

[0009] In some embodiments, an acoustic bird-repelling module is also included, which outputs ultrasonic signals and / or audio signals simulating predator calls to enhance the bird-repelling effect through auditory stimulation.

[0010] In some implementations, the AI ​​visual analysis module employs a lightweight YOLOv5 model, which features a feature pyramid structure in its backbone network and an attention mechanism module in its neck network. The AI ​​visual analysis module incorporates a Focal Loss function module during training to adjust sample weights.

[0011] In some implementations, the AI ​​visual analysis module integrates the DeepSORT multi-target tracking algorithm, which combines Kalman filtering for state estimation and uses an LSTM neural network to model historical motion sequences to predict the flight path trend of the target bird.

[0012] In some implementations, it also includes a power module, a communication module, and a control module. The power module provides power. The control module connects the image acquisition module, the AI ​​visual analysis module, the laser emission module, the high-intensity light module, the sound wave bird deterrent module, and the communication module. It is used to receive data and dynamically generate control commands according to a preset strategy, and to schedule the bird deterrent modules to work together.

[0013] In some embodiments, a power amplifier module is also included, which is connected to both the control module and the sound wave bird deterrent module, to amplify the original audio electrical signal output by the control module in order to enhance the effect of the sound waves emitted by the sound wave bird deterrent module.

[0014] In some embodiments, an installation frame is also included, on which the image acquisition module, the laser emission module, the high-intensity light module, the power supply module, the control module, the sound wave bird deterrent module, and the communication module are all integrated.

[0015] To achieve the second objective mentioned above, the present invention provides the following technical solution: The bird-repelling method based on the above-mentioned AI vision-based integrated intelligent laser bird-repelling device includes: The image acquisition module collects image data around the power transmission line and transmits the image data to the AI ​​visual analysis module. Based on the image data, the AI ​​visual analysis module identifies and obtains the number, trajectory, and dwelling behavior of bird targets, generating AI visual analysis data. Based on AI visual analysis data instructions, the laser emission module projects a laser point near the target bird or its flight path, causing the target bird to actively fly away from the protected airspace. The image acquisition module continuously monitors bird targets. When the flock of birds is dense, the repelling efficiency of the high-intensity light module is insufficient, or when targeting specific birds that are not sensitive to lasers, the high-intensity light module is instructed to be triggered to emit high-intensity light towards the target birds until the bird repelling task is completed.

[0016] In some implementations, a sound wave bird deterrent module emits sound waves to the target birds, using ultrasonic signals and / or simulated predator calls in conjunction with the laser emission module and the high-intensity light module to drive away the target birds.

[0017] The beneficial effects of this invention, a comprehensive intelligent laser bird-repelling device and method based on AI vision, are as follows: This invention relates to an AI-based intelligent laser bird deterrence device. An image acquisition module collects environmental images, while an AI vision analysis module identifies, tracks, and predicts the flight trajectories of birds in the images, generating analytical data including location, speed, number, and behavioral type. Based on this data, a laser emission module projects green laser dots towards the target birds, effectively stimulating their instinctive fear and forcing them to actively fly away from the protected airspace, thus achieving the bird deterrence goal. The green laser dots do not harm the birds and effectively prevent them from entering the protected area. When the laser emission module's deterrence efficiency is insufficient, a high-intensity light module emits strong light towards the target birds to supplement the laser effect, achieving efficient, precise, and adaptive bird deterrence performance. Furthermore, it can dynamically adjust the intervention parameters of the laser and high-intensity light modules, achieving closed-loop optimization of the bird deterrence strategy. This not only significantly improves the accuracy and response speed of bird deterrence but also reduces power consumption through dynamic energy allocation, extending the device's lifespan while avoiding excessive intervention. This ensures a safe, environmentally friendly, and long-term effective bird deterrence process, providing a reliable bird protection solution for critical facilities such as power transmission lines.

[0018] An electronic device is also provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the above-described method when executing the computer program. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the integrated intelligent laser bird deterrent device based on AI vision in this embodiment.

