An autonomous precision landing system for UAVs at fixed airports

Abstract

The present application belongs to the field of drone control and management technology, and in particular, relates to an autonomous precision landing system for drones at fixed airports. It includes a positioning and attitude detection module for obtaining drone positioning data, adjusting and collecting drone flight attitudes; a distance sensor for measuring the distance between the drone and the ground and fixed airport points; an image acquisition module for collecting images of fixed airports; and a rectangular marking frame is provided on the fixed airport. The present application is used for the recovery and landing of various types of drones, and is used to enhance the recovery efficiency of drones, improve the accuracy of landing, and ensure that drones can be quickly recovered and reused, so that when the overall speed of the drone is increased, it can better avoid crashes or collisions during the recovery process due to position deviations, and provide a basis for improving the recovery efficiency of drones and reducing operation time.

Description

Technical Field

[0001] The present application belongs to the technical field of unmanned aerial vehicle control and management, and in particular, relates to an autonomous precision landing system for unmanned aerial vehicles at fixed airports. Background Art

[0002] With the rapid development of drone design and manufacturing technology, drone systems are being used in more and more life and work scenarios. In more and more occasions, by configuring fixed airports for guiding and recovering drones (including various airports located on ground carrier platforms), the recovery and delivery efficiency of drones can be improved, and the efficiency and stability of drone daily operations can be ensured. Thanks to the image acquisition and positioning navigation systems equipped on most drones, during the landing process of drones, machine recognition and positioning guidance technology can be used to help drones quickly determine the airport location and achieve recovery. However, at present, the overall speed of drones approaching and recovering airports is relatively slow. With the increase of drone flight speed and control efficiency, the slow recovery process has an increasingly serious impact on the further improvement of overall operations. Summary of the invention

[0003] The purpose of this application is to provide a drone fixed airport autonomous precision landing system for drone recovery and delivery, which helps to improve drone recovery efficiency and landing accuracy.

[0004] To achieve the above objectives, this application adopts the following technical solutions.

[0005] An autonomous precision landing system for a UAV at a fixed airport, comprising a positioning and attitude detection module, a distance sensor, and an image acquisition module arranged on the UAV; a machine recognition module and a command control module arranged on the control end;

[0006] The positioning and attitude detection module is used to obtain the positioning data of the UAV, adjust and collect the flight attitude of the UAV, and transmit the corresponding information to the control end; the distance sensor is used to measure the distance between the UAV and the ground and the fixed airport points; the image acquisition module is used to collect images of fixed airports; the machine recognition module is used to process the image data obtained by the image acquisition module; the command control module is used to issue UAV flight control commands;

[0007] A rectangular marking frame is provided on the fixed airport, and the rectangular marking frame is used as a positioning identification element for the positioning and navigation of the UAV. For the convenience of identification, the interior of the rectangular marking frame in this embodiment is light-colored and the frame is dark-colored; for the convenience of determining the direction of the fixed airport, the size of the rectangular marking frame should ensure that when the UAV flies over the fixed airport, the ratio of the corresponding length to the pixel points of the frame in the image collected is not less than 2, and the thickness of the frame is not less than the pixel width that can be directly detected by the machine identification;

[0008] The autonomous precision landing system operates based on the following principles and steps:

[0009] Step 1: Use the positioning module and the control center to perform preliminary navigation based on the location of the fixed airport and the positioning of the UAV, so that the UAV is close to the fixed airport, and use the image acquisition module to obtain the image of the fixed airport. If the image of the fixed airport cannot be detected, lower the altitude and detect again;

[0010] Step 2: perform boundary detection based on the fixed airport image, delete unnecessary elements, and improve the recognition efficiency of the rectangular mark frame. Specifically, considering that the rectangular mark frame appears as an obvious boundary in the image, the original image is trimmed through the boundary recognition and shrinkage scheme, and the unnecessary element area of ​​the image is cut by shrinking from the outside to the inside;

[0011] Step 3: Continuously acquire fixed airport images based on the oscillation return method, adjust the UAV's flight attitude and flight distance, and make the UAV approach the fixed airport; specifically:

[0012] Step 4: When the drone is close enough to the fixed airport, control the drone to fly to the top of the rectangular positioning frame for vertical landing.

