Unmanned aerial vehicle visual positioning method and device in satellite denial environment

Abstract

This application discloses a visual positioning method and apparatus for unmanned aerial vehicles (UAVs) in satellite-denied environments, belonging to the field of visual positioning technology. It includes: acquiring digital orthophotos of the UAV's flight area; dividing the digital orthophotos of the UAV's flight area into adaptation regions based on an image matching algorithm to obtain adaptation region division results; planning flight paths based on the adaptation region division results to obtain the UAV's flight curve; preparing and binding a reference base map of the UAV's flight curve to obtain a reference base map and uploading it to the UAV's onboard operating system; cropping the reference base map based on the UAV's pose to obtain a reference map; matching the reference map and a real-time map based on an image matching algorithm to obtain the matching position of the center point of the real-time map on the reference map; and obtaining the UAV's real-time spatial coordinates based on the matching position, the UAV's current attitude, and its relative ground altitude. This method improves the accuracy of UAV visual positioning.

Description

Technical Field

[0001] This application belongs to the field of visual positioning technology, and in particular relates to a visual positioning method and device for unmanned aerial vehicles in a satellite-denied environment. Background Technology

[0002] Positioning technology is the core support for autonomous flight and mission execution of unmanned aerial vehicles (UAVs). Currently, the mainstream positioning methods heavily rely on satellite positioning systems such as GPS and BeiDou. However, in satellite-denied environments, such as areas with strong electromagnetic interference, areas where satellite signals are blocked, or areas with complex terrain obstruction, satellite positioning signals may be interrupted, distorted, or experience a significant decrease in accuracy. This can lead to UAVs being unable to achieve accurate positioning, or even experiencing flight loss of control and mission failure.

[0003] Existing UAV positioning solutions for satellite-denied environments mostly employ a combination of inertial navigation and visual assistance. However, inertial navigation suffers from cumulative errors, with positioning deviations increasing over time. Traditional visual positioning methods directly match images of the entire flight area without considering the adaptability of images from different regions. This leads to matching errors in areas with uniform textures and similar terrain, making it difficult to guarantee positioning accuracy. Furthermore, existing solutions lack customized trajectory planning and baseline map preparation for visual positioning requirements. The excessive amount of image processing data on the airborne end results in poor real-time positioning performance and a lack of precise compensation for attitude and altitude errors, further reducing the stability and accuracy of positioning.

[0004] In satellite-denied environments, unmanned aerial vehicles (UAVs) lack sufficient environmental awareness, making it difficult to quickly and accurately acquire and calculate their own positions. Existing positioning technologies suffer from cumulative errors in inertial navigation systems, poor robustness of visual SLAM-based algorithms in dynamic environments, and high costs for lidar solutions, failing to meet the low-cost, high-precision positioning requirements of small UAVs. In summary, existing technologies for UAV positioning in satellite-denied environments suffer from low positioning accuracy, poor real-time performance, error accumulation, and weak environmental adaptability, failing to meet the precise autonomous flight requirements of UAVs in such environments. Summary of the Invention

[0005] This application aims to address at least one of the technical problems existing in the prior art. To this end, this application proposes a visual positioning method and apparatus for unmanned aerial vehicles (UAVs) in satellite-denied environments. This method eliminates the dependence on satellite positioning systems and improves the accuracy, real-time performance, and environmental adaptability of UAV positioning in satellite-denied environments through adaptation zone division, customized trajectory planning, reference map preparation, and precise image matching and error compensation.

[0006] In a first aspect, this application provides a visual positioning method for unmanned aerial vehicles in a satellite-denied environment, the method comprising: A digital orthophoto of the unmanned aerial vehicle (UAV) flight area is acquired. An adaptation region is then divided into the digital orthophoto of the UAV flight area based on an image matching algorithm to obtain the adaptation region division result. The adaptation region includes a first adaptation region, a second adaptation region, and a third adaptation region. Based on the adaptation zone division results, flight trajectory planning is performed to obtain the flight curve of the unmanned aerial vehicle; A baseline map is prepared and bound for the flight curve of the unmanned aerial vehicle, and the baseline map is then uploaded to the onboard operating system of the unmanned aerial vehicle. The baseline image is obtained by cropping the baseline image based on the pose of the UAV, and the real-time image captured by the camera directly below the UAV is obtained. The baseline image and the real-time image are matched based on the image matching algorithm to obtain the matching position of the center point of the real-time image on the baseline image. Based on the matched location, the current attitude of the UAV, and its relative altitude to the ground, the real-time spatial coordinates of the UAV are obtained.

[0007] According to one embodiment of this application, the adaptation region division of the digital orthophoto of the unmanned aerial vehicle's flight area based on the image matching algorithm to obtain the adaptation region division result includes: The digital orthophoto of the unmanned aerial vehicle's flight area is preprocessed into blocks to obtain multiple image sub-blocks; A deep learning classification neural network model is constructed, and the feature adaