An unmanned aerial vehicle autonomous removal device and method suitable for power line bird nests

Abstract

The application discloses a kind of unmanned aerial vehicle autonomous removal device and method suitable for transmission line bird nest, including unmanned aerial vehicle platform and the bird nest removal device being arranged below unmanned aerial vehicle platform, bird nest removal device includes mechanical arm suspension rod, balance component, motor fixing mechanism, mechanical arm motor, motor rotating mechanism, connecting rod and mechanical gripper, mechanical arm suspension rod is arranged at the bottom of unmanned aerial vehicle platform, and mechanical arm motor is fixed on it by motor fixing mechanism, mechanical arm motor is connected with connecting rod by motor rotating mechanism, and the end of connecting rod is provided with mechanical gripper, balance component is arranged on mechanical arm suspension rod, located the opposite side of mechanical arm motor and mechanical gripper, so that unmanned aerial vehicle platform keeps balance when mechanical gripper works.The present application can realize stubborn bird damage on line tower unmanned intelligent identification and accurate removal.

Description

Technical Field

[0001] The present invention relates to an autonomous removal device and method of a drone for clearing bird nests on power transmission lines, which are used in the field of power transmission line obstacle removal. Background Art

[0002] In recent years, with the continuous improvement of my country's natural environment, the bird population has increased year by year. In particular, large birds of prey such as ospreys, hawks, and falcons prefer to build nests on power transmission lines, significantly increasing the risk of short circuits and damage. Furthermore, the presence of bird nests on power lines increases the risk of leakage during rainy days. Therefore, developing effective devices to remove bird nests from power lines and minimize their impact is crucial.

[0003] Traditional manual cleaning of bird nests on power lines is the simplest and most direct method, but it requires manual labor at height, presents safety risks, and is inefficient. Using high-pressure water jets from high-pressure water guns to clean bird nests is fast and efficient, clearing obstructions. However, this requires a drone-mounted device to provide sufficient water pressure and volume at high altitude, which is challenging, consumes water resources, and may impact the surrounding environment and equipment. Drop-and-pull drone cleaning methods are portable and easy to assemble and disassemble, offering promising application prospects. However, this method requires precise drop accuracy, requiring multiple drops to completely remove a bird nest. If a nest becomes lodged in the steel frame of a power tower, this significantly increases the risk of the operation. Laser technology can be used to cut or burn bird nests, offering precise control and remote operation. However, since these devices require the nest to be dry during ignition, they cannot be used when the nest is wet, such as during rainy days. Furthermore, a fire could pose a significant risk to surrounding power equipment and the environment. While drone-mounted robotic grippers exist for bird nest removal, current grippers lack vertical movement, limited by the drone's limited attitude control and autonomous operation capabilities. Summary of the Invention

[0004] The purpose of the present invention is to overcome the shortcomings of the existing technology and provide a drone autonomous removal device and method suitable for bird nests on transmission lines. Based on airborne multi-degree-of-freedom mechanical grasping technology and AI image recognition means, it can realize unmanned intelligent identification and precise removal of stubborn bird pests on line towers, which can greatly reduce the workload of manual inspections. In the future, it can also be extended to the early detection and rapid treatment of hazards caused by foreign objects such as kites, greenhouse plastics, balloons, etc. on lines that are easier to identify, thereby comprehensively ensuring the safe operation of the power grid.

[0005] A technical solution to achieve the above-mentioned purpose is: a drone autonomous removal device for bird nests on transmission lines, including a drone platform and a bird nest removal device arranged under the drone platform.

[0006] The bird's nest removal device includes a robotic arm suspension rod, a balancing assembly, a motor fixing mechanism, a robotic arm motor, a motor rotation mechanism, a connecting rod and a robotic claw. The robotic arm suspension rod is arranged at the bottom of the UAV platform, and the robotic arm motor is fixed thereon through the motor fixing mechanism. The robotic arm motor is connected to the connecting rod through the motor rotation mechanism. A robotic claw is provided at the end of the connecting rod. The balancing assembly is arranged on the robotic arm suspension rod and is located on the opposite side of the robotic arm motor and the robotic claw, so that the UAV platform remains balanced when the robotic claw is operating.

[0007] Furthermore, the balancing assembly includes a counterweight, a servo screw and a servo motor. The counterweight is connected to the servo motor through the servo screw and moves horizontally with the help of the screw under the action of the servo motor. The servo motor is connected to the robotic arm motor to read the torque of the robotic arm motor.

[0008] Furthermore, a camera is provided below the UAV platform on one side of the robotic arm motor and the robotic gripper.

[0009] Furthermore, the UAV platform is an X-shaped six-rotor UAV, and the robotic arm suspension rods are arranged front to back along the tripod direction of the UAV platform.

[0010] Further, the following steps are included:

[0011] Step 1: The drone platform equipped with the bird nest removal device approaches the transmission line tower where the bird nest is located. During the approach, the drone platform uses a camera to monitor the bird nest.

[0012] Step 2: The drone platform hovers near the bird's nest, and the camera captures the nest image. The target detection algorithm is used to obtain the position information of the nest relative to the drone.

[0013] Step 3: Calculate the movement of the drone and robotic arm based on the pose information, control the movement of the drone and robotic arm, and extend the robotic claw to the bird's nest;

[0014] Step 4: The robotic claws close to grab the bird's nest, and the drone and robotic arm retract;

[0015] Step 5: Use the camera to detect the bird's nest. If the nest is cleaned, leave the work area. If there are any remnants of the nest, repeat steps 2 to 5.

[0016] Furthermore, in step 1, the UAV platform is equipped with a high-precision positioning and orientation system to accurately obtain the position and control the flight. The onboard computer of the UAV platform sends the coordinates of the tower that needs to clear the bird's nest to the UAV flight control. The flight control controls the UAV to fly to the corresponding position and hover stably. During this process, the onboard computer uses the YOLO v8 target detection algorithm to detect the bird's nest obtained by the camera to ensure that the bird's nest is always in the field of view. When the bird's nest deviates to one side of the field of view, the onboard computer will send a correction instruction to the flight control to make the bird's nest return to the field of view close to the center. This feedback control method uses the PID control algorithm.

[0017] Furthermore, step 2 specifically includes:

[0018] Step 2.1: The drone hovers near the bird's nest, obtains high-precision pose information of the current drone, takes the drone's current position as the origin, and establishes a local rectangular coordinate system o-xyz. W is the camera's heading angle in the local coordinate system o-xyz.

[0019] Step 2.2: Take a picture and use YOLO v8 for target detection. The target detection result is the pixel coordinates of the top left, top right, bottom left, and bottom right vertices of the bird's nest image in the two-dimensional image. These four points constrain the position of the bird's nest image in the camera's field of view.

[0020] In step 2.3, based on the pixel coordinates, camera parameters, and camera pose information, the camera's pinhole imaging model can be used to determine the direction vectors of the lines containing the four vertices in the local rectangular coordinate system o-xyz. Point A is the camera's optical center, and point C is a point in the camera's field of view. Given the pixel coordinates of point C and the camera parameters, and AB being the optical axis direction of the drone's camera, the direction vector of line L containing point AC can be obtained.

[0021] In step 2.4, use the Depth Anything V2 model to analyze the image and obtain the image's depth information, specifically the distance from point C to the camera's optical center, point A. By combining the direction of vector AC, we can estimate the coordinates of point C in the o-xyz rectangular coordinate system. Similarly, we can obtain the positions of the four vertices in the target detection results of the bird's nest image and calculate the nest's position relative to the drone in space.

[0022] Furthermore, step 3 is specifically as follows:

[0023] For the task of grabbing a bird's nest with a drone-mounted robotic arm, the end-effector gripper requires at least five degrees of freedom: three positional degrees of freedom and two angular degrees of freedom: heading and pitch. The three positional degrees of freedom and heading are provided by the drone body, meaning the drone flies to a specified coordinate point and hovers in a specified heading; the pitch angle is provided by the robotic arm; the drone remains level, and the angle of the robotic arm is read by the motor encoder.

