A Differentiated Cooperative Removal System and Method for Small Animals in Substations

Abstract

This invention proposes a differentiated collaborative animal repelling system and method for substations. The system includes a front-end detection module that monitors the substation's monitoring area, collecting the target's location, movement trajectory, morphological features, sound characteristics, thermal imaging information, and image information. An AI fusion control module acquires the target's location, movement trajectory, and thermal imaging information, sequentially determining whether the target has invaded, whether the invading target is an animal, and the initial target classification result. After determining that an animal has invaded the substation area, a preliminary repelling strategy is determined based on the initial classification result. The front-end detection module then acquires the animal target's morphological features, sound characteristics, and image information again, and based on these, determines the animal target's species type. Furthermore, based on the animal target's species type and the species repelling strategy database containing the animal target's visual sensitivity, acoustic sensitivity, and auditory fear sources, a final repelling strategy is determined for the current animal target. The repelling execution module then repelles the animal target according to the preliminary and final repelling strategies determined by the AI ​​fusion control module.

Description

Technical Field

[0001] This invention relates to the field of power equipment protection technology, and in particular to a differentiated and collaborative system and method for repelling small animals in substations. Background Technology

[0002] As the core hub of the power system, the safe and stable operation of substations is directly related to the reliability of regional power supply. Open-type substations expose critical equipment such as main transformers, busbars, and insulators to the natural environment, making them vulnerable to invasion by small animals such as rats, cats, birds, and weasels. These animals can cause serious incidents such as short circuits, flashovers, and tripping by gnawing on cables, building nests, and polluting the equipment with their feces, posing a serious threat to power grid safety.

[0003] Currently, common small animal protection measures in substations mainly include physical isolation (such as cat spikes, rat traps, and electronic fences) and single repelling technologies (such as green laser bird repellers, fixed-frequency ultrasonic rodent repellers, and bird alarms). However, these existing technologies have many drawbacks: First, the protection is not targeted enough. Different small animals have significantly different physiological characteristics. For example, sparrows are sensitive to green lasers, while doves are basically unresponsive to them; rats are sensitive to high-frequency ultrasound, while cats are more likely to be frightened by certain sound effects. Existing repelling devices mostly use a single technology or fixed parameter output, which cannot be adapted to the sensitivity characteristics of different species, resulting in poor repelling effects for some small animals.

[0004] Secondly, small animals are prone to adaptation. When fixed patterns of bird deterrence are used for a long time, small animals will gradually adapt to the deterrence signals. For example, after a year of deployment, green laser bird deterrents no longer have a significant deterrent effect on sparrows, and fixed-frequency ultrasonic devices become ineffective against rodents after 3-5 days, which greatly reduces the long-term effectiveness of the protection system.

[0005] Third, the protection range is limited and there are blind spots. The protective area of ​​physical isolation measures is small and it is difficult to cover the entire outdoor equipment area of ​​the substation; the sensing range and effective distance of traditional bird deterrent devices are limited. For example, bird alarms can only cover a small surrounding area and cannot achieve full-area monitoring and deterrence without blind spots.

[0006] Fourth, there is a lack of coordination and linkage mechanisms. Most existing protection systems operate independently, and there is no effective linkage between monitoring equipment and de-intrusion equipment. It is impossible to dynamically adjust the de-intrusion strategy according to the real-time status of the intrusion target, and it is impossible to synchronize alarm information to the central control station in a timely manner, which makes it impossible for maintenance personnel to take proactive measures.

[0007] To address this problem, the present invention provides a differentiated and collaborative animal repelling system and method for substations, thereby solving at least one technical problem existing in the prior art. Summary of the Invention

[0008] To address the problems existing in the prior art, this invention innovatively proposes a differentiated collaborative expulsion system and method for small animals in substations, effectively solving the problem of low expulsion reliability in substations caused by existing technologies and effectively improving the reliability of expulsion in substations.

[0009] The first aspect of this invention provides a differentiated collaborative animal removal system for substations, comprising a front-end detection module, a removal execution module, and an AI fusion control module. The front-end detection module monitors the substation monitoring area and collects the target's location, movement trajectory, morphological features, sound features, thermal imaging information, and image information. The AI ​​fusion control module acquires the target's location, movement trajectory, and thermal imaging information, and sequentially determines whether the target has invaded, whether the invading target is an animal, and the initial classification result of the target. After determining that an animal target has invaded the substation area, it determines a preliminary removal strategy based on the initial classification result. The front-end detection module then acquires the animal target's morphological features, sound features, and image information again, and based on these features, determines the species type of the animal target. Based on the species type, the visual sensitivity, acoustic sensitivity, and sound fear sources of the animal target's species type in the species removal strategy database, it determines the final removal strategy corresponding to the current animal target. The removal execution module removes the animal target according to the preliminary and final removal strategies determined by the AI ​​fusion control module.

[0010] Optionally, it also includes a linkage module, which is used to connect the substation data monitoring system and the central control station data network system to realize data linkage and remote alarm.

[0011] Optionally, the front-end detection module includes a radar unit, a thermal imaging unit, a sound acquisition unit, and a camera. The radar unit is used to monitor the substation monitoring area and acquire the location and movement trajectory of the target. The thermal imaging unit is used to acquire a thermal image of the target after the radar unit detects a moving target. The sound acquisition unit is used to acquire the sound information of the animal target after the thermal image acquired by the thermal imaging unit identifies the animal target. The camera is used to acquire the image information of the animal target after the thermal image acquired by the thermal imaging unit identifies the animal target.

[0012] Optionally, the deportation execution module includes an ultrasonic deportation unit, an intelligent voice deportation unit, a green laser deportation unit, and a high-pressure blasting deportation unit. The ultrasonic deportation unit outputs ultrasonic waves in different frequency ranges based on the classification results of different animal targets and the acoustic sensitivity characteristics of different species. The intelligent voice deportation unit has built-in predator calls and cries of other species corresponding to different species, and outputs predator calls and cries of other species according to the predator characteristics and sound pressure sensitivity range of different species. The green laser deportation unit outputs a laser beam based on the classification results of different animal targets and the visual sensitivity characteristics of different species. The high-pressure blasting deportation unit outputs a blasting sound and a momentary flash based on the classification results of different animal targets and the blasting sound sensitivity characteristics of different species.

[0013] Furthermore, after identifying the animal target's intrusion into the substation area, the initial deportation strategy based on the initial target classification results is as follows: Based on the hot spot temperature, hot spot distribution area, and movement trajectory of the animal target in the thermal imaging image, the animal target is determined to be a bird or a mammal. If the animal target is a bird, the initial deterrent strategy is to output a green laser, and the green laser output method is a fan-shaped scan. If the animal target is a mammal, the corresponding initial deterrent strategy is to output random low-frequency ultrasound.

[0014] Optionally, the species repulsion strategy library includes fear source parameters, priority repulsion technology combinations, parameter optimization rules, and anti-adaptive rotation schemes for each species type; the fear source parameters are used to describe the sensitivity of different species to ultrasound, sound effects, lasers, strong light, and loud noises; the priority repulsion technology combinations are used to determine the combination of sound and light repulsion strategies based on the species' sensitivity characteristics; the parameter optimization rules are used to describe the repulsion parameter adjustment methods under different environmental conditions and target behavior states, and to optimize the repulsion parameters based on the historical repulsion strategy effects; the anti-adaptive rotation scheme sets a rotation cycle and periodically updates the combination of repulsion units in the repulsion strategy; wherein, the repulsion parameters include the scanning method, combination method, implementation duration, and number of implementations per unit time for different repulsion units in the repulsion execution module.

[0015] Furthermore, based on the species type of the animal target, and the visual sensitivity, acoustic sensitivity, and auditory fear sources of the animal target's species type in the species repulsion strategy database, the specific final repulsion strategy corresponding to the current animal target is determined, including: The substation area is divided into protection zones of different risk levels according to equipment type or distance from the substation; Determine the corresponding sound fear source, visual sensitivity, sound wave sensitivity, and popping sound sensitivity based on the species type of the animal target; Based on the protection zone level to which the animal target's current location belongs, determine a deportation strategy with the same level as the current protection zone. Among these strategies, the deportation strategy level for a single deportation unit is lower than the deportation strategy level for different units. Furthermore, the more deportation units there are, the higher the deportation strategy level. Within the same level, the deportation priority for sound-based fear sources, visual sensitivity, sound wave sensitivity, and loud noise sensitivity decreases in that order. The number of deportation strategy levels is greater than the number of protection zone levels.

[0016] Furthermore, based on the species type of the animal target, and the visual sensitivity, acoustic sensitivity, and auditory fear sources of the animal target's species type in the species repulsion strategy database, the final repulsion strategy corresponding to the current animal target specifically includes: The effectiveness of the current expulsion strategy is judged. If the current expulsion strategy is ineffective, the current strategy is upgraded step by step. If the current expulsion strategy is effective, the current expulsion strategy is evaluated based on the animal target's response time, stay time, number of animals expelled, and expulsion success rate. When expulsion strategies of the same level are in effect, the expulsion strategy with the higher evaluation score is given priority. If the animal target is still not removed even with the highest level of removal strategy, or if the animal identification result is human, then the current removal strategy will be suspended and an alert will be sent to the maintenance personnel.

