Intelligent bird repelling robot

Abstract

The invention provides an intelligent bird repelling robot, and relates to the technical field of inspection robots, the intelligent bird repelling robot comprises a hardware part and a software part of the intelligent bird repelling robot, the hardware part comprises a main control system, a motion chassis and a bird repelling holder, and the software part comprises a detection algorithm and a control strategy. A main control system in the hardware provides a communication interface for the motion chassis and the bird repelling holder to realize cooperative control, a detection algorithm in the software provides bird target information, and a control strategy drives the motion chassis and the bird repelling holder according to a detection result to emit laser and ultrasound to repel birds. Through the synergistic effect of hardware and software, full-map bird real-time detection and repelling can be effectively realized, and an effective solution is provided for mobile intelligent bird repelling.

Description

Technical Field

[0001] The present invention relates to the field of intelligent inspection robots, and in particular to an intelligent bird-repelling robot. Background Art

[0002] At present, the main bird repellent technologies include shape repellent, sound repellent, ultrasonic repellent and laser repellent. Shape repellent uses the instinctive fear of birds to natural enemies and humans, and arranges dummies, ribbons, eagles, bird-proof spikes and other devices in farmlands or orchards to achieve the purpose of repelling birds. However, birds are easily adapted to the above devices, so the effectiveness of shape repellent is poor. Sound repellent drives birds away by emitting sounds of specific frequencies or rhythms, imitating the calls of natural enemies or producing unpleasant noises. Since noise can also affect human health, sound repellent is usually only used in sparsely populated open areas. Ultrasonic bird repellent uses ultrasonic signals of 16KHz to 25KHz to stimulate the nervous system and physiological system of birds, and scares birds by constantly changing the emission frequency to achieve the effect of repelling them. Laser bird repellent emits a green laser beam with a wavelength of 532nm to illuminate birds. When birds are illuminated by the moving laser beam, they will be frightened and flee. Laser bird repellent (also known as bird repellent pan / tilt) is generally built into the pan / tilt camera, and a scanning path must be preset when used. Since the PTZ camera is fixed, the range of the laser bird repellent is limited by the effective distance of the laser beam. Buildings in the environment may block the laser. For large sites such as airports and substations, the equipment must be deployed at multiple angles to achieve full coverage. In addition, the laser bird repellent utilizes the dual-axis digital control function of the PTZ camera, but does not effectively utilize the image acquisition function of the PTZ camera, and lacks the ability to intelligently identify and automatically repel birds. Summary of the invention

[0003] In view of the existing problems, the purpose of the present invention is to provide an intelligent bird-repellent robot, which can effectively realize real-time detection and repelling of birds on the entire map through the synergy of hardware and software, providing an effective solution for mobile intelligent bird repellent.

[0004] In order to realize the above technical features, the object of the present invention is achieved as follows: an intelligent bird-repelling robot, the intelligent bird-repelling robot is composed of a hardware part and a software part, the hardware part includes a main control system, a motion chassis and a bird-repelling pan-tilt platform; The software part includes a detection algorithm and a control strategy. The main control system in the hardware part provides a communication interface for the motion chassis and the bird-repelling pan-tilt platform to achieve coordinated control. The detection algorithm in the software part provides bird target information. The control strategy drives the motion chassis and the bird-repelling pan-tilt platform according to the detection results, and emits laser and ultrasound to drive away the birds. Through the synergistic effect of the hardware part and the software part, real-time detection and expulsion of birds in the entire map can be achieved.

[0005] Preferably, the main control system adopts Jeston Orin NX as the main control platform. The main control system provides a communication interface for the motion chassis and the bird-repelling gimbal to achieve coordinated control, and provides computing power for the implementation of detection algorithms and control strategies.

[0006] Preferably, the bird-repelling pan-tilt is connected to the main control system via an RJ-45 network interface, supports the Onvif control protocol, and is respectively equipped with an infrared camera module, a visible light camera module, a laser module, a first ultrasonic module, and a second ultrasonic module. The bird-repelling pan-tilt can rotate 0-360° around the vertical axis and ±90° around the horizontal axis, and can quickly collect video streams within the inspection area. When the visible light camera module captures the image of the bird, the control unit of the main control system controls the laser module, the first ultrasonic module, and the second ultrasonic module to work, so that the laser module, the first ultrasonic module, and the second ultrasonic module emit laser beams and ultrasonic waves in the direction of the birds to achieve the bird-repelling effect.

[0007] Preferably, the motion chassis is equipped with an STM32 single-chip microcomputer, a laser radar and a GNSS module. The STM32 single-chip microcomputer receives a signal from the main control system through a USB serial port to drive the first steering servo, the second steering servo, the first drive motor and the second drive motor, and cooperates with the three-dimensional point cloud map of the main control system and the robot position information to realize full-map unmanned driving.

[0008] Preferably, the bird detection algorithm adopts an improved YOLOv5 neural network model to read the visible light camera module video stream for real-time reasoning to determine whether there are birds in the inspection area of ​​the bird-repellent robot; by integrating the TPH module, the ultra-small target prediction head and the CBAM attention mechanism module, high-precision recognition of a large number of densely arranged small-volume bird flocks can be achieved.

[0009] Preferably, the control strategy includes four threads that execute simultaneously: (1) Bird-repelling PTZ inspection thread: Based on the Onvif control protocol, the bird-repelling PTZ is driven to rotate 360° around the vertical axis, and the visible light camera module 12 collects the video stream of the surrounding environment of the bird-repelling robot; (2) Target detection thread: The YOLOv5 network model reads the video stream collected by the visible light camera module and performs real-time inference to determine whether there are birds in the environment. If birds are found, the location information of the detected birds is passed to the PID control thread and the bird-repelling PTZ inspection thread is interrupted. (3) PID control thread: Based on the position information of the target detection thread, the error between the image center of the bird-repelling gimbal and the center of the detected bird target is calculated, and the bird-repelling gimbal is driven to move to the center of the detected bird target based on the Onvif control protocol; (4) Bird repelling thread: Receives the detection results of the target detection thread and simultaneously turns on the laser module, the first ultrasonic module and the second ultrasonic module to repel the detected birds.

[0010] Preferably, in the bird-repelling pan-tilt inspection thread, the bird-repelling pan-tilt pauses for 1 second every time it rotates 20 degrees around the vertical axis to ensure that the visible light camera module can obtain a clear video stream.

[0011] Preferably, the bird bounding box in the target detection thread is enlarged to 1.2 times to provide a certain allowable error for bird target tracking in the subsequent PID control thread.

[0012] Preferably, the video stream image height of the bird-repelling pan-tilt visible light camera is H ,Width W , the coordinates of the upper left corner of the prediction box are ( x 1 , y 1 ), the coordinates of the upper left corner are ( x 2 ,y 2 ), the target value of PID control is the coordinate of the center point of the image ( ), the error value from the center point of the prediction box to the center point of the image can be calculated as: ; ; in, They are the horizontal error value and vertical error value from the center point of the prediction box to the center point of the image respectively; The control quantity can be calculated according to the PID calculation formula; ; in, Output control quantity for PID control algorithm, is the error between the bird target and the center of the video, is the proportional control coefficient, is the integral control coefficient, is the differential control coefficient.

[0013] Preferably, the initial PID parameters in the PID control thread are =0.1, =0.1, =0.1, based on the position information of the target detection thread, calculate the proportion of the bird boundary box magnified 1.2 times in the picture, and dynamically adjust the PID parameters according to the proportion of the recognition box in the picture; (1) When the recognition frame area accounts for less than 10% of the screen, increase the PID Parameters increased by 50%; (2) When the recognition frame area accounts for 10% to 30% of the screen, keep the current PID parameters unchanged; (3) When the recognition frame area accounts for more than 30% of the screen, reduce the PID The parameter is reduced by 40%, and the integral parameter is increased to 50%.

[0014] The present invention has the following beneficial effects: Through the synergy of hardware and software, the present invention can effectively realize real-time detection and expulsion of birds on the entire map, providing an effective solution for mobile intelligent bird repellent. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0016] Figure 1 The figure is a schematic structural diagram of an intelligent bird-repelling person and an intelligent bird-repelling robot according to a control strategy of the present invention.