[0024] At the same time, in order to balance the center of gravity of the drone, the counterweight needs to be moved when the robotic arm moves. When the drone takes off and is balanced, the robotic arm rotates the motor to adjust the center of gravity of the drone. When the robotic arm is unfolded, the torque applied to the robotic arm's rotating motor is read, recorded as M, and the gravity applied to the counterweight is G. The required movement distance L = M / G. After moving the counterweight, the center of gravity of the drone can be brought close to the center.

[0025] Furthermore, step 4 is specifically as follows: after the mechanical claws close and grab the bird's nest, the drone and the mechanical arm take the bird's nest away from the tower. During the movement of the mechanical arm, the torque applied to the motor is calculated in real time to move the balance block.

[0026] Furthermore, the drone returns to the hovering position in step 2, takes a photo of the bird's nest, and calls the YOLO v8 target detection algorithm again to detect the bird's nest. If the bird's nest cannot be detected, it leaves the operation area. If there are any remnants of the bird's nest, steps 2 to 5 are repeated.

[0027] The advantages of the present invention are:

[0028] 1. By adopting a vertically movable robotic arm, the problem of the robotic claw locating bird nests in complex locations is solved. Combined with the drone's own horizontal freedom, it can grasp bird nests at different angles, achieving the purpose of accurately grasping bird nests.

[0029] 2. By carrying the counterweight of the motor, the imbalance problem of the drone when the robotic arm moves is solved, so that the center of gravity of the drone when grabbing the bird's nest is always kept in the center of the device, increasing the stability of the drone.

[0030] 3. By adopting a double-headed opening and closing mechanical claw, the disadvantageous situation of the claw getting stuck when grabbing the bird's nest is avoided, and the risk of the drone falling is reduced.

[0031] 4. Artificial intelligence image recognition technology and machine learning algorithms are used to process images captured by high-definition cameras on large-load drone platforms to accurately identify hidden bird nests, reducing the workload of manual alignment and improving work efficiency. The relative position of the mechanical claw and the bird nest is continuously corrected in a dynamic environment, achieving precise control in a dynamic environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1This is a schematic structural diagram of a drone autonomous removal device for bird nests on power transmission lines according to the present invention;

[0033] Figure 2 This is a schematic structural diagram of a bird's nest removal device of an autonomous drone-assisted bird's nest removal device suitable for use on power transmission lines according to the present invention;

[0034] Figure 3 A spatial diagram of the object detection algorithm for bird nest images. DETAILED DESCRIPTION

[0035] In order to better understand the technical solution of the present invention, the following is a detailed description through specific embodiments:

[0036] See also Figure 1 and Figure 2 The present invention provides an autonomous drone removal device for bird nests on power transmission lines, comprising a drone platform 1 and a bird nest removal device arranged below the drone platform.

[0037] The UAV platform 1 is composed of a multi-rotor UAV equipped with a bird nest removal device. It can accurately move to the designated location according to the instructions and hover stably, providing a mobile and stable working platform for other components. Figure 1 The X-shaped hexacopter drone in the image above has its robotic arms arranged forward and backward, along the platform's legs. The drone is controlled using a Pixhawk flight controller and equipped with a UM982-based GNSS module. This module not only provides a positioning error of less than 2.5 cm, but its dual-antenna design also provides a heading error of less than 0.1 degrees. A camera 6 is located below the platform, next to the robotic arm's motor and gripper.

[0038] The bird's nest removal device, consisting of a robotic arm suspension rod 3, a balancing assembly, a motor mounting mechanism 4, a robotic arm motor 5, a motor rotation mechanism 7, a connecting rod 8, and a robotic gripper 9, serves as the system's primary actuator. The motor mounting mechanism, motor, and connecting rod adjust the gripper's extension angle, enabling it to coordinate with the drone's movements. The gripper, the end actuator, is responsible for grasping the bird's nest. The robotic arm suspension rod and balancing assembly are used to balance the drone's flight posture. A drone's center of gravity is sensitive in mid-air, and relying solely on propeller adjustment is insufficient for stable operation. Therefore, a balancing device is employed to maintain the drone's center of gravity as close to its geometric center as possible during robotic arm movement. In this example, an integrated servo motor is employed. Its electric drive not only provides stable operation control but also accurately measures torque. The balancing device utilizes a servo screw structure to precisely move the counterweight 2.