[0017] Optionally, the adaptive rotation scheme further includes adjusting the rotation cycle based on the historical effectiveness of different expulsion strategies at the same level, specifically including: Obtain the historical effects of different expulsion strategies at the same level within the current rotation period. If the historical effect of the expulsion strategy at the same level within the current rotation period is greater than the preset effect threshold, increase the execution time of the expulsion strategy or increase the expulsion parameters in the expulsion strategy according to the preset time step. If the historical effect of the expulsion strategy at the same level within the current rotation period is not greater than the preset effect threshold, decrease the execution time of the expulsion strategy or decrease the expulsion parameters in the expulsion strategy according to the preset time step. The historical effect is the number of animals that leave when the expulsion strategy is executed, or the reciprocal of the animal's stay time, or the response time of the animal leaving when the expulsion strategy is executed.

[0018] The second aspect of this invention also provides a differentiated cooperative method for repelling small animals in substations, based on a differentiated cooperative system for repelling small animals in substations provided in the first aspect of this invention, comprising: The front-end detection module monitors the substation's monitoring area, collecting the target's location, movement trajectory, and thermal imaging information. The AI ​​fusion control module acquires the target's location, movement trajectory, and thermal imaging information, and sequentially determines whether the target has invaded, whether the invading target is an animal, and the initial classification result of the target; and after determining that an animal target has invaded the substation area, it determines the initial deportation strategy based on the initial classification result of the target. The expulsion execution module is used to initially expel the animal target based on the preliminary expulsion strategy corresponding to the current animal target determined by the AI ​​fusion control module. The front-end detection module collects morphological features, sound features, and image information of animal targets; The AI ​​fusion control module is used to acquire the morphological features, sound features, and image information of animal targets. Based on the morphological features, sound features, and image information of animal targets, it determines the species type to which the animal targets belong. Based on the species type to which the animal targets belong, the visual sensitivity, sound wave sensitivity, and sound fear sources of the species type to which the animal targets belong in the species repulsion strategy library, it determines the final repulsion strategy corresponding to the current animal target. The expulsion execution module is used to expel the animal target according to the final expulsion strategy corresponding to the current animal target determined by the AI ​​fusion control module.

[0019] The technical solution adopted in this invention has the following technical effects: 1. In the technical solution of this invention, the front-end detection module collects the target's location, movement trajectory, morphological features, sound features, thermal imaging information, and image information respectively; the AI ​​fusion control module is used to acquire the target's location, movement trajectory, and thermal imaging information, and sequentially determine whether the target has invaded, whether the invading target is an animal, and the initial classification result of the target; after determining that the animal target has invaded the substation area, a preliminary deportation strategy is determined based on the initial classification result of the target; the front-end detection module again acquires the animal target's morphological features, sound features, and image information, and based on the animal target's morphological features, sound features, and image information, determines the species type of the animal target, and determines the final deportation strategy corresponding to the current animal target based on the species type of the animal target, the visual sensitivity, sound wave sensitivity, and sound fear source of the animal target's species type in the species deportation strategy library; the deportation execution module is used to deport the target according to the preliminary deportation strategy and the final deportation strategy corresponding to the current target determined by the AI ​​fusion control module, effectively solving the problem of low reliability of substation deportation caused by existing technology, and effectively improving the reliability of substation deportation.

[0020] 2. The technical solution of the present invention also includes a linkage module, which is used to connect the substation data monitoring system and the central control station data network system to realize data linkage and remote alarm.

[0021] 3. The front-end detection module in this invention includes a radar unit, a thermal imaging unit, a sound acquisition unit, and a camera. The radar unit is used to monitor the substation monitoring area and collect the location and movement trajectory of the target. The thermal imaging unit is used to collect the thermal image of the target after the radar unit detects a moving target. The sound acquisition unit is used to collect the sound information of the animal target after the thermal image collected by the thermal imaging unit is identified as an animal target. The camera is used to collect the image information of the animal target after the thermal image collected by the thermal imaging unit is identified as an animal target. By using the radar unit and the thermal imaging unit to classify whether an intrusion has occurred and whether an intrusion has occurred, the sound information and image information of the animal target are obtained, which improves the accuracy of small animal target identification and thus ensures the reliability of driving away different types of small animals in the substation.

[0022] 4. The repulsion execution module described in the technical solution of this invention includes an ultrasonic repulsion unit, an intelligent voice repulsion unit, a green laser repulsion unit, and a high-pressure explosion repulsion unit; the species repulsion strategy library includes fear source parameters for each species type, priority repulsion technology combinations, parameter optimization rules, and anti-adaptation rotation schemes; it can effectively combine multiple repulsion units according to different types of small animals; it achieves differentiated repulsion of different types of small animals, and at the same time, through the strategy rotation mechanism, it avoids small animals from developing adaptation, thereby improving the long-term effectiveness and reliability of the protection system.

[0023] 5. In the technical solution of this invention, the animal target is determined to be a bird or a mammal based on the hot spot temperature, hot spot distribution area, and movement trajectory of the animal target in the thermal imaging image. If the animal target is a bird, the corresponding preliminary deterrence strategy is to output a green laser, and the output mode of the green laser is fan-shaped scanning. If the animal target is a mammal, the corresponding preliminary deterrence strategy is to output random low-frequency ultrasound. Different preliminary deterrence strategies can be executed according to the initial classification results of the animal target, ensuring the reliability of safety protection in the substation area.

[0024] 6. In the technical solution of this invention, the substation area is divided into protection zones of different risk levels according to the equipment type or the distance from the substation; the corresponding sound fear source, visual sensitivity, sound wave sensitivity, and blast sound sensitivity are determined according to the species type of the animal target; and a repulsion strategy of the same level as the current protection zone level is determined according to the protection zone level of the current location of the animal target. This allows the corresponding level of repulsion strategy to be determined according to the equipment type and the protection zone level of the animal target within the substation area, which not only achieves effective repulsion of the animal target, but also reduces harm to the animal and energy consumption.

[0025] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 This is a schematic diagram of the system structure in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram showing the distribution of the front-end detection module and the expulsion execution module in the system of Embodiment 1 of the present invention; Figure 3 This is a schematic diagram of the detection and early warning distribution of the front-end detection module in the system of Embodiment 1 of the present invention; Figure 4 This is a flowchart illustrating the method of Embodiment 2 in the present invention. Detailed Implementation

[0028] To clearly illustrate the technical features of this solution, the invention will be described in detail below through specific embodiments and in conjunction with the accompanying drawings. The following disclosure provides many different embodiments or examples for implementing different structures of the invention. To simplify the disclosure of the invention, components and arrangements of specific examples are described below. Furthermore, reference numerals and / or letters may be repeated in different examples. This repetition is for simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or arrangements discussed. It should be noted that the components illustrated in the drawings are not necessarily drawn to scale. Descriptions of well-known components, processing techniques, and processes are omitted in this invention to avoid unnecessarily limiting the invention.

[0029] Example 1 like Figure 1As shown, this invention provides a substation small animal differentiated collaborative repelling system, including a front-end detection module, a repelling execution module, and an AI fusion control module. The front-end detection module is used to monitor the substation monitoring area, collecting the target's location, movement trajectory, morphological features, sound features, thermal imaging information, and image information. The AI ​​fusion control module is used to acquire the target's location, movement trajectory, and thermal imaging information, sequentially determining whether the target has invaded, whether the invading target is an animal, and the initial classification result of the target. After determining that an animal target has invaded the substation area, a preliminary repelling strategy is determined based on the initial classification result. The front-end detection module then acquires the animal target's morphological features, sound features, and image information again, and based on these features, determines the species type of the animal target. Based on the animal target's species type, and the visual sensitivity, acoustic sensitivity, and sound fear sources of the animal target's species type in the species repelling strategy library, the final repelling strategy corresponding to the current animal target is determined. The repelling execution module (… Figure 2 The integrated defense system is used to drive away animal targets based on the initial and final drive-away strategies determined by the AI ​​fusion control module.

[0030] Among them, such as Figures 2-3 As shown, the front-end detection module is used to achieve all-weather, no-blind-spot detection of small animals within the substation monitoring area, collecting the target's location, movement trajectory, morphological characteristics, and thermal imaging information to provide data support for subsequent identification and removal. This front-end detection module includes a radar unit, a thermal imaging unit, a sound acquisition unit, and a camera.

