Deep intelligent learning multispectral imaging and bird identification method and device

Abstract

The invention relates to the technical field of bird feature recognition, in particular to a deep intelligent learning multispectral imaging and bird recognition method and device, which comprises a mounting tower, a fixed tower frame arranged on the outer side of the mounting tower, a solar power supply assembly arranged on the outer side of the fixed tower frame, and a dual-spectrum Mini holder arranged above the mounting tower. A central processing unit is arranged in the dual-spectrum Mini holder, and a bird repelling module is arranged on the outer side of the dual-spectrum Mini holder, so that the flight path, attitude and type of birds can be captured and analyzed in real time through an integrated high-resolution camera, an image processing algorithm and an intelligent recognition technology; and by analyzing data such as bird activity rules, environmental factors and working states of the bird repelling device, key factors influencing the bird repelling effect are identified, and a bird repelling strategy is adjusted, so that the bird repelling effect is continuously optimized, intelligent identification of birds is realized, false triggering of equipment and invalid repelling are reduced, and the efficiency and accuracy of the bird repelling device are improved.

Description

Technical Field

[0001] This invention relates to the field of bird feature recognition technology, and in particular to a deep intelligent learning multispectral imaging and bird recognition method and apparatus. Background Technology

[0002] In recent years, with the continuous improvement of the ecological environment, the number of bird species and their activity range in nature have gradually increased. Incidents of power facility grounding short circuits and insulation flashovers caused by bird damage are common. Currently, the main types of bird repelling devices on the market are as follows: Sound wave bird repelling devices: Sound wave bird repelling devices are a type of intelligent bird repelling technology. They utilize birds' sensitivity to specific sound wave frequencies, emitting sound waves of specific frequencies to drive birds away from the repelling area by simulating warning or attack sounds. This device typically uses electronic devices to generate sound waves, which are then transmitted to the target area through speakers or other propagation media. Laser bird repelling devices: Birds are most sensitive to vision, especially to green lasers with a wavelength of 532nm. Laser bird repelling devices simulate the visual reflection conditions of living organisms. When a 532nm green laser bar is shone, birds perceive it as a large green stick, thus achieving the bird repelling effect. Besides sound wave and laser bird repelling devices, there are many other types of bird repelling devices on the market. There are wind-powered bird deterrents, loud-sound bird deterrents, and some physical deterrents, such as bird nets, bird spikes, and motorized bird deterrents, which are suitable for different application scenarios.

[0003] However, in the aforementioned existing technologies, birds are highly intelligent and will gradually adapt to static and repetitive bird deterrent methods, such as fixed reflective strips and single sounds, which make bird deterrents effective in the early stages, but birds may no longer be afraid after a few weeks. Summary of the Invention

[0004] The purpose of this invention is to provide a method and device for multispectral imaging and bird identification using deep intelligent learning, which solves the technical problem in the prior art that birds are highly intelligent and will gradually adapt to static and repetitive bird deterrence methods, such as fixed reflective strips and single sounds, which make bird deterrence devices effective in the initial stage, but birds may no longer be afraid after a few weeks.

[0005] To achieve the above objectives, the present invention employs a deep intelligent learning multispectral imaging and bird identification device, comprising a mounting tower, a fixed tower frame disposed on the outer side of the mounting tower, a solar power supply component disposed on the outer side of the fixed tower frame, a dual-spectral mini gimbal disposed above the mounting tower, a central processing unit disposed inside the dual-spectral mini gimbal, and a bird deterrent module disposed on the outer side of the dual-spectral mini gimbal. The device is characterized in that... The dual-spectrum Mini gimbal includes an adjustment unit, a thermal imaging camera, and a high-definition visible light camera. The adjustment unit is fixedly connected to the mounting tower and located above the mounting tower. The thermal imaging camera is fixedly connected to the adjustment unit and located outside the adjustment unit. The high-definition visible light camera is fixedly connected to the adjustment unit and located on the side of the thermal imaging camera away from the adjustment unit. The bird deterrent module includes a protective shell, a laser bird deterrent, a blue light bird deterrent, and an ultrasonic bird deterrent. The protective shell is fixedly connected to the dual-spectrum Mini gimbal and is located below the laser bird deterrent. The laser bird deterrent is fixedly connected to the protective shell and is located inside the protective shell. The blue light bird deterrent is detachably connected to the dual-spectrum Mini gimbal and is located below the high-definition visible light camera. The ultrasonic bird deterrent is fixedly connected to the dual-spectrum Mini gimbal and is located above the dual-spectrum Mini gimbal.