[0020] Figure 2 This is a flowchart of the bird-repelling method of the AI ​​vision-based integrated intelligent laser bird-repelling device in this embodiment.

[0021] Figure Labels 1. Antenna module; 2. Sound wave bird deterrent module; 3. Button module; 4. High-intensity light module; 5. Control module; 6. Laser high-voltage board; 7. Image acquisition module; 8. Power amplifier module; 9. Controller module; 10. Laser emission module; 11. Power supply module. Detailed Implementation

[0022] Preferred embodiments of the invention will now be described in more detail. While preferred embodiments of the invention have been shown, it should be understood that the invention can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that the invention will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.

[0023] The terminology used in this invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The singular forms “a” and “the” as used in this invention and the appended claims are intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0024] It should be understood that although the terms "first," "second," "third," etc., may be used in this invention to describe various information, this information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this invention, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Thus, features defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0025] Transmission lines are often located in open, complex environments, making them ideal habitats for birds to perch and nest on conductors and towers. This poses a serious threat to line safety and power supply stability, potentially causing outages, short circuits, and other accidents. Therefore, it is essential to develop an intelligent bird deterrence system suitable for power applications. This phase will focus on developing AI hardware and software systems adapted for outdoor visual recognition tasks, possessing strong robustness and real-time performance, capable of accurately identifying birds and their behavior. A controllable laser and acoustic-optical combined bird deterrence module will be developed, balancing deterrence effectiveness with safety for birds and personnel, ensuring compliance with power industry standards. A high-precision, stable gimbal platform will be designed, combined with motor control circuitry, to achieve rapid response and precise target locking, laying the foundation for subsequent automatic tracking and intelligent bird deterrence.

[0026] To solve the above-mentioned technical problems, the following embodiments are disclosed: Example

[0027] This embodiment discloses a comprehensive intelligent laser bird deterrent device based on AI vision. Please refer to [link to relevant documentation]. Figure 1 ,include: Image acquisition module 7 is used to acquire environmental images and transmit them to the AI ​​visual analysis module. Specifically, the image acquisition module 7 collects high-definition videos of the surrounding environment in real time, providing raw visual data to the backend AI visual analysis module.

[0028] The AI ​​visual analysis module, based on a lightweight target detection model and multi-target tracking algorithm, identifies, tracks, and predicts the flight trajectories of birds in images, generating analytical data including location, speed, number, and behavior type. Specifically, the AI ​​visual analysis module can accurately identify and track bird targets, and analyze their numbers, trajectories, and dwelling behaviors. Based on this intelligent perception, the device can decide whether to trigger other modules.

[0029] The laser emission module can project green laser dots towards target birds.

[0030] Specifically, the laser emission module 10 receives instructions from the AI ​​visual analysis module, converting the invisible control signal into a visible, non-harmful green laser beam. This beam is then precisely projected onto the vicinity of the target bird or its flight path via a precision gimbal that rotates horizontally and vertically, effectively stimulating its instinctive fear and forcing it to fly away from the protected airspace. Furthermore, this precision gimbal is a two-dimensional servo gimbal, capable of rotating ±180° horizontally and ±90° vertically.

[0031] High-intensity light module 4: When the laser emission module 11 is not efficient enough in driving away birds, the high-intensity light module 4 can emit high-intensity light towards the target birds.

[0032] Specifically, the high-intensity light module 4 serves as an effective supplement to laser-based precise bird deterrence. When encountering dense flocks of birds, insufficient laser deterrence efficiency, or specific birds that are not sensitive to lasers, this module is instantly triggered by instructions from the AI ​​visual analysis module, emitting a high-intensity, specific-frequency flashing beam of light. This sudden burst of intense light can simulate natural predators or danger signals, causing strong fright and interference to the birds' visual systems, thus inducing instinctive panic and prompting them to quickly flee the area. This significantly improves the overall deterrence efficiency and reliability of the device in complex scenarios.

[0033] In this embodiment, a sound wave bird repelling module 2 is also included. The sound wave bird repelling module 2 outputs ultrasonic signals and / or audio signals that simulate the calls of natural enemies, thereby enhancing the bird repelling effect through auditory stimulation.

