Rotor unmanned aerial vehicle autonomous wireless charging system with coil positioning function

Abstract

The invention discloses a rotor unmanned aerial vehicle autonomous wireless charging system with a coil positioning function, and relates to the field of wireless charging, and the system comprises a positioning module, a flight control module and a wireless charging module. An unmanned aerial vehicle roughly positions a charging base station through a GPS, then rapidly detects a landing identifier of the charging base station through a dynamic ROI, moves to the position above the identifier, sequentially captures and positions AprilTag identifiers from outside to inside in the descending process, then accurately lands right above the inner layer AprilTag, and finally carries out wireless charging through a solenoid type coil assembly. The visual navigation technology and the wireless electric energy transmission technology are combined, the endurance problem of the unmanned aerial vehicle is solved, autonomous electric energy supply of the unmanned aerial vehicle is achieved, manual interference is reduced, and the working efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of wireless charging, and particularly to an autonomous wireless charging system for a rotor unmanned aerial vehicle with a coil positioning function. Background Art

[0002] In the civilian field, most unmanned aerial vehicles are low-altitude small unmanned aerial vehicles. Due to their small size, simple operation, strong mobility and other advantages, they are highly favored by the majority of users and have a wide range of applications. They can be used in agriculture, forestry planning and surveying, mapping, fire fighting and search and rescue, bridge inspection, power line inspection, traffic supervision and other aspects. Nowadays, most civilian unmanned aerial vehicles on the market use lithium batteries as the power source. Lithium batteries have the advantages of long cycle life, small size, light weight, relatively mature technology, safety and reliability, and high energy density. At the same time, when the lithium battery is in its non-working state, the energy loss is small, that is, the self-discharge rate is low. However, the battery capacity is limited, and the flight time is restricted, which in turn affects the operation range and time of the unmanned aerial vehicle. Therefore, the endurance time is an important factor in measuring the performance of the unmanned aerial vehicle. Currently, the general small multi-rotor unmanned aerial vehicles on the market generally maintain about 30 minutes. Without doubt, long endurance will become a key point and a difficult point in the future development of unmanned aerial vehicle technology. In order to improve the endurance of the unmanned aerial vehicle, the current measures such as reducing the aircraft weight, increasing the battery capacity, and optimizing high-power-consuming devices do not have obvious effects, and the problem of insufficient energy carried during flight still exists. However, carrying multiple lithium batteries will increase the weight of the unmanned aerial vehicle, which is not conducive to improving its endurance. Moreover, when the battery energy is exhausted, it is more necessary to manually replace the battery, which will also increase the labor cost during the use of the unmanned aerial vehicle, and the manual interference will greatly affect the work efficiency of the unmanned aerial vehicle.

[0003] The charging methods of unmanned aerial vehicles can be divided into two categories: wired charging and wireless charging. The wired charging of unmanned aerial vehicles includes charging at a charging station, charging using a power transmission line, etc., but inevitably, there is equipment wear during the plugging and unplugging process before and after charging. While wireless charging not only eliminates the charging plugging and unplugging actions and avoids the wear caused by the plugging and unplugging process, but also can more conveniently achieve autonomous endurance and complete operations independently.

[0004] Magnetic resonance wireless charging requires the airborne receiving coil and the base station transmitting coil to be exactly aligned, otherwise high-efficiency charging cannot be achieved, or even charging cannot be carried out, which in turn drags down the progress of the unmanned aerial vehicle operation. It can be seen that to achieve efficient autonomous wireless charging of an unmanned aerial vehicle, the unmanned aerial vehicle needs to accurately land above the base station coil. However, due to the influence of various environmental factors, it is difficult to achieve very accurate landing. Therefore, at the same time, the offset tolerance should be improved, that is, the anti-offset ability of the coil should be improved. The present invention starts from the accurate landing of the unmanned aerial vehicle and the device design, studies the efficient autonomous power supply of the unmanned aerial vehicle, reduces the interference of manual operation, and improves the operation efficiency of the unmanned aerial vehicle. Summary of the Invention

[0005] The present invention provides an autonomous wireless charging system for a rotor UAV with a coil positioning function to solve the technical problems mentioned in the background art.

[0006] An autonomous wireless charging system for a rotor UAV with a coil positioning function, the system comprising: a UAV body, a landing gear, an on-board camera, a wireless charging receiver, a landing marker, and a wireless charging transmitter; the landing gear and the on-board camera are connected below the UAV body; the wireless charging receiver is in a circular ring shape and is fixed outside the landing gear; the wireless charging transmitter is placed on the ground, and its outer shell is made of a rigid plastic box; the landing marker is a double-layer nested AprilTag, and the landing marker is placed directly above the wireless charging transmitter with the center aligned; the coupling coils for wireless charging are two single-layer solenoid coils, namely a receiving coil and a transmitting coil, the diameter of the receiving coil is smaller than that of the transmitting coil, and they are respectively placed inside the wireless charging receiver and the wireless charging transmitter; the system module includes a positioning module, a flight control module, and a wireless charging module; the positioning module is used to quickly capture and locate the charging position; the flight control module is to adjust the pose of the UAV in real time during the flight of the UAV through PID control to achieve precise alignment and stable descent; both ends of the wireless charging module adopt solenoid coils to adapt to the UAV model and achieve wireless charging after alignment.

[0007] Preferably, the landing marker includes an outer-layer AprilTag with a large size and an inner-layer AprilTag with a small size, and the inner-layer AprilTag is located at the central position inside the outer-layer AprilTag to solve the problems that when the UAV lands, the on-board camera has different fields of view at different heights, resulting in unclear and incomplete recognition.

[0008] Preferably, the positioning module includes GPS positioning, dynamic ROI (Region of Interest) target detection, and double-layer nested AprilTag visual positioning. The GPS positioning is used for the UAV to search for the approximate position of a nearby charging base station for preliminary positioning; the dynamic ROI target detection is used to reduce the search range of the UAV, quickly detect the landing marker in the camera view, and optimize the AprilTag algorithm; the double-layer nested AprilTag visual positioning includes image acquisition, image preprocessing, contour detection, quadrilateral fitting, decoding, and pose estimation.

[0009] Preferably, the dynamic ROI target detection refers to predicting the position and range of a target (such as an AprilTag) in the next frame of image, narrowing the detection area, reducing the calculation amount, and improving the real-time performance. The method steps are as follows: S1: Obtain the current frame through the image, detect the target, and record the position and size. S2: Predict the target area of the next frame based on a target motion model (such as Kalman filtering); S3: Generate an ROI according to the prediction result to limit the search range of the next frame; S4: If the detection fails within the ROI, gradually expand the ROI or switch to full-image search.

[0010] Preferably, the image preprocessing includes grayscale conversion and binarization; the grayscale conversion means converting the input image into a grayscale image to simplify subsequent processing; the binarization steps are as follows: S1: Divide the grayscale image into several local windows; S2: Calculate the threshold for each window according to the threshold of the window where the current pixel is located; S3: Determine whether it belongs to 0 (black) or 255 (white), and the determination method is: ; where I(x, y) is the pixel grayscale value at the position (x, y) in the grayscale image, Binary(x, y) is the pixel value at the position (x, y) after binarization, and T(x, y) is the threshold of the window where it is located.

[0011] Preferably, the contour detection refers to performing image processing by using the method of continuous boundary segmentation after obtaining the grayscale image. The main principle is to obtain the gradient of each pixel point by analyzing the change of the pixel grayscale value in the grayscale image, and then cluster the gradient information to generate a series of line segments.

[0012] Preferably, the quadrilateral fitting is to perform linear fitting on each contour pixel point to find a candidate convex quadrilateral as the candidate area of the AprilTag.

[0013] Preferably, the decoding steps are as follows: S1: Use affine transformation to map the quadrilateral into a regular square and divide it into several blocks for easy decoding; S2: Convert each block into a binary sequence by judging the pixel value of the block after mapping; S3: Perform an exclusive OR operation on the detected sequence and the AprilTag label sequence to verify the correctness of the label.

[0014] Preferably, the pose estimation refers to calculating the position and orientation of the detected AprilTag in three-dimensional space. Through these steps, the double-layer nested AprilTag visual positioning can accurately detect and locate the AprilTag in the image and provide its 3D position and orientation relative to the camera.