[0039] The onboard computer system includes the onboard computer, camera, and communication module. The onboard computer and communication module are located inside the aircraft. The onboard computer is responsible for collecting camera data, performing target detection and motion planning, and sending control commands to the robotic arm, gripper, flight control system, and other components, controlling the movement of the aircraft, robotic arm, gripper, and other components according to the corresponding algorithms. The onboard computer uses the RK3588 main controller and has an NPU, which provides the conditions for the deployment of artificial intelligence algorithms. The onboard camera is equipped with a gimbal, which not only increases shooting stability but also provides camera posture information.

[0040] The autonomous clearing method of the present invention specifically includes the following steps.

[0041] Step 1: A drone equipped with a bird's nest removal robot and a camera approaches the power transmission line tower where the bird's nest is located. During the approach, the drone uses the camera to monitor the bird's nest.

[0042] The drone is equipped with an RTK high-precision positioning and orientation system, enabling accurate location and flight control. The onboard computer transmits the coordinates of the tower where the bird's nest needs to be cleared to the drone's flight control, which then controls the drone to the corresponding location and maintains a stable hover. During this process, the onboard computer uses the YOLO v8 target detection algorithm to detect the bird's nest captured by the camera, ensuring it remains in view. If the nest deviates to one side of the field of view, the onboard computer sends correction instructions to the flight control, returning the nest to a position closer to the center of the field of view. This feedback control method uses the classic PID control algorithm.

[0043] Step 2: The drone hovers near the bird's nest, and the camera captures the image of the nest. The target detection algorithm is used to obtain the position information of the nest relative to the drone. Figure 3 , specifically including:

[0044] Step 2.1: The drone hovers near the bird's nest, obtains high-precision pose information of the current drone, takes the drone's current position as the origin, and establishes a local rectangular coordinate system o-xyz. W is the camera's heading angle in the local coordinate system o-xyz.

[0045] Step 2.2: Take a picture and use YOLO v8 for target detection. The target detection result is the pixel coordinates of the top left, top right, bottom left, and bottom right vertices of the bird's nest image in the two-dimensional image. These four points constrain the position of the bird's nest image in the camera's field of view.

[0046] In step 2.3, based on the pixel coordinates, camera parameters, and camera pose information, the camera's pinhole imaging model can be used to determine the direction vectors of the lines containing the four vertices in the local rectangular coordinate system o-xyz. Point A is the camera's optical center, and point C is a point in the camera's field of view. Given the pixel coordinates of point C and the camera parameters, and AB being the optical axis direction of the drone's camera, the direction vector of line L containing point AC can be obtained.

[0047] In step 2.4, use the Depth Anything V2 model to analyze the image and obtain the image's depth information, specifically the distance from point C to the camera's optical center, point A. By combining the direction of vector AC, we can estimate the coordinates of point C in the o-xyz rectangular coordinate system. Similarly, we can obtain the positions of the four vertices in the target detection results of the bird's nest image and calculate the nest's position relative to the drone in space.

[0048] Step 3: Based on the pose information, the movement of the drone and robotic arm is calculated, and the drone and robotic arm are controlled to extend the robotic claw to the nest. Specifically, for the drone-mounted robotic arm to grasp the nest, the end effector robotic claw requires at least five degrees of freedom: three positional degrees of freedom (x, y, z) and two angular degrees of freedom: heading and pitch. The three positional degrees of freedom and the heading angle are provided by the drone body, meaning the drone flies to the specified coordinate point and hovers in the specified heading. The pitch angle is provided by the robotic arm. The drone remains level, and the robotic arm's angle is read by the motor encoder. Simultaneously, to balance the drone's center of gravity, a counterweight is moved as the robotic arm moves. When the drone takes off and balances, the robotic arm's rotating motor adjusts the drone's center of gravity. When the robotic arm is deployed, the torque applied to the arm's rotating motor is read, denoted as M. The weight acting on the counterweight is G. The required movement distance, L, equals M / G. Moving the counterweight brings the drone's center of gravity closer to the center.