[0006] The solar power supply component includes a mounting bracket and a hardware clamp. The mounting bracket is detachably connected to the fixed tower and is located above the fixed tower. The hardware clamp is fixedly connected to the mounting bracket and is located on the outside of the fixed tower.

[0007] The solar power supply component further includes a mounting sleeve and a solar cell array. The mounting sleeve is fixedly connected to the mounting bracket and is located outside the mounting bracket. The solar cell array is fixedly connected to the mounting sleeve and is located inside the mounting sleeve.

[0008] The solar power supply component also includes a solar panel and a battery. The solar panel is electrically connected to the solar cell array and is located outside the solar cell array. The battery is electrically connected to the solar cell array and is located outside the mounting sleeve.

[0009] The adjusting part includes a housing and a partition plate. The housing is fixedly connected to the mounting tower and located above the mounting tower. The partition plate is fixedly connected to the housing and located inside the housing.

[0010] The adjustment unit further includes a motor, a driving gear, and a driven gear. The output end of the motor is fixedly connected to the driving gear and is located above the partition plate. The driving gear meshes with the driven gear and is located below the partition plate. The driven gear is rotatably connected to the housing and is located outside the driving gear.

[0011] The adjustment unit further includes a drive rod and a bushing. The drive rod is fixedly connected to the driven gear and passes through the housing. The bushing is fixedly connected to the housing and is located outside the drive rod.

[0012] The adjustment unit further includes an adjustment box, which is fixedly connected to the drive rod and located above the drive rod.

[0013] The adjustment unit further includes a dual-axis motor and two drive shafts. The dual-axis motor is fixedly connected to the adjustment box and located inside the adjustment box. One end of the two drive shafts is fixedly connected to the output end of the dual-axis motor, and the other end of the two drive shafts is fixedly connected to the thermal imaging camera and the high-definition visible light camera, respectively.

[0014] The present invention also provides a construction method, comprising the following steps; S1. Install the device in the target area using the fixed tower and the hardware clamps, start the solar power supply component, and ensure that the battery can maintain the continuous operation of the equipment for 7 days under no-light conditions; adjust the dual-spectrum Mini gimbal to the initial monitoring angle so that it can rotate 360° horizontally and adjust vertically within the range of -90° to +90°, and the thermal imaging camera and the high-definition visible light camera enter the working state. S2. Real-time acquisition of dual-spectrum images: The thermal imaging camera optimizes the imaging algorithm for the isothermal characteristics of birds and automatically adjusts the contrast to capture clear thermal images in low-visibility environments such as night, heavy rain, and heavy fog; the high-definition visible light camera simultaneously acquires visible light images, and the data from both are transmitted to the central processing unit in real time. S3.AI Intelligent Recognition and Positioning: The central processing unit calls the YOLOv8n-seg deep learning model to perform fusion analysis on the dual-spectral images, accurately identify bird targets and distinguish their species, flight trajectories and attitudes, eliminate interference items such as birds and insects, determine the bird's location coordinates, and trigger dual-light video recording to record the entire process from detection to flight away; S4. Dynamic Bird Repelling Strategy Execution: Based on the recognition results, the central processing unit schedules and integrates the bird repelling modules, adopting a three-dimensional collaborative strategy of "laser + blue light + ultrasound": green laser, combined with dynamic scanning by the gimbal, forms a physical deterrent; the laser bird repeller and the blue light bird repeller emit specific wavelengths of light to stimulate the birds' vision; and the ultrasound bird repeller releases sound waves in frequencies sensitive to birds. Depending on the bird species and behavioral characteristics, bird repelling methods are activated individually or in combination to prevent birds from developing adaptation. S5. Data Feedback and Strategy Optimization: The back-end control system receives device operation data via wireless network, including bird activity patterns, bird deterrence effects, and environmental parameters. It uses reinforcement learning algorithms to analyze historical data and automatically adjust parameters such as laser scanning angle, sound wave frequency, and bird deterrence timing. It also supports remote control, allowing users to view logs and set working periods through a mobile app or the back-end system to continuously optimize bird deterrence strategies.