[0031] The radar unit employs advanced radar modulation systems and radar imaging signal processing technology, achieving rapid scanning based on electronic scanning principles. Scanning speeds reach millisecond levels, far superior to traditional mechanical scanning radars. Its key parameters are as follows: detection range of 10m-200m, covering the entire outdoor equipment area and surrounding protective zone of the substation; angular coverage of 120°×±30°, achieving 360° full-range coverage through multi-unit deployment; distance resolution ≤0.5m, velocity resolution ≤0.1m / s, capable of accurately capturing the subtle movements of small animals; supports simultaneous tracking of ≥50 targets, meeting monitoring needs in cluster intrusion scenarios; refresh rate of 10Hz, ensuring real-time performance. The unit utilizes wide-temperature technology, operating within a temperature range of -40℃ to 70℃, unaffected by rain, fog, sandstorms, or other harsh weather conditions, and possesses strong anti-electromagnetic interference capabilities, enabling stable operation in the complex electromagnetic environment of substations.

[0032] Thermal imaging is a technology that detects the infrared radiation (heat) emitted by objects and converts it into a visual image. Deploying thermal imaging in substations allows for real-time monitoring of heat changes in boundary areas, timely detection of small animals such as cats and mice climbing over walls and intruding, and coordinated with defense equipment for early warning and response. The thermal imaging unit works in conjunction with the radar unit; when the radar detects a moving target, it immediately triggers the thermal imaging unit to turn towards the target area and acquire a thermal image. The thermal imaging unit captures changes in the heat of a target by detecting its infrared radiation, effectively identifying nocturnal animals such as cats and mice, and can operate normally even in the absence of light.

[0033] The camera supports small animal tracking and capture, obtaining clear images of the animal targets. Specifically, high-definition visible light PTZ cameras with supplementary lighting can be deployed around key equipment points in substations (such as main transformers and cable joints). Normally, it does not operate; only when the thermal imaging unit detects an animal target in the substation area and provides an initial classification signal does the AI ​​fusion control module guide its turning, zooming, and supplementary lighting for animal type identification. It can extract information such as the target's body shape, outline, and morphological features, complementing radar and sound data to improve subsequent identification accuracy.

[0034] The sound acquisition unit can be a microphone or sound collector, capable of capturing the sounds emitted by small animals. Specifically, sound acquisition units can be deployed around key equipment points in substations (such as main transformers and cable joints). Normally, it does not operate; only when the thermal imaging unit detects an animal target in the substation area and provides an initial classification signal does the AI ​​fusion control module guide it to acquire sound. This process extracts the animal target's vocal characteristics (such as timbre, audio frequency, and tone), complementing radar and image data to improve subsequent recognition accuracy.

[0035] The AI ​​fusion control module is used to acquire the target's location, movement trajectory, and thermal imaging information, and sequentially determine whether the target has invaded, whether the invading target is an animal, and the initial classification result of the target. After determining that an animal target has invaded the substation area, it determines the initial deportation strategy based on the initial classification result. The front-end detection module then acquires the animal target's morphological characteristics, sound characteristics, and image information, and determines the species type of the animal target based on these characteristics. Based on the species type of the animal target and the visual sensitivity, sound sensitivity, and sound fear source of the animal target's species type in the species deportation strategy database, it determines the final deportation strategy corresponding to the current animal target.

[0036] Specifically, 3D information on intruding foreign objects provided by radar units, sound acquisition units, thermal imaging units, and cameras can be used to conduct image recognition and video tracking, providing a basis for decision-making in subsequent handling. An initial classification knowledge base is built based on information such as the location, movement trajectory, and thermal imaging information of small animals captured by radar and thermal imaging units to determine whether the intruding target is a bird or a mammal. A reclassification knowledge base is then built based on information such as the location, movement trajectory, sound characteristics, and image information of small animals provided by radar units, sound acquisition units, and cameras to determine the species type of the animal target. Data is continuously optimized through AI training to enrich the information base, improve recognition accuracy, and implement targeted protective measures to enhance effectiveness.

[0037] The AI ​​fusion control module includes an AI fusion processing host, which, based on a high-performance processor and deep learning algorithms, possesses powerful data processing and analysis capabilities. It receives radar data, image data, and target acoustic feature information transmitted from the front-end detection module. Through multi-source data fusion technology, it accurately extracts parameters such as the target's position, velocity, shape, and acoustic features (which may also include thermal feature values ​​and distribution), providing support for identification and decision-making. Simultaneously, the host can receive feedback data from the expulsion execution module in real time, dynamically evaluate the expulsion effect, and adjust the expulsion strategy based on the evaluation results.

[0038] The small animal identification model is built using deep learning algorithms (such as Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs), and optimized through extensive sample training. The training sample library contains morphological features (outlines), behavioral patterns (aerial or ground-based), thermal imaging data (thermal imaging temperature values ​​and distribution), and sound feature data (timbre, audio, tone, etc.) corresponding to more than 25 common substation intrusion animals, including sparrows, magpies, doves, cats, mice, weasels, and falcons. These features are used as input data, and the animal species type is used as the output data for training the animal identification model. During training, transfer learning techniques are employed to improve the accuracy of identifying species in small samples, and data augmentation techniques are used to expand sample diversity. Testing shows that the model achieves an accuracy rate of ≥98% for identifying common small animals, quickly distinguishing different species from non-target objects (such as leaves and dust), and effectively reducing the false alarm rate.

[0039] The expulsion execution module includes an ultrasonic expulsion unit, an intelligent voice expulsion unit, a green laser expulsion unit, and a high-pressure explosion expulsion unit.

[0040] The ultrasonic repelling unit outputs ultrasonic waves in different frequency ranges based on the auditory sensitivity characteristics of different species. Specifically, the ultrasonic repelling unit utilizes the auditory sensitivities of different small animals to achieve wide-range frequency conversion output of ultrasonic waves. The main operating frequency is set above 20kHz to minimize potential impact on maintenance personnel, ensuring effective animal repellency while avoiding disruption to substation maintenance staff. By preset sensitive frequency ranges for different species, targeted ultrasonic waves can be output precisely: 25-35kHz for mice (sensitive range 25-55kHz), 20-25kHz for cats (sensitive range 20-50kHz), 20-25kHz for weasels (sensitive range 20-60kHz), 20-22kHz for sparrows (sensitive range 18-22kHz), and 20-22kHz for magpies (sensitive range 18-22kHz). Furthermore, the unit supports dynamic frequency switching to prevent animals from developing adaptive frequencies. The sound wave sensitivity data of different animals can be obtained from existing scientific research and publicly available literature, such as by searching papers in the fields of animal behavior, auditory physiology, pest control in agriculture, forestry, animal husbandry and fishery, bird strike control at airports, and pest control in cities; they can also be obtained from guidance documents or research reports on the habits and control technologies of specific animals (such as rodents and birds) issued by agricultural departments, forestry departments, and civil aviation management departments; or they can be obtained through behavioral experiments. The method of acquisition and the specific range of data are not limited in this invention.

[0041] Humans can perceive sound waves in the range of 20Hz–20kHz. However, the hearing range varies greatly among small animals. Some birds are extremely sensitive to ultrasound waves of 16–25kHz; these high-frequency waves strongly stimulate their auditory nerves, causing anxiety and discomfort, forcing them to actively flee the sound source area. Cats are extremely sensitive to ultrasound waves of 20–25kHz. This device can achieve wide-range frequency conversion for separation within the 20kHz–25kHz range. To prevent the ultrasound waves from affecting human activities, the sound wave range is set to 20kHz–25kHz.

[0042] Small animals are only sensitive to specific sound wave frequencies. The effect of random ultrasound playback on repelling small animals is also random. Furthermore, small animals become more tolerant of sound wave equipment that runs for a long time, and the repelling effect gradually weakens. At the same time, high-intensity, long-term sound wave equipment can cause irreversible damage to small animals, which is not in line with the concept of eco-friendliness.

[0043] The intelligent voice deterrence unit has built-in predator calls and cacophony sounds of the same species corresponding to different types of species. Based on the predator characteristics and sound pressure sensitivity range of different types of species, it outputs predator calls and cacophony sounds of the same species accordingly. Specifically, the intelligent voice-activated decoy unit has a built-in library of over 200 animal predator calls, cries of other animals, and loud noise samples, creating a rich sound effects library. Based on the characteristics of different small animals' predators and their sound pressure sensitivity ranges, it outputs targeted sound effects: for sparrows (whose predators are large birds of prey, with a sound pressure sensitivity range of 80-100dB), it plays varying raptor calls, with sound pressure controlled at 80-100dB and dynamically changing; for magpies (whose predators are large birds of prey, with a sound pressure sensitivity range of 90-100dB), it plays eagle calls, accompanied by random popping sounds, with a sound effect frequency of 200-500Hz; for doves (whose predators are large birds of prey and owls, with a sound pressure sensitivity range of 100-120dB), it plays owl calls, with a sound pressure of 1... 00-120dB; For cats (predators are dogs and snakes, sound pressure sensitivity range 85-95dB), play dog ​​barking + 8-10kHz snake hissing sound, with increasing sound pressure gradient; For mice (predators are cats and owls, sound pressure sensitivity range >85dB), play owl calls, sound pressure >85dB; For weasels (predators are dogs, sound pressure sensitivity range 80-90dB), play dog ​​barking + loud noise, sound pressure 80-90dB; For falcons (predators are large eagles, sound pressure sensitivity range >110dB), play osprey calls, sound pressure >120dB. This unit supports random switching and combination of sound effects to enhance the unpredictability of the deterrent effect. To avoid noise pollution, the maximum sound pressure level does not exceed 130dB.