[0049] For example, when the robotic arm extends forward and the motor reads a torque value of 10 N·m, and the counterweight weighs 2 kg, that is, the gravity is 20 N, the servo screw is controlled to make the counterweight move backward 0.5 m. At this time, the drone is close to a balanced state.

[0050] Step 4: After the claw closes and grasps the nest, the drone and robotic arm lift the nest off the tower. As the robotic arm moves, as in Step 3, it calculates the torque acting on the motor in real time and moves the balance weight. For example, when the claw is closed, the drone moves backward 1 meter, and the claw releases the nest, completing the grasping motion.

[0051] Step 5: The drone returns to the hovering position in step 2, takes a photo of the bird's nest, and calls the YOLO v8 target detection algorithm to detect the bird's nest again. If the bird's nest cannot be detected, it leaves the operation area. If there are any remnants of the bird's nest, repeat steps 2 to 5.

[0052] The airborne mechanical claw designed in the present invention can quickly grasp and clean bird nests through a flexible mechanical gripper. It can grasp at multiple angles and is not restricted by the posture of the drone, which improves cleaning efficiency and reduces cleaning time. Through machine learning algorithms and artificial intelligence image recognition technology, the drone can quickly and autonomously identify bird nests on power transmission lines, further increasing cleaning efficiency. In addition, the use of an airborne mechanical claw can avoid manual climbing operations and reduce safety risks for personnel. The mechanical gripper can accurately control and position, reducing the risk of power outages caused by misoperation. This device can adapt to the shapes and structures of different bird nests and has strong adaptability. The mechanical gripper can adjust its operating mode adaptively according to the differentiated appearance and placement of different bird nests to achieve better cleaning results. The airborne mechanical gripper can perform cleaning operations through remote control or autonomous judgment, ensuring that operations are performed within a safe distance.

[0053] Those skilled in the art should recognize that the above embodiments are merely intended to illustrate the present invention and are not intended to limit the present invention. As long as they are within the spirit of the present invention, any changes or modifications to the above embodiments will fall within the scope of the claims of the present invention.

Claims

1. A method for autonomously clearing a drone autonomously clearing device, wherein: The drone autonomous removal device is suitable for removing bird nests on power transmission lines. It includes a drone platform and a bird nest removal device arranged below the drone platform. The bird nest removal device includes a mechanical arm suspension rod, a balancing assembly, a motor fixing mechanism, a mechanical arm motor, a motor rotation mechanism, a connecting rod and a mechanical gripper. The mechanical arm suspension rod is arranged at the bottom of the drone platform. The mechanical arm motor is fixed thereon by the motor fixing mechanism. The mechanical arm motor is connected to the connecting rod by the motor rotation mechanism. The end of the connecting rod is provided with a mechanical gripper. The balancing assembly is arranged on the mechanical arm suspension rod and is located on the opposite side of the mechanical arm motor and the mechanical gripper, so that the drone platform maintains balance when the mechanical gripper is operating. The autonomous clearing method is characterized by comprising the following steps: Step 1: The drone platform equipped with the bird nest removal device approaches the transmission line tower where the bird nest is located. During the approach, the drone platform uses a camera to monitor the bird nest. Step 2: The drone platform hovers near the bird's nest, and the camera captures the nest image. The target detection algorithm is used to obtain the position information of the nest relative to the drone. Step 3: Calculate the movement of the drone and robotic arm based on the pose information, control the movement of the drone and robotic arm, and extend the robotic claw to the bird's nest; Step 4: The robotic claws close to grab the bird's nest, and the drone and robotic arm retract; Step 5: Use the camera to detect the bird's nest. If the nest is cleaned, leave the work area. If there are any remaining bird's nests, repeat steps 2 to 5. Step 2 specifically includes: Step 2.1: The drone hovers near the bird's nest, obtains high-precision pose information of the current drone, takes the drone's current position as the origin, and establishes a local rectangular coordinate system o-xyz. W is the camera's heading angle in the local coordinate system o-xyz. Step 2.2: Take a picture and use YOLO v8 for target detection. The target detection result is the pixel coordinates of the top left, top right, bottom left, and bottom right vertices of the bird's nest image in the two-dimensional image. These four points constrain the position of the bird's nest image in the camera's field of view. In step 2.3, based on the pixel coordinates, camera parameters, and camera pose information, the camera's pinhole imaging model can be used to determine the direction vectors of the lines containing the four vertices in the local rectangular coordinate system o-xyz. Point A is the camera's optical center, and point C is a point in the camera's field of view. Given the pixel coordinates of point C and the camera parameters, and AB being the optical axis direction of the drone's camera, the direction vector of line L containing point AC can be obtained. In step 2.4, use the Depth Anything V2 model to analyze the image and obtain the image's depth information, specifically the distance from point C to the camera's optical center, point A. By combining the direction of vector AC, we can estimate the coordinates of point C in the o-xyz rectangular coordinate system. Similarly, we can obtain the positions of the four vertices in the target detection results of the bird's nest image and calculate the nest's position relative to the drone in space.

2. The autonomous cleaning method of a drone autonomous cleaning device according to claim 1, characterized in that: The balancing assembly includes a counterweight, a servo screw and a servo motor. The counterweight is connected to the servo motor through the servo screw and moves horizontally with the help of the screw under the action of the servo motor. The servo motor is connected to the robotic arm motor to read the torque of the robotic arm motor.

3. The autonomous cleaning method of a drone autonomous cleaning device according to claim 1, characterized in that: A camera is installed below the UAV platform on one side of the robotic arm motor and robotic claw.

4. The autonomous cleaning method of a drone autonomous cleaning device according to claim 1, characterized in that: The UAV platform is an X-shaped six-rotor UAV, and the robotic arm suspension rods are arranged front and back along the direction of the UAV platform's tripod.

5. The autonomous cleaning method of a drone autonomous cleaning device according to claim 1, characterized in that: Step 3 is as follows: For the task of grabbing a bird's nest with a drone-mounted robotic arm, the end-effector gripper requires at least five degrees of freedom: three positional degrees of freedom and two angular degrees of freedom: heading and pitch. The three positional degrees of freedom and heading are provided by the drone body, meaning the drone flies to a specified coordinate point and hovers in a specified heading; the pitch angle is provided by the robotic arm; the drone remains level, and the angle of the robotic arm is read by the motor encoder. At the same time, in order to balance the center of gravity of the drone, the counterweight needs to be moved when the robotic arm moves. When the drone takes off and is balanced, the robotic arm rotates the motor to adjust the center of gravity of the drone. When the robotic arm is unfolded, the torque applied to the robotic arm's rotating motor is read, recorded as M, and the gravity applied to the counterweight is G. The required movement distance L = M / G. After moving the counterweight, the center of gravity of the drone can be brought close to the center.

6. The autonomous cleaning method of a drone autonomous cleaning device according to claim 1, characterized in that: Step 4 is as follows: After the mechanical claws close and grab the bird's nest, the drone and the mechanical arm take the bird's nest away from the tower. During the movement of the mechanical arm, the torque acting on the motor is calculated in real time and the balance block is moved.

7. The autonomous cleaning method of a drone autonomous cleaning device according to claim 1, characterized in that: The drone returns to the hovering position in step 2, takes a photo of the bird's nest, and calls the YOLO v8 target detection algorithm to detect the bird's nest again. If the bird's nest cannot be detected, it leaves the operation area. If there are any remnants of the bird's nest, repeat steps 2 to 5.

8. The autonomous cleaning method of a drone autonomous cleaning device according to claim 1, characterized in that: In step 1, the UAV platform is equipped with a high-precision positioning and orientation system to accurately obtain the position and control the flight. The UAV platform's onboard computer sends the coordinates of the tower where the bird's nest needs to be cleared to the UAV flight control. The flight control controls the UAV to fly to the corresponding position and hover stably. During this process, the onboard computer uses the YOLO v8 target detection algorithm to detect the bird's nest obtained by the camera to ensure that the bird's nest is always in the field of view. If the bird's nest deviates to one side of the field of view, the onboard computer will send a correction instruction to the flight control to make the bird's nest return to the field of view near the center. This feedback control method uses the PID control algorithm.

Images