[0015] This invention discloses a deep intelligent learning multispectral imaging and bird identification device. In practical use, a solar power supply component is fixed above the fixed tower to power the device. This allows the device to operate normally for at least 7 days without sunlight or other power supplementation. The adjustment unit can adjust the thermal imaging camera and the high-definition visible light camera within a horizontal range of 0°~360° and a vertical range of -90°~+90°. The thermal imaging camera provides clear thermal images through infrared thermal imaging, improving bird identification rates. The high-definition visible light camera is equipped with a 25x optical zoom lens, automatically focusing to provide clear visible light images, which are transmitted to the central processing unit. According to the scheduling of the central processing unit, the laser bird repeller, blue light bird repeller, and ultrasonic bird repeller are driven individually or together to drive away birds, thus preventing bird adaptation. This solves the technical problem that birds, being highly intelligent, gradually adapt to static and repetitive bird-repelling methods. Attached Figure Description

[0016] 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, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the structure of a deep intelligent learning multispectral imaging and bird recognition device according to the present invention.

[0018] Figure 2 This is the invention Figure 1 Enlarged view of the local structure at point A.

[0019] Figure 3 This is a front view of a deep intelligent learning multispectral imaging and bird recognition device according to the present invention.

[0020] Figure 4 This is a side view of a deep intelligent learning multispectral imaging and bird recognition device according to the present invention.

[0021] Figure 5 This is a schematic diagram of the bird-repelling module of the present invention.

[0022] Figure 6 This is a partial structural schematic diagram of the dual-spectrum Mini gimbal of the present invention.

[0023] 1-Mounting tower, 2-Fixed tower, 3-Solar power supply component, 301-Mounting bracket, 302-Hardware clamp, 303-Mounting sleeve, 304-Solar cell array, 305-Solar panel, 306-Battery, 4-Dual-spectrum Mini gimbal, 401-Housing, 402-Divider plate, 403-Motor, 404-Drive gear, 405-Driven gear, 406-Drive rod, 407-Shaft sleeve, 408-Adjustment box, 409-Dual-axis motor, 410-Drive shaft, 411-Thermal imaging camera, 412-High-definition visible light camera, 5-Central processing unit, 6-Bird deterrent module, 601-Protective housing, 602-Laser bird deterrent, 603-Blue light bird deterrent, 604-Ultrasonic bird deterrent. Detailed Implementation

[0024] The embodiments of the present invention are described in detail below. Examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, but should not be construed as limiting the present invention.

[0025] Please see Figures 1 to 6 ,in Figure 1 This is a schematic diagram of the structure of a deep intelligent learning multispectral imaging and bird recognition device according to the present invention. Figure 2 This is the invention Figure 1 Enlarged view of the local structure at point A. Figure 3 This is a front view of a deep intelligent learning multispectral imaging and bird recognition device according to the present invention. Figure 4 This is a side view of a deep intelligent learning multispectral imaging and bird recognition device according to the present invention. Figure 5 This is a schematic diagram of the bird-repelling module of the present invention. Figure 6 This is a partial structural schematic diagram of the dual-spectrum Mini gimbal of the present invention.