[0015] Preferably, the flight control module refers to using a PID controller to adjust the pose of the drone in real time according to the recognition result of the AprilTag, ensuring that it accurately aligns with the landing target and descends stably; PID control is a feedback control algorithm widely used in industrial control systems. By dynamically adjusting the three parameters of proportional, integral, and differential, the system output is adjusted to minimize the landing system error of the drone, achieve fast and stable control, and ensure that it accurately aligns with the landing target and descends stably.

[0016] Beneficial effects achieved by the present invention: 1. The present invention combines vision navigation technology and wireless power transmission technology to solve the battery life problem of drones, achieve autonomous power supply for drones, reduce manual interference, and improve their operation efficiency; 2. The present invention adopts a double-layer nested AprilTag identifier, which can achieve accurate positioning at different heights of the drone's on-board camera, is stable and has high fault tolerance, ensures the accurate landing of the drone, and thus realizes efficient and autonomous wireless charging of the drone; 3. The present invention uses dynamic ROI target detection to optimize the AprilTag detection process, reduces the detection area, reduces interference and calculation amount, and improves the real-time performance of recognition; 4. The present invention adopts a solenoid-type wireless charging coil. Compared with the commonly used planar coil, it is more suitable for the compact model of the drone, and provides a convenient and efficient charging method without affecting the normal operation of the drone. Description of the Drawings

[0017] Figure 1 It is a flowchart of autonomous wireless charging of a drone with a coil positioning function.

[0018] Figure 2 It is a schematic diagram of the autonomous wireless charging system of the drone in the embodiment of the present invention.

[0019] Figure 3 It is a schematic diagram of the landing identifier of the charging base station in the embodiment of the present invention.

[0020] Figure 4 It is a schematic diagram of the charging coil in the embodiment of the present invention.

[0021] Annotation of reference numerals in the drawings: 1 - drone body, 2 - landing gear, 3 - on-board camera, 4 - wireless power receiving end, 5 - landing identifier, 6 - wireless power transmitting end, 7 - outer AprilTag, 8 - inner AprilTag, 9 - receiving coil, 10 - transmitting coil. Detailed implementation manners

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of them. Therefore, the detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents the selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0023] In the description of the present invention, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present invention, "a plurality of" means two or more, unless otherwise specifically defined.

[0024] Please refer to Figure 2 , an autonomous wireless charging system for a rotor unmanned aerial vehicle with a coil positioning function is provided in an embodiment of the present invention. The system includes: An unmanned aerial vehicle body 1, a landing gear 2, an on-board camera 3, a wireless charging receiving end 4, a landing mark 5, and a wireless charging transmitting end 6; the landing gear 2 and the on-board camera 3 are connected below the unmanned aerial vehicle body 1; the wireless charging receiving end 4 is in a circular shape and is fixed to the outside of the landing gear 2; the wireless charging transmitting end 6 is placed on the ground, and its shell is made of a rigid plastic box; the landing mark 5 is a double-layer nested AprilTag, and the landing mark 5 is placed directly above the wireless charging transmitting end 6 with the centers aligned; as Figure 3 shown, the coupling coils for wireless charging are two single-layer solenoid coils, namely a receiving coil 9 and a transmitting coil 10 respectively. The diameter of the receiving coil 9 is smaller than that of the transmitting coil 10. As Figure 4 shown, they are respectively placed inside the wireless charging receiving end 4 and the wireless charging transmitting end 6; the system module includes a positioning module, a flight control module, and a wireless charging module; the positioning module includes GPS rough positioning, dynamic ROI (Region of Interest) target detection, and double-layer nested AprilTag visual positioning to achieve rapid capture and positioning of the charging position; the flight control module adjusts the pose of the unmanned aerial vehicle in real time during the flight of the unmanned aerial vehicle through PID control to achieve precise alignment and stable descent; both ends of the wireless charging module adopt solenoid coils to adapt to the unmanned aerial vehicle model and achieve wireless charging after alignment.

[0025] In this embodiment, the landing identifier 5 includes a large-sized outer AprilTag 7 and a small-sized inner AprilTag 8. The inner AprilTag 8 is located at the central position inside the outer AprilTag 7 to solve the problem that when the drone lands, the onboard camera 3 has different fields of view at different heights, resulting in unclear and incomplete recognition.

