Real-time visual identification and positioning method and system between multi-modules of split flying vehicle

Abstract

The application discloses a split type flying vehicle multi-module real-time visual identification positioning method and system, calibrates camera internal and external parameters for image data, obtains the internal and external parameters of the camera and a scale transformation matrix, carries out image preprocessing for the image data to obtain a threshold grayscale image, then carries out contour extraction, quadrilateral screening, perspective transformation and internal code identification processing to obtain a front view with marks and corner point coordinates, carries out edge thinning processing to obtain a high-accuracy positioning contour, and based on the real-time detection positioning result, uses conversion matrices between a pixel coordinate system, an image coordinate system, a camera coordinate system and a world coordinate system to obtain the relative position of a camera lens center point to a center point of a region where an ArUco code arranged in advance and a stationary module is located and the heading deviation angle of a docking moving module relative to the stationary module, that is, the precise real-time visual identification positioning between the multi-modules of the flying vehicle is realized.

Description

Technical Field

[0001] This invention belongs to the field of unmanned vehicles, specifically relating to a method and system for real-time visual recognition and positioning between multiple modules of a split-type flying vehicle. Background Technology

[0002] Currently, with the rapid development of artificial intelligence, semiconductors, and sensors, modular flying vehicles are constantly expanding their capabilities and application areas through rapid technological iteration. During the precise docking of multiple modules within a modular flying vehicle, real-time identification and positioning are required based on the deployment of its sensors. However, different sensors are limited by their respective environments, necessitating different solutions for different environments to achieve accurate real-time identification and positioning. Due to the unique application environment of modular flying vehicles, high positioning and navigation accuracy is required during multi-module docking, typically at the centimeter or even millimeter level. GPS (Global Positioning System), under good signal conditions, generally achieves meter-level positioning accuracy, but GPS signal strength attenuates significantly during docking, and may even be completely blocked, resulting in poor navigation performance and failing to meet application requirements. LiDAR can accurately measure distances and locate the surrounding environment to complete the intended positioning and navigation tasks, but LiDAR is expensive and requires additional structures to provide positioning information. Inertial navigation elements can measure relative displacement, but the measured data exhibits some drift, requiring collaboration with other sensors to meet accuracy requirements, making the algorithm complex. ArUco markers are binary square reference markers used for camera identification and positioning. Their main advantages are simple and fast detection, and strong robustness. An ArUco marker is a square marker consisting of a wide black border and an internal binary matrix that defines its identifier (id). The black border facilitates rapid detection in an image, while the internal binary encoding is used for marker identification and error detection and correction. The size of the ArUco marker determines the size of the internal matrix; for example, a 4x4 marker consists of 16 bits of binary data. Because a single ArUco marker can provide sufficient correspondence—for example, with four distinct corner points and internal binary encoding—ArUco markers are widely used to increase the amount of information when mapping from the 2D world to the 3D world, facilitating the discovery of projection relationships between the two-dimensional and 3D worlds, thereby enabling applications such as calculating camera pose.

[0003] Patent CN106969766A provides an indoor autonomous navigation method based on monocular vision and QR code road signs. This method solves the problems of accuracy in indoor autonomous navigation and computational complexity of binocular cameras.

[0004] The existing technology does not refine the method according to the characteristics of the scenario. Specifically, it does not consider the working conditions such as large changes in lighting conditions when setting the environment, which may lead to inaccurate positioning in different environments.

[0005] The existing technology has insufficient number of QR codes, and when the target is lost, it is necessary to continuously scan QR codes in other directions, so there is still room for improvement in terms of computational efficiency.

[0006] Existing technology requires QR codes to be placed in advance according to the driving route, and the layout of the QR codes must strictly follow certain rules, requiring accurate measurement in advance, which is greatly affected by the environment. Summary of the Invention

[0007] In order to solve the technical problems existing in the background art, the present invention aims to provide a method and system for real-time visual recognition and positioning between multiple modules of a split-type flying vehicle.

[0008] To solve the technical problem, the technical solution of the present invention is as follows:

[0009] A real-time visual recognition and positioning method for multiple modules of a split-type flying vehicle, the method comprising:

[0010] The image data is calibrated by adjusting the camera's intrinsic and extrinsic parameters to obtain the mapping relationship from world coordinates to pixel coordinates.

[0011] Based on the mapping relationship from world coordinates to pixel coordinates, camera images are acquired and image preprocessing is performed to obtain a thresholded grayscale image;

[0012] The thresholded grayscale image is processed by contour extraction, quadrilateral filtering, perspective transformation and internal coding recognition to obtain a front view with markings and corner coordinates.

[0013] The front view with markings and corner coordinates is subjected to edge refinement processing to obtain an image with a contour containing high positioning accuracy;

[0014] Based on images containing high-accuracy positioning contours, pose calculations are performed using the mapping relationship between world coordinates and pixel coordinates. This yields the relative position of the camera lens center point with respect to the pre-arranged center point of the ArUco code area at the bottom of the stationary module, as well as the heading deviation angle of the docking motion module relative to the stationary module. This achieves precise real-time visual recognition and positioning between multiple modules of the flying vehicle.

[0015] Furthermore, image data is collected by using an upward-looking onboard camera module mounted on a split-type flying vehicle.

[0016] Furthermore, the calibration process for the camera's intrinsic and extrinsic parameters specifically includes:

[0017] A rigid body transformation is performed on the coordinate system of the image data to obtain the mapping relationship between points in the world coordinate system and points in the camera coordinate system;

[0018]

[0019] By applying pinhole imaging to the coordinate system that maps points in the world coordinate system to points in the camera coordinate system, the mapping relationship between the camera module coordinate system and the image coordinate system is obtained.

[0020]

[0021] The coordinate system of the mapping relationship between the camera module coordinate system and the image coordinate system is transformed by unit conversion and installation deviation and lens distortion are introduced to obtain the transformation result between the world coordinate system and the pixel coordinate system, that is, the mapping relationship from world coordinates to pixel coordinates.

[0022]

[0023] Furthermore, the image preprocessing specifically includes:

[0024] The image data is converted to grayscale to obtain a grayscale image;

[0025] Gray(x,y)=0.299R(x,y)+0.578G(x,y)+0.114B(x,y);

[0026] Where R(x,y), G(x,y), and B(x,y) are the values ​​of the three components of the RGB image, respectively;

[0027] The grayscale image is thresholded to obtain a thresholded grayscale image;

[0028]

[0029]

[0030] T(x,y)=WA s×s (x,y)-p.

[0031] Furthermore, the edge refinement process for the front view with markings and corner coordinates specifically includes:

[0032] The ArUco detection algorithm is used to perform contour detection and ID recognition on the front view to obtain the initial contour result;

[0033] By using the cost function, we can calculate the cost function for all points on the initial contour result and obtain a high-accuracy contour that minimizes the cost function.

[0034]

[0035] Where R(p) and E(p,x) i The actual and expected values ​​of the line at image location p are given by xi; the predicted value is obtained through a given edge model using the distance d(p,x) from the given point p to the initial line. i ).

[0036] Furthermore, based on images containing contours with high positioning accuracy, pose calculation is performed using the mapping relationship from world coordinates to pixel coordinates, specifically including:

[0037] Calculate the relative position of the parking module of the split-type flying vehicle with respect to the stationary module;

[0038] The center coordinates of each ArUco marker are calculated based on the quadrilateral contour of the image containing a high-accuracy positioning contour.

[0039] By introducing the mapping relationship from world coordinates to pixel coordinates into the following formula, the relative position of the ground parking module of the split flying vehicle relative to the stationary module can be calculated.

[0040] vector_center=vector_original-vector_camera;

[0041] Where vector_original represents the position vector of a certain Aruco marker center point in the world coordinate system; vector_camera represents the position vector of a certain Aruco marker center point in the pixel coordinate system; and vector_center represents the position vector of the camera module coordinate system origin relative to the world coordinate system origin.

[0042] Calculate the heading deviation angle of the parking module of the split-type flying vehicle relative to the stationary module;

[0043] When the detection algorithm detects two or more ArUco markers in the processed input image, it inputs the center point coordinates and ID values ​​of each paired ArUco marker into the heading angle calculation module. This module calculates the linear equations of any pair of ArUco markers in both the world coordinate system and the pixel coordinate system. By using the mapping relationship from world coordinates to pixel coordinates, the vehicle heading angle can be calculated.

[0044]

[0045] Where α represents the angle between the line and the positive x-axis in the world coordinate system; β represents the angle between the line and the positive x-axis in the pixel coordinate system. The equation of the line is obtained by marking the coordinates of the center points of each pair of ArUco markers.

[0046] A real-time visual recognition and positioning system for multiple modules of a split-type flying vehicle, the system comprising:

[0047] The calibration module is used to calibrate the camera's intrinsic and extrinsic parameters on the image data, and to obtain the mapping relationship from world coordinates to pixel coordinates.

[0048] The first processing module is used to acquire camera images and perform image preprocessing based on the mapping relationship from world coordinates to pixel coordinates to obtain a thresholded grayscale image.

[0049] The second processing module is used to perform contour extraction, quadrilateral filtering, perspective transformation and internal encoding recognition processing on the thresholded grayscale image to obtain a front view with markings and corner coordinates.

[0050] The third processing module is used to refine the edges of the front view with markings and corner coordinates to obtain an image with a high positioning accuracy contour.

[0051] The pose calculation module is used to perform pose calculation based on an image containing a high-accuracy contour and by using the mapping relationship between world coordinates and pixel coordinates. It obtains the relative position of the camera lens center point with respect to the center point of the pre-arranged ArUco code area at the bottom of the stationary module and the heading deviation angle of the docking motion module relative to the stationary module, thus realizing accurate real-time visual recognition and positioning between multiple modules of the flying vehicle.

[0052] A computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the method of any one of claims 1 to 6.

[0053] Compared with the prior art, the advantages of the present invention are as follows:

[0054] By using a large number of ArUco markers with sizes matching the distance between the camera and the marker plane, the target will not be lost during target identification, thus ensuring the stability of the identification and positioning process.

[0055] By preprocessing the input image data, noise is eliminated and background and target labels are segmented, thereby improving the speed of the detection algorithm and achieving an average computation speed of 5.23ms / frame.

[0056] By using edge refinement methods to optimize the initial detection image, the algorithm can achieve a maximum positioning error of 3.31 mm while maintaining a high detection rate. Attached Figure Description

[0057] Figure 1 The following is a flowchart of the real-time visual recognition and positioning method and system for multi-module split-type flying vehicles provided in this embodiment of the invention;

[0058] Figure 2 A schematic diagram of the signboard used in the embodiments of the present invention;

[0059] Figure 3 A schematic diagram of the AGV verification platform used in the embodiments of the present invention;

[0060] Figure 4 A schematic diagram of the edge refinement model used in the embodiments of the present invention;

[0061] Figure 5 The following are schematic diagrams of the detection results under different light intensity conditions provided in the embodiments of the present invention. Detailed Implementation

[0062] The specific implementation of the present invention is described below with reference to embodiments:

[0063] It should be noted that the structures, proportions, sizes, etc. shown in this specification are only used to complement the content disclosed in the specification for those skilled in the art to understand and read, and are not intended to limit the conditions under which the present invention can be implemented. Any modifications to the structure, changes in the proportions, or adjustments to the size, without affecting the effects and objectives that the present invention can produce, should still fall within the scope of the technical content disclosed in the present invention.

[0064] Furthermore, the terms such as "upper," "lower," "left," "right," "middle," and "one" used in this specification are merely for clarity of description and are not intended to limit the scope of the invention. Any changes or adjustments to their relative relationships, without substantially altering the technical content, should also be considered within the scope of the invention.

[0065] Example 1:

[0066] The camera-based real-time identification and positioning method provided in this invention is applied to a split-type flying vehicle. An upward-looking onboard camera module is deployed on the split-type flying vehicle to collect images in real time for image preprocessing. The preprocessed images are then subjected to ArUco marker detection, and the detected contours are edge-refined to obtain marker contours with higher positioning accuracy. This provides position information for the flying vehicle to perform precise docking, which is beneficial for the ground module to dock with the aircraft after parking.

[0067] Figure 1 This is a flowchart illustrating a method for real-time stitching of bird's-eye view images and parking space detection for a split-type flying vehicle, as provided in an embodiment of the present invention. Figure 1 As shown, the method includes:

[0068] Step 101: Calibrate the camera's intrinsic and extrinsic parameters.