[0026] This invention provides a deep intelligent learning multispectral imaging and bird identification device, including a mounting tower 1, a fixed tower 2, a solar power supply component 3, a dual-spectrum mini gimbal 4, a central processing unit 5, and a bird deterrent module 6. The solar power supply component 3 includes a mounting bracket 301, a hardware clamp 302, a mounting sleeve 303, a solar cell array 304, a solar panel 305, and a battery 306. The dual-spectrum mini gimbal 4 includes an adjustment unit, a thermal imaging camera 411, and a high-definition visible light camera 412. The adjustment unit includes a housing 401. The bird deterrent module 6 includes a partition plate 402, a motor 403, a drive gear 404, a driven gear 405, a drive rod 406, a bushing 407, an adjustment box 408, a dual-shaft motor 409, and two drive shafts 410. The bird deterrent module 6 also includes a protective shell 601, a laser bird deterrent 602, a blue light bird deterrent 603, and an ultrasonic bird deterrent 604. This addresses the problem in existing technologies where birds, being highly intelligent, gradually adapt to static and repetitive bird deterrent methods, such as fixed reflective strips and single sounds, which are effective initially but may become less frightening after a few weeks.

[0027] This specific embodiment includes an installation tower 1, a fixed tower frame 2 on the outer side of the installation tower 1, a solar power supply component 3 on the outer side of the fixed tower frame 2, a dual-spectrum mini gimbal 4 on top of the installation tower 1, a central processing unit 5 inside the dual-spectrum mini gimbal 4, and a bird-repelling module 6 on the outer side of the dual-spectrum mini gimbal 4. The dual-spectrum mini gimbal 4 includes an adjustment unit, a thermal imaging camera 411, and a high-definition visible light camera 412. The adjustment unit is fixedly connected to the installation tower 1 and located above it. The thermal imaging camera 411 is fixedly connected to the adjustment unit and located outside the adjustment unit. The high-definition visible light camera 412 is fixedly connected to the adjustment unit and located on the side of the thermal imaging camera 411 away from the adjustment unit. The bird-repelling module 6 includes a protective shell 601, a laser bird repeller 602, a blue light bird repeller 603, and an ultrasonic bird repeller 604. The protective shell 601 is fixedly connected to the dual-spectrum Mini gimbal 4 and located below the laser bird repeller 602. The laser bird repeller 602 is fixedly connected to the protective shell 601 and located inside the protective shell 601. The blue light bird repeller 603 is detachably connected to the dual-spectrum Mini gimbal 4 and located below the high-definition visible light camera 412. The ultrasonic bird repeller 604 is fixedly connected to the dual-spectrum Mini gimbal 4 and located above the dual-spectrum Mini gimbal 4. The solar power supply component 3 is fixed above the fixed tower 2 to power the equipment and can maintain the bird identification device's normal operation for at least 7 days under conditions of continuous no light and no other power supplementation. The adjustment unit can adjust the thermal imaging camera 411 and the high-definition visible light camera 412 in a horizontal range of 0°~360° and a vertical range of -90°~+90°. The thermal imaging camera 411 uses infrared thermal imaging and employs 384 With a resolution of 288 and a 24° standard lens, this high-definition visible light camera 412 utilizes an optimized imaging algorithm to automatically adjust image contrast, taking into account the fact that birds are warm-blooded animals whose body temperature typically remains between 38.5-42°C. This provides clear thermal images and improves bird recognition rates. A 1080p, 2-megapixel high-definition visible light sensor, equipped with a 25x optical zoom lens and autofocus, provides clear visible light images and transmits them to the central processing unit 5. According to the scheduling of the central processing unit 5, the laser bird repeller 602, blue light bird repeller 603, and ultrasonic bird repeller 604 are driven individually or together to repel birds, thereby preventing bird adaptation. The laser bird repeller 602, blue light bird repeller 603, and ultrasonic bird repeller 604 are all known prior art and will not be described in detail. The protective housing 601 fixes the laser bird repeller 602 and blue light bird repeller 603 to the underside of the thermal imaging camera 411 and the high-definition visible light camera 412, respectively. The central processing unit 5 is responsible for the operation of the overall system. Its main function is to control the bird-repelling module 6 to perform bird-repelling functions based on the results of AI intelligent recognition. Other functions include: self-detection function, providing self-detection information on the device's operating status and recording fault logs; self-recovery function, enabling the device to automatically resume normal operation after situations such as abnormal power supply termination, without losing stored data and enabling automatic resending; dynamic response to remote time query / setting, data request, restart, etc.; modification of device operating parameters according to remote commands; password authentication and remote program update, with a reliable update mechanism and method; periodic self-detection function and ability to output self-detection information; error self-recovery and timed restart functions; and hibernation function. The system supports sleep / wake-up mode; the device defaults to no sleep mode. The central processing unit 5's AI intelligent recognition function utilizes the YOLOv8n-seg deep learning model, a version of the YOLOv8 series from Ultralytics focused on instance segmentation. Based on the YOLOv8 backbone network structure, it achieves the fusion of target detection and segmentation by adding an additional feature map generation module. YOLOv8n-seg can simultaneously complete target detection and instance segmentation tasks. Through the high-definition visible light camera 412 and the thermal imaging camera 411, the system accurately detects bird activity and determines its location, providing this information to the intelligent recognition module for AI recognition. Once the system recognizes a bird, it immediately records dual-light video. The entire process of bird detection, identification, deterrence, and flight facilitates verification of deterrence effectiveness and tracing of bird incidents. Utilizing wireless network communication, it supports access to monitoring platforms such as the State Grid. Users can remotely control the bird deterrence device via wireless network, enabling real-time monitoring and control. The device can also be remotely started, stopped, or adjusted via a mobile application or computer interface. Users can preset the device's operating time periods for automated control. Furthermore, it features data recording and analysis functions, allowing users to view the device's logs and environmental data remotely, providing a basis for subsequent maintenance and optimization. This addresses the technical challenge of birds, being highly intelligent, gradually adapting to static and repetitive bird deterrence methods.

[0028] The solar power supply component 3 includes a mounting bracket 301 and a hardware clamp 302. The mounting bracket 301 is detachably connected to the fixed tower 2 and is located above the fixed tower 2. The hardware clamp 302 is fixedly connected to the mounting bracket 301 and is located on the outside of the fixed tower 2. The mounting bracket 301 is made of Q235 carbon steel with hot-dip galvanized anti-corrosion treatment. It is installed above the fixed tower 2 by the hardware clamp 302, thus eliminating the need for drilling and making it suitable for installation with various mounting towers 1.

[0029] Secondly, the solar power supply component 3 also includes a mounting sleeve 303 and a solar cell array 304. The mounting sleeve 303 is fixedly connected to the mounting bracket 301 and is located on the outside of the mounting bracket 301. The solar cell array 304 is fixedly connected to the mounting sleeve 303 and is located inside the mounting sleeve 303. The mounting sleeve 303 is fixed to the outside of the mounting bracket 301, and the solar cell array 304 is disposed inside the mounting sleeve 303.

[0030] Meanwhile, the solar power supply component 3 also includes a solar panel 305 and a battery 306. The solar panel 305 is electrically connected to the solar cell array 304 and is located outside the solar cell array 304. The battery 306 is electrically connected to the solar cell array 304 and is located outside the mounting sleeve 303. The solar panel 305 stores the light energy through the solar cell array 304 in the battery 306. The battery 306 uses a lithium iron phosphate battery, which has excellent high and low temperature discharge performance, a unique SOC algorithm, and more accurate power display. The solar power supply component 3 achieves off-grid power self-sufficiency. It is designed that under continuous no sunlight and no other power supplementation, the battery should be able to maintain the bird deterrent device's normal operation for at least 7 days. It has a reasonable circuit design, excellent anti-shading ability, reduces the risk of heat spots, and improves product reliability. The solar power supply component 3 also has compatible LED indicator lights and buttons for easy querying of SOC, charging CHG, and discharging DHG status.

