UAV precise guidance landing method and system based on cooperative goals

Abstract

The present invention relates to the field of unmanned aerial vehicle technology, and provides a method and system for accurately guiding the landing of unmanned aerial vehicles based on cooperative targets, the system comprising a landing identification device with multiple layers of nested coding identification, a visual acquisition module, a visual processing module and a flight control computer. The method comprises: S1, a landing identification device is set at the landing site of the unmanned aerial vehicle, and multiple layers of nested coding identification are set on the landing identification device; S2, when the unmanned aerial vehicle approaches the unmanned aerial vehicle site, the coding identification from the outermost layer to the innermost layer is sequentially identified, the visual position and attitude of the unmanned aerial vehicle is solved, and the confidence test is performed on the positioning data, and the unmanned aerial vehicle is guided to land accurately according to the positioning data after the confidence test. The present invention greatly expands the application scenarios and fields of unmanned aerial vehicles, improves the environmental adaptability of unmanned aerial vehicles, improves the navigation accuracy of unmanned aerial vehicle take-off and landing, and reduces the requirements of unmanned aerial vehicle take-off and landing on the site environment, which is particularly suitable for remote border areas where satellite positioning is not ideal and urban areas with complex electromagnetic environments.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicles, and in particular to a method and system for accurately guiding the landing of an unmanned aerial vehicle based on a cooperative target. Background Art

[0002] Drone takeoff and landing are the basic links and necessary functions for drones to perform flight missions, and have higher requirements for navigation and control accuracy. Drone vision-guided landing technology can reduce the requirements for the site for drone takeoff and landing, and can assist drones to achieve autonomous takeoff and landing in environments such as poor satellite positioning accuracy, satellite positioning interference, and magnetic field interference, thereby ensuring the safety of drone takeoff and landing.

[0003] In order to improve the accuracy of the visual positioning of the drone, the existing technology will place multiple coded markers at the landing site. The visual positioning processing module on the drone will parse and process the multiple coded markers in the image one by one to obtain the position coordinates of the landing site relative to the camera. The drone will guide the landing flight according to this relative position coordinate. In order to improve the accuracy of the analysis of the landing site position coordinates, the existing technology needs to place as many coded markers as possible at the landing site, which will increase the computing pressure of the visual processing equipment on the drone, resulting in the problem that the positioning data output by the visual positioning equipment is poor in real time and has a low effective accuracy rate.

[0004] The existing technology has the problems of poor real-time performance and low frame rate of visual positioning output data, as well as low visual positioning accuracy. Summary of the invention

[0005] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0006] In view of the above and / or existing problems, the present invention proposes a method and system for accurately guiding the landing of an unmanned aerial vehicle based on a cooperative target, which improves the real-time and accuracy of visual positioning data, so that even if there are fewer coded markers (cooperative targets) placed on the landing site, more accurate positioning information can be obtained, thereby improving the safety of visually guided flight of the unmanned aerial vehicle.

[0007] The present invention adopts the following technical solution:

[0008] On the one hand, the present invention provides a method for accurately guiding a UAV to land based on a cooperative target, comprising:

[0009] S1. A landing identification device (for example, a landing blanket) is provided at the landing site of the UAV, and a plurality of nested coding identifications are provided on the landing identification device, and each layer of coding identifications has the same positive direction, a unified rectangular coordinate system and a coordinate origin;

[0010] S2. When the UAV approaches the landing site from a high altitude, the outermost coding mark is first identified and located, and the position and attitude of the UAV relative to the landing mark device is solved. The UAV guides the landing flight according to the position and attitude data obtained by the identification and solution; during the landing process, the UAV successively identifies and locates the innermost coding mark, and re-solves the position and attitude of the UAV relative to the landing mark device, guiding the UAV to land on the landing mark device more accurately;

[0011] During the process of coding identification positioning and UAV position and attitude solution, the UAV position and attitude data is tested for confidence, and the UAV is guided to land based on the position and attitude data after the confidence test.

[0012] Any possible implementation as described above, further provides an implementation, in step S1, the coded identification uses three to five layers of nested identification, each layer of coded identification uses four coded identification images of the same size as graphic data for visual guidance positioning; the outermost layer of identification image size enables clear imaging of the UAV at an altitude of more than 30 meters, and the degree of image distortion is within a set threshold; the innermost layer of coded identification image size enables the camera to have the same imaging effect when the UAV is within 0.5 meters in altitude.

[0013] According to any possible implementation described above, a further implementation is provided, in step S2, for the coded identifiers of the same layer identified at the same time, a parallel processing thread is used to implement the processing of the single coded identifier image information, specifically including:

[0014] S2.1 identifies the coding mark from the coding mark image captured by the camera, and obtains the identification pixel coordinates of the coding mark;

[0015] S2.2 Calculate the rotation matrix and translation matrix of the center of the landing mark device relative to the center of the camera based on the known coordinate values ​​of the coded mark corner points on the landing mark device relative to the center of the landing mark device;

[0016] S2.3 obtains the calculated pixel coordinates of the coded mark in the image according to the rotation matrix and translation matrix obtained in step S2.2 and the coordinate value of the center of the coded mark in the central coordinate system of the landing mark device; subtracts the calculated pixel coordinates from the identified pixel coordinates obtained in step S2.1 to obtain a pixel difference; when the pixel difference exceeds a set threshold, the corresponding UAV position and attitude data is discarded, and the accuracy check of the UAV position and attitude data is completed;

[0017] S2.4 Calculate the confidence value of the drone position and attitude data based on the image quality of the coded mark, the accuracy of the position and attitude data, and the changes in the coded mark in the time series;

[0018] S2.5 weights and fuses the confidence parameters of the drone position and attitude data output by each thread to calculate the accurate drone position and attitude data.

[0019] Any possible implementation as described above, further provides an implementation, in step S2.3, in the position and attitude data accuracy check, the calculated position of the coded mark in the camera coordinate system and the camera attitude angle, the camera calibration parameters, and the position of the coded mark in the landing mark device coordinate system are fused to calculate the calculated pixel coordinates of the coded mark on the landing mark device in the camera image.

[0020] Any possible implementation as described above further provides an implementation, in which in step S2.4, the grayscale histogram distribution and image distortion of the coding mark image are used to measure the quality of the coding mark image; whether the change in the coding mark position within a set time exceeds a set threshold is used as a measure of the change in the coding mark in the time series.

[0021] Any possible implementation described above further provides an implementation, in step S2.4, the confidence value is calculated as follows:

[0022] W=(a*Q n +b*Rate acc )*J;

[0023] Where: W is the confidence value; Q n is the image quality parameter; Rate acc is the accuracy parameter of the pose data; J is the measurement parameter of the jump situation of the coding mark positioning; a and b are weight parameters.

[0024] On the other hand, the present invention also provides a cooperative target-based UAV precise guidance landing system, comprising a landing identification device provided with multi-layer nested coding identification, a visual acquisition module, a visual processing module and a flight control computer;

[0025] The multi-layer nested coding mark, each layer of coding mark uses four coding mark images of the same size as the graphic data for visual guidance positioning; the outermost layer of the mark image size enables the drone to have clear imaging above 30 meters, and the image distortion is within the set threshold; the innermost layer of the coding mark image size enables the drone to have the same imaging effect as a camera within 0.5 meters;

[0026] The visual acquisition module is arranged on the UAV, and is used to collect image data of each layer of the multi-layer nested coded identification, and transmit the image data to the visual processing module;

[0027] The visual processing module is used to receive the image data of the coded mark collected by the visual acquisition module, calculate the relative posture information between the UAV and the mark image in real time, perform accuracy check and confidence calculation on the relative posture information, and feed back the results of the accuracy check and confidence calculation to the flight control computer;

[0028] The flight control computer exchanges information with the visual processing module and guides the UAV to land accurately according to the information provided by the visual processing module.

[0029] As any possible implementation described above, an implementation is further provided, wherein the visual acquisition module is a fixed-focus visible light camera, for example, a 6mm focal length lens may be selected.

[0030] Any possible implementation as described above, further provides an implementation, wherein the visual acquisition module acquires RGB image data of a set frame rate and a set resolution, and transmits the output to the visual processing module through an image transmission cable (such as a MIPI cable), and the visual processing module is responsible for image data filtering, grayscale conversion, edge detection, clustering processing, two-dimensional code decoding, and callback detection of the displacement matrix, and fuses the solved relative positioning data with the inertial navigation, and sends the fused positioning data and confidence value to the flight control computer through an output transmission cable (such as a serial port cable).

[0031] As any possible implementation described above, a further implementation is provided, wherein the frame rate of the image captured by the visual acquisition module is 1 to 60 fps, and the resolution can be, for example, 2560*1920.

[0032] The beneficial effects of the present invention are:

[0033] The present invention will greatly expand the application scenarios and fields of drones, and improve the environmental adaptability and market value of drones. The drone precision guidance and landing technology based on cooperative goals can improve the navigation accuracy of drone takeoff and landing, and reduce the requirements of the site positioning environment for drone takeoff and landing. It is especially suitable for drone applications in remote border areas where satellite positioning is not ideal and urban areas with complex electromagnetic environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0035] Figure 1 The present invention is a schematic structural diagram of a UAV precision guidance landing system based on cooperative targets according to an embodiment of the present invention.

[0036] Figure 2 Schematic diagram of the process of visual positioning in an embodiment of the present invention.

[0037] Figure 3 Shown is a flowchart of the parallel processing thread workflow in an embodiment.

[0038] Figure 4 The figure shows an example of a multi-layer nested coding identification selected in the embodiment. DETAILED DESCRIPTION

[0039] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0040] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0041] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0042] like Figure 1 As shown, an embodiment of the present invention is a UAV precise guidance landing system based on cooperative targets, including a landing identification device provided with multi-layer nested coding identifications, a visual acquisition module, a visual processing module and a flight control computer;

[0043] The multi-layer nested coding mark, each layer of coding mark uses four coding mark images of the same size as the graphic data for visual guidance positioning; the outermost layer of the mark image size enables the drone to have clear imaging above 30 meters without serious distortion; the innermost layer of the coding mark image size enables the drone to have the same imaging effect as a camera within 0.5 meters;

[0044] The visual acquisition module is arranged on the drone and is used to collect image data of each layer of the multi-layer nested coded identifiers and transmit the image data to the visual processing module;

[0045] The visual processing module is used to receive the image data of the coded mark collected by the visual acquisition module, calculate the relative posture information between the UAV and the mark image in real time, perform accuracy check and confidence calculation on the relative posture information, and feed back the results of the accuracy check and confidence calculation to the flight control computer;

[0046] The flight control computer exchanges information with the visual processing module and guides the UAV to land accurately according to the information provided by the visual processing module.

[0047] The visual acquisition module and the visual processing module constitute the visual positioning system of the UAV. The visual acquisition module can collect RGB image data with a set frame rate (such as 1 to 60fps) and a set resolution (such as 2560*1920), and transmit the output to the visual processing module through an image transmission cable (such as a MIPI cable). The visual processing module is responsible for image data filtering, grayscale conversion, edge detection, clustering processing, QR code decoding, and callback detection of the displacement matrix, and fuses the solved relative positioning data with the inertial navigation, and sends the fused positioning data and confidence to the flight control computer through an output transmission cable (such as a serial port cable) to complete the auxiliary positioning function of the UAV.

[0048] In a specific embodiment, the multi-layer nested coding identifier is as follows: Figure 4 As shown, other coding marks with different patterns and arrangements may also be used. Figure 4 This is for illustration purposes only.

[0049] In a specific embodiment, the visual transmission module is a fixed-focus visible light camera, and the lens can be a 6mm focal length lens, which is used to collect image information of fixed ground markers in real time and transmit it back to the visual processing module.

[0050] like Figure 2 , Figure 3 As shown, an embodiment of the present invention provides a method for accurately guiding a UAV to land based on a cooperative target, including:

[0051] S1. A landing identification device is set up at the landing site of the UAV. The landing identification device is provided with multiple layers of nested coded identifications. Each layer of coded identifications has the same positive direction, a unified rectangular coordinate system and a coordinate origin.

[0052] S2. When the UAV approaches the landing site from a high altitude, the outermost coding mark is first identified and located, and the position and attitude of the UAV relative to the landing mark device is solved. The UAV guides the landing flight according to the position and attitude data obtained by the identification and solution; during the landing process, the UAV successively identifies and locates the innermost coding mark, and re-solves the position and attitude of the UAV relative to the landing mark device, guiding the UAV to land on the landing mark device more accurately;

[0053] During the process of coding identification positioning and UAV position and attitude solution, the UAV position and attitude data is tested for confidence, and the UAV is guided to land based on the position and attitude data after the confidence test.

[0054] In a specific embodiment, in step S1, the coding mark adopts three to five layers of nested marks, and each layer of coding mark adopts four coding mark images of the same size as graphic data for visual guidance positioning (the code can be a QR code); the outermost layer of the mark image size enables the drone to have clear imaging above 30 meters without serious distortion; the innermost layer of the coding mark image size achieves the same imaging effect as a camera within 0.5 meters for the drone.

[0055] For example, when three-layer coding labels are used, the drone can accurately visually identify the outer, middle and inner layers of coding labels when flying at high altitude, low altitude and close proximity.

[0056] In order to improve the accuracy of position estimation, the prior art will identify many coded logos at one time. The present invention can ensure that logo images of different nested layers can be identified at different flight altitudes, thereby ensuring the effect of accurate positioning. By using a landing logo device with nested coded logos, the number of logos to be identified at the same altitude is reduced, thereby reducing the amount of calculation for image operations.

[0057] In a specific embodiment, in step S2, for the coded marks of the same layer identified at the same time, a parallel processing thread is used to implement the processing of the single coded mark image information, specifically including:

[0058] S2.1 identifies the coding mark from the coding mark image captured by the camera, and obtains the identification pixel coordinates of the coding mark;

[0059] S2.2 Calculate the rotation matrix and translation matrix of the center of the landing mark device relative to the center of the camera based on the known coordinate values ​​of the coded mark corner points on the landing mark device relative to the center of the landing mark device;

[0060] S2.3 obtains the calculated pixel coordinates of the coded mark in the image according to the rotation matrix and translation matrix obtained in step S2.2 and the coordinate value of the center of the coded mark in the central coordinate system of the landing mark device; subtracts the calculated pixel coordinates from the identified pixel coordinates obtained in step S2.1 to obtain a pixel difference; when the pixel difference exceeds a set threshold, the corresponding UAV position and attitude data is discarded, and the accuracy check of the UAV position and attitude data is completed;

[0061] S2.4 Calculate the confidence value of the drone position and attitude data based on the image quality of the coded mark, the accuracy of the position and attitude data, and the changes in the coded mark in the time series;

[0062] S2.5 weights and fuses the confidence parameters of the drone position and attitude data output by each thread to calculate the accurate drone position and attitude data.

[0063] In a specific embodiment, in step S2.3, in the posture data accuracy check, the three types of data, namely, the calculated position of the coded marker in the camera coordinate system and the camera attitude angle, the camera calibration parameters, and the position of the coded marker in the landing marker device coordinate system, are fused to calculate the pixel coordinates of the coded marker on the landing marker device in the camera image.

[0064] The three types of data are fused and processed using the position of the coded mark in the camera coordinate system, the camera attitude angle, the camera calibration parameters, and the position of the coded mark in the landing mark device coordinate system. The implementation principle is as follows:

[0065] P pixel =M cam *(R 3*3 *P w +T 3*1 )

[0066] Among them, P pixel is the pixel coordinate value [P x P y 1] T ;M cam is the camera intrinsic parameter matrix; R 3*3 is the calculated rotation matrix of the landing marker relative to the camera; T 3*1 is the calculated translation matrix of the landing marker relative to the camera; P w It is the coordinates of the corner point or center point of the coded mark in the center coordinate system of the landing mark device.

[0067] Since the coordinates of the corner points or center points of the coded mark on the landing mark device relative to the center of the landing mark device are known, that is, P w It is known that after the above calculation, it can be concluded that: using the currently estimated rotation matrix R3*3 and the translation matrix T 3*1 The calculated pixel coordinates P pixel This pixel P pixel and the pixel coordinates P obtained by identifying the image rec_pixel The pixel difference err between them pixel =(err x err y 1), and use this pixel difference to calculate the pose data accuracy parameter Rate acc =err pixel / ThPIX max , where ThPIX max is the maximum allowable pixel difference threshold, and the setting of this threshold can be adjusted according to the maximum resolution of the camera and the actual flight situation of the aircraft. For example, ThPIX max =10, and Rate acc has a value range of (0 to 1.0).

[0068] In a specific embodiment, in step S2.4, the gray histogram distribution and image distortion of the encoded identification image are used to measure the quality of the encoded identification image; whether the change amount of the encoded identification position exceeds the set threshold within the set time is used as a measure of the change of the encoded identification in the time series.

[0069] The quality of the encoded identification image on the landing identification device can be obtained by weighted calculation through two image evaluation methods. One is to construct a gray histogram after graying the identification image and evaluate the distribution of the histogram. For example, calculate the average pixel gray value B av and count "the proportion A of the number of pixels with gray levels less than 255*a% in the gray histogram" and "the proportion B of the number of pixels with gray levels greater than 255*b% in the gray histogram" (a < b), perform image brightness evaluation calculation, and obtain the identification image brightness parameter G. For example: if A > 30%, B < 20% and B av < 100, it is considered that the image is too dark; if A < 30%, B > 20% and B av > 150, it is considered that the image is too bright; if A < 30%, B < 20% and B av ∈(100, 150), it is considered that the image brightness is appropriate. The image brightness parameter G of an over-bright or over-dark image will be smaller, and only when A, B and B av are within the appropriate range, the value of G will be close to 1, and G takes values from 0 to 1.0.

[0070] The other is the evaluation of the image distortion of the identification image. Measure the actual length-width ratio Ra of each encoded identification on the landing identification device in advance n , and calculate the length-width ratio ra of the identified encoded identification target nNote: If other images are used, the same method can be used. For example, the circular cooperation logo can calculate Ra by the ratio of the longest axis and the shortest axis from the center of the circle to the arc. n andra n .

[0071] Calculate the image distortion value E of the coded mark on the landing mark device n :

[0072] E n =1-|ra n -Ra n | / Ra n

[0073] The more serious the distortion of the logo image is, the higher the E n The smaller;

[0074] Image quality parameter Q of landing marking device n =α*G n +β*E n The weight factors α and β are adjusted according to the actual situation.

[0075] Record the relative position data calculated by each coding mark within a period of time T [T x1～N T y1～N T z1～N ] n Where n is the coding identification number, and 1 to N are the data numbers in the T time series.

[0076] Real-time determination of the relative position data T between the center of the newly calculated landing mark device and the center of the camera x0 、T y0 、T z0 Compared with the previous calculation of T x1 、T y1 、T z1 The difference eT x0 , eT y0 , eT z0 Are they all less than the threshold parameter T? thx 、T thy 、T thz .

[0077] If yes, then the calculated T x0 、T y0 、T z0 The data does not have a large jump, and this data is updated to the time series [T x1～N T y1～N T z1～N ] n In the example, the metric parameter J for determining the jump situation of the coding identification positioning is 1.

[0078] If not, determine the newly calculated T x0 , T y0 , T z0 The data is abnormal and unavailable, that is, the confidence value of the positioning data is 0, and this abnormal value is not saved in the time series [T x0～N T y0～N T z0～N ], the metric parameter J for determining the jump of the coding mark positioning is 0. The next newly calculated positioning data T x0 , T y0 , T z0 With T x1 , T y1 , T z1 The judgment threshold becomes eT x0 *2. eT y0 *2. eT z0 *2. If the data jump abnormality is judged again, the threshold value eT will be increased according to the number of judgments. x0 *m、eT y0 *m、eT z0 *m, where m is the number of judgments.

[0079] The existing drone landing technology uses a combination of satellite and inertial navigation data for positioning. The longitude and latitude of the drone landing point are preset in advance. During the drone landing process, the longitude and latitude of the drone are controlled to be close to the longitude and latitude of the landing point. This embodiment uses a fusion positioning method of visual positioning and inertial navigation data, and uses a Kalman filter algorithm to fuse visual positioning data with inertial navigation data.

[0080] The relative position of the drone and the landing point can be calculated by subtracting the current latitude and longitude of the drone from the latitude and longitude of the landing point [dt x dt y dt z ]. Due to satellite positioning interference or satellite signal blocking, there is a deviation in this relative position. The method of integrating visual relative positioning data with inertial navigation data can better make up for this deficiency.

[0081] The visual relative positioning reliability parameters can be calculated using the above parameters:

[0082] W=(a*Q n +b*Rate acc )*J

[0083] The weight parameters a and b can be adjusted according to the actual situation, such as a = 0.5, b = 0.5; Q n The image quality parameter of the landing marking device; Rate acc is the accuracy parameter of the pose data; J is the measurement parameter of the jump situation of the coding mark positioning.

[0084] The innovative point of the fusion positioning of visual relative positioning and inertial navigation data of the present invention is the use of a confidence parameter W for adjusting the weights of the visual relative positioning data and the inertial navigation positioning data in the visually guided landing positioning data result.

[0085] Vision-guided landing positioning data:

[0086] P xyz =W*[T x T y T z ]+(1-W)*[INP x INP y INP z ].

[0087] Among them, [VL x V L y V L z ] is the visual relative positioning; [INP x INP y INP z ] is the inertial navigation positioning data. The accuracy of the positioning data and the stability of the equipment can be improved by weighted calculation of the visual relative positioning data and the inertial navigation positioning data.

[0088] In the geodetic coordinate system, the position data of the landing marker relative to the drone: Pe xyz =C 3*3 *P xyz +P_inst xyz . Among them, C 3*3 P_inst is the rotation matrix from the UAV body coordinate system to the earth coordinate system. xyz is the installation error of the camera lens relative to the center of the drone. The position of the landing mark device relative to the drone in the geodetic coordinate system Pe xyz The visual positioning device ultimately sends the UAV’s guided landing flight data to the flight control computer.

[0089] The working principle of the present invention is as follows:

[0090] The parallel processing thread realizes the processing of a single coded mark image information, including: calculating the pixel coordinates of the mark, calculating the position coordinates of the coded mark in the camera coordinate system and the pitch, roll and heading angles of the camera relative to the landing mark device, completing the accuracy check of the position and attitude data, and calculating the confidence value of the visual positioning data according to the image quality of the coded mark, the accuracy of the posture data and the change of the coded mark in the time series, and packaging and integrating the confidence value and the calibrated coded mark position coordinates, attitude angle and other information. The positioning and attitude information output by each thread is weighted and fused with the confidence parameters to calculate the position and attitude data of the landing mark device.

[0091] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for accurately guiding the landing of an unmanned aerial vehicle based on cooperative goals, characterized in that: The method comprises: S1. A landing identification device is set up at the landing site of the UAV. The landing identification device is provided with multiple layers of nested coded identifications. Each layer of coded identifications has the same positive direction, a unified rectangular coordinate system and a coordinate origin. S2. When the UAV approaches the landing site from a high altitude, the outermost coding mark is first identified and located, and the position and attitude of the UAV relative to the landing mark device is solved. The UAV guides the landing flight according to the position and attitude data obtained by the identification and solution; during the landing process, the UAV successively identifies and locates the innermost coding mark, and re-solves the position and attitude of the UAV relative to the landing mark device, guiding the UAV to land on the landing mark device more accurately; During the process of coding identification positioning and UAV position and attitude calculation, the UAV position and attitude data is tested for confidence, and the UAV is guided to land based on the position and attitude data after the confidence test; In step S1, the coding mark adopts three to five layers of nested marks, and each layer of coding mark adopts four coding mark images of the same size as the graphic data for visual guidance positioning; the outermost layer of the mark image size enables clear imaging when the drone is above 30 meters, and the image distortion degree is within the set threshold; the innermost layer of the coding mark image size enables the camera to have the same imaging effect when the drone is within 0.5 meters in height; In step S2, for the coded marks of the same layer identified at the same time, a parallel processing thread is used to implement the processing of the single coded mark image information, specifically including: S2.1 identifies the coding mark from the coding mark image captured by the camera, and obtains the identification pixel coordinates of the coding mark; S2.2 Calculate the rotation matrix and translation matrix of the center of the landing mark device relative to the center of the camera based on the known coordinate values ​​of the coded mark corner points on the landing mark device relative to the center of the landing mark device; S2.3 obtains the calculated pixel coordinates of the coded mark in the image according to the rotation matrix and translation matrix obtained in step S2.2 and the coordinate value of the center of the coded mark in the central coordinate system of the landing mark device; subtracts the calculated pixel coordinates from the identified pixel coordinates obtained in step S2.1 to obtain a pixel difference; when the pixel difference exceeds a set threshold, the corresponding UAV position and attitude data is discarded, and the accuracy check of the UAV position and attitude data is completed; S2.4 Calculate the confidence value of the drone position and attitude data based on the image quality of the coded mark, the accuracy of the position and attitude data, and the changes in the coded mark in the time series; S2.5 weights and fuses the confidence parameters of the drone position and attitude data output by each thread to calculate the accurate drone position and attitude data.

2. The method for accurately guiding the landing of a UAV based on cooperative objectives as claimed in claim 1, characterized in that: In step S2.3, in the accuracy check of the position and attitude data, the three types of data, namely the calculated position of the coded marker in the camera coordinate system and the camera attitude angle, the camera calibration parameters, and the position of the coded marker in the landing mark device coordinate system, are fused to calculate the calculated pixel coordinates of the coded marker on the landing mark device in the camera image.

3. The method for accurately guiding the landing of a UAV based on cooperative objectives as claimed in claim 1, characterized in that: In step S2.4, the grayscale histogram distribution and image distortion of the coding mark image are used to measure the quality of the coding mark image; whether the change in the coding mark position within a set time exceeds a set threshold is used as a measure of the change in the coding mark in the time series.

4. The method for accurately guiding the landing of a UAV based on cooperative objectives as claimed in claim 1, characterized in that: In step S2.4, the confidence value is calculated as follows: W=(a*Q n +b*Rate acc )*J; Where: W is the confidence value; Q n is the image quality parameter; Rate acc is the accuracy parameter of the pose data; J is the measurement parameter of the jump situation of the coding mark positioning; a and b are weight parameters.

5. A UAV precision guidance landing system based on cooperative goals, used to implement the method according to any one of claims 1 to 4, characterized in that: The system includes a landing identification device provided with multi-layer nested coding identification, a visual acquisition module, a visual processing module and a flight control computer; The multi-layer nested coding mark, each layer of coding mark uses four coding mark images of the same size as the graphic data for visual guidance positioning; the outermost layer of the mark image size enables the drone to have clear imaging above 30 meters, and the image distortion is within the set threshold; the innermost layer of the coding mark image size enables the drone to have the same imaging effect as a camera within 0.5 meters; The visual acquisition module is arranged on the UAV, and is used to collect image data of each layer of the multi-layer nested coded identification, and transmit the image data to the visual processing module; The visual processing module is used to receive the image data of the coded mark collected by the visual acquisition module, calculate the relative posture information between the UAV and the mark image in real time, perform accuracy check and confidence calculation on the relative posture information, and feed back the results of the accuracy check and confidence calculation to the flight control computer; The flight control computer exchanges information with the visual processing module and guides the UAV to land accurately according to the information provided by the visual processing module.

6. The cooperative target-based UAV precise guidance landing system according to claim 5, characterized in that: The visual acquisition module is a fixed-focus visible light camera.

7. The cooperative target-based UAV precision guidance landing system according to claim 5, characterized in that: The visual acquisition module acquires RGB image data of a set frame rate and a set resolution, and transmits the output to the visual processing module through an image transmission cable. The visual processing module is responsible for image data filtering, grayscale conversion, edge detection, clustering processing, two-dimensional code decoding, and callback detection of the displacement matrix, and fuses the solved relative positioning data with the inertial navigation, and sends the fused positioning data and confidence value to the flight control computer through the output transmission cable.

8. The cooperative target-based UAV precise guidance landing system according to claim 6, characterized in that: The frame rate of the image collected by the visual acquisition module is 1 to 60 fps.

Images