[0026] In this embodiment, as Figure 1 shown, the autonomous wireless charging process of the drone is as follows: The drone body 1 first roughly locates the charging base station through GPS, then identifies the landing identifier 5 of the charging base station, adjusts the yaw angle, and moves above the landing identifier 5; During the descent, the onboard camera 3 sequentially captures and locates the outer AprilTag 7 and the inner AprilTag 8 identifiers from the outside to the inside, and then precisely lands directly above the inner AprilTag 8.

[0027] Finally, wireless charging is performed through the solenoid coil group. The present invention combines visual navigation technology and wireless power transmission technology to solve the drone endurance problem, achieve autonomous power supply for the drone, reduce manual interference, and improve its operation efficiency.

[0028] Since the field of view of the onboard camera 3 is limited during the landing of the drone, it is difficult to fully recognize the landing identifier 5, thus making it difficult to determine the landing center point. Therefore, the present invention uses a double-layer nested AprilTag as the landing identifier 5, enabling the drone to lock the center of the identifier during near-ground landing, perform precise landing, and align the coils to achieve autonomous and efficient wireless charging of the drone.

[0029] This example uses the AprilTag of "Tag36h11, id = 77" with a large blank space in the middle as the outer identifier, and the AprilTag of "Tag36h11, id = 56" as the inner identifier, as Figure 3 shown. The side length of the outer identifier is 5 times that of the inner identifier, and the small-sized inner identifier is placed at the exact center of the outer identifier. Since there are 6 blocks in a single row of the AprilTag in the Tag36h11 family, the inner identifier occupies 4 central blocks of the outer identifier, which is less than the recognition tolerance of the Tag36h11 family. Therefore, the double-layer nested design does not affect the recognition of the outer identifier.

[0030] In this embodiment, the positioning module includes GPS positioning, dynamic ROI (Region of Interest) target detection, and double-layer nested AprilTag visual positioning. The GPS positioning is used for the drone to search for the approximate location of the nearby charging base station for preliminary positioning. The dynamic ROI target detection is used to reduce the search range of the drone, quickly detect the landing marker in the camera's view, and optimize the AprilTag algorithm. The double-layer nested AprilTag visual positioning includes image acquisition, image preprocessing, contour detection, quadrilateral fitting, decoding, and pose estimation.

[0031] In this embodiment, the flight control module refers to using a PID controller to adjust the pose of the drone in real time according to the recognition result of the AprilTag, ensuring that it accurately aligns with the landing target and descends stably. PID control is a feedback control algorithm widely used in industrial control systems. By dynamically adjusting the three parameters of proportional, integral, and differential, it adjusts the system output to minimize the landing system error of the drone, achieving fast and stable control, and ensuring that it accurately aligns with the landing target and descends stably.