[0017] Figure 2 The figure is a schematic diagram of a bird-repelling robot motion chassis of an intelligent bird-repelling person and control strategy of the present invention.

[0018] Figure 3 The present invention provides a simplified flow chart of the multi-threaded control strategy of an intelligent bird-repellent person and control strategy.

[0019] Figure 4 The figure is a calculation diagram of the PID control thread of an intelligent bird-repellent person and control strategy of the present invention.

[0020] Figure 5 This is a detection result diagram of the target detection algorithm of an intelligent bird-repellent person and control strategy of the present invention.

[0021] Figure 6 This is a detection result diagram of the target detection algorithm of an intelligent bird-repellent person and control strategy of the present invention.

[0022] In the figure, 1. bird-repelling gimbal; 11. first ultrasonic module; 12. visible light camera module; 13. laser module; 14. infrared camera module; 15. second ultrasonic module; 2. motion chassis; 21. laser radar; 221. first steering servo; 222. second steering servo; 231. first drive motor; 232. second drive motor; 24. GNSS module; 25. STM32 microcontroller; 3. main control system. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0024] An intelligent bird-repelling robot and control strategy, such as Figure 1-Figure 2 The hardware part of the intelligent bird-repelling robot shown includes: a main control system 3, a bird-repelling platform 1 and a motion chassis 2.

[0025] The main control system 3 uses Jeston Orin NX as the main control platform. The main control system 3 provides a communication interface for the motion chassis 2 and the bird-repelling gimbal 1 to achieve collaborative control, and provides computing power support for the implementation of detection algorithms and control strategies.

[0026] The bird-repelling pan-tilt 1 is connected to the main control system 3 via an RJ-45 network interface (not shown in the figure), supports the Onvif control protocol, and is equipped with an infrared camera module 14, a visible light camera module 12, a laser module 13, a first ultrasonic module 11, and a second ultrasonic module 15. The bird-repelling pan-tilt 1 can rotate around the vertical axis 0-360° and the horizontal axis ±90°, and can quickly collect video streams within the inspection area. When the visible light camera module 12 captures a bird image, the laser module 13, the first ultrasonic module 11, and the second ultrasonic module 15 can be controlled by the control unit to work, so that the laser module 13, the first ultrasonic module 11, and the second ultrasonic module 15 emit laser beams and ultrasonic waves toward the direction of the bird to achieve the bird-repelling effect.

[0027] In specific implementation, the laser module 13 emits a green laser beam with a wavelength of 532nm to illuminate the birds, and the birds will be frightened and flee when illuminated by the moving laser beam. At the same time, the first ultrasonic module 11 and the second ultrasonic module 15 emit 16KHz to 25KHz constantly changing ultrasonic signals to stimulate the nerves and physiological systems of the birds, thereby achieving the effect of driving away birds.

[0028] The motion chassis 2 is equipped with an STM32 single-chip microcomputer 25, a laser radar 21 and a GNSS module 24. The STM32 single-chip microcomputer 25 receives signals from the main control system 3 through a USB serial port to drive the first steering servo 221, the second steering servo 222, the first drive motor 231 and the second drive motor 232, and cooperates with the three-dimensional point cloud map of the main control system 3 and the robot position information to realize full-map unmanned driving.

[0029] The bird detection algorithm uses an improved YOLOv5 neural network model to read the video stream of the visible light camera module 12 for real-time reasoning to determine whether there are birds in the inspection area of ​​the bird-repelling robot. Figure 5-Figure 6 As shown in the figure, by integrating the TPH module, the ultra-small target prediction head and the CBAM attention mechanism module, high-precision recognition of a large number of densely arranged small-volume bird flocks is achieved.

