Unmanned aerial vehicle navigation landing method and system for mobile landing platform

Abstract

The invention provides an unmanned aerial vehicle navigation landing method and system for a mobile landing platform, and the method comprises the steps: capturing a two-dimensional code image of a vehicle roof airport landing platform in real time through a sensor (such as a high-definition camera, an infrared sensor and the like) carried by an unmanned aerial vehicle, and comparing the two-dimensional code image with a pre-stored reference image library through a scene matching algorithm; and determining the accurate position of the unmanned aerial vehicle. Meanwhile, a specific identifier, such as a two-dimensional code, on an airport landing platform is quickly identified by using a coded image identification technology, and the specific identifier is used as accurate guidance of a landing point. And the precise landing of the unmanned aerial vehicle in the non-cooperative environment is realized by fusing the scene matching result and the coded image recognition result.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicles, and in particular to a navigation and landing method and system for an unmanned aerial vehicle (UAV) on a mobile landing platform. Background Art

[0002] As an emerging aircraft, drones have shown great application potential in military reconnaissance, civil aerial photography, logistics and distribution, etc. However, in non-cooperative environments, such as complex urban areas and unknown areas in the wild, the precise landing of drones faces many challenges, such as GPS signal obstruction and unclear ground reference objects. Therefore, it is of great significance to study a drone precise landing navigation method that does not rely on external infrastructure, has strong adaptability and high accuracy.

[0003] At present, the main positioning and navigation methods for drones include the Global Navigation Satellite System (GNSS), Inertial Navigation System (INS), and Ultra Wide Band (UWB). The global satellite navigation system relies on good communication with satellites and is easily interfered by factors such as occlusion and electromagnetic signals. It has poor adaptability in non-cooperative environments. Inertial navigation positioning errors accumulate rapidly over time and cannot be used for long-distance navigation and positioning. Radio positioning often requires the deployment of additional equipment and cannot be applied to non-cooperative environments. Summary of the invention

[0004] The present invention provides a method and system for unmanned aerial vehicle navigation and landing on a mobile landing platform, which are used to solve the defect that the positioning and navigation of unmanned aerial vehicles on mobile landing platforms in the prior art are not accurate enough.

[0005] In a first aspect, the present invention provides a method for navigation and landing of an unmanned aerial vehicle (UAV) on a mobile landing platform, comprising: Acquire a logo image of the mobile landing platform; Using the SURF-BRISK algorithm, real-time scene capture and matching are performed on the identification image and a pre-established reference image library to obtain a scene matching result; Performing coded image recognition on the scene matching result to obtain a coded image result; Based on the scene matching result and the coded image result, a navigation path plan for a mobile landing platform is obtained, and navigation flight control is performed according to the navigation path plan for the mobile landing platform.

[0006] According to a method for navigation and landing of a drone on a mobile landing platform provided by the present invention, before acquiring an identification image of the mobile landing platform, the method further includes: Setting the identification image on the mobile landing platform according to environmental factors; Complete the system configuration of the mobile landing platform and the UAV respectively; A reference image library of the identification image is established.

[0007] According to a method for navigation and landing of an unmanned aerial vehicle on a mobile landing platform provided by the present invention, a reference image library of the identification image is established, comprising: Collect multiple images of the mobile landing platform relative to the landing area of ​​the drone at different angles; Performing denoising, contrast enhancement and color correction on the multiple images to obtain preprocessed images; The preprocessed image is subjected to edge detection and contour extraction based on the Canny operator and the Freeman chain code to obtain the reference image library.

[0008] According to a method for navigation and landing of a UAV on a mobile landing platform provided by the present invention, a SURF-BRISK algorithm is used to perform real-time scene capture and matching on the identification image and a pre-established reference image library to obtain a scene matching result, including: Obtain multi-layer nested images to be registered; Using the SURF algorithm to extract feature points of the image to be registered and the reference image in the reference image library respectively; Using the BRISK feature point descriptor to generate a binary descriptor corresponding to the feature point; Performing rough feature point matching on the feature points and the binary descriptors based on the fast nearest neighbor search package FLANN algorithm to obtain an initial matching point set between the image to be registered and the reference image; Based on the random sample consistent RANSAC algorithm, the initial matching point set is precisely matched with feature points to obtain an optimized matching point set; The corresponding transformation relationship matrix between the image to be registered and the reference image is determined according to the optimized matching point set, and the scene matching result is obtained by mapping the corresponding transformation relationship matrix.

[0009] According to a method for unmanned aerial vehicle navigation and landing on a mobile landing platform provided by the present invention, a SURF algorithm is used to extract feature points of the image to be registered and the reference image in the reference image library, respectively, including: Calculating an integral image of the image to be registered and the reference image; Detecting image feature points of the integral image using a Hessian matrix; Convolving the image with box filters of different sizes to construct a scale space, wherein the scale space includes a plurality of layers, and each four layers constitute a group; Positioning the image feature points in the scale space to determine preliminary feature points; Determine the main direction of the feature point by statistically analyzing the Harr wavelet features in the neighborhood of the preliminary feature point; A feature descriptor is constructed based on the main direction of the feature point.

[0010] According to a method for unmanned aerial vehicle navigation and landing on a mobile landing platform provided by the present invention, encoding image recognition is performed on the scene matching result to obtain an encoding image result, including: Using YOLOv8 to locate the scene matching result, detect the identification image and output a target detection frame, and crop the identification image according to the target detection frame to obtain a cropped image; Identify identification information and corner point coordinates of the cropped image; Perform PnP solution on the corner point coordinates to obtain the encoded image result.

[0011] According to a method for navigation and landing of a drone with a mobile landing platform provided by the present invention, based on the scene matching result and the coded image result, a navigation path planning of the mobile landing platform is obtained, and navigation flight control is performed according to the navigation path planning of the mobile landing platform, including: Determine an optimal flight path of the UAV from a current position to a landing point according to the scene matching result and the coded image result; The flight attitude and speed of the UAV are controlled according to the optimal flight path through the flight control system of the UAV.

[0012] In a second aspect, the present invention further provides a UAV navigation and landing system for a mobile landing platform, comprising: An acquisition module, used to acquire a logo image of the mobile landing platform; A scene matching module, for performing real-time scene capture and matching of the identification image and a pre-established reference image library using a SURF-BRISK algorithm to obtain a scene matching result; A coding recognition module, used for performing coding image recognition on the scene matching result to obtain a coding image result; A navigation control module is used to obtain a navigation path plan for a mobile landing platform based on the scene matching result and the coded image result, and perform navigation flight control according to the navigation path plan for the mobile landing platform.

[0013] In a third aspect, the present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, an unmanned aerial vehicle navigation and landing method for a mobile landing platform as described in any one of the above-mentioned methods is implemented.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for navigation and landing of a drone for a mobile landing platform.

[0015] The method and system for the navigation and landing of a UAV on a mobile landing platform provided by the present invention uses the sensor carried by the UAV to capture the mobile landing platform identification image in real time, and compares it with the pre-stored reference image library through the scene matching algorithm to determine the precise position of the UAV. At the same time, the coded image recognition technology is used to quickly identify the specific identification on the mobile landing platform, such as the QR code, as a precise guide to the landing point, and the accurate landing of the UAV in a non-cooperative environment is achieved by integrating the results of scene matching and coded image recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 It is a flow chart of a method for navigation and landing of a UAV on a mobile landing platform provided by the present invention; Figure 2 It is a block diagram of the SURF-BRISK algorithm provided by the present invention; Figure 3 It is a structural schematic diagram of a UAV navigation and landing system for a mobile landing platform provided by the present invention; Figure 4 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] Figure 1 FIG. 1 is a flow chart of a method for navigation and landing of a UAV on a mobile landing platform provided by an embodiment of the present invention. Figure 1 As shown, including: Step 100: Acquire a logo image of a mobile landing platform; Step 200: using the SURF-BRISK algorithm to perform real-time scene capture and matching on the identification image and a pre-established reference image library to obtain a scene matching result; Step 300: performing coded image recognition on the scene matching result to obtain a coded image result; Step 400: Based on the scene matching result and the encoded image result, obtain a navigation path plan for a mobile landing platform, and perform navigation flight control according to the navigation path plan for the mobile landing platform.

[0020] It should be noted that in the implementation of the present invention, the mobile landing platform is usually a vehicle-mounted mobile platform for use in starting, stopping and landing the UAV.