[0069] like Figure 3 As shown, a camera module is positioned above the front axle of the split-type flying vehicle, with its optical axis perpendicular to the ground and pointing upwards. Before parameter calibration, the resolution and frame rate of the selected camera need to be determined. Figure 2 This is a schematic diagram of the marking board used in an embodiment of the present invention. In this example, each ArUco mark on the marking board is 30mm*30mm in size, arranged in a row of 7 and a column of 11, with a 5mm interval between each pair of ArUco marks. First, let's define the coordinate systems we use. The world coordinate system is established with the target as its origin. Therefore, in this paper, the origin of the world coordinate system is placed at the center of the bottom of the stationary module of the split-type flying vehicle. The Y-axis is directly in front of the stationary module, and the Z-axis is perpendicular to the ground plane and upwards, conforming to the right-hand rule. The camera module coordinate system has the optical center of the camera as its origin. The Z-axis coincides with the optical axis of the camera module and is perpendicular to the imaging plane, with the camera's orientation as the positive direction. It also conforms to the right-hand rule. The image coordinate system has the center of the captured image as its origin, with the X and Y axes parallel to the two perpendicular sides of the image plane, respectively. The pixel coordinate system has its origin at the upper left corner of the CCD image, and the u-axis and v-axis of the pixel coordinate system are parallel to the x-axis and Y-axis of the image coordinate system, respectively. The conversion between world coordinates and pixel coordinates requires the following process:

[0070] (a) Achieve the transformation from a point in the world coordinate system to a point in the camera coordinate system through rigid body transformation, where R is a 3*3 rotation matrix and T is a 3*1 translation matrix.

[0071]

[0072] (b) The transformation between the camera module coordinate system and the image coordinate system is obtained by transforming the above formula using the principle of pinhole imaging, where f is the focal length of the camera module.

[0073]

[0074] (c) Since the image coordinate system uses mm and the pixel coordinate system uses px, unit conversion is required. Wherein, d x d y These are the dimensions of each pixel on the x and y axes, respectively. u0 and v0 are variables introduced to eliminate mounting errors and lens distortion.

[0075]

[0076] By modifying the above equation, we get:

[0077]

[0078] Combining the above equations yields the transformation expression between the world coordinate system and the pixel coordinate system, which allows us to determine the camera's intrinsic and extrinsic parameter correction parameters based on actual working conditions.

[0079]

[0080] Step 102, Image preprocessing.

[0081] To some extent, grayscale images are a special representation of color images. Grayscale images also reflect the overall and local features of an image, and using grayscale images can reduce the computational load in image processing. This embodiment performs grayscale processing on the input image according to the following weighted average formula.

[0082] Gray(x,y)=0.299R(x,y)+0.578G(x,y)+0.114B(x,y);

[0083] Where R(x,y), G(x,y), and B(x,y) are the values ​​of the three components of the RGB image, respectively.

[0084] The input image contains the target image, complex background information, and noisy data. Image thresholding can effectively remove noise and segment the background and target regions, thereby improving the speed of the detection algorithm. This embodiment uses inverse binary thresholding. Since the experiment was conducted outdoors and was greatly affected by lighting conditions, local adaptive thresholding was used to select the threshold for image preprocessing. Experimental results show that this processing helps improve the robustness of the detection algorithm to changes in lighting.

[0085]

[0086]

[0087] T(x,y)=WA s×s (x,y)-p;

[0088] Where x and y are the coordinates of any point in the image, and G(x,y) is the Gaussian value WA of any point. s×s (x,y) is the Gaussian weighted average value of any point (x,y) within the s×s matrix centered at (x,y), and T(x,y) is the selected threshold.

[0089] Step 103, Identification detection and edge refinement.

[0090] like Figure 4As shown, to further improve the localization accuracy of ArUco marker detection, we propose a novel optimization-based edge refinement method. Each edge of the detected quadrilateral is described by a straight-line model. The initial parameters of each line of the quadrilateral are obtained from the initially detected corner points and then refined by minimizing the following cost function.

[0091]

[0092] Where R(p) and E(p,x) i ) represents the actual and expected values ​​at image position p, given by x. i The line represents the initial line. The predicted value is obtained using a given edge model, such as the sigmoid model shown in the figure, using the distance d(p, x) from a given point p to the initial line. i ) Result in refined edge lines Then, the positions of each corner point are inferred from the intersection of the intersecting edge lines.

[0093] Step 104, vehicle pose calculation.

[0094] (a) Calculation of relative position

[0095] like Figure 5 As shown, based on the above method, a certain number of quadrilateral contours with high positioning accuracy are obtained, thereby calculating the center coordinates of each ArUco marker. The correction parameters obtained after camera intrinsic and extrinsic parameter calibration are introduced into the following formula to calculate the relative position of the ground parking module of the split-type flying vehicle relative to the stationary module (in the world coordinate system).

[0096] cector_center=vector_original-vector_camera;

[0097] Where vector_original represents the position vector of a certain Aruco marker center point in the world coordinate system; vector_camera represents the position vector of a certain Aruco marker center point in the pixel coordinate system; and vector_center represents the position vector of the camera module coordinate system origin relative to the world coordinate system origin.

[0098] (b) Calculation of vehicle heading angle

[0099] The automatically generated coordinate codes proposed in this paper can be used to obtain the coordinates of each ArUco marker on the printed marker board in the stationary module (world coordinate system). When the detection algorithm detects two or more ArUco markers in the input processed image, it inputs the center point coordinates and ID values ​​of each paired ArUco marker into the heading angle calculation module. This module calculates the linear equations of any two ArUco markers in the world coordinate system and the pixel coordinate system, and then calculates the vehicle heading angle using the correction parameters obtained after camera intrinsic and extrinsic parameter calibration.

[0100]

[0101] Where α represents the angle between the line connecting any two ArUco markers in the world coordinate system and the X-axis, and β represents the angle between the line connecting any two ArUco markers in the pixel coordinate system and the X-axis.

[0102] References (e.g., patents / papers / standards)

[0103] [1] Beijing PinChuang Intelligent Technology Co., Ltd. An indoor autonomous navigation method based on monocular vision and QR code road signs: CN201710172646.1[P]. 2017-07-21.

[0104] [2] Shanghai Intelligent Connected Vehicle Technology Center Co., Ltd. An Augmented Reality Marker Structure Integrating Multiple ArUco Codes: CN202122061386.5 [P]. 2022-01-28.

[0105] [3]Kallwies J,Forkel B,Wuensche H J.Determining and improving the localization accuracy of AprilTag detection[C] / / 2020IEEE International Conference on Robotics and Automation(ICRA).IEEE,2020:8288-8294.

[0106] [4]Romero-Ramirez FJ, -Salinas R,Medina-Carnicer R.Speeded updetection of squared fiducial markers[J].Image and vision Computing, 2018,76:38-47.

[0107] The preferred embodiments of the present invention have been described in detail above. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

[0108] Many other changes and modifications can be made without departing from the concept and scope of this invention. It should be understood that this invention is not limited to the specific embodiments, and the scope of this invention is defined by the appended claims.

Claims

1. A real-time visual recognition and positioning method for multiple modules of a split-type flying vehicle, characterized in that, The method includes: The image data is calibrated using camera intrinsic and extrinsic parameters to obtain the mapping relationship from world coordinates to pixel coordinates. Based on the mapping relationship from world coordinates to pixel coordinates, camera images are acquired and image preprocessing is performed to obtain a thresholded grayscale image; The thresholded grayscale image is processed by contour extraction, quadrilateral filtering, perspective transformation and internal coding recognition to obtain a front view with markings and corner coordinates. The front view with markings and corner coordinates is subjected to edge refinement processing to obtain an image with a high positioning accuracy contour; Based on images containing high-accuracy positioning contours, pose calculations are performed using the mapping relationship between world coordinates and pixel coordinates. This yields the relative position of the camera lens center point with respect to the pre-arranged center point of the ArUco code area at the bottom of the stationary module, as well as the heading deviation angle of the docking motion module relative to the stationary module. This achieves precise real-time visual recognition and positioning between multiple modules of the flying vehicle.

2. The real-time visual recognition and positioning method among multiple modules of a split-type flying vehicle according to claim 1, characterized in that, Image data is collected by using an upward-looking onboard camera module mounted on a split-type flying vehicle.

3. The real-time visual recognition and positioning method among multiple modules of a split-type flying vehicle according to claim 1, characterized in that, The calibration process for camera intrinsic and extrinsic parameters specifically includes: A rigid body transformation is performed on the coordinate system of the image data to obtain the mapping relationship between points in the world coordinate system and points in the camera coordinate system; By applying pinhole imaging to the coordinate system that maps points in the world coordinate system to points in the camera coordinate system, the mapping relationship between the camera module coordinate system and the image coordinate system is obtained. The coordinate system of the mapping relationship between the camera module coordinate system and the image coordinate system is transformed by unit conversion and installation deviation and lens distortion are introduced to obtain the transformation result between the world coordinate system and the pixel coordinate system, that is, the mapping relationship from world coordinates to pixel coordinates.

4. The real-time visual recognition and positioning method among multiple modules of a split-type flying vehicle according to claim 1, characterized in that, The image preprocessing specifically includes: The image data is converted to grayscale to obtain a grayscale image; Gray(x,y)=0.299R(x,y)+0.578G(x,y)+0.114B(x,y); Where R(x,y), G(x,y), and B(x,y) are the values ​​of the three components of the RGB image, respectively; The grayscale image is thresholded to obtain a thresholded grayscale image; T(x,y)=WA s×s (x,y)-p。 5. The real-time visual recognition and positioning method among multiple modules of a split-type flying vehicle according to claim 1, characterized in that, The edge refinement process for the front view with markings and corner coordinates specifically includes: The ArUco detection algorithm is used to perform contour detection and ID recognition on the front view to obtain the initial contour result; By using the cost function, we can calculate the cost function for all points on the initial contour result and obtain a high-accuracy contour that minimizes the cost function. Where R(p) and E(p,x) i ) represents the actual and expected values ​​at image position p, given by x. i The line represents the initial line; the predicted value is obtained through a given edge model, using the distance d(p,x) from a given point p to the initial line. i ).

6. The real-time visual recognition and positioning method among multiple modules of a split-type flying vehicle according to claim 1, characterized in that, Based on images containing contours with high positioning accuracy, pose calculation is performed using the mapping relationship between world coordinates and pixel coordinates, specifically including: Calculate the relative position of the parking module of the split-type flying vehicle with respect to the stationary module; The center coordinates of each ArUco marker are calculated based on the quadrilateral contour of the image containing a high-accuracy positioning contour. By introducing the mapping relationship from world coordinates to pixel coordinates into the following formula, the relative position of the ground parking module of the split flying vehicle relative to the stationary module can be calculated. vector_center=vector_original-vector_camera; Where vector_original represents the position vector of a certain Aruco marker center point in the world coordinate system; vector_camera represents the position vector of a certain Aruco marker center point in the pixel coordinate system; and vector_center represents the position vector of the camera module coordinate system origin relative to the world coordinate system origin. Calculate the heading deviation angle of the parking module of the split-type flying vehicle relative to the stationary module; When the detection algorithm detects two or more ArUco markers in the processed input image, it inputs the center point coordinates and ID values ​​of each paired ArUco marker into the heading angle calculation module. This module calculates the linear equations of any pair of ArUco markers in both the world coordinate system and the pixel coordinate system. By using the mapping relationship from world coordinates to pixel coordinates, the vehicle heading angle can be calculated. Where α represents the angle between the line and the positive x-axis in the world coordinate system; β represents the angle between the line and the positive x-axis in the pixel coordinate system. The equation of the line is obtained by marking the coordinates of the center points of each pair of ArUco markers.

7. A real-time visual recognition and positioning system between multiple modules of a split-type flying vehicle, characterized in that, The system includes: The calibration module is used to calibrate the camera's intrinsic and extrinsic parameters on the image data, and to obtain the mapping relationship from world coordinates to pixel coordinates. The first processing module is used to acquire camera images and perform image preprocessing based on the mapping relationship from world coordinates to pixel coordinates to obtain a thresholded grayscale image. The second processing module is used to perform contour extraction, quadrilateral filtering, perspective transformation and internal encoding recognition processing on the thresholded grayscale image to obtain a front view with markings and corner coordinates. The third processing module is used to refine the edges of the front view with markings and corner coordinates to obtain an image with a high positioning accuracy contour. The pose calculation module is used to perform pose calculation based on an image containing a high-accuracy contour and by using the mapping relationship between world coordinates and pixel coordinates. It obtains the relative position of the camera lens center point with respect to the center point of the pre-arranged ArUco code area at the bottom of the stationary module and the heading deviation angle of the docking motion module relative to the stationary module, thus realizing accurate real-time visual recognition and positioning between multiple modules of the flying vehicle.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method of any one of claims 1 to 6.

Images