[0031] Additionally, the adjustment unit includes a housing 401 and a partition plate 402. The housing 401 is fixedly connected to the mounting tower 1 and located above the mounting tower 1. The partition plate 402 is fixedly connected to the housing 401 and located inside the housing 401. The housing 401 is located above the mounting tower 1, the partition plate 402 is located inside the housing 401, and the central processing unit 5 is located outside the partition plate 402.

[0032] The adjustment unit further includes a motor 403, a driving gear 404, and a driven gear 405. The output end of the motor 403 is fixedly connected to the driving gear and is located above the partition plate 402. The driving gear 404 meshes with the driven gear 405 and is located below the partition plate 402. The driven gear 405 is rotatably connected to the housing 401 and is located outside the driving gear 404. The motor 403 is disposed above the partition plate 402. The motor 403 drives the driving gear 404 to rotate. Through the meshing of the driving gear 404 and the driven gear 405, the driving gear 404 drives the driven gear 405 to rotate.

[0033] Secondly, the adjustment part also includes a drive rod 406 and a bushing 407. The drive rod 406 is fixedly connected to the driven gear 405 and passes through the housing 401. The bushing 407 is fixedly connected to the housing 401 and is located outside the drive rod 406. The driven gear 405 drives the drive rod 406 to rotate, and the bushing 407 ensures the stability of the rotation of the drive rod 406.

[0034] Furthermore, the adjustment unit also includes an adjustment box 408, which is fixedly connected to the drive rod 406 and located above the drive rod 406. The drive rod 406 drives the adjustment box 408 to achieve horizontal rotation adjustment within a range of 0° to 360°.

[0035] In addition, the adjustment unit also includes a dual-axis motor 409 and two drive shafts 410. The dual-axis motor 409 is fixedly connected to the adjustment box 408 and located inside the adjustment box 408. One end of the two drive shafts 410 is fixedly connected to the output end of the dual-axis motor 409, and the other end of the two drive shafts 410 is fixedly connected to the thermal imaging camera 411 and the high-definition visible light camera 412, respectively. The dual-axis motor 409 is disposed inside the adjustment box 408. The dual-axis motor 409 drives the drive shafts 410 to rotate. The drive shafts 410 drive the thermal imaging camera 411 and the high-definition visible light camera 412 to move within a vertical range of -90° to +90°, enabling large-area monitoring of targets. It supports up to 200 preset positions and can perform cruise scanning in both horizontal and vertical directions, achieving real-time monitoring of birds without blind spots.

[0036] This embodiment of a deep intelligent learning multispectral imaging and bird identification device, by setting up the solar power supply component 3, the dual-spectral Mini gimbal 4, the central processing unit 5, and the bird deterrent module 6, in specific use, the solar power supply component 3 is fixed above the fixed tower 2 to power the device, and can maintain the bird identification device's normal operation for at least 7 days under continuous conditions without sunlight and other power supplementation. The motor 403 drives the drive gear 404 to rotate, and the drive gear 404 interacts with the slave... The meshing of the driven gear 405 causes the driving gear 404 to drive the driven gear 405 to rotate, the driven gear 405 to drive the drive rod 406 to rotate, and the drive rod 406 to drive the adjustment box 408 to rotate and adjust, thereby driving the thermal imaging camera 411 and the high-definition visible light camera 412 to rotate and adjust horizontally within a range of 0° to 360°. The dual-axis motor 409 drives the drive shaft 410 to rotate, and the drive shaft 410 drives the thermal imaging camera 411 and the high-definition visible light camera 412 to rotate vertically within a range of -90° to +90°. With a 90° field of view, the system can monitor bird targets over a wide area. The thermal imaging camera 411 automatically adjusts image contrast to provide clear thermal images. The high-definition visible light camera 412 is equipped with a 25x optical zoom lens, automatically focuses, and provides clear visible light images, which are then transmitted to the central processing unit 5. The system accurately detects bird activity and determines its location, providing this information to the intelligent recognition module for AI identification. Once a bird is detected, the system immediately records dual-light video, documenting the entire process of bird detection, identification, shooing away, and flight. The laser is then driven according to the scheduling of the central processing unit 5. Bird repeller 602, blue light bird repeller 603, and ultrasonic bird repeller 604, individually or together, can drive away birds. By integrating high-resolution cameras and image processing algorithms, intelligent recognition technology can capture and analyze the flight trajectory, posture, and species of birds in real time. Through the analysis of data such as bird activity patterns, environmental factors, and the working status of bird repeller devices, key factors affecting the bird repeller effect can be identified, and bird repeller strategies can be adjusted to continuously optimize the bird repeller effect. This achieves intelligent bird recognition, reduces false triggering and ineffective repellering, and improves the efficiency and accuracy of bird repeller devices.

[0037] The above description discloses only one preferred embodiment of the present invention, and should not be construed as limiting the scope of the present invention. Those skilled in the art will understand that all or part of the processes of the above embodiments can be implemented, and equivalent changes made in accordance with the claims of the present invention are still within the scope of the invention.

Claims

1. A deep intelligent learning multispectral imaging and bird identification device, comprising a mounting tower, a fixed tower frame disposed on the outer side of the mounting tower, a solar power supply component disposed on the outer side of the fixed tower frame, a dual-spectral mini gimbal disposed above the mounting tower, a central processing unit disposed inside the dual-spectral mini gimbal, and a bird deterrent module disposed on the outer side of the dual-spectral mini gimbal, characterized in that, The dual-spectrum Mini gimbal includes an adjustment unit, a thermal imaging camera, and a high-definition visible light camera. The adjustment unit is fixedly connected to the mounting tower and located above the mounting tower. The thermal imaging camera is fixedly connected to the adjustment unit and located outside the adjustment unit. The high-definition visible light camera is fixedly connected to the adjustment unit and located on the side of the thermal imaging camera away from the adjustment unit. The bird deterrent module includes a protective shell, a laser bird deterrent, a blue light bird deterrent, and an ultrasonic bird deterrent. The protective shell is fixedly connected to the dual-spectrum Mini gimbal and is located below the laser bird deterrent. The laser bird deterrent is fixedly connected to the protective shell and is located inside the protective shell. The blue light bird deterrent is detachably connected to the dual-spectrum Mini gimbal and is located below the high-definition visible light camera. The ultrasonic bird deterrent is fixedly connected to the dual-spectrum Mini gimbal and is located above the dual-spectrum Mini gimbal.

2. The deep intelligent learning multispectral imaging and bird identification device as described in claim 1, characterized in that, The solar power supply component includes a mounting bracket and a hardware clamp. The mounting bracket is detachably connected to the fixed tower and is located above the fixed tower. The hardware clamp is fixedly connected to the mounting bracket and is located on the outside of the fixed tower.

3. The deep intelligent learning multispectral imaging and bird identification device as described in claim 2, characterized in that, The solar power supply assembly also includes a mounting sleeve and a solar cell array. The mounting sleeve is fixedly connected to the mounting bracket and is located outside the mounting bracket. The solar cell array is fixedly connected to the mounting sleeve and is located inside the mounting sleeve.

4. The deep intelligent learning multispectral imaging and bird identification device as described in claim 3, characterized in that, The solar power supply assembly also includes a solar panel and a battery. The solar panel is electrically connected to the solar cell array and is located outside the solar cell array. The battery is electrically connected to the solar cell array and is located outside the mounting sleeve.