[0021] Specifically, the sensors carried by the drone (such as high-definition cameras, infrared sensors, etc.) are used to capture the QR code image of the rooftop airport landing platform in real time, and the image is compared with the pre-stored reference image library through the scene matching algorithm to determine the precise location of the drone. At the same time, the coded image recognition technology is used to quickly identify specific signs on the airport landing platform, such as QR codes, as precise guidance for the landing point. By integrating the results of scene matching and coded image recognition, the drone can achieve precise landing in a non-cooperative environment.

[0022] The vehicle-mounted UAV precise landing navigation method based on scene matching and coded image recognition provided by the present invention can realize the precise landing of the vehicle-mounted UAV in an unfavorable environment, and can provide accurate guidance on the landing point in real time during the flight of the UAV, thereby shortening the landing time and improving the landing efficiency.

[0023] Based on the above embodiment, before step 100, the following steps are further included: Setting the identification image on the mobile landing platform according to environmental factors; Complete the system configuration of the mobile landing platform and the UAV respectively; A reference image library of the identification image is established.

[0024] Wherein, establishing the reference image library of the identification image includes: Collect multiple images of the mobile landing platform relative to the landing area of ​​the drone at different angles; Performing denoising, contrast enhancement and color correction on the multiple images to obtain preprocessed images; The preprocessed image is subjected to edge detection and contour extraction based on the Canny operator and the Freeman chain code to obtain the reference image library.

[0025] Specifically, the embodiments of the present invention, as preliminary preparations for implementation, include coded image design and layout and vehicle-mounted platform and UAV system configuration.

[0026] In the design and layout of coded images, a unique and highly recognizable QR code image is designed for the drone to recognize and decode the location information. The coded image is arranged in the predetermined landing area to ensure that the drone can clearly capture the image during the landing process. When arranging, the effects of factors such as lighting conditions and changes in viewing angle on image recognition need to be considered.

[0027] In the configuration of the vehicle-mounted platform and the UAV system, the vehicle-mounted platform needs to be equipped with a stable and reliable power supply system, wireless communication module, data processing unit, etc. to support the take-off, landing and real-time data transmission of the UAV; the UAV system needs to integrate high-definition cameras, image processing modules, navigation control units, etc. to realize scene acquisition, image recognition, path planning and flight control and other functions.

[0028] To further establish the benchmark image library, the first step is data collection. It is necessary to use drones to take aerial photos of the vehicle landing areas at different angles and collect QR code images. Then, image preprocessing is performed on the collected scene images, such as denoising, contrast enhancement, and color correction. The main processes include: image filtering, grayscale, thresholding, etc. The median filter algorithm is used to filter the original image to suppress noise and retain image details. The color image is grayscaled to simplify the three-component value into a single component value. The weighted average method is selected to calculate the average value of the three-component value of the pixel point according to a certain weight as the grayscale value of the pixel point. Through grayscale processing, the value range of each pixel in the image is narrowed to 0~255. The Otsu adaptive thresholding method is used to threshold the cooperation sign image to obtain a black and white binary image. On the basis of retaining the image feature information, the redundant data of the image is further reduced, which is convenient for subsequent visual processing applications. Finally, image edge detection and contour extraction are performed. The contour of the collected QR code image is extracted by drone edge detection technology to improve image quality and reduce recognition errors. The embodiment of the present invention uses the Canny operator to perform image edge detection. The Canny edge detection uses a two-dimensional Gaussian function and discrete derivatives to obtain the image edge based on the image gradient size and gradient direction. After the Canny edge detector calculates the gradient size and direction, it first takes a gradient threshold to generate a binary image with sufficiently strong edge candidates, and then discretizes the edge direction to determine whether each pixel in the binary image is located in a consistent direction area. If so, the pixel point is regarded as an edge point and stored. After obtaining the edge points in the image through Canny edge detection, the Freeman chain code is used to extract and save the contour to obtain the contour of the cooperation logo. According to the results of edge detection, the connected region analysis is used to extract the complete contour of the QR code.

[0029] Based on the above embodiment, step 200 includes: Obtain multi-layer nested images to be registered; Using the SURF algorithm to extract feature points of the image to be registered and the reference image in the reference image library respectively; Using the BRISK feature point descriptor to generate a binary descriptor corresponding to the feature point; Performing rough feature point matching on the feature points and the binary descriptors based on the fast nearest neighbor search package FLANN algorithm to obtain an initial matching point set between the image to be registered and the reference image; Based on the random sample consistent RANSAC algorithm, the initial matching point set is precisely matched with feature points to obtain an optimized matching point set; The corresponding transformation relationship matrix between the image to be registered and the reference image is determined according to the optimized matching point set, and the scene matching result is obtained by mapping the corresponding transformation relationship matrix.

[0030] The SURF algorithm is used to extract feature points of the image to be registered and the reference image in the reference image library, respectively, including: Calculating an integral image of the image to be registered and the reference image; Detecting image feature points of the integral image using a Hessian matrix; Convolving the image with box filters of different sizes to construct a scale space, wherein the scale space includes a plurality of layers, and each four layers constitute a group; Positioning the image feature points in the scale space to determine preliminary feature points; Determine the main direction of the feature point by statistically analyzing the Harr wavelet features in the neighborhood of the preliminary feature point; A feature descriptor is constructed based on the main direction of the feature point.

[0031] Specifically, during the flight, the drone obtains a hierarchical landing marker image of multiple layers of nested two-dimensional codes through an onboard camera.

[0032] The real-time image is compared with the images in the reference image library to find the most similar image and determine the precise location of the drone. The scene matching algorithm can be implemented by extracting feature points from the real-time image and the reference image and matching them using a feature-based approach. During the matching process, it is necessary to consider the impact of different time, lighting, angle and other conditions on the scene, and make corresponding corrections and optimizations.

[0033] For UAV remote sensing image registration, the Speeded Up Robust Features (SURF) algorithm is improved, and the SURF feature point extraction is combined with the Binary Robust Invariant Scalable Keypoints (BRISK) feature point descriptor to propose the SURF-BRISK algorithm. This algorithm improves the running speed of the algorithm and achieves a fast and accurate image registration process. The Fast Library for Approximate Nearest Neighbors (FLANN) algorithm is used for rough matching of feature points, and then the Random Sample Consensus (RANSAC) algorithm is used to further eliminate mismatched points, effectively reducing the mismatch rate.

[0034] Among them, the main steps of the SURF algorithm include: (1) Calculate the integral image. Constructing the integral image is a very important step in the entire SURF algorithm process. By using the integral image, the SURF algorithm reduces the complexity of the convolution operation and speeds up the feature point detection.

[0035] (2) Use the Hessian matrix to detect feature points. The SURF algorithm detects image feature points through the Hessian matrix, and the detected feature points have higher accuracy.

[0036] (3) Constructing scale space. In order to detect the scale invariance of the feature point detection algorithm, image feature points need to be detected in different scale spaces, so it is necessary to construct scale space. The SURF algorithm constructs scale space by using box filters of different sizes to convolve the same image. The scale space of the SURF algorithm consists of several layers, and every 4 layers constitute a group. In the SURF algorithm, the size of the smallest scale box filter is 9×9, and the sizes of box filters in other scale spaces are all expanded from 9×9 box filters. The operation speed is independent of the size of the box filter template.

[0037] (4) Locating feature points. After the pixel point is processed by the Hessian matrix, it needs to be compared with the 26 points in its 3D neighborhood. If the point is the maximum value point among these 26 points, it will be used as the preliminarily determined feature point.

[0038] (5) Determine the main direction of the feature point. To ensure the rotation invariance of the algorithm, the main direction of the feature point needs to be determined. In the SURF algorithm, the main direction of the feature point is determined by statistically analyzing the Harr wavelet features in the neighborhood of the feature point. The sector direction with the largest modulus value in the circular neighborhood is the only main direction of the feature point.

[0039] (6) Construct feature descriptors. Take a square frame around the feature point and divide the square into 16 sub-regions when generating feature descriptors, with 25 sampling points in each sub-region.

[0040] like Figure 2 As shown, the scene matching steps based on the SURF-BRISK algorithm are as follows: (1) Feature point detection and description 1) Extract feature points. The feature detector in the algorithm is used to extract image features. The number of detected feature points needs to meet certain requirements, and the number of redundant feature points should be as small as possible. SURF, as a feature detector, has good robustness and can more quickly obtain feature points with scale invariance, rotation invariance, and noise resistance, and the acquired image feature points are relatively stable.

[0041] 2) Generate binary descriptors. The BRISK operator is a binary operator that takes up less computing resources and describes feature points in a uniform sampling mode. It takes less time to generate descriptors and has a faster matching speed. At the same time, the BRISK descriptor also has good rotation and scaling invariance and noise resistance.