[0034] Specifically, the acoustic bird-repelling module 2 works in conjunction with the laser emission module 11 and the high-intensity light module 4. The acoustic bird-repelling module 2 selectively emits two types of sound waves: one is high-frequency ultrasound, exemplarily 20kHz to 25kHz, which, although inaudible to humans, causes strong discomfort to the auditory system of birds; the other is pre-recorded deterrent sound waves such as calls of bird predators, terrified cries of birds, or plosive sounds, directly triggering their instinctive fear. The acoustic bird-repelling module is particularly suitable for scenarios with poor visibility (or dense vegetation that affects optical methods). This forms a three-dimensional, composite bird-repelling strategy, effectively preventing birds from adapting to a single repelling method, thereby significantly improving the long-lasting repelling effectiveness of the entire system.

[0035] In this embodiment, the AI ​​visual analysis module adopts a lightweight YOLOv5 (You Only Look Once Version 5, real-time object detection algorithm) model. This lightweight YOLOv5 model has a feature pyramid structure in its backbone network and an attention mechanism module in its neck network. The YOLOv5 model is used for bird detection. It is optimized for small targets by introducing an improved feature pyramid structure and attention mechanism module. The loss function uses Focal Loss to alleviate sample imbalance.

[0036] Feature pyramid structure fusion formula: Let shallow feature map and deep feature maps Feature map after fusion for: , in, Dimensionality reduction for 1x1 convolutions; This is an upsampling operation.

[0037] Attention mechanism: Channel attention: , Spatial attention: , in, Input features; For average pooling output; This is for max-pooling output; , For fully connected weights; For ReLU functions; It is the sigmoid function; Loss function: , in, Predict probabilities for the model; Category weights; It is a regulating factor.

[0038] The AI ​​visual analysis module incorporates a Focal Loss function module during training to adjust sample weights.

[0039] In some implementations, the AI ​​visual analysis module integrates the DeepSORT multi-target tracking algorithm, which combines Kalman filtering for state estimation and utilizes an LSTM (Long Short-Term Memory) neural network to model historical motion sequences to predict flight path trends.

[0040] Specifically, the dynamic tracking algorithm uses YOLOv5 and DeepSORT (Deep Simple Online and Realtime Tracking, a multi-target tracking algorithm) to achieve multi-target tracking, introduces Kalman filtering to predict motion state, uses a lightweight ReID (Re-identification) network for appearance matching, and uses LSTM to predict trajectory.

[0041] Kalman Filter Prediction and Update: State prediction: , ; Status Update: , , ; in, This is the state transition matrix; The observation matrix; and For noise covariance; These are the observed values.

[0042] Trajectory prediction LSTM model: Let the input sequence be Output predicted trajectory : , ; It is in a hidden state; , These are the output layer parameters.

[0043] In some implementations, it also includes a power module, a communication module, and a control module. The power module provides power to the device; The control module is connected to the image acquisition module, the AI ​​visual analysis module, the laser emission module, the high-intensity light module, the sound wave bird deterrent module, and the communication module. It is used to receive and analyze data and dynamically generate control commands according to preset strategies, and to schedule the bird deterrent modules to work together.

[0044] This control module enables the modules to operate collaboratively, rather than independently, by running a set of cooperative control algorithms to achieve closed-loop optimization of perception, decision-making, and execution. The algorithm described below dynamically adjusts the bird-repelling strategy and energy allocation based on the output of the AI ​​vision and tracking modules.

[0045] Target threat assessment model: The system first performs a threat assessment on identified bird targets to determine the intensity and priority of intervention. The threat level T is determined by a combination of target distance, behavior, and species. The threat level T is determined by the following formula: ; in, The real-time estimated distance (in meters) between the target and the transmission line conductor; For the maximum effective range of the system, when At that time, the distance term is 0; Rate the behavior (stopping to build a nest, flying slowly, traversing quickly); Risk type rating (large birds of prey, medium birds of prey, small birds of prey); α, β, γ are weighting coefficients, and It is used to adjust the influence of different factors.

[0046] When the threat level T exceeds the preset threshold, the device will immediately start to drive away birds.

[0047] Laser irradiation coordinated control: The laser beam is precisely directed to the target location. Let the coordinates of the target identified by AI vision in the camera image coordinate system be (u,v). Through camera calibration and coordinate transformation, these coordinates are converted into the pitch angle (θ) and yaw angle (φ) of the gimbal.

[0048] Coordinate transformation formula: , , in, The pixel coordinates of the camera's optical center in the image; This is the equivalent focal length of the camera; , This is the mechanical mounting offset compensation angle between the laser module and the camera module.

[0049] Dynamic tracking and prediction: For moving targets, a predicted value needs to be added. The target position in the next frame is predicted using a Kalman filter. The actual pointing angle should be: , ; in, Proportional control factor (used to smooth gimbal movement and avoid excessive shaking).

[0050] Formula for dynamic adjustment of multimodal intervention parameters: The strength of the bird deterrent module is not fixed, but depends on the threat level. Distance to target Dynamic adjustments are made to achieve precise and energy-efficient removal.

[0051] laser power : ,in, This is the maximum safe power for the laser; , This is the adjustment coefficient.

[0052] To ensure safety, the closer the distance and the greater the threat, the higher the laser power should be.

[0053] Sound Boeing quantity : , in, This is a time-based sequence function. It is used to implement "random playback" of sound wave types, preventing birds from developing adaptations.

[0054] High-intensity light module : T, of which The readings are normalized for the ambient light sensor (0 for complete darkness, 1 for strong light).

[0055] It automatically reduces the brightness of the supplemental light when the daytime light is strong, and reaches its maximum brightness at night when the threat is high.

[0056] Energy management and dispatch strategies The device is powered by solar energy and requires intelligent management of energy consumption. The system defines an energy state index. : , , in, Remaining battery power; The battery is fully charged; is the current sunlight intensity; is the solar charging efficiency coefficient.

[0057] Adjust the system working mode according to Estate: Mode 1 (Estate > 0.7Estate > 0.7): Full-function mode, all algorithms and bird repellent modules operate at full capacity.

[0058] Mode 2 (0.3 < Estate ≤ 0.70.3 < Estate ≤ 0.7): Energy-saving mode, appropriately reduce the AI recognition frame rate, and preferentially use the laser module (because its energy consumption is lower than that of sound waves).

[0059] Mode 3 (Estate ≤ 0.3Estate ≤ 0.3): Standby mode, only maintain basic communication and key sensors running, and wait for charging.

[0060] In this embodiment, it further includes a power amplifier module, which is respectively connected to the control module and the sound wave bird repellent module, and is used to amplify the original audio electrical signal output by the control module to enhance the effect of the sound wave emitted by the sound wave bird repellent module.

[0061] Specifically, the power amplifier module is the power guarantee for the sound wave bird repellent module to effectively play a deterrent role. It receives the weak original audio electrical signal sent by the control module, and greatly amplifies its power and voltage to drive the high-power sound wave bird repellent module. Ensure that the finally generated deterrent sound wave or ultrasonic wave has sufficient loudness, intensity and propagation distance, so as to form an effective auditory deterrence range in the vast transmission line environment, which directly determines the final effect and coverage of sound wave bird repellent.

[0062] In this embodiment, it further includes an installation frame, and the image acquisition module, the laser emission module, the strong light module, the power supply module, the control module, the sound wave bird repellent module, and the communication module are all integrated on the installation frame.

[0063] Specifically, as Figure 1 shown, this communication module is the antenna module 1, the image acquisition module 7, the laser emission module 10, the strong light module 4, the power supply module 11, the control module 5, the sound wave bird repellent module 2, where the sound wave bird repellent module 2 is set as a circle to transmit sound wave signals in all directions to enhance the sound wave deterrence effect. A key module 3 and a laser high-voltage board 6 are also provided, as well as a controller module 9.

[0064] Embodiment 2 This embodiment discloses a bird repellent method for an integrated intelligent laser bird repellent device based on AI vision, as Figure 2 shown, including: The image acquisition module collects image data around the power transmission line and transmits the image data to the AI ​​visual analysis module. Based on the image data, the AI ​​visual analysis module identifies the number, trajectory, and dwelling behavior of bird targets, generating AI visual analysis data. Based on AI visual analysis data instructions, the laser emission module projects a laser point near the target bird or its flight path, causing the target bird to actively fly away from the protected airspace. The image acquisition module continuously monitors bird targets. When the flock of birds is dense, the repelling efficiency of the high-intensity light module is insufficient, or when targeting specific birds that are not sensitive to lasers, the high-intensity light module is instructed to be triggered to emit high-intensity light towards the target birds until the bird repelling task is completed.

[0065] It also emits sound waves to target birds through a sound wave bird deterrent module, using specific frequency ultrasonic waves and / or deterrent sound waves in conjunction with the laser emission module and the high-intensity light module to drive away the target birds.

[0066] Specifically, The image acquisition module collects images of the surrounding environment in real time and transmits the collected image data to the AI ​​visual analysis module.

[0067] In the AI ​​visual analysis module, a lightweight target detection model is used to identify birds in images. The lightweight target detection model is based on the YOLOv5 architecture. A feature pyramid structure is set in the backbone network to enhance the multi-scale feature extraction capability. An attention mechanism module is introduced in the neck network to improve the perception accuracy of key areas. The FocalLoss function is combined to dynamically adjust the weights of positive and negative samples during the model training process to alleviate the class imbalance problem, thereby improving the detection accuracy of small target birds.

[0068] The DeepSORT multi-target tracking algorithm is used to perform continuous frame correlation processing on the received detection results to achieve stable tracking of multiple individual birds. Kalman filtering is used to predict and update the motion state of each bird, and LSTM neural network is combined to model the historical trajectory sequence to predict its future flight path trend, generating comprehensive analysis data including position, speed, number and behavior type.

[0069] The threat level of birds is determined based on the analysis of data. The threat level is comprehensively assessed based on bird species, flight direction, approach speed, and flock density.

[0070] The control module receives the analysis data and dynamically generates control commands based on the preset bird deterrence strategy, scheduling one or more modules in the multimodal bird deterrence execution unit to work together.

[0071] When the threat level is low to medium and only a few birds enter the warning area, the laser emission module is activated. The green semiconductor laser is driven by a two-dimensional servo gimbal to project a 532nm visible laser into the space near the bird or in front of its predicted flight path after passing through an optical collimation component. This induces the bird to change its flight direction and fly away from the target area. The two-dimensional servo gimbal supports ±180° horizontal rotation and ±90° pitch rotation to achieve omnidirectional coverage.

[0072] When the density of birds exceeds the set threshold or a specific high-risk bird species is detected, the high-intensity light module is activated, which emits high-intensity pulsed flashing light from a high-brightness LED array. The flashing frequency varies randomly within the range of 3Hz to 8Hz, creating visual interference and serving as a supplementary means to laser bird deterrence to enhance the deterrence effect.

[0073] When ambient lighting conditions are poor or when working at night, activate the sound wave bird deterrence module and select the output mode according to the bird type: output ultrasonic signals with a frequency range of 20kHz to 25kHz to the birds, or play deterrent audio signals such as simulated raptor calls or gunshots through an audio speaker to further enhance the bird deterrence response through auditory stimulation.

[0074] The power module provides a stable power supply for the entire system. The power module includes a solar photovoltaic panel and a lithium iron phosphate battery. The solar photovoltaic panel is used to collect solar energy during the day and store it in the lithium iron phosphate battery to achieve all-weather autonomous operation. The communication module establishes a connection with the remote management platform through 4G LTE or LoRa wireless means, and reports the device's operating status, alarm events and environmental data on a regular basis. It also receives control parameter updates and strategy adjustment instructions from the platform to realize remote monitoring and intelligent operation and maintenance.

Claims

1. A comprehensive intelligent laser bird deterrent device based on AI vision, characterized in that, include The image acquisition module is used to acquire environmental images and transmit them to the AI ​​visual analysis module. The AI ​​visual analysis module, based on a target detection model and multi-target tracking algorithm, identifies, tracks, and predicts the flight trajectories of birds in images, generating analytical data including location, speed, number, and behavior type. Laser emitting module, used to project laser points. High-intensity light module, used to emit strong light; The power module is used to provide power for the operation of the device; A gimbal mechanism, wherein the laser emitting module and the high-intensity light module are mounted on the gimbal mechanism, and are used to drive the laser emitting module and the high-intensity light module to point towards the target bird; The control module is used to control the gimbal mechanism to aim at the target bird based on the analysis data. It is also used to calculate the repulsion efficiency and determine whether the repulsion efficiency of the laser emission module meets the repulsion threshold. If it does, the laser emission module is controlled to operate. If not, the high-intensity light module and the laser emission module are controlled to operate simultaneously.

2. The integrated intelligent laser bird deterrent device based on AI vision according to claim 1, characterized in that, It also includes a sonic bird deterrent module, which drives away target birds by outputting ultrasonic signals and / or audio signals that simulate the calls of natural enemies.

3. The integrated intelligent laser bird deterrent device based on AI vision according to claim 1, characterized in that, The AI ​​visual analysis module employs a lightweight YOLOv5 model, which features a feature pyramid structure in its backbone network and an attention mechanism module in its neck network. The AI ​​visual analysis module incorporates a Focal Loss function module during training to adjust sample weights.

4. The integrated intelligent laser bird deterrent device based on AI vision according to claim 3, characterized in that, The AI ​​visual analysis module integrates the DeepSORT multi-target tracking algorithm, which combines Kalman filtering for state estimation and uses an LSTM neural network to model historical motion sequences to predict the flight path trend of target birds.

5. The integrated intelligent laser bird deterrent device based on AI vision according to claim 2, characterized in that, It also includes a communication module; The communication module is used to establish a connection between the device and the remote management platform; The control module connects the image acquisition module, the AI ​​visual analysis module, the laser emission module, the high-intensity light module, the sound wave bird deterrent module, and the communication module. It is used to receive data and dynamically generate control commands according to a preset strategy, and to schedule the bird deterrent modules to work together.

6. The integrated intelligent laser bird deterrent device based on AI vision according to claim 5, characterized in that, It also includes a power amplifier module, which is electrically connected to the control module and the sound wave bird deterrent module respectively, and is used to amplify the original audio electrical signal output by the control module to enhance the sound wave effect emitted by the sound wave bird deterrent module.

7. The integrated intelligent laser bird deterrent device based on AI vision according to claim 5, characterized in that, It also includes an installation frame, on which the image acquisition module, the laser emission module, the high-intensity light module, the power supply module, the control module, the sound wave bird deterrent module, and the communication module are all integrated.

8. A bird-repelling method based on the AI ​​vision-based integrated intelligent laser bird-repelling device according to any one of claims 1 to 7, characterized in that, include: The image acquisition module collects image data around the power transmission line and transmits the image data to the AI ​​visual analysis module. Based on the image data, the AI ​​visual analysis module identifies and obtains the number, trajectory, and dwelling behavior of bird targets, generating AI visual analysis data. Based on AI visual analysis data instructions, the laser emission module projects a laser point near the target bird or its flight path, causing the target bird to actively fly away from the protected airspace. The image acquisition module continuously monitors bird targets. When the flock of birds is dense, the repelling efficiency of the laser emission module is insufficient, or when targeting specific birds that are not sensitive to lasers, the high-intensity light module is triggered to emit high-intensity light towards the target birds until the bird repelling task is completed.

9. The bird-repelling method according to claim 8, characterized in that, It also emits sound waves to target birds through a sound wave bird-repelling module, using ultrasonic signals and / or simulated predator calls in conjunction with the laser emission module and the high-intensity light module to drive away the target birds.

10. An electronic device, characterized in that, The invention includes a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the computer program, implements the method as described in any one of claims 8 or 9.

Images