5. The deep intelligent learning multispectral imaging and bird identification device as described in claim 4, characterized in that, The adjustment unit includes a housing and a partition plate. The housing is fixedly connected to the mounting tower and located above the mounting tower. The partition plate is fixedly connected to the housing and located inside the housing.

6. The deep intelligent learning multispectral imaging and bird identification device as described in claim 5, characterized in that, The adjustment unit also includes a motor, a drive gear, and a driven gear. The output end of the motor is fixedly connected to the drive gear and is located above the partition plate. The drive gear meshes with the driven gear and is located below the partition plate. The driven gear is rotatably connected to the housing and is located outside the drive gear.

7. The deep intelligent learning multispectral imaging and bird identification device as described in claim 6, characterized in that, The adjustment part further includes a drive rod and a bushing. The drive rod is fixedly connected to the driven gear and passes through the housing. The bushing is fixedly connected to the housing and is located outside the drive rod.

8. The deep intelligent learning multispectral imaging and bird identification device as described in claim 7, characterized in that, The adjustment unit also includes an adjustment box, which is fixedly connected to the drive rod and located above the drive rod.

9. The deep intelligent learning multispectral imaging and bird identification device as described in claim 8, characterized in that, The adjustment unit also includes a dual-axis motor and two drive shafts. The dual-axis motor is fixedly connected to the adjustment box and located inside the adjustment box. One end of the two drive shafts is fixedly connected to the output end of the dual-axis motor, and the other end of the two drive shafts is fixedly connected to the thermal imaging camera and the high-definition visible light camera, respectively.

10. A deep intelligent learning-based multispectral imaging and bird identification method, characterized in that, The multispectral imaging and bird identification device for deep intelligent learning as described in claim 9, S1. Install the device in the target area using the fixed tower and the hardware clamps, start the solar power supply component, and ensure that the battery can maintain the continuous operation of the equipment for 7 days under no-light conditions; adjust the dual-spectrum Mini gimbal to the initial monitoring angle so that it can rotate 360° horizontally and adjust vertically within the range of -90° to +90°, and the thermal imaging camera and the high-definition visible light camera enter the working state. S2. Real-time acquisition of dual-spectrum images: The thermal imaging camera optimizes the imaging algorithm for the isothermal characteristics of birds, automatically adjusts the contrast, and captures clear thermal images in low-visibility environments; the high-definition visible light camera simultaneously acquires visible light images, and the data from both are transmitted to the central processing unit in real time. S3.AI Intelligent Recognition and Positioning: The central processing unit calls the YOLOv8n-seg deep learning model to perform fusion analysis on the dual-spectral images, accurately identify bird targets and distinguish their species, flight trajectories and attitudes, eliminate interference from birds and insects, determine the bird's location coordinates, and simultaneously trigger dual-light video recording to record the entire process from detection to flight away; S4. Dynamic Bird Repelling Strategy Execution: Based on the recognition results, the central processing unit schedules and integrates the bird repelling modules, adopting a three-dimensional collaborative strategy of "laser + blue light + ultrasound": green lasers work in conjunction with dynamic scanning by the gimbal to form a physical deterrent; the laser bird repeller and the blue light bird repeller emit specific wavelengths of light to stimulate the birds' vision; and the ultrasound bird repeller releases sound waves in frequencies sensitive to birds. Depending on the bird species and behavioral characteristics, bird repelling methods are activated individually or in combination to prevent birds from developing adaptation. S5. Data Feedback and Strategy Optimization: The back-end control system receives device operation data via wireless network, including bird activity patterns, bird deterrence effects, and environmental parameters. It uses reinforcement learning algorithms to analyze historical data and automatically adjusts laser scanning angle, sound wave frequency, and bird deterrence timing parameters. It also supports remote control, allowing users to view logs and set working periods through a mobile app or the back-end system to continuously optimize bird deterrence strategies.

Images