[0044] The green laser repelling unit outputs laser beams based on the visual sensitivity characteristics of different species. Specifically, the green laser deterrence unit outputs a green laser beam with a wavelength of 520nm-532nm. This wavelength has a strong visual stimulating effect on birds such as sparrows and magpies, which perceive it as a physical threat. Through beam expansion and collimation techniques, the laser beam can form a uniform light band, covering the target area. This unit supports randomized adjustment of the laser path, allowing for various modes such as straight-line scanning, fan-shaped scanning, and irregular scanning, preventing small animals from adapting to fixed scanning trajectories. For highly sensitive species such as sparrows and magpies, laser deterrence is the preferred technology; for less sensitive species such as doves, mice, weasels, and falcons, it is used only as an auxiliary deterrence method in conjunction with other units.

[0045] The high-pressure detonation unit outputs detonation sounds and flashes of light based on the detonation sensitivity characteristics of different species.

[0046] Specifically, the high-pressure detonation deterrent unit, referencing the principle of airport gas cannons, stores electrical energy electronically and releases it instantaneously, breaking down the air to produce a strong detonation sound and flash of light greater than 130 dB. Table 1 below shows the sensitivity of different small animals to the flash and detonation sound.

[0047] Table 1. Statistical table of typical animal sensitivities to flashing lights and popping sounds.

[0048] Research (based on a comprehensive analysis of publicly available animal behavior literature and industry experience reports on pest control, combined with preliminary field observations) shows that small birds exhibit a strong sense of panic from flashing lights and loud noises, causing them to flee quickly, with a long adaptation period. Larger birds show higher adaptability to flashing lights and loud noises, but the repelling effect is generally moderate. Reptiles are highly sensitive to flashing lights and loud noises, but their adaptation time is shorter. This high-pressure loud noise repelling unit is significantly effective in repelling species such as sparrows, magpies, mice, and weasels, as its flashing lights and loud noises can induce a strong panic response in these animals, and the adaptation period is long. For moderately sensitive species such as cats and falcons, the triggering frequency should be controlled to avoid overstimulation. This unit is usually used in conjunction with other repelling units to enhance the repelling effect.

[0049] Specifically, after identifying animal intrusion targets in the substation area, the initial deportation strategy based on the initial target classification results is as follows: Based on the hot spot temperature, hot spot distribution area, and movement trajectory of the animal target in the thermal imaging image, the animal target is determined to be a bird or a mammal. Specifically, based on the heat spot temperature, birds (high body temperature, ~40℃) and mammals* (low body temperature, ~37℃) can be roughly distinguished. Based on the heat spot area, small targets (birds) and medium-sized targets (cats, dogs, foxes) can be distinguished. Combining the animal target trajectory can help determine whether it is "flying or in the air" (high probability of birds) or "sneaking / climbing on the ground" (high probability of mammals).

[0050] If the animal target is a bird, the initial deterrent strategy is to output a green laser, and the green laser output method is a fan-shaped scan. If the animal target is a mammal, the corresponding initial repelling strategy is to output random low-frequency ultrasound (20-30kHz).

[0051] Preferably, the expulsion strategy library is a knowledge base built based on a large amount of experimental data and field survey results, which includes fear source parameters for each type of small animal, priority expulsion technology combinations, parameter optimization rules, collaborative strategies, and anti-adaptive rotation schemes.

[0052] The fear source parameter clarifies the sensitivity of various small animals to stimuli such as ultrasound, sound effects, lasers, bright light, and loud noises (strong, moderate, average; or high, moderate, low, or other custom-defined descriptions). The combination of priority removal techniques is determined based on the species' sensitivity characteristics, thus defining the combination of removal units. Regarding small animal intrusions, technologies such as ultrasound, green lasers, intelligent voice, flashing lights, and popping sounds all have specific application ranges and widespread adaptability. A single protective technology is difficult to effectively or permanently repel small animals. Repelling units can be combined. For example, green lasers combined with varied raptor calls can be used for sparrows, while variable frequency ultrasound combined with owl calls can be used for rats. For flocks of birds, green laser scanning combined with peregrine falcon calls and random popping sounds can be used. For highly intelligent birds such as magpies, sudden changes in raptor sound effects combined with randomized laser paths can be used.

[0053] The parameter optimization rules clarify the methods for adjusting the repulsion parameters under different environmental conditions (light, temperature, humidity, wind speed) and target behavior states, and optimize the repulsion parameters based on the effects of historical repulsion strategies. For example, at night, the ultrasonic frequency is switched to 35kHz for mice, and the ultrasonic frequency is turned off for doves on cloudy or rainy days. Specifically, the deportation parameters include the scanning methods of different deportation units in the deportation execution module (such as linear scanning, fan-shaped scanning, irregular scanning, etc.), the parameter values ​​of different deportation units (such as ultrasonic frequency, sound pressure, etc.), the combination method, the implementation duration (5 min or 10 min), and the number of implementations per unit time (3 or 4 times per hour, etc.). A historical deportation effect database corresponding to different deportation parameters can be established, and the deportation effect can be evaluated (such as the number of small animals leaving when the deportation strategy is implemented, or the reciprocal of the duration of the small animals' stay, or the response time of the small animals leaving when the deportation strategy is implemented, etc.). Deportation parameters with good deportation effect (high success rate) are given bonuses or weights, while deportation parameters with poor deportation effect (low success rate) are given deductions or weights, so that deportation parameter combinations with high historical deportation effect can be prioritized in the future.

[0054] The collaborative strategy specifies the linkage method of multiple deterrent units, such as synchronously playing the calls of natural enemies during laser scanning to strengthen fear association; Specifically, enhanced technological means can be established to reject inefficient measures: This can be achieved by combining intrusion-target fear techniques, such as visual-based deterrence combined with laser-based methods, to strengthen the effectiveness. Furthermore, a principle of technological complementarity can be established to enhance strategic synergy: by analyzing typical characteristics of targets, such as intelligence, nocturnal habits, and gregariousness, and by analyzing the different characteristics of target species' adaptability, periodicity, and anti-deception capabilities to deterrence measures, anti-adaptive, rotational, and random deterrence strategies can be developed.

[0055] The anti-adaptive rotation scheme sets a rotation cycle (days, weeks, etc., which can be customized according to actual conditions) and regularly updates the repulsion strategy combination to prevent small animals from developing adaptation.

[0056] The AI ​​fusion control module determines the final removal strategy for the current animal target based on its species type, visual sensitivity, acoustic sensitivity, and auditory fear sources in the species removal strategy library. Specifically, this includes: The substation area is divided into protection zones of different risk levels according to equipment type or distance from the substation; Specifically, the lowest first risk level can be defined as 30-40 meters away from the substation within the substation area, the second lowest risk level as 10-30 meters away, and the highest third risk level as within 10 meters. The specific distance classification can be flexibly distinguished according to the actual situation. If the equipment type within the substation area is a main transformer, then the corresponding risk level is increased by 1.

[0057] Determine the corresponding sound fear source, visual sensitivity, sound wave sensitivity, and popping sound sensitivity based on the species type of the animal target; Based on the protection zone level to which the animal target's current location belongs, determine a deportation strategy with the same level as the current protection zone. Among these strategies, the deportation strategy level for a single deportation unit is lower than the deportation strategy level for different units. Furthermore, the more deportation units there are, the higher the deportation strategy level. Within the same level, the deportation priority for sound-based fear sources, visual sensitivity, sound wave sensitivity, and loud noise sensitivity decreases in that order. The number of deportation strategy levels is greater than the number of protection zone levels.

[0058] Specifically, in this embodiment, there are 4 deportation units, so there are 4 deportation strategy levels. That is, when the deportation strategy level is 1, there is 1 deportation unit; when the deportation strategy level is 2, there are 2 deportation units; when the deportation strategy level is 3, there are 3 deportation units; and when the deportation strategy level is 4, there are 4 deportation units. When the protection zone level is the lowest first risk level, the deportation strategy level executed is also 1, that is, one deportation unit is called. The priority order of calling is as follows: sound fear source (intelligent voice deportation unit), visual sensitivity (green laser deportation unit), sound wave sensitivity (ultrasonic deportation unit), and explosion sound sensitivity (high-pressure explosion deportation unit). When the protected area is at the second-lowest risk level, the deportation strategy level is also 2, which means two deportation units are activated. The priority is given to a combination of sound-based fear sources (intelligent voice deportation unit) and visual sensitivity (green laser deportation unit), then the priority is gradually reduced (e.g., sound-based fear sources (intelligent voice deportation unit) + sound wave sensitivity (ultrasonic deportation unit)), finally down to sound wave sensitivity (ultrasonic deportation unit) and then explosion sensitivity (high-pressure explosion deportation unit). When the protected area is at the second-lowest risk level, the deportation strategy level is also 3, which means three deportation units are activated. The priority is given to a combination of sound-based fear sources (intelligent voice deportation unit) + visual sensitivity (green laser deportation unit) + sound wave sensitivity (ultrasonic deportation unit), then the priority is gradually reduced (e.g., sound-based fear sources (intelligent voice deportation unit) + explosion sensitivity (high-pressure explosion deportation unit)), finally down to visual sensitivity (green laser deportation unit), sound wave sensitivity (ultrasonic deportation unit), and then explosion sensitivity (high-pressure explosion deportation unit).

[0059] Preferably, determining the final removal strategy corresponding to the current animal target based on the species type of the animal target, the visual sensitivity, acoustic sensitivity, and auditory fear sources of the animal target's species type in the species removal strategy database, specifically further includes: Determine the effectiveness of the current expulsion strategy. If the current expulsion strategy is ineffective, upgrade the current strategy step by step. If the current expulsion strategy is effective, evaluate the current expulsion strategy based on the animal target's response time, stay time, number of animals expelled, and expulsion success rate. When expulsion strategies of the same level are in effect, prioritize the expulsion strategy with the higher evaluation score (the combination of expulsion units). The criteria for successful or effective repulsion are as follows: If the target begins to leave the protected area within 30 seconds and leaves completely within 1 minute, the repulsion is deemed effective, and the current repulsion strategy is maintained until 1 minute after the target leaves, and then the monitoring status is restored; if the target does not leave within 30 seconds, or returns after leaving, the repulsion is deemed ineffective, and the number of repulsion units is increased. For example, for magpies that have not left, green laser scanning is added to the existing eagle calls and popping sounds.

[0060] If the animal target is still not removed even with the highest level of removal strategy, or if the animal identification result is human, the current removal strategy will be suspended, and an enhanced alarm will be sent to the central control station through the remote linkage unit to prompt maintenance personnel to intervene.

[0061] Preferably, the adaptive rotation scheme further includes adjusting the rotation cycle based on the historical effectiveness of different expulsion strategies at the same level, specifically including: Obtain the historical effects of different expulsion strategies at the same level within the current rotation period. If the historical effect of the expulsion strategy at the same level within the current rotation period is greater than the preset effect threshold, increase the execution time of the expulsion strategy or increase the expulsion parameters in the expulsion strategy according to the preset time step. If the historical effect of the expulsion strategy at the same level within the current rotation period is not greater than the preset effect threshold, decrease the execution time of the expulsion strategy or decrease the expulsion parameters in the expulsion strategy according to the preset time step. The historical effect is the number of animals that leave when the expulsion strategy is executed, or the reciprocal of the animal's stay time, or the response time of the animal leaving when the expulsion strategy is executed.

[0062] Preferably, a three-dimensional redundant information database can be established by analyzing the visual sensitivity, auditory sensitivity, and core sound-based fear sources of small animals, optimizing the "blind driving" measures in traditional expulsion strategies. The first step is to establish targeted expulsion methods to avoid blind action: by analyzing the fear factors of intruding targets, the system automatically executes corresponding measures, enhancing targeting. The second step is to establish enhanced technical means to avoid inefficient measures: by combining intruding target fear techniques, such as visual-based expulsion combined with laser-based expulsion modes, the effectiveness is enhanced. The third step is to establish the principle of technological complementarity to enhance strategy synergy: by analyzing typical characteristics of targets, such as intelligence, nocturnal behavior, and herd behavior, the different characteristics of target species' adaptability, periodicity, and anti-deception to expulsion measures are analyzed, and anti-adaptive, rotational, and random expulsion strategies are formulated. The fourth step is to establish a rotation cycle to enhance the sustainability of measures: in this scheme, strategies are formulated weekly and rotated monthly to avoid species adaptability and ensure the long-term sustainable effectiveness of expulsion measures. For example, when a sparrow is detected in a low-light environment at night, the repelling strategy is determined to be a green laser (random path scanning) + a variable raptor call (sound pressure 80-100dB, dynamic frequency); when a mouse is detected at night, the repelling strategy is determined to be a variable frequency ultrasound (25-35kHz) + an owl call (sound pressure > 75dB).

[0063] Preferably, the system can also evaluate the effectiveness of the repulsion and optimize the strategy. Specifically, the AI ​​fusion control module evaluates the repulsion effect based on the feedback target movement data. If the target begins to leave the protected area within 30 seconds and completely leaves within 1 minute, the repulsion is deemed effective, and the current repulsion strategy is maintained until 1 minute after the target leaves, then monitoring is resumed. If the target does not leave within 30 seconds, or returns after leaving, the repulsion effect is deemed poor, and a collaborative enhancement mode is activated, adding complementary repulsion units. For example, for magpies that have not left, green laser scanning is added to the existing eagle calls and popping sounds. If the repulsion is still ineffective after collaborative enhancement, the system sends a strengthened alarm to the central control station via a remote linkage unit, prompting maintenance personnel to intervene.

[0064] Preferably, adaptive rotation and model optimization can also be performed. Under the same level of expulsion strategy, taking a week and a combination of two expulsion units as an example, the expulsion strategy combination is updated regularly. Week 1: Birds are mainly treated with laser + 20kHz low-frequency ultrasound, and mammals are mainly treated with popping sounds + 30kHz mid-frequency ultrasound; Week 2: Birds are treated with predator sound effects + 25kHz high-frequency ultrasound, and mammals are treated with dog barking sounds + flashing lights; Week 3: Birds are treated with popping sounds + laser scanning, and mammals are treated with 25-35kHz variable frequency ultrasound + laser scanning; Week 4: Both birds and mammals are treated with random combinations (random combinations of any two expulsion units out of the four expulsion units). At the same time, the system records relevant data for each expulsion process (species, environment, strategy, effect), and regularly optimizes and updates the small animal recognition model and expulsion strategy library to improve the long-term adaptability and expulsion effect of the system.

[0065] To more clearly illustrate the strategies for repelling different types of small animals, the following explanations focus on several typical animal groups: I. Bird Repelling Strategies: 1. Core fears and sensitivities: (1) Visual sensitivity: Some birds are most sensitive to dynamic green laser (532nm) and regard it as a physical threat.

[0066] (2) Hearing sensitivity: loud noises (>120dB) cause the flock to take flight; the calls of predators (such as falcons and eagles) trigger instinctive escape.

[0067] (3) Reduced adaptability: The use of variable frequency ultrasonic design and the combination of sound and light can effectively delay the adaptation period.

[0068] 2. Common bird differentiation strategies are shown in Table 2 below: Table 2. Common Bird Differentiation Strategies and Improvement Methods:

[0069] II. Mammal Removal Strategies: 1. Core fears and sensitivities: (1) Auditory dominance: Rats are sensitive to ultrasound (15kHz-65kHz) and cat meows; cats are afraid of dog barking; weasels are afraid of dog barking and loud noise.

[0070] (2) Visual aids: Nocturnal species (rats, weasels) avoid strong light; cats have high tolerance to dynamic lasers.

[0071] (3) Adaptation risk: Rodents can adapt to a fixed sound source in 3-5 days, but require frequency conversion and intermittent stimulation.

[0072] 2. Common differentiation strategies in mammals are shown in Table 3 below: Table 3. Common Differentiation Strategies in Mammals:

[0073] III. Multi-technology Collaboration and Rotation Strategy: 1. Complementary technology design: For specific small animal invasion scenarios, we analyze their typical characteristics and propose corresponding solutions. The complementary technologies are shown in Table 4 below.

[0074] Table 4: Technology Complementary Design Table

[0075] 2. The rotation cycle and intensity adjustment are shown in Table 5 below: Table 5. Rotation cycle and intensity adjustment of bird deterrence strategies:

[0076] By establishing functional zones for early warning, identification, and response, a differentiated, collaborative, and sustainable three-dimensional de-radicalization system is constructed to ensure the long-term effectiveness of the system.

[0077] Differentiation: Birds mainly use lasers and predator sound effects, while mammals mainly use variable frequency ultrasound and loud noises.

[0078] Synergy: Sound and light technologies are triggered synchronously (such as playing a falcon cry when a laser is being scanned) to enhance fear association.

[0079] Sustainability: Monthly rotation avoids adaptability, while real-time monitoring and identification ensure targeted effectiveness.

[0080] Preferably, the substation small animal differentiated collaborative repelling system provided in this embodiment also includes a linkage module. The linkage module is used to connect to the substation data monitoring system and the central control station data network system to realize data linkage and remote alarm, improve the convenience of operation and maintenance and the efficiency of handling, including a data linkage unit and a remote linkage unit.

[0081] The data linkage unit communicates with the thermal imaging unit and the existing monitoring system of the substation. When the front-end detection module detects an intrusion target, it immediately triggers the thermal imaging unit to track the target, collect clear data images, and upload the data information to the AI ​​fusion control module and the central control station in real time.

[0082] The remote linkage unit connects to the substation's central control station via a data network. When the system detects a small animal intrusion and initiates a removal process, it automatically sends an alarm signal to the central control station. The alarm information includes the intrusion time, target species, location, and removal strategy. Upon receiving the alarm, the central control station can display the corresponding data screen, allowing maintenance personnel to remotely view and intervene in the removal process. If the removal is ineffective, personnel can be dispatched to handle the situation on-site.

[0083] To further illustrate the technical solution of this invention, the following detailed explanation is provided in conjunction with an application strategy for an open substation.

[0084] An open-type substation measures 152 meters wide (north-south) and 184 meters long (east-west). The outdoor equipment area measures 110 meters east-west and 152 meters north-south. The surrounding area consists of farmland and trees, and is frequented by birds, rats, cats, weasels, and other small animals. The deployment plan for this system at this substation is as follows: Front-end detection module: Four radar units are deployed, installed at the four corners of the substation. Each radar covers a 120° scanning angle, and together they achieve 360° full-range coverage. Four thermal imaging units are deployed, each corresponding to the radar coverage area, to achieve accurate target capture.

[0085] Removal Execution Module: Deploy 4 integrated removal systems. Each integrated removal system integrates an ultrasonic removal unit, an intelligent voice removal unit, an anti-adaptive rotation system, a green laser removal unit, and a high-voltage explosion removal unit, which are installed around the outdoor equipment area and around the core equipment (main transformer and busbar).

[0086] AI Fusion Control Module: Deploy one AI fusion processing host in the substation protection control room, load the small animal recognition model and expulsion strategy library, and connect it to the front-end detection module and expulsion execution module via optical fiber.

[0087] Linkage module: Connects to the existing data monitoring system of the substation and the data network of the central control station to realize data linkage and remote alarm.

[0088] The system operates as follows: At 3:00 AM one day, the radar unit detected a moving target on the northwest side of the substation, with location coordinates (X: 35m, Y: 120m) and speed of 0.5m / s. This immediately triggered other imaging units to acquire thermal imaging images of the target, showing that the target was small in size, had a body temperature of about 37℃, and its movement trajectory was meandering.

[0089] After receiving radar data and thermal imaging images, the AI ​​fusion control module performs preliminary classification analysis and initially determines that the target is a mammal. It then controls the ultrasonic repelling unit to output random intermediate frequency ultrasonic waves. Finally, it controls the triggering to collect the sound and image information of the animal target. The AI ​​fusion control module receives radar data, sound information of the animal target, and image information (which may also include temperature values ​​and temperature distribution contours in thermal imaging images). Through analysis using an animal recognition model, it determines the target to be a mouse (98.2% confidence level). The module then retrieves the repulsion strategy library. The core fear sources for mice are ultrasound (25-55kHz) and owl calls. The preferred repulsion technology combination is variable-frequency ultrasound + owl calls. Given the current environment (nighttime, light intensity < 5 lux), temperature 15℃, and humidity 60%, and based on parameter optimization rules, the ultrasound frequency is set to 35kHz, the sound pressure level to 80dB, and the owl call sound pressure level to 90dB.

[0090] The AI-integrated control module sends a command to the northwest integrated deterrence system. The ultrasonic deterrence unit activates its 35kHz frequency conversion output, while the intelligent voice deterrence unit simultaneously plays an owl call. Ten seconds after the deterrence, the radar detects the target moving towards the substation's perimeter, increasing its speed to 1 m / s. Thirty seconds after the deterrence, the target leaves the substation's protected area. The system continuously monitors for one minute without detecting the target returning, thus determining the deterrence successful, stopping the deterrence output, and resuming monitoring. Simultaneously, the linkage module sends an alarm message to the central control station ("At 3:05 AM on [Date], a rat intrusion was detected in the northwest equipment area. It was successfully deterred using 35kHz frequency conversion ultrasonic waves and an owl call"), and uploads relevant data images.

[0091] In this invention, the front-end detection module collects the target's location, movement trajectory, morphological features, sound features, thermal imaging information, and image information. The AI ​​fusion control module acquires the target's location, movement trajectory, and thermal imaging information, and sequentially determines whether the target has invaded, whether the invading target is an animal, and the initial classification result of the target. After determining that an animal target has invaded the substation area, a preliminary removal strategy is determined based on the initial classification result. The front-end detection module then acquires the animal target's morphological features, sound features, and image information again, and based on these, determines the species type of the animal target. Based on the species type of the animal target and the visual sensitivity, acoustic sensitivity, and sound fear sources of the animal target's species type in the species removal strategy library, the final removal strategy corresponding to the current animal target is determined. The removal execution module removes the target according to the preliminary and final removal strategies determined by the AI ​​fusion control module. This effectively solves the problem of low reliability of substation removal caused by existing technologies and effectively improves the reliability of substation removal.

[0092] The technical solution of this invention also includes a linkage module, which is used to connect the substation data monitoring system and the central control station data network system to realize data linkage and remote alarm.

[0093] The front-end detection module of this invention includes a radar unit, a thermal imaging unit, a sound acquisition unit, and a camera. The radar unit is used to monitor the substation monitoring area and collect the location and movement trajectory of the target. The thermal imaging unit is used to collect the thermal image of the target after the radar unit detects a moving target. The sound acquisition unit is used to collect the sound information of the animal target after the thermal image collected by the thermal imaging unit identifies the animal target. The camera is used to collect the image information of the animal target after the thermal image collected by the thermal imaging unit identifies the animal target. By using the radar unit and the thermal imaging unit to classify whether an intrusion has occurred and whether an intrusion has occurred, and then obtaining the sound and image information of the animal target, the accuracy of small animal target identification is improved, thereby ensuring the reliability of driving away different types of small animals in the substation.

[0094] The repulsion execution module described in this invention includes an ultrasonic repulsion unit, an intelligent voice repulsion unit, a green laser repulsion unit, and a high-pressure explosion repulsion unit. The species repulsion strategy library includes fear source parameters for each species type, priority repulsion technology combinations, parameter optimization rules, and anti-adaptation rotation schemes. It can effectively combine multiple repulsion units according to different types of small animals, achieving differentiated repulsion for different types of small animals, while avoiding the development of adaptation in small animals through a strategy rotation mechanism, thereby improving the long-term effectiveness and reliability of the protection system.

[0095] In this invention, the animal target is determined to be a bird or a mammal based on the hot spot temperature, hot spot distribution area, and movement trajectory in the thermal imaging image. If the animal target is a bird, the corresponding initial deterrent strategy is to output a green laser in a fan-shaped scanning manner. If the animal target is a mammal, the corresponding initial deterrent strategy is to output random low-frequency ultrasound. Different initial deterrent strategies can be executed according to the initial classification results of the animal target, ensuring the reliability of safety protection in the substation area.

[0096] In this invention, the substation area is divided into protection zones of different risk levels according to equipment type or distance from the substation; the corresponding sound fear sources, visual sensitivity, sound wave sensitivity, and blast sound sensitivity are determined based on the species type of the animal target; and a repulsion strategy of the same level as the current protection zone is determined based on the protection zone level of the animal target's current location. This allows for the determination of a corresponding level of repulsion strategy based on the equipment type and the protection zone level of the animal target within the substation area, which not only effectively drives away the animal target but also reduces harm to the animal and energy consumption.

[0097] Example 2 like Figure 4As shown, the present invention also provides a differentiated collaborative repelling method for small animals in substations, which is based on a differentiated collaborative repelling system for small animals in substations in Embodiment 1, and includes: S1, the front-end detection module monitors the substation monitoring area and collects the target's location, movement trajectory, and thermal imaging information respectively; S2, the AI ​​fusion control module acquires the target's location, movement trajectory and thermal imaging information, and sequentially determines whether the target has invaded, whether the invading target is an animal, and the initial classification result of the target; and after determining that the animal target has invaded the substation area, it determines the initial deportation strategy based on the initial classification result of the target; S3, the expulsion execution module is used to initially expel the animal target according to the initial expulsion strategy corresponding to the current animal target determined by the AI ​​fusion control module; S4, the front-end detection module collects the morphological features, sound features, and image information of the animal target; S5, the AI ​​fusion control module is used to acquire the morphological features, sound features, and image information of animal targets, and based on the morphological features, sound features, and image information of animal targets, determine the species type to which the animal targets belong, and determine the final removal strategy corresponding to the current animal target based on the species type to which the animal targets belong, the visual sensitivity, sound wave sensitivity, and sound fear source of the species type to which the animal targets belong in the species removal strategy library. S6, the expulsion execution module is used to expel the animal target according to the final expulsion strategy corresponding to the current animal target determined by the AI ​​fusion control module.

[0098] Step S1 corresponds to system initialization and monitoring startup. After the system is powered on, the front-end detection module starts the all-weather real-time monitoring mode. The radar unit continuously scans the substation monitoring area according to preset parameters, and the thermal imaging unit is in standby mode, ready to respond to radar trigger signals at any time. The sound acquisition unit and camera are in standby mode, ready to respond to thermal imaging unit trigger signals at any time. The AI ​​fusion control module loads the small animal recognition model and the repulsion strategy library and completes initialization. Target detection and data acquisition. During scanning, when the radar unit detects a moving target, it determines the target distance by calculating the time difference between the emission and return of electromagnetic waves. Simultaneously, the radar beam points in the direction of the target, and the Doppler effect is used to determine the target's velocity. The system calculates the target's position and velocity to determine whether it has entered a preset protection zone. A 30-meter radius around the substation serves as the early warning zone for equipment protection. If a target enters the protection zone, the radar unit sends a trigger signal to the thermal imaging unit. The thermal imaging unit quickly turns towards the target area and acquires thermal imaging data (thermal temperature values ​​and distribution). The data is simultaneously transmitted to the AI ​​fusion control module. Step S2 corresponds to the initial target classification. The AI ​​fusion control module acquires the target's location, movement trajectory, and thermal imaging information, and sequentially determines whether the target has invaded, whether the invading target is an animal, and the initial target classification result. After determining that an animal target has invaded the substation area, a preliminary deportation strategy is determined based on the initial target classification result. If the animal target is a bird, the corresponding preliminary deportation strategy is to output a green laser, and the green laser output mode is a fan-shaped scan. If the animal target is a mammal, the corresponding preliminary deportation strategy is to output random intermediate frequency ultrasound. Step S3: The expulsion execution module is used to initially expel the animal target according to the initial expulsion strategy corresponding to the current animal target determined by the AI ​​fusion control module.

[0099] In step S4, the front-end detection module collects the morphological features, sound features, and image information of the animal target. The sound acquisition unit is used to collect the sound information of the animal target after the thermal imaging image collected by the thermal imaging unit is identified as an animal target. The camera is used to collect the image information of the animal target after the thermal imaging image collected by the thermal imaging unit is identified as an animal target.

[0100] S5, Target Recognition and Species Classification. The AI ​​fusion control module receives front-end detection data (morphological features, auditory features, and image information of the animal target) and inputs it into the small animal recognition model. The model extracts the target's morphological features (body size, outline), thermal features (body temperature distribution), and motion features (movement speed, trajectory), compares them with data in the training sample library, achieves accurate classification of the target species, and outputs the recognition result (species name, confidence level). If the confidence level is below 95%, the system initiates a second review, increases data collection time, and optimizes recognition accuracy; if the confidence level is ≥95%, it proceeds to the next step. Specifically, the substation area is divided into protection zones of different risk levels according to equipment type or distance from the substation; the corresponding sound fear sources, visual sensitivities, acoustic sensitivities, and blast sensitivity are determined based on the species type of the animal target; based on the protection zone level of the animal target's current location, a repulsion strategy of the same level as the current protection zone level is determined; the repulsion strategy level of a single repulsion unit is lower than the repulsion strategy level of multiple units; and the more repulsion units there are, the higher the repulsion strategy level; within the same level, the repulsion priority of sound fear sources, visual sensitivities, acoustic sensitivities, and blast sensitivity decreases in that order; the number of repulsion strategy levels is greater than the number of protection zone levels. When assessing the effectiveness of the current expulsion strategy, if the current strategy is ineffective, the strategy is progressively upgraded. If the current strategy is effective, it is evaluated based on the animal target's response time, dwell time, number of animals expelled, and expulsion success rate. Among expulsion strategies of the same level, the strategy with the higher evaluation score is prioritized. If the animal target is still not expelled even with the highest-level expulsion strategy, or if the animal identification result is human, the current expulsion strategy is paused and an alarm is sent to maintenance personnel.

[0101] Step S6 corresponds to the differentiated expulsion execution. The AI ​​fusion control module sends the expulsion command of the final expulsion strategy to the expulsion execution module. The expulsion execution module starts the corresponding expulsion unit according to the command and outputs the differentiated expulsion signal.

[0102] Preferably, the system may also include step S7 (not shown in the figure), where the AI ​​fusion control module evaluates the deterrence effect and optimizes the strategy. Specifically, the AI ​​fusion control module evaluates the deterrence effect based on the feedback target movement status data. If the target begins to leave the protected area within 30 seconds and completely leaves within 1 minute, the deterrence is deemed effective, and the current deterrence strategy is maintained until 1 minute after the target leaves, then the monitoring state is restored. If the target does not leave within 30 seconds, or returns after leaving, the deterrence effect is deemed poor, and a collaborative enhancement mode is activated, adding complementary deterrence units. For example, for magpies that have not left, green laser scanning is added to the existing eagle calls and popping sounds. If the deterrence is still ineffective after collaborative enhancement, the system sends an enhanced alarm to the central control station through the remote linkage unit, prompting maintenance personnel to intervene.

[0103] Preferably, step S8 (not shown in the figure) may also be included, where the AI ​​fusion control module performs adaptive rotation and model optimization. Under the same level of expulsion strategy, taking a week and a combination of two expulsion units as an example, the expulsion strategy combination is updated regularly. Week 1: Birds are mainly treated with laser + 20kHz low-frequency ultrasound, and mammals are mainly treated with popping sounds + 30kHz mid-frequency ultrasound; Week 2: Birds are treated with predator sound effects + 25kHz high-frequency ultrasound, and mammals are treated with dog barking sounds + flashing lights; Week 3: Birds are treated with popping sounds + laser scanning, and mammals are treated with 25-35kHz variable frequency ultrasound + laser scanning; Week 4: Both birds and mammals are treated with random combinations (random combinations of any two expulsion units out of the four expulsion units). At the same time, the system records relevant data (species, environment, strategy, effect) for each expulsion process, and regularly optimizes and updates the small animal recognition model and expulsion strategy library to improve the long-term adaptability and expulsion effect of the system.

[0104] Preferably, the process may also include step S9 (not shown in the figure), where the AI ​​fusion control module performs alarm and linkage processing through the linkage module. During the removal process, the data linkage unit continuously tracks the target and uploads data images; the remote linkage unit sends alarm signals to the central control station, which displays alarm information and data images, supporting remote viewing and control by maintenance personnel. In case of an emergency (such as when a small animal has come into contact with the equipment), maintenance personnel can remotely activate the emergency removal mode, with all removal units working simultaneously to quickly remove the target.

[0105] In this invention, the front-end detection module monitors small animals within the substation's monitoring area, collecting the target's location, movement trajectory, morphological features, and thermal imaging information. The AI ​​fusion control module acquires the target's location, movement trajectory, morphological features, and thermal imaging information. Based on the target's location, movement trajectory, morphological features, thermal imaging information, and animal recognition models, it determines the target's species type. Based on the target's species type and the visual sensitivity, auditory sensitivity, and sound fear sources of the target's species type in the species repulsion strategy library, it determines the corresponding repulsion strategy for the current target. The repulsion execution module, based on the repulsion strategy determined by the AI ​​fusion control module, repels the target, effectively solving the problem of low reliability in substation repulsion caused by existing technologies and significantly improving the reliability of substation repulsion.

[0106] The technical solution of this invention also includes a linkage module, which is used to connect the substation data monitoring system and the central control station data network system to realize data linkage and remote alarm.

[0107] The front-end detection module in the technical solution of this invention includes a radar unit and a thermal imaging unit. It can not only collect the position and movement trajectory of the target, but also collect thermal imaging images of the target. Based on the thermal imaging images, the morphological characteristics of the target can be determined, which improves the accuracy of small animal target identification and thus ensures the reliability of driving away different types of small animals in substations.

[0108] The repulsion execution module described in this invention includes an ultrasonic repulsion unit, an intelligent voice repulsion unit, a green laser repulsion unit, and a high-pressure blast repulsion unit. The ultrasonic repulsion unit outputs ultrasonic waves in different frequency ranges based on the auditory sensitivity characteristics of different species. The intelligent voice repulsion unit incorporates predator calls and cries of other species corresponding to different species, outputting these sounds according to the predator characteristics and sound pressure sensitivity range of each species. The green laser repulsion unit outputs laser beams based on the visual sensitivity characteristics of different species. The high-pressure blast repulsion unit outputs a loud blast and flashes of light based on the acoustic sensitivity characteristics of different species, allowing for the effective combination of multiple repulsion units to be executed according to different types of small animals.

[0109] The species repulsion strategy library in this invention includes fear source parameters, priority repulsion technology combinations, parameter optimization rules, and an anti-adaptation rotation scheme for each species type. The fear source parameters describe the sensitivity of different species to ultrasound, sound effects, lasers, strong light, and loud noises. The priority repulsion technology combinations determine the optimal combination of sound and light repulsion strategies based on the species' sensitivity characteristics. The parameter optimization rules describe the adjustment methods for repulsion parameters under different environmental conditions and target behavior states, and optimize the repulsion parameters based on historical repulsion strategy effects. The anti-adaptation rotation scheme sets a rotation cycle and periodically updates the repulsion strategy combinations. Through multi-module collaborative work, it can achieve accurate identification and differentiated repulsion of different types of small animals, while the strategy rotation mechanism prevents small animals from developing adaptations, improving the long-term effectiveness and reliability of the protection system.

[0110] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A differentiated and collaborative small animal repelling system for substations, characterized in that, It includes a front-end detection module, a deportation execution module, and an AI fusion control module. The front-end detection module is used to monitor the substation monitoring area and collect the target's location, movement trajectory, morphological features, sound features, thermal imaging information, and image information. The AI ​​fusion control module is used to acquire the target's location, movement trajectory, and thermal imaging information, and sequentially determine whether the target has invaded, whether the invading target is an animal, and the initial classification result of the target. After identifying the animal targets that have invaded the substation area, a preliminary deportation strategy is determined based on the initial classification results of the targets. The front-end detection module acquires the morphological, auditory, and image information of the animal target again. Based on the morphological, auditory, and image information of the animal target, the species type of the animal target is determined. Based on the species type of the animal target, the visual sensitivity, acoustic sensitivity, and auditory fear source of the species type of the animal target in the species repulsion strategy library, the final repulsion strategy corresponding to the current animal target is determined. The expulsion execution module is used to expel the animal target according to the preliminary expulsion strategy and the final expulsion strategy corresponding to the current animal target determined by the AI ​​fusion control module.

2. The substation small animal differentiated cooperative repelling system according to claim 1, characterized in that, It also includes a linkage module, which is used to connect the substation data monitoring system and the central control station data network system to realize data linkage and remote alarm.

3. The substation small animal differentiated cooperative repelling system according to claim 1, characterized in that, The front-end detection module includes a radar unit, a thermal imaging unit, a sound acquisition unit, and a camera. The radar unit is used to monitor the substation monitoring area and acquire the location and movement trajectory of the target. The thermal imaging unit is used to acquire the thermal image of the target after the radar unit detects the moving target. The sound acquisition unit is used to acquire the sound information of the animal target after the thermal imaging image acquired by the thermal imaging unit is identified as an animal target; the camera is used to acquire the image information of the animal target after the thermal imaging image acquired by the thermal imaging unit is identified as an animal target.

4. A substation small animal differentiated cooperative repelling system according to claim 1, characterized in that, The expulsion execution module includes an ultrasonic expulsion unit, an intelligent voice expulsion unit, a green laser expulsion unit, and a high-pressure explosion expulsion unit; the ultrasonic expulsion unit outputs ultrasonic waves in different frequency ranges based on the classification results of different animal targets and the acoustic sensitivity characteristics of different species. The intelligent voice deterrence unit has built-in predator calls and cacophony sounds of the same species corresponding to different types of species. Based on the predator characteristics and sound pressure sensitivity range of different types of species, it outputs predator calls and cacophony sounds of the same species accordingly. The green laser deterrence unit outputs laser beams based on the classification results of different animal targets and the visual sensitivity characteristics of different species. The high-pressure detonation unit outputs a detonation sound and a momentary flash of light based on the classification results of different animal targets and the detonation sound sensitivity characteristics of different species.

5. A substation small animal differential cooperative repelling system according to claim 4, characterized in that, After identifying the animal target's intrusion into the substation area, the initial deportation strategy based on the initial target classification results is as follows: Based on the hot spot temperature, hot spot distribution area, and movement trajectory of the animal target in the thermal imaging image, the animal target is determined to be a bird or a mammal. If the animal target is a bird, the initial deterrent strategy is to output a green laser, and the green laser output method is a fan-shaped scan. If the animal target is a mammal, the corresponding initial decoy strategy is to output random intermediate frequency ultrasound.

6. A substation small animal differential cooperative repelling system according to claim 4, characterized in that, The species repulsion strategy library includes fear source parameters, priority repulsion technology combinations, parameter optimization rules, and anti-adaptive rotation schemes for each species type. The fear source parameters describe the sensitivity of different species to ultrasound, sound effects, lasers, strong light, and loud noises. The priority repulsion technology combinations determine the combination of audio-visual repulsion strategies based on the species' sensitivity characteristics. The parameter optimization rules describe the adjustment methods for repulsion parameters under different environmental conditions and target behavior states, and optimize the repulsion parameters based on historical repulsion strategy effects. The anti-adaptive rotation scheme sets a rotation cycle, periodically updating the combination of repulsion units in the repulsion strategy. The repulsion parameters include the scanning method, combination method, implementation duration, and number of implementations per unit time for different repulsion units in the repulsion execution module.

7. A substation small animal differential cooperative repelling system according to claim 6, characterized in that, Based on the species type of the animal target, and the visual sensitivity, acoustic sensitivity, and auditory fear sources of the animal target's species type in the species removal strategy database, the final removal strategy corresponding to the current animal target is determined, specifically including: The substation area is divided into protection zones of different risk levels according to equipment type or distance from the substation; Determine the corresponding sound fear source, visual sensitivity, sound wave sensitivity, and popping sound sensitivity based on the species type of the animal target; Based on the protection zone level to which the animal target's current location belongs, determine a deportation strategy with the same level as the current protection zone. Among these strategies, the deportation strategy level for a single deportation unit is lower than the deportation strategy level for different units. Furthermore, the more deportation units there are, the higher the deportation strategy level. Within the same level, the deportation priority for sound-based fear sources, visual sensitivity, sound wave sensitivity, and loud noise sensitivity decreases in that order. The number of deportation strategy levels is greater than the number of protection zone levels.

8. A substation small animal differential cooperative repelling system according to claim 7, characterized in that, Based on the species type of the animal target, and the visual sensitivity, acoustic sensitivity, and auditory fear sources of the animal target's species type in the species removal strategy database, the final removal strategy corresponding to the current animal target is determined, which also includes: The effectiveness of the current expulsion strategy is judged. If the current expulsion strategy is ineffective, the current strategy is upgraded step by step. If the current expulsion strategy is effective, the current expulsion strategy is evaluated based on the animal target's response time, stay time, number of animals expelled, and expulsion success rate. When expulsion strategies of the same level are in effect, the expulsion strategy with the higher evaluation score is given priority. If the animal target is still not removed even with the highest level of removal strategy, or if the animal identification result is human, then the current removal strategy will be suspended and an alert will be sent to the maintenance personnel.

9. A substation small animal differentiated cooperative repelling system according to claim 6, characterized in that, The adaptive rotation scheme also includes adjusting the rotation cycle based on the historical effectiveness of different expulsion strategies at the same level, specifically including: Obtain the historical effects of different expulsion strategies at the same level within the current rotation period. If the historical effect of the expulsion strategy at the same level within the current rotation period is greater than the preset effect threshold, increase the execution time of the expulsion strategy or increase the expulsion parameters in the expulsion strategy according to the preset time step. If the historical effect of the expulsion strategy at the same level within the current rotation period is not greater than the preset effect threshold, decrease the execution time of the expulsion strategy or decrease the expulsion parameters in the expulsion strategy according to the preset time step. The historical effect is the number of animals that leave when the expulsion strategy is executed, or the reciprocal of the animal's stay time, or the response time of the animal leaving when the expulsion strategy is executed.

10. A differentiated and collaborative method for repelling small animals in a substation, characterized in that, Based on the substation small animal differentiated collaborative repelling system described in any one of claims 1-9, it includes: The front-end detection module monitors the substation's monitoring area, collecting the target's location, movement trajectory, and thermal imaging information. The AI ​​fusion control module acquires the target's location, movement trajectory, and thermal imaging information, and sequentially determines whether the target has invaded, whether the invading target is an animal, and the initial classification result of the target; and after determining that an animal target has invaded the substation area, it determines the initial deportation strategy based on the initial classification result of the target. The expulsion execution module is used to initially expel the animal target based on the preliminary expulsion strategy corresponding to the current animal target determined by the AI ​​fusion control module. The front-end detection module collects morphological features, sound features, and image information of animal targets; The AI ​​fusion control module is used to acquire the morphological features, sound features, and image information of animal targets. Based on the morphological features, sound features, and image information of animal targets, it determines the species type to which the animal targets belong. Based on the species type to which the animal targets belong, the visual sensitivity, sound wave sensitivity, and sound fear sources of the species type to which the animal targets belong in the species repulsion strategy library, it determines the final repulsion strategy corresponding to the current animal target. The expulsion execution module is used to expel the animal target according to the final expulsion strategy corresponding to the current animal target determined by the AI ​​fusion control module.

Images