[0030] like Figure 3 As shown in the figure, the bird-repelling robot's bird-repelling control strategy contains four threads that execute simultaneously: (1) Bird-repellent PTZ inspection thread: Based on the Onvif control protocol, the bird-repellent PTZ 1 is driven to rotate 360° around the vertical axis, and the visible light camera module 12 collects the video stream of the environment around the bird-repellent robot; (2) Target detection thread: The YOLOv5 network model reads the video stream collected by the visible light camera module 12 and performs real-time reasoning to determine whether there are birds in the environment. If birds are found, the location information of the detected birds is passed to the PID control thread and the bird-repelling PTZ inspection thread is interrupted; (3) PID control thread: Based on the position information of the target detection thread, the error between the image center of the bird-repelling gimbal 1 and the center of the detected bird target is calculated, and the bird-repelling gimbal is driven to move to the center of the detected bird target based on the Onvif control protocol; (4) Bird driving away thread: receiving the detection result of the target detection thread, and simultaneously turning on the laser module 13, the first ultrasonic module 11 and the second ultrasonic module 15 to drive away the detected birds.

[0031] In the bird-repelling pan-tilt inspection thread, the bird-repelling pan-tilt 1 pauses for 1 second every time it rotates 20 degrees around the vertical axis to ensure that the visible light camera module 12 can obtain a clear video stream.

[0032] The bird bounding box in the target detection thread is enlarged to 1.2 times to provide a certain allowable error for the subsequent bird target tracking of the PID control thread.

[0033] like Figure 4 As shown, the video stream image height of the visible light camera module 12 is H ,Width W , the coordinates of the upper left corner of the prediction box are ( x 1 , y 1 ), the coordinates of the upper left corner are ( x 2 ,y 2 ). The target value of PID control is the coordinate of the center point of the image ( ), the error value from the center point of the prediction box to the center point of the image can be calculated as: ; ; in, They are the horizontal error and vertical error from the center of the prediction box to the center of the image.

[0034] The control quantity can be calculated according to the PID calculation formula.

[0035] ; in, Output control quantity for PID control algorithm, is the error between the bird target and the center of the video, is the proportional control coefficient, is the integral control coefficient, is the differential control coefficient.

[0036] The initial PID parameters in the PID control thread are =0.1, =0.1, =0.1, based on the position information of the target detection thread, calculate the proportion of the bird boundary box magnified 1.2 times in the picture, and dynamically adjust the PID parameters according to the proportion of the recognition box in the picture.

[0037] (1) When the recognition frame area accounts for less than 10% of the screen, increase the PID Parameter to 0.15; (2) When the recognition frame area accounts for 10% to 30% of the screen, keep the current PID parameters unchanged; (3) When the recognition frame area accounts for more than 30% of the screen, reduce the PID Parameter to 0.06, increase Parameter to 0.15.

[0038] The above description is only a specific implementation of the present application, so that those skilled in the art can understand or implement the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest range consistent with the principles and novel features applied for herein.

Claims

1. An intelligent bird-repelling robot, characterized in that: The intelligent bird-repelling robot is composed of a hardware part and a software part, wherein the hardware part includes a main control system, a motion chassis and a bird-repelling pan-tilt platform; The software part includes a detection algorithm and a control strategy. The main control system in the hardware part provides a communication interface for the motion chassis and the bird-repelling pan-tilt platform to achieve coordinated control. The detection algorithm in the software part provides bird target information. The control strategy drives the motion chassis and the bird-repelling pan-tilt platform according to the detection results, and emits laser and ultrasound to drive away the birds. Through the synergistic effect of the hardware part and the software part, real-time detection and expulsion of birds in the entire map can be achieved.

2. The intelligent bird-repelling robot according to claim 1, characterized in that: The main control system uses JestonOrin NX as the main control platform. The main control system provides a communication interface for the motion chassis and the bird-repelling gimbal to achieve collaborative control, and provides computing power for the implementation of detection algorithms and control strategies.

3. The intelligent bird-repelling robot according to claim 1, characterized in that: The bird-repelling pan-tilt is connected to the main control system via an RJ-45 network interface, supports the Onvif control protocol, and is respectively equipped with an infrared camera module, a visible light camera module, a laser module, a first ultrasonic module, and a second ultrasonic module. The bird-repelling pan-tilt can rotate 0-360° around the vertical axis and ±90° around the horizontal axis, and can quickly collect video streams within the inspection area. When the visible light camera module captures a bird image, the laser module, the first ultrasonic module, and the second ultrasonic module are controlled by the control unit of the main control system to work, so that the laser module, the first ultrasonic module, and the second ultrasonic module emit laser beams and ultrasonic waves in the direction of the birds to achieve the bird-repelling effect.

4. The intelligent bird-repelling robot according to claim 3, characterized in that: The sports chassis is equipped with an STM32 single-chip microcomputer, a laser radar and a GNSS module. The STM32 single-chip microcomputer receives a signal from the main control system through a USB serial port to drive the first steering servo, the second steering servo, the first drive motor and the second drive motor, and cooperates with the three-dimensional point cloud map of the main control system and the robot position information to realize full-map unmanned driving.

5. The intelligent bird-repelling robot according to claim 4, characterized in that: The bird detection algorithm adopts an improved YOLOv5 neural network model to read the visible light camera module video stream for real-time reasoning to determine whether there are birds in the inspection area of ​​the bird-repellent robot; by integrating the TPH module, the ultra-small target prediction head and the CBAM attention mechanism module, high-precision recognition of a large number of densely arranged small-volume bird flocks can be achieved.

6. The intelligent bird-repelling robot according to claim 5, characterized in that: The control strategy consists of four threads that execute simultaneously: (1) Bird-repelling PTZ inspection thread: Based on the Onvif control protocol, the bird-repelling PTZ is driven to rotate 360° around the vertical axis, and the visible light camera module 12 collects the video stream of the surrounding environment of the bird-repelling robot; (2) Target detection thread: The YOLOv5 network model reads the video stream collected by the visible light camera module and performs real-time inference to determine whether there are birds in the environment. If birds are found, the location information of the detected birds is passed to the PID control thread and the bird-repelling PTZ inspection thread is interrupted. (3) PID control thread: Based on the position information of the target detection thread, the error between the image center of the bird-repelling gimbal and the center of the detected bird target is calculated, and the bird-repelling gimbal is driven to move to the center of the detected bird target based on the Onvif control protocol; (4) Bird repelling thread: Receives the detection results of the target detection thread and simultaneously turns on the laser module, the first ultrasonic module and the second ultrasonic module to repel the detected birds.

7. The intelligent bird-repelling robot according to claim 6, characterized in that: In the bird-scaring PTZ inspection thread, the bird-scaring PTZ pauses for 1 second every 20 degrees of rotation around the vertical axis to ensure that the visible light camera module can obtain a clear video stream.

8. The intelligent bird-repelling robot according to claim 6, characterized in that: The bird bounding box in the target detection thread is enlarged to 1.2 times to provide a certain allowable error for the bird target tracking in the subsequent PID control thread.

9. The intelligent bird-repelling robot according to claim 6, characterized in that: The video stream image of the bird-repelling PTZ visible light camera is H ,Width W , the coordinates of the upper left corner of the prediction box are ( x 1 ,y 1), the coordinates of the upper left corner are ( x 2 ,y 2), the target value of PID control is the coordinate of the center point of the image ( ), the error value from the center point of the prediction box to the center point of the image can be calculated as: ； ； in, They are the horizontal error value and vertical error value from the center point of the prediction box to the center point of the image respectively; The control quantity can be calculated according to the PID calculation formula; ； in, Output control quantity for PID control algorithm, is the error between the bird target and the center of the video, is the proportional control coefficient, is the integral control coefficient, is the differential control coefficient.

10. The intelligent bird-repelling robot according to claim 6, characterized in that: The initial PID parameters in the PID control thread are =0.1, =0.1, =0.1, based on the position information of the target detection thread, calculate the proportion of the bird boundary box magnified 1.2 times in the picture, and dynamically adjust the PID parameters according to the proportion of the recognition box in the picture; (1) When the recognition frame area accounts for less than 10% of the screen, increase the PID Parameters increased by 50%; (2) When the recognition frame area accounts for 10% to 30% of the screen, keep the current PID parameters unchanged; (3) When the recognition frame area accounts for more than 30% of the screen, reduce the PID The parameter is reduced by 40%, and the integral parameter is increased to 50%.

Images