[0032] In this example, the operating steps of the autonomous wireless charging system for the rotor drone are as follows: Step 1: Send a charging request When the drone's battery is low, the main controller sends a charging request. The ground charging base station receives the request signal and sends a charging permission signal to the drone. The drone receives the signal and starts to enter the autonomous landing mode. Step 2: Rough GPS positioning The drone roughly locates the charging base station that sent the charging permission signal through GPS and makes a horizontal yaw movement in the direction of the charging base station. Step 3: Ground image acquisition, AprilTag recognition and positioning The drone uses the on-board camera to collect ground images in real time, search for the charging position, capture, recognize, and locate the landing marker. Among them, the AprilTag visual positioning technology includes image acquisition, image preprocessing, contour detection, quadrilateral fitting, decoding, and pose estimation. And through the dynamic ROI target detection technology, the AprilTag algorithm is optimized to reduce useless images and interference, improve the system calculation efficiency, and improve the real-time performance. ROI, that is, the region of interest, is usually a target area containing the object to be processed in one frame of the image. Dynamic ROI target detection refers to predicting the position and range of the target (such as AprilTag) in the next frame of the image, and finding the target area closest to the prediction result in the next frame, which is considered the ROI area, and thus the tracking is completed. The steps of the dynamic ROI target detection are as follows: S1: Initial detection; In the first frame, a full-image search is conducted. After detecting the AprilTag, record its bounding box coordinates (x0, y0, w0, h0), where x0 and y0 are the positions of the bounding box, and w0 and h0 are the dimensions of the bounding box respectively; In this example, the focal length f = 800 pixels, and the resolution is 1280×720 pixels; The size of the inner AprilTag to be detected is 0.3m×0.3m, and the bounding box (x0, y0, w0, h0) = (600, 350, 100, 100) is set; S2: Motion prediction; Take the Kalman filter as an example: The state matrix is: ; Where x and y are the coordinate values of the border position, v x , v y are the relative velocities, and d t is the position and velocity of the current identification center; In this example, the initial state is x = 600 pixels, y = 350 pixels, v x = 10 pixels / frame, v y = -5 pixels / frame, then ; The prediction equation is: ; Where d t+1 is the predicted state, d t is the current state, and F is the state transition matrix; In this example, a uniform motion model is set, so , ; Meanwhile, the prediction result can be corrected according to the noise covariance R (set according to the detection accuracy) update equation; S3: ROI selection; Generate an ROI according to the prediction result to limit the search range of the next frame; Expand the area according to the predicted position coordinates (x t+1 , y t+1 ) = (610, 345): ; Where w t , h t are the sizes of the current frame area, k is the expansion coefficient, usually used in dynamic ROI target detection when the drone has not reached directly above the identification, taking 1.5~2 to cope with prediction errors; The size of the current frame (w t , h t ) = (100, 100), and the expansion coefficient k takes the empirical value of 1.5, then ; If the ROI exceeds the image range, it is truncated to the valid area: 0≤x<1280, 0≤y<720, the ROI is valid; crop the ROI area, cut a 300 pixel × 300 pixel sub-image from the original image, successfully detect AprilTag in the sub-image, and the local coordinates of the new bounding box relative to the ROI are (150,145,110,110); convert the local coordinates to global coordinates (150+460, 145+195, 110, 110) = (610, 340, 110, 110); iterate and calculate each subsequent frame. If the detection confidence is high (such as 5 consecutive successful frames), reduce k to 1.2 to generate a smaller ROI; At the same time, the k value is dynamically adjusted according to the drone height or target size. In the high-altitude search stage, the drone is far away from the target and the tag moves relatively fast, so the ROI size range needs to be expanded, that is, the k value needs to be increased. In the near-ground alignment stage, the tag moves relatively slowly, so the ROI needs to be reduced to improve accuracy, that is, the k value needs to be reduced. S4: Continuous tracking and failure recovery; After 5 consecutive frames are successfully detected, the system enters the "high confidence" mode and reduces the ROI. If no target is detected in the ROI for 3 consecutive frames, the ROI range is expanded. After 3 failures, the system resets to the full-image search mode. Dynamic ROI target detection optimizes the AprilTag algorithm to avoid traversing and solving the pixel values of all pixels on the screen, reduce the interference of screen lines, alleviate the system calculation pressure, and improve the real-time recognition performance; especially in the high-altitude search stage, the QR code occupies a small proportion of the screen, so it is particularly important for ROI to quickly determine the target area; AprilTag recognition and positioning include image preprocessing, contour detection, quadrilateral fitting, decoding and pose estimation; The image preprocessing includes graying and binarization; the graying refers to converting the input image into a grayscale image to simplify subsequent processing; the binarization step is: S1: Divide the grayscale image into several local windows; S2: Calculate the threshold for each window. According to the threshold of the window where the current pixel is located;

[0033] S3: Determine whether it is black or white. The determination method is: ; Where I(x, y) is the grayscale value of the pixel at position (x, y) in the grayscale image, Binary(x, y) is the pixel value at position (x, y) after binarization, and T(x, y) is the threshold of the window. Calculate the threshold according to the maximum between-class variance method, calculate the threshold for each window, and compare the threshold of each pixel with the corresponding window: if it is less than or equal to the threshold, set Binary(x, y) to 0, representing black; if it is greater than the threshold, set Binary(x, y) to 255, representing white.

[0034] The contour detection refers to performing image processing using the continuous boundary segmentation method after obtaining the grayscale image. The main principle is to obtain the gradient of each pixel point by analyzing the change of pixel grayscale values in the grayscale image. Then, cluster the gradient information to generate a series of line segments;

[0035] The quadrilateral fitting performs line fitting on each contour pixel point to find candidate convex quadrilaterals as candidate regions for AprilTag; The steps of the decoding are as follows: S1: Use affine transformation to map the quadrilateral to a regular square and divide it into several blocks for easy decoding; S2: Convert each block into a binary sequence by judging the pixel values of the mapped blocks; S3: Perform an exclusive OR operation on the detected sequence and the AprilTag label sequence to verify the correctness of the label; The pose estimation refers to calculating the position and orientation of the detected AprilTag in the three-dimensional space according to its geometric information. Through these steps, the AprilTag recognition system can accurately detect and locate the AprilTag in the image and provide its 3D position and orientation relative to the camera;

[0036] The flight control refers to using a PID controller to adjust the pose of the drone in real time according to the recognition result of AprilTag to ensure that it accurately aligns with the landing target and descends stably. PID control is a feedback control algorithm widely used in industrial control systems. By dynamically adjusting the three parameters of proportional, integral, and derivative to adjust the system output, it minimizes the drone landing system error to achieve fast and stable control, ensuring that it accurately aligns with the landing target and descends stably;

[0037] Step Four: Drone Flight Control By identifying the pose information of the identifier, the main control system of the drone will calculate the yaw motion and the yaw angle and speed of the vertical landing, and control the position and motion attitude of the fuselage; The drone needs to control the following degrees of freedom: Horizontal position (X, Y): Align the center of AprilTag with the center of the identifier; Height (Z): Adjust the descent speed according to the size of the identifier; Yaw angle (Yaw): Make the drone approach the identifier; Taking the control of the horizontal position X as an example: S1: Error calculation; The error of the horizontal position X is: ; where e x is the error of the horizontal position X, x tag is the x-axis coordinate in the tag obtained from the AprilTag detection result, and x center is the x-axis coordinate of the image center; S2: PID output; The output of the drone's horizontal X-direction speed is: ; where K p 、K i 、K d are the proportional, integral, and differential coefficients respectively, T is the sampling period, e x,prev is the error of the horizontal position X in the previous sampling period, and v x is the output speed of the drone's horizontal X direction; In this example, the vertical landing process of the drone: Initial search: The drone hovers, the camera detects the AprilTag, and calculates the horizontal error e x = 50 pixels; Proportional control: Adjust the parameters, set Kp = 0.1, pixels / frame; Approaching the target: e x decreases to 5 pixels, and the differential term suppresses overshoot; S3: Control execution; Send the PID control command to the drone flight controller, adjust the motor thrust to achieve horizontal movement, and move towards the center of the marker; Step Five: Hierarchical landing Through yaw movement, when the drone is directly above the marker, its camera will locate the outer AprilTag in the charging area for preliminary landing; when descending to a certain height, the field of view of the on-board camera will become narrow and it is difficult to fully capture the outer AprilTag. At this time, capture the inner AprilTag for identification and positioning, and accurately land at the center of the marker; Step Six: Landing for wireless charging The drone lands directly above the transmitting coil, the two coils are aligned correctly, the engine is turned off, and wireless charging is carried out.

[0038] The above description of the disclosed embodiments is for those skilled in the art to understand or apply. Various modifications to the above embodiments are easy for those skilled in the art. Therefore, the present invention will not be limited to the above embodiments, but rather to the broadest scope consistent with the principles and novel features disclosed in the present invention.

[0039] It should be noted that, in this article, the term "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or device comprising that element.

[0040] The above are only the preferred embodiments of the present invention, and do not limit the patent scope of the present invention accordingly. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall be equally included in the patent protection scope of the present invention.

Claims

1. An autonomous wireless charging system for a rotor unmanned aerial vehicle with a coil positioning function, characterized in that The system includes: an unmanned aerial vehicle (UAV) body (1), landing gear (2), an airborne camera (3), a wireless charging receiver (4), a landing identifier (5), and a wireless charging transmitter (6); the landing gear (2) and the airborne camera (3) are connected below the UAV body (1), and the wireless charging receiver (4) is fixed to the outside of the landing gear (2); the wireless charging transmitter (6) is placed on the ground, the landing identifier (5) is a double-layer nested AprilTag, the landing identifier (5) is placed directly above the wireless charging transmitter (6) with the center aligned; the coupling coils for wireless charging are two single-layer solenoid coils, namely a receiving coil (9) and a transmitting coil (10), which are respectively placed inside the wireless charging receiver (4) and the wireless charging transmitter (6); the system module includes a positioning module, a flight control module, and a wireless charging module.

2. The autonomous wireless charging system for a rotor UAV with a coil positioning function according to claim 1, characterized in that, The landing identifier (5) includes a large-sized outer-layer AprilTag (7) and a small-sized inner-layer AprilTag (8), and the inner-layer AprilTag (8) is located at the central position inside the outer-layer AprilTag (7).

3. The autonomous wireless charging system for a rotor UAV with a coil positioning function according to claim 1, characterized in that, The positioning module includes GPS positioning, dynamic ROI target detection, and double-layer nested AprilTag visual positioning. The GPS positioning is used for the UAV to search for the approximate position of a nearby charging base station for preliminary positioning; the dynamic ROI target detection is used to reduce the search range of the UAV and quickly detect the landing identifier in the camera view, optimizing the AprilTag algorithm; the double-layer nested AprilTag visual positioning includes image acquisition, image preprocessing, contour detection, quadrilateral fitting, decoding, and pose estimation.

4. The autonomous wireless charging system for a rotor UAV with a coil positioning function according to claim 3, characterized in that, The dynamic ROI target detection refers to reducing the detection area and computational amount and improving real-time performance by predicting the position and range of the target in the next frame of the image. The method steps are as follows: S1: Obtain the current frame through the image, detect the target, and record the position and size. S2: Predict the target area in the next frame based on the target motion model. S3: Generate an ROI according to the prediction result to limit the search range of the next frame. S4: If the detection fails within the ROI, gradually expand the ROI or switch to full-image search.

5. The autonomous wireless charging system for a rotor UAV with a coil positioning function according to claim 3, characterized in that, The image preprocessing includes grayscale conversion and binarization; the grayscale conversion refers to converting the input image into a grayscale image to simplify subsequent processing; the binarization steps are as follows: S1: Divide the grayscale image into several local windows. S2: Calculate the threshold for each window according to the threshold of the window where the current pixel is located. S3: Judge whether it belongs to 0 or 255. The judgment method is as follows: ； where I(x, y) is the pixel grayscale value at the position (x, y) in the grayscale image, Binary(x, y) is the pixel value at the position (x, y) after binarization, and T(x, y) is the threshold of the window where it is located.

6. The autonomous wireless charging system for a rotor unmanned aerial vehicle with a coil positioning function according to claim 3, characterized in that, The contour detection refers to performing image processing using the method of continuous boundary segmentation after obtaining the grayscale image. The principle is to obtain the gradient of each pixel point by analyzing the change of the pixel grayscale value in the grayscale image, and then cluster the gradient information to generate a series of line segments.

7. The autonomous wireless charging system for a rotor UAV with a coil positioning function according to claim 3, wherein The quadrilateral fitting is to perform linear fitting on each contour pixel point to find candidate convex quadrilaterals as the candidate regions of the AprilTag.

8. The autonomous wireless charging system for a rotor UAV with a coil positioning function according to claim 3, wherein, The steps of decoding are as follows: S1: Use affine transformation to map the quadrilateral into a regular square and divide it into several blocks for easy decoding; S2: Convert each block into a binary sequence by judging the pixel values of the blocks after mapping; S3: Perform an exclusive OR operation on the detected sequence and the AprilTag label sequence to verify the correctness of the label.

9. The autonomous wireless charging system for a rotor UAV with a coil positioning function according to claim 3, characterized in that, The pose estimation refers to calculating the position and orientation of the detected AprilTag in the three-dimensional space according to its geometric information. Through these steps, the double-layer nested AprilTag visual positioning accurately detects and locates the AprilTag in the image and provides its 3D position and orientation relative to the camera.

10. The autonomous wireless charging system for a rotor UAV with a coil positioning function according to claim 1, characterized in that, The flight control module is to use a PID controller to adjust the pose of the UAV in real time according to the recognition result of the AprilTag to ensure that it accurately aligns with the landing target and descends stably; PID control is a feedback control algorithm widely used in industrial control systems. By dynamically adjusting the three parameters of proportional, integral, and differential, the system output is adjusted to minimize the UAV landing system error, achieve fast and stable control, and ensure that it accurately aligns with the landing target and descends stably.

Images