[0013] A further improvement or preferred implementation scheme of the aforementioned UAV fixed airport autonomous precision landing system, wherein step one specifically includes:

[0014] S1. Marking of calibration points based on fixed airports, specifically: using the diagonal endpoints of the rectangular marking frame as calibration points for positioning, and using the positioning system to determine the geodetic coordinates of the calibration points (x a ,y a 、z a )、(x b ,y b 、z b );

[0015] S2. Establish a coordinate system based on the relative position of the airport and the drone, specifically: take the planned landing point o as the coordinate origin, the ray from the calibration point a to the calibration point b as the positive direction of the z-axis, the opposite direction of gravity as the z-axis, and the positive direction of the z-axis rotated 90° counterclockwise around the z-axis as the positive direction of the y-axis to establish a positioning coordinate system X'Y'Z'; take the planned landing point o as the coordinate origin, and the axis direction of the geodetic coordinate system as the axis direction of the coordinate system to establish a relative coordinate system X"Y"Z";

[0016] S3. Complete the coordinate system conversion, specifically, determine the coordinate conversion ratio and coordinate value conversion relationship, including:

[0017] a. Complete the coordinate conversion of the calibration point in the relative coordinate system. Assume that X 1 Axis rotates counterclockwise to X0 The rotation angle of the axis is α; then the coordinates of the calibration point a in the relative coordinate system (x' a ,y' a )satisfy: Among them l oa is the straight-line distance between the planned landing point o and the calibration point a;

[0018] b. Complete the ratio conversion between the relative coordinate system and the geodetic coordinate system.

[0019] l ab is the straight-line distance between calibration point a and calibration point b;

[0020] c. Complete the conversion of calibration points between coordinate systems. The coordinates of calibration point a in the relative coordinate system are obtained by magnifying the coordinates of the geodetic coordinate system by k times. At the same time, the coordinates in the geodetic coordinate system are obtained by translating the coordinates of the relative coordinate system to the geodetic coordinate system. The coordinates of the planned landing point o in the geodetic coordinate system (x o ,y o ) can be expressed as The Z-direction distance between the calibration point a and the relative coordinate system is where β oa is the angle between the ray connecting the planned landing point o and the calibration point a and the horizontal plane; in the relative coordinate system, the Z coordinates of the planned landing point o and the calibration point a are consistent, that is,

[0021] S4. Determine the rotation angle of the coordinate system based on coordinate conversion, specifically, define the rotation angle of the X' axis counterclockwise to the X" axis as γ; the angle φ between the line connecting the calibration points a and b and the X" axis; the angle ∠oab formed by the line connecting the calibration point and the planned landing point o is λ; then the rotation angle γ is converted as follows:

[0022] but

[0023] S5. Determine the relative coordinates (x o ,y o ,z o ):

[0024] x o =x a +L a cosμ·K y o =y a +L a sinμ·K z o =z a +L a K;

[0025] in,

[0026]

[0027]

[0028] is the angle between the ray connecting the planned landing point o and the calibration point a and the horizontal plane; where l oa is the straight-line distance between the planned landing point o and the calibration point a; is the angle between the ray connecting the planned landing point o and the calibration point b and the horizontal plane; where l ob is the straight-line distance between the planned landing point o and the calibration point b; It is the angle formed by the projection of ∠oab in the UAV coordinate system;

[0029] S6. Based on the relative coordinates (x o ,y o ,z o ) Continuously correct the flight direction and altitude of the drone until the drone approaches a fixed airport, and use the image acquisition module to obtain the image of the fixed airport. If the image of the fixed airport cannot be detected, lower the altitude and detect again;

[0030] A further improvement or preferred implementation scheme of the aforementioned UAV fixed airport autonomous precision landing system is that if the image cannot be detected after repeated detection or after descending to the minimum altitude, the current attitude and position of the UAV is maintained and an error message is fed back until further operation instructions are obtained or a preset operation is performed; the preset operation refers to one of the following operations:

[0031] 1) Locate and navigate to the next available fixed airport;

[0032] 2) locate and navigate to the nearest available fixed airport;

[0033] 3) Return to the starting point.

[0034] A further improvement or preferred implementation scheme of the aforementioned UAV fixed airport autonomous precision landing system, wherein the step 2 specifically includes:

[0035] 2.1 Obtain fixed airport images, convert them into grayscale images, adjust the pixel resolution of all images to M*M; divide the images into square blocks with a pixel size of m*m; obtain OK Column Total Blocks are divided into rows and marked in the order of top to bottom and left to right.

[0036] 2.2 Calculate the grayscale mean of each block. Based on each column of blocks, analyze the blocks in the column from top to bottom. If a block M j If the grayscale mean value is lower than the preset threshold, block M j Block M on the lower side j+1 As the upper boundary, if block M j If the grayscale mean value is not lower than the preset threshold, block M j As the upper boundary, continue to observe the next block M j+1 ; If block M j+1 If the grayscale mean value is lower than the preset threshold, block M j+1 Block M on the lower side j+2 As the upper boundary, if block M j+1 If the grayscale mean value is not lower than the preset threshold, block M j+1 As the lower boundary; if the grayscale mean values ​​of all blocks in the column are lower than the preset threshold or exceed the preset advance, no upper and lower boundaries are set;

[0037] Based on each row of blocks, analyze each block in the column from left to right. If a block M j If the grayscale mean value is lower than the preset threshold, block M j Block M on the right j+1 As the left boundary, if block M j If the grayscale mean value is not lower than the preset threshold, block M j As the left boundary, continue to observe the right block M j+1 ; If block M j+1 If the grayscale mean value is lower than the preset threshold, block M j+1 Block M on the right j+2 As the left boundary, if block M j+1 If the grayscale mean value is not lower than the preset threshold, block M j+1 as the right boundary; if the grayscale mean values ​​of all blocks in the column are lower than the preset threshold or exceed the preset advance, the left and right boundaries are not set; connect all the upper boundaries with a straight line, connect all the lower boundaries with a straight line, connect all the left boundaries with a straight line, connect all the right boundaries with a straight line, intercept the closed area surrounded by the upper, lower, left and right boundaries, retain the elements inside the closed area as valid data to obtain the required rectangular element image.

[0038] A further improvement or preferred implementation scheme of the aforementioned UAV fixed airport autonomous precision landing system is to set the phase pixel size parameter m by using an active threshold in the process of boundary detection, where Among them, k m To adjust the scale parameter, m max is the pixel size parameter, H iThe flying altitude of the drone.

[0039] A further improvement or preferred implementation scheme of the aforementioned UAV fixed airport autonomous precision landing system, wherein step three specifically includes:

[0040] 3.1 Continuously acquire rectangular element images, perform rectangular fitting according to the closed areas determined in each rectangular element image, obtain the fitted rectangular area corresponding to the closed area in each rectangular element image, screen the fitted rectangular area according to the actual aspect ratio of the rectangular marking frame, and delete the closed area with inconsistent aspect ratio; based on the current UAV flight altitude and the scaling ratio of the image acquisition module, calculate the pixel area interval of the rectangular marking frame in the image, and delete the fitted rectangular area whose pixel area is outside the interval;

[0041] 3.2 Lower or raise the flight altitude and repeat step 3.1 until a unique fitting rectangular area is obtained. In this process, use the circumscribed circle to intercept the rectangular element image so that the circumscribed circle can cover all the remaining fitting rectangular areas. According to the position of the center of the circumscribed circle in the rectangular element image, adjust the front, back, left, and right flight distance of the drone so that the vertical projection of the center of the drone is close to the center of the circumscribed circle.

[0042] A further improvement or preferred implementation scheme of the aforementioned UAV fixed airport autonomous precision landing system is as follows: during the flight distance adjustment process, according to the horizontal distance S between the vertical projection of the center of the UAV and the center of the circumscribed circle i And the height of the drone H i , adjust the horizontal displacement distance of the drone according to the proportion

[0043] Where S i H is the horizontal distance between the vertical projection of the center of the UAV detected during the i-th adjustment and the center of the circumscribed circle; i ΔH is the current height of the drone at the i-th adjustment; i is the adjustment amount of the UAV's flight altitude during the i-th adjustment; k i is the adjustment ratio, which is a decimal whose value decreases as the adjustment number i increases.

[0044] Further improvement or preferred implementation scheme of the aforementioned UAV fixed airport autonomous precision landing system, k i As the average flight speed of the drone increases, the ratio k is adjusted i There is a shrinking trend.

[0045] Its beneficial effects are:

[0046] The drone fixed airport autonomous precision landing system of the present application is mainly used for the recovery and landing of various types of drones, for enhancing the recovery efficiency of drones, improving the accuracy of landing, ensuring that drones can be quickly recovered and reused, and increasing the overall speed of the drone. It can better avoid problems such as crashes or collisions during the recovery process due to position deviations, providing a basis for improving the recovery efficiency of drones and reducing operation time. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a schematic diagram of the implementation steps of the autonomous precision landing system for drones at fixed airports;

[0048] Figure 2 Schematic diagram of the initial navigation process;

[0049] Figure 3 It is a schematic diagram of the process of boundary detection and oscillation centering based on fixed airport images. DETAILED DESCRIPTION

[0050] The present application is described in detail below in conjunction with specific embodiments.

[0051] The UAV fixed airport autonomous precision landing system of the present application is mainly used as a supporting system for various types of UAV equipment, used to assist UAV equipment in quickly locating the landing point, improving the accuracy of the landing process, and reducing the landing time.

[0052] The system mainly includes a positioning and attitude detection module, a distance sensor, and an image acquisition module installed on the drone; a machine recognition module and a command control module installed on the control end;

[0053] The positioning and attitude detection module is used to obtain the positioning data of the UAV, adjust and collect the flight attitude of the UAV, and transmit the corresponding information to the control end; the distance sensor is used to measure the distance between the UAV and the ground and the fixed airport points; the image acquisition module is used to collect images of fixed airports; the machine recognition module is used to process the image data obtained by the image acquisition module; the command control module is used to issue UAV flight control commands;

[0054] A rectangular marking frame is provided on the fixed airport. The rectangular marking frame is used as a positioning identification element for the positioning and navigation of the UAV. For the convenience of identification, the interior of the rectangular marking frame in this embodiment is light-colored and the frame is dark-colored. For the convenience of determining the orientation of the fixed airport, the size of the rectangular marking frame should ensure that when the UAV flies over the fixed airport, the ratio of its corresponding length to the pixel points of the frame in the image collected is not less than 2, and the thickness of the frame is not less than the pixel width that can be directly detected by the machine identification.

[0055] The operating principle and operating steps of the autonomous precision landing system of this application are as follows: Figure 1The specific steps are described as follows.

[0056] Step 1: Use the positioning module and the control center to perform preliminary navigation based on the location of the fixed airport and the positioning of the UAV, so that the UAV is close to the fixed airport, and use the image acquisition module to obtain the image of the fixed airport. If the image of the fixed airport cannot be detected, lower the altitude and detect again;

[0057] In the actual implementation process, due to the different natures of fixed airports and UAV equipment, the two generally use different positioning systems. Fixed airports are generally based on geocentric coordinate systems, paracentric coordinate systems, and regional independent coordinate systems established based on surface parameters such as the earth's longitude and latitude or meridians. The coordinates of each point at the airport are measured using different geodetic coordinate references, while mobile devices (including UAVs) mainly use GPS, Beidou and other satellite navigation systems for positioning and navigation, which are not consistent with each other. In order to meet the needs of multiple devices and fast positioning at any location within a fixed airport, during the operation of this system, the coordinate conversion and positioning steps are established using benchmark calibration points, and the calibration points within the fixed airport are used to guide the conversion of three-dimensional coordinates of any landing point within the fixed airport, so that it can be associated with the UAV's own coordinate system, so that the UAV's flight attitude and target can be determined based on the coordinates of two points in the UAV's coordinate system. The specific steps are as follows: Figure 2 shown, including:

[0058] S1. Marking of calibration points based on fixed airports, specifically: using the diagonal endpoints of the rectangular marking frame as calibration points for positioning, and using the positioning system to determine the geodetic coordinates of the calibration points (x a ,y a 、z a )、(x b ,y b 、z b );

[0059] S2. Establish a coordinate system based on the relative position of the airport and the drone, specifically: take the planned landing point o as the coordinate origin, the ray from the calibration point a to the calibration point b as the positive direction of the z-axis, the opposite direction of gravity as the z-axis, and the positive direction of the z-axis rotated 90° counterclockwise around the z-axis as the positive direction of the y-axis to establish a positioning coordinate system X'Y'Z'; take the planned landing point o as the coordinate origin, and the axis direction of the geodetic coordinate system as the axis direction of the coordinate system to establish a relative coordinate system X"Y"Z";

[0060] S3. Complete the coordinate system conversion, specifically, determine the coordinate conversion ratio and coordinate value conversion relationship, including:

[0061] a. Complete the coordinate conversion of the calibration point in the relative coordinate system. Assume that X 1 Axis rotates counterclockwise to X 0The rotation angle of the axis is α; then the coordinates of the calibration point a in the relative coordinate system (x' a ,y' a )satisfy: Among them l oa is the straight-line distance between the planned landing point o and the calibration point a;

[0062] b. Complete the ratio conversion between the relative coordinate system and the geodetic coordinate system.

[0063] l ab is the straight-line distance between calibration point a and calibration point b;

[0064] c. Complete the conversion of calibration points between coordinate systems. The coordinates of calibration point a in the relative coordinate system are obtained by magnifying the coordinates of the geodetic coordinate system by k times. At the same time, the coordinates in the geodetic coordinate system are obtained by translating the coordinates of the relative coordinate system to the geodetic coordinate system. The coordinates of the planned landing point o in the geodetic coordinate system (x o ,y o ) can be expressed as The Z-direction distance between the calibration point a and the relative coordinate system is where β oa is the angle between the ray connecting the planned landing point o and the calibration point a and the horizontal plane; in the relative coordinate system, the Z coordinates of the planned landing point o and the calibration point a are consistent, that is,

[0065] S4. Determine the rotation angle of the coordinate system based on coordinate conversion, specifically, define the rotation angle of the X' axis counterclockwise to the X" axis as γ; the angle φ between the line connecting the calibration points a and b and the X" axis; the angle ∠oab formed by the line connecting the calibration point and the planned landing point o is λ; then the rotation angle γ is converted as follows:

[0066] but

[0067] S5. Determine the relative coordinates (x o ,y o ,z o ):

[0068] x o =x a +L a cosμ·K y o =y a +L a sinμ·K z o =z a +L a K;

[0069] in,

[0070]

[0071]

[0072] is the angle between the ray connecting the planned landing point o and the calibration point a and the horizontal plane; where l oa is the straight-line distance between the planned landing point o and the calibration point a; is the angle between the ray connecting the planned landing point o and the calibration point b and the horizontal plane; where l ob is the straight-line distance between the planned landing point o and the calibration point b; It is the angle formed by the projection of ∠oab in the UAV coordinate system;

[0073] S6. Based on the relative coordinates (x o ,y o ,z o ) Continuously correct the flight direction and altitude of the drone until the drone approaches a fixed airport, and use the image acquisition module to obtain the image of the fixed airport. If the image of the fixed airport cannot be detected, lower the altitude and detect again;

[0074] In particular, to avoid situations where detection cannot be made due to unexpected circumstances, such as changes in the structure of a fixed airport, if the image cannot be detected after repeated detection or after descending to the minimum altitude, the current attitude and position of the drone will be maintained and an error message will be fed back until further operation instructions are obtained or a preset operation is performed; the preset operation refers to one of the following operations:

[0075] 1) Locate and navigate to the next available fixed airport;

[0076] 2) locate and navigate to the nearest available fixed airport;

[0077] 3) Return to the starting point;

[0078] Step 2: perform boundary detection based on the fixed airport image, delete unnecessary elements, and improve the recognition efficiency of the rectangular marking box;

[0079] During the entire positioning process, the visual detection system and the control center need to continuously obtain images containing fixed airports collected by the drone to make factual judgments on the relative position relationship between the drone and the fixed airport in order to control the flight posture of the drone. In this process, a large amount of image data processing is involved, but in actual images, the data of fixed airports often only involves a small part, and the higher the flying altitude of the drone, the higher the proportion of non-essential data. In order to reduce the amount of data processing and improve the efficiency of machine recognition, step two of this application is used to implement image boundary shrinkage processing. According to the strong edge characteristics of the rectangular marking box of the fixed airport in the image, useless data is deleted through boundary shrinkage and interception. Specifically, considering that the rectangular marking box appears as an obvious boundary in the image, the original image is trimmed through the boundary recognition and shrinkage scheme, and the non-essential element area of ​​the image is cut by shrinking from the outside to the inside; the specific steps are as follows Figure 3 shown, including:

[0080] 2.1 Obtain fixed airport images, convert them into grayscale images, adjust the pixel resolution of all images to M*M; divide the images into square blocks with a pixel size of m*m; obtain OK Column Total Blocks are divided into rows and marked in the order of top to bottom and left to right.

[0081] 2.2 Calculate the grayscale mean of each block. Based on each column of blocks, analyze the blocks in the column from top to bottom. If a block M j If the grayscale mean value is lower than the preset threshold, block M j Block M on the lower side j+1 As the upper boundary, if block M j If the grayscale mean value is not lower than the preset threshold, block M j As the upper boundary, continue to observe the next block M j+1 ; If block M j+1 If the grayscale mean value is lower than the preset threshold, block M j+1 Block M on the lower side j+2 As the upper boundary, if block M j+1 If the grayscale mean value is not lower than the preset threshold, block M j+1 As the lower boundary; if the grayscale mean values ​​of all blocks in the column are lower than the preset threshold or exceed the preset advance, no upper and lower boundaries are set;

[0082] Based on each row of blocks, analyze each block in the column from left to right. If a block M j If the grayscale mean value is lower than the preset threshold, block M j Block M on the rightj+1 As the left boundary, if block M j If the grayscale mean value is not lower than the preset threshold, block M j As the left boundary, continue to observe the right block M j+1 ; If block M j+1 If the grayscale mean value is lower than the preset threshold, block M j+1 Block M on the right j+2 As the left boundary, if block M j+1 If the grayscale mean value is not lower than the preset threshold, block M j+1 as the right boundary; if the grayscale mean values ​​of all blocks in the column are lower than the preset threshold or exceed the preset advance, then no left and right boundaries are set; connect all upper boundaries with a straight line, connect all lower boundaries with a straight line, connect all left boundaries with a straight line, connect all right boundaries with a straight line, intercept the closed area surrounded by the upper, lower, left and right boundaries, and retain the elements inside the closed area as valid data to obtain the required rectangular element image;

[0083] In particular, in the process of boundary detection in the above steps, the accuracy of boundary detection can be controlled by setting the pixel size m*m of the block. Generally, the smaller the pixel size of the block, the higher the efficiency of boundary detection. However, the corresponding amount of calculation increases, and the cost requirements for equipment hardware and software increase. In order to achieve the coordination of implementation cost and effect, considering that as the distance between the drone and the fixed airport decreases, the boundary size and clarity of the rectangular marking box in the image will continue to increase, the active threshold method can be used to set the pixel size parameter m, where Among them, k m To adjust the scale parameter, m max is the pixel size parameter, H i The flight altitude of the drone;

[0084] Step 3: Continuously acquire fixed airport images based on the oscillation return method, adjust the UAV's flight attitude and flight distance, and make the UAV approach the fixed airport; specifically:

[0085] 3.1 Continuously acquire rectangular element images, perform rectangular fitting according to the closed areas determined in each rectangular element image, obtain the fitted rectangular area corresponding to the closed area in each rectangular element image, screen the fitted rectangular area according to the actual aspect ratio of the rectangular marking frame, and delete the closed area with inconsistent aspect ratio; based on the current UAV flight altitude and the scaling ratio of the image acquisition module, calculate the pixel area interval of the rectangular marking frame in the image, and delete the fitted rectangular area whose pixel area is outside the interval;

[0086] 3.2 Lower or raise the flight altitude and repeat step 3.1 until a unique fitting rectangular area is obtained. In this process, use the circumscribed circle to intercept the rectangular element image so that the circumscribed circle can cover all the remaining fitting rectangular areas. According to the position of the center of the circumscribed circle in the rectangular element image, adjust the front, back, left, and right flight distance of the drone so that the vertical projection of the center of the drone is close to the center of the circumscribed circle.

[0087] In particular, as the attitude of the drone is adjusted, the drone gradually approaches the landing point, and the horizontal distance of the drone from the landing point in the direction of the drone's flight continues to shrink until it reaches zero. However, due to the inertia of the drone in this process, it may cause offside. To avoid this phenomenon, the horizontal distance is steadily reduced each time the drone is adjusted, and the adjustment target of the attitude control is controlled by adjusting the ratio. Specifically, according to the horizontal distance S between the vertical projection of the center of the drone and the center of the circumscribed circle, i And the height of the drone H i , adjust the horizontal displacement distance of the drone according to the proportion

[0088] Where S i H is the horizontal distance between the vertical projection of the center of the UAV detected during the i-th adjustment and the center of the circumscribed circle; i ΔH is the current height of the drone at the i-th adjustment; i is the adjustment amount of the UAV's flight altitude during the i-th adjustment; k i To adjust the proportion;

[0089] Furthermore, as the UAV gets closer to the landing point, in order to further improve the control accuracy, the active adjustment ratio is used to control the adjustment target of the attitude control, where the adjustment ratio parameter k i is a decimal whose value keeps decreasing as the number of adjustments i increases; and as the average flight speed of the drone increases, the adjustment ratio k i The trend is shrinking; adjust the ratio k i The actual value is set according to the current system control accuracy and drone control performance.

[0090] Step 4: When the drone is close enough to the fixed airport, control the drone to fly to the top of the rectangular positioning frame for vertical landing.

[0091] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present application, rather than to limit the scope of protection of the present application. Although the present application has been described in detail with reference to the preferred embodiments, ordinary technicians in this field should understand that the technical solution of the present application can be modified or replaced by equivalents without departing from the essence and scope of the technical solution of the present application.

Claims

1. An autonomous precision landing system for UAVs at fixed airports, characterized in that: It includes a positioning and attitude detection module, a distance sensor, and an image acquisition module installed on the drone; a machine recognition module and a command control module installed on the control end; The positioning and attitude detection module is used to obtain the positioning data of the UAV, adjust and collect the flight attitude of the UAV, and transmit the corresponding information to the control end; the distance sensor is used to measure the distance between the UAV and the ground and the fixed airport points; the image acquisition module is used to collect images of fixed airports; the machine recognition module is used to process the image data obtained by the image acquisition module; the command control module is used to issue UAV flight control commands; A rectangular marking frame is provided on the fixed airport, and the rectangular marking frame is used as a positioning identification element for the positioning and navigation of the UAV. For the convenience of identification, the interior of the rectangular marking frame is light-colored and the frame is dark-colored. For the convenience of determining the direction of the fixed airport, the size of the rectangular marking frame should ensure that when the UAV flies over the fixed airport, the ratio of the corresponding length to the pixel points of the frame in the image collected is not less than 2, and the thickness of the frame is not less than the pixel width that can be directly detected by the machine identification. The autonomous precision landing system operates based on the following principles and steps: Step 1: Use the positioning module and the control center to perform preliminary navigation based on the location of the fixed airport and the positioning of the UAV, so that the UAV is close to the fixed airport, and use the image acquisition module to obtain the image of the fixed airport. If the image of the fixed airport cannot be detected, lower the altitude and detect again; Step 2: perform boundary detection based on the fixed airport image, delete unnecessary elements, and improve the recognition efficiency of the rectangular mark frame. Specifically, considering that the rectangular mark frame appears as an obvious boundary in the image, the original image is trimmed through the boundary recognition and shrinking scheme, and the unnecessary element area of ​​the image is cut by shrinking from the outside to the inside; specifically, it includes: 2.1 Obtain fixed airport images, convert them into grayscale images, adjust the pixel resolution of all images to M*M; divide the images into square blocks with a pixel size of m*m; obtain OK Column Total Blocks are divided into rows and marked as M in the order from top to bottom and from left to right. i , 2.2 Calculate the grayscale mean of each block. Based on each column of blocks, analyze the blocks in the column from top to bottom. If a block M j If the grayscale mean value is lower than the preset threshold, block M j Block M on the lower side j+1 As the upper boundary, if block M j If the grayscale mean value is not lower than the preset threshold, block M j As the upper boundary, continue to observe the next block M j+1 ; If block M j+1 If the grayscale mean value is lower than the preset threshold, block M j+1 Block M on the lower side j+2 As the upper boundary, if block M j+1 If the grayscale mean value is not lower than the preset threshold, block M j+1 As the lower boundary; if the grayscale mean values ​​of all blocks in the column are lower than the preset threshold or all exceed the preset threshold, no upper and lower boundaries are set; Based on each row of blocks, analyze each block in the column from left to right. If a block M j If the grayscale mean value is lower than the preset threshold, block M j Block M on the right j+1 As the left boundary, if block M j If the grayscale mean value is not lower than the preset threshold, block M j As the left boundary, continue to observe the right block M j+1 ; If block M j+1 If the grayscale mean value is lower than the preset threshold, block M j+1 Block M on the right j+2 As the left boundary, if block M j+1 If the grayscale mean value is not lower than the preset threshold, block M j+1 as the right boundary; if the grayscale mean values ​​of all blocks in the column are lower than the preset threshold or exceed the preset advance, then no left and right boundaries are set; connect all upper boundaries with a straight line, connect all lower boundaries with a straight line, connect all left boundaries with a straight line, connect all right boundaries with a straight line, intercept the closed area surrounded by the upper, lower, left and right boundaries, and retain the elements inside the closed area as valid data to obtain the required rectangular element image; In the process of boundary detection, the active threshold method is used to set the phase pixel size parameter m, where Among them, k m To adjust the scale parameter, m max is the pixel size parameter, H i The flight altitude of the drone; Step 3: Continuously acquire fixed airport images based on the oscillation return method, adjust the UAV's flight attitude and flight distance, and make the UAV approach the fixed airport; Step 4: When the drone is close enough to the fixed airport, control the drone to fly to the top of the rectangular positioning frame for vertical landing.

2. The UAV fixed airport autonomous precision landing system according to claim 1 is characterized in that: The step 1 specifically includes: S1. Marking of calibration points based on fixed airports, specifically: using the diagonal endpoints of the rectangular marking frame as calibration points for positioning, and using the positioning system to determine the geodetic coordinates of the calibration points (x a ,y a 、z a )、(x b ,y b 、z b ); S2. Establish a coordinate system based on the relative position of the airport and the drone, specifically: take the planned landing point o as the coordinate origin, the ray from the calibration point a to the calibration point b as the positive direction of the z-axis, the opposite direction of gravity as the z-axis, and the positive direction of the z-axis rotated 90° counterclockwise around the z-axis as the positive direction of the y-axis to establish a positioning coordinate system X'Y'Z'; take the planned landing point o as the coordinate origin, and the axis direction of the geodetic coordinate system as the axis direction of the coordinate system to establish a relative coordinate system X"Y"Z"; S3. Complete the coordinate system conversion, specifically, determine the coordinate conversion ratio and coordinate value conversion relationship, including: a. Complete the coordinate conversion of the calibration point in the relative coordinate system. Assume that X 1 Axis rotates counterclockwise to X 0 The rotation angle of the axis is α; then the coordinates of the calibration point a in the relative coordinate system (x' a ,y' a )satisfy: Among them l oa is the straight-line distance between the planned landing point o and the calibration point a; b. Complete the ratio conversion between the relative coordinate system and the geodetic coordinate system. l ab is the straight-line distance between calibration point a and calibration point b; c. Complete the conversion of calibration points between coordinate systems. The coordinates of calibration point a in the relative coordinate system are obtained by magnifying the coordinates of the geodetic coordinate system by k times. At the same time, the coordinates in the geodetic coordinate system are obtained by translating the coordinates of the relative coordinate system to the geodetic coordinate system. Then the coordinates of the planned landing point o in the geodetic coordinate system (x o ,y o ) can be expressed as The Z-direction distance between the calibration point a and the relative coordinate system is where β oa is the angle between the ray connecting the planned landing point o and the calibration point a and the horizontal plane; in the relative coordinate system, the Z coordinates of the planned landing point o and the calibration point a are consistent, that is, S4. Determine the rotation angle of the coordinate system based on coordinate conversion, specifically, define the rotation angle of the X' axis counterclockwise to the X" axis as γ; the angle φ between the line connecting the calibration points a and b and the X" axis; the angle ∠oab formed by the line connecting the calibration point and the planned landing point o is λ; then the rotation angle γ is converted as follows: but S5. Determine the relative coordinates (x o ,y o ,z o ): x o =x a +L a cosμ·K y o =y a +L a sinμ·K z o =z a +L a ·K; in, is the angle between the ray connecting the planned landing point o and the calibration point a and the horizontal plane; where l oa is the straight-line distance between the planned landing point o and the calibration point a; is the angle between the ray connecting the planned landing point o and the calibration point b and the horizontal plane; where l ob is the straight-line distance between the planned landing point o and the calibration point b; It is the angle formed by the projection of ∠oab in the UAV coordinate system; S6. Based on the relative coordinates (x o ,y o ,z o ) Continue to correct the flight direction and altitude of the drone until the drone approaches a fixed airport, and use the image acquisition module to obtain the image of the fixed airport. If the image of the fixed airport cannot be detected, lower the altitude and detect again.

3. The UAV fixed airport autonomous precision landing system according to claim 2 is characterized in that: If the image cannot be detected after repeated detection or descending to the minimum altitude, the current attitude and position of the drone will be maintained and an error message will be fed back until further operation instructions are obtained or a preset operation is performed; the preset operation refers to one of the following operations: 1) locating and navigating to the next available fixed airport; 2) locating and navigating to the nearest available fixed airport; 3) returning to the starting point.

4. The UAV fixed airport autonomous precision landing system according to claim 1 is characterized in that: The step three specifically includes: 3.1 Continuously acquire rectangular element images, perform rectangular fitting according to the closed areas determined in each rectangular element image, obtain the fitted rectangular area corresponding to the closed area in each rectangular element image, screen the fitted rectangular area according to the actual aspect ratio of the rectangular marking frame, and delete the closed area with inconsistent aspect ratio; based on the current UAV flight altitude and the scaling ratio of the image acquisition module, calculate the pixel area interval of the rectangular marking frame in the image, and delete the fitted rectangular area whose pixel area is outside the interval; 3.2 Lower or increase the flight altitude and repeat step 3.1 until a unique fitting rectangular area is obtained. In this process, use the circumscribed circle to intercept the rectangular element image so that the circumscribed circle can include all remaining fitting rectangular areas. According to the position of the center of the circumscribed circle in the rectangular element image, adjust the front, back, left and right flight distance of the UAV so that the vertical projection of the center of the UAV is close to the center of the circumscribed circle.

5. The UAV fixed airport autonomous precision landing system according to claim 4 is characterized in that: During the flight distance adjustment process, the horizontal distance S between the vertical projection of the center of the drone and the center of the circumscribed circle is calculated. i And the height of the drone H i , adjust the horizontal displacement distance of the drone according to the proportion Where S i H is the horizontal distance between the vertical projection of the center of the UAV detected during the i-th adjustment and the center of the circumscribed circle; i ΔH is the current height of the drone at the i-th adjustment; i is the adjustment amount of the UAV's flight altitude during the i-th adjustment; k i is the adjustment ratio, which is a decimal whose value decreases as the adjustment number i increases.

6. The UAV fixed airport autonomous precision landing system according to claim 5 is characterized in that: k i As the average flight speed of the drone increases, the ratio k is adjusted i There is a shrinking trend.

Images