[0042] (2) Feature matching After completing the feature extraction and description of the image to be registered and the reference image, it is necessary to perform a matching operation on the feature description points to obtain similar feature point pairs.

[0043] As a commonly used image feature point matching method, FLANN can perform fast nearest neighbor search for large data sets and high-dimensional features. This method can improve the accuracy of image matching by adjusting parameters and setting thresholds, and has a faster speed. The FLANN algorithm is used to perform rough matching operations on feature points of UAV remote sensing images to obtain the initial matching point set of the image to be registered and the reference image.

[0044] (3) Feature matching point optimization In order to solve the problem of mismatching of image feature points, this patent uses the RANSAC algorithm to eliminate mismatching points. The RANSAC algorithm is a commonly used matching point purification method, which mainly screens and eliminates mismatching feature point pairs in the initial matching point set of the image to be registered and the reference image, thereby improving the accuracy of registration. It has good robustness and can optimize the matching results to a great extent. The RANSAC algorithm adopts a sampling and verification strategy. It can iteratively find the optimal parameter model in the data set containing mismatching points to remove mismatches and improve the accuracy of matching.

[0045] (4) Scene recognition and positioning Finally, the corresponding relationship between the scene currently photographed by the drone and the reference scene is identified based on the successfully matched feature point pairs. The precise position of the drone in the target area is determined by combining the flight data of the drone and the scene recognition results.

[0046] Based on the above embodiment, step 300 includes: Using YOLOv8 to locate the scene matching result, detect the identification image and output a target detection frame, and crop the identification image according to the target detection frame to obtain a cropped image; Identify identification information and corner point coordinates of the cropped image; Perform PnP solution on the corner point coordinates to obtain the encoded image result.

[0047] Specifically, a unique coded image (such as a QR code) is placed in the target landing area as an accurate identifier of the landing point. The coded image should have anti-occlusion and anti-deformation properties to ensure accurate identification in complex environments. At the same time, the location and layout of the coded image should be reasonably designed so that the drone can quickly find and identify the landing point.

[0048] On the basis of scene matching, the coded image in the image is further identified. By using image recognition and matching, the relative position between the UAV and the target is accurately calculated, and the UAV is automatically adjusted in flight attitude and angle to achieve precise landing of the UAV.

[0049] By using the YOLOv8 framework to identify the QR code, the corresponding corner point coordinates are obtained and the position of the drone is decoded; the Kalman filter is used to fuse the position data and IMU data to obtain the accurate three-dimensional world coordinates of the drone in the QR code coordinate system; combined with the layered landing strategy, the drone can achieve autonomous landing.

[0050] 1) QR code detection. YOLOv8 is used to pre-process the environment image captured by the downward-looking camera, locate and detect the QR code in the image, and output the target detection frame; the QR code is cropped from the image according to the target detection frame to achieve QR code extraction. When blocked by shadows, the QR code features will be affected to a certain extent, but the YOLOv8 model can still accurately locate and detect the QR code in the image.

[0051] 2) QR code recognition: QR code feature positioning and information recognition are performed through QR code recognition technology to obtain QR code information and corner point coordinates.

[0052] 3) Data fusion. Use the coordinates of the QR code corner points output in the above steps to perform PnP solution to obtain the drone's position information. Using the Kalman filter to fuse IMU information and PnP solution can more accurately describe the drone's motion and obtain the drone's precise three-dimensional world coordinates.

[0053] 4) Autonomous landing strategy. The drone can identify the QR code image of the airport landing platform on the vehicle to obtain the accurate landing position, realize autonomous and precise landing of the drone, and improve operation efficiency and safety.

[0054] The horizontal deviation allowed for the drone to descend is proportional to the altitude of the drone. If the horizontal deviation of the drone exceeds the set value at a certain descent altitude, the drone will stop descending and continue to perform the descent mission after adjusting the horizontal deviation below the corresponding set value.

[0055] Based on the above embodiment, step 400 includes: Determine an optimal flight path of the UAV from a current position to a landing point according to the scene matching result and the coded image result; The flight attitude and speed of the UAV are controlled according to the optimal flight path through the flight control system of the UAV.

[0056] Specifically, the first is flight path planning. Based on the results of scene matching and coded image recognition, the optimal flight path from the current position of the drone to the landing point is planned. The flight path planning algorithm should be able to consider factors such as obstacle avoidance, energy consumption, and time, and generate a safe and efficient flight trajectory. At the same time, the algorithm should have real-time and dynamic adjustment capabilities to cope with emergencies that may occur during the flight.

[0057] The second step is to achieve flight control. Through the UAV's flight control system, the UAV's flight attitude and speed are controlled according to the planned path. The flight control system should be able to monitor the UAV's position and attitude information in real time, and make adjustments as needed to ensure that the UAV can land stably and accurately at the designated location. During the flight, the impact of environmental factors such as wind and temperature on the UAV's flight also needs to be considered.

[0058] The precise landing navigation method for vehicle-mounted UAV based on scene matching and coded image recognition provided by the present invention can realize the precise landing of vehicle-mounted UAV in unfavorable environment. Compared with other control technologies, the precise landing navigation method provided by the present invention has the following advantages: (1) High-precision positioning: Combining scene matching and coded image recognition technology, it is possible to achieve high-precision positioning of UAVs in non-cooperative environments. The scene matching algorithm can handle complex and changing ground scenes, while the coded image recognition technology can provide fast and accurate location guidance.

[0059] (2) Fast identification: Encoded image recognition technology is fast and accurate, and can provide precise guidance on the landing point in real time during the flight of the UAV, shortening the landing time and improving the landing efficiency.

[0060] The following is a description of the UAV navigation and landing system for a mobile landing platform provided by the present invention. The UAV navigation and landing system for a mobile landing platform described below and the UAV navigation and landing method for a mobile landing platform described above can be referenced to each other.

[0061] Figure 3 FIG. 1 is a schematic diagram of a structure of a UAV navigation and landing system for a mobile landing platform provided by an embodiment of the present invention. Figure 3 As shown, it includes: an acquisition module 31, a scene matching module 32, a code recognition module 33 and a navigation control module 34, wherein: The acquisition module 31 is used to acquire the identification image of the mobile landing platform; the scene matching module 32 is used to use the SURF-BRISK algorithm to perform real-time scene capture and matching on the identification image and a pre-established reference image library to obtain a scene matching result; the code recognition module 33 is used to perform code image recognition on the scene matching result to obtain a code image result; the navigation control module 34 is used to acquire the navigation path planning of the mobile landing platform based on the scene matching result and the code image result, and perform navigation flight control according to the navigation path planning of the mobile landing platform.

[0062] Figure 4 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 4As shown, the electronic device may include: a processor 410, a communication interface 420, a memory 430 and a communication bus 440, wherein the processor 410, the communication interface 420 and the memory 430 communicate with each other through the communication bus 440. The processor 410 may call the logic instructions in the memory 430 to execute the navigation and landing method of the unmanned aerial vehicle for the mobile landing platform, the method comprising: obtaining the identification image of the mobile landing platform; using the SURF-BRISK algorithm to perform real-time scene capture and matching of the identification image and the pre-established reference image library to obtain a scene matching result; performing coded image recognition on the scene matching result to obtain a coded image result; based on the scene matching result and the coded image result, obtaining the navigation path planning of the mobile landing platform, and performing navigation flight control according to the navigation path planning of the mobile landing platform.

[0063] In addition, the logic instructions in the above-mentioned memory 430 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0064] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the unmanned aerial vehicle navigation and landing method for a mobile landing platform provided by the above-mentioned methods, the method comprising: obtaining an identification image of the mobile landing platform; using the SURF-BRISK algorithm to perform real-time scene capture and matching of the identification image and a pre-established reference image library to obtain a scene matching result; performing coded image recognition on the scene matching result to obtain a coded image result; based on the scene matching result and the coded image result, obtaining a navigation path planning of the mobile landing platform, and performing navigation flight control according to the navigation path planning of the mobile landing platform.

[0065] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0066] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0067] Finally, 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 aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for unmanned aerial vehicle navigation and landing on a mobile landing platform, characterized in that: include: Acquire a logo image of the mobile landing platform; Using accelerated robust features and a binary robust invariant scalable key point SURF-BRISK algorithm, real-time scene capture and matching are performed on the identification image and a pre-established reference image library to obtain a scene matching result; Performing coded image recognition on the scene matching result to obtain a coded image result; Based on the scene matching result and the coded image result, a navigation path plan for a mobile landing platform is obtained, and navigation flight control is performed according to the navigation path plan for the mobile landing platform.

2. The method for unmanned aerial vehicle navigation and landing for a mobile landing platform according to claim 1, characterized in that: Before obtaining the identification image of the mobile landing platform, it also includes: Setting the identification image on the mobile landing platform according to environmental factors; Complete the system configuration of the mobile landing platform and the UAV respectively; A reference image library of the identification image is established.

3. The method for unmanned aerial vehicle navigation and landing for a mobile landing platform according to claim 2, characterized in that: Establishing a reference image library of the identification image includes: Collect multiple images of the mobile landing platform relative to the landing area of ​​the drone at different angles; Performing denoising, contrast enhancement and color correction on the multiple images to obtain preprocessed images; The preprocessed image is subjected to edge detection and contour extraction based on the Canny operator and the Freeman chain code to obtain the reference image library.

4. The method for unmanned aerial vehicle navigation and landing for a mobile landing platform according to claim 1, characterized in that: The SURF-BRISK algorithm is used to perform real-time scene capture and matching on the identification image and the pre-established reference image library to obtain scene matching results, including: Obtain multi-layer nested images to be registered; Using the SURF algorithm to extract feature points of the image to be registered and the reference image in the reference image library respectively; Using the BRISK feature point descriptor to generate a binary descriptor corresponding to the feature point; Performing rough feature point matching on the feature points and the binary descriptors based on the fast nearest neighbor search package FLANN algorithm to obtain an initial matching point set between the image to be registered and the reference image; Based on the random sample consistent RANSAC algorithm, the initial matching point set is precisely matched with feature points to obtain an optimized matching point set; The corresponding transformation relationship matrix between the image to be registered and the reference image is determined according to the optimized matching point set, and the scene matching result is obtained by mapping the corresponding transformation relationship matrix.

5. The method for unmanned aerial vehicle navigation and landing for a mobile landing platform according to claim 4, characterized in that: The SURF algorithm is used to extract feature points of the image to be registered and the reference image in the reference image library, respectively, including: Calculating an integral image of the image to be registered and the reference image; Detecting image feature points of the integral image using a Hessian matrix; Convolving the image with box filters of different sizes to construct a scale space, wherein the scale space includes a plurality of layers, and each four layers constitute a group; Positioning the image feature points in the scale space to determine preliminary feature points; Determine the main direction of the feature point by statistically analyzing the Harr wavelet features in the neighborhood of the preliminary feature point; A feature descriptor is constructed based on the main direction of the feature point.

6. The method for unmanned aerial vehicle navigation and landing for a mobile landing platform according to claim 1, characterized in that: Performing coded image recognition on the scene matching result to obtain a coded image result includes: Using YOLOv8 to locate the scene matching result, detect the identification image and output a target detection frame, and crop the identification image according to the target detection frame to obtain a cropped image; Identify identification information and corner point coordinates of the cropped image; Perform PnP solution on the corner point coordinates to obtain the encoded image result.

7. The method for unmanned aerial vehicle navigation and landing for a mobile landing platform according to claim 1, characterized in that: Based on the scene matching result and the coded image result, obtaining a navigation path plan for a mobile landing platform, and performing navigation flight control according to the navigation path plan for the mobile landing platform, including: Determine an optimal flight path of the UAV from a current position to a landing point according to the scene matching result and the coded image result; The flight attitude and speed of the UAV are controlled according to the optimal flight path through the flight control system of the UAV.

8. A UAV navigation and landing system for a mobile landing platform, characterized in that: include: An acquisition module, used to acquire a logo image of the mobile landing platform; A scene matching module, for performing real-time scene capture and matching of the identification image and a pre-established reference image library using a SURF-BRISK algorithm to obtain a scene matching result; A coding recognition module, used for performing coding image recognition on the scene matching result to obtain a coding image result; A navigation control module is used to obtain a navigation path plan for a mobile landing platform based on the scene matching result and the coded image result, and perform navigation flight control according to the navigation path plan for the mobile landing platform.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the unmanned aerial vehicle navigation and landing method for a mobile landing platform as described in any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the unmanned aerial vehicle navigation and landing method for a mobile landing platform as described in any one of claims 1 to 7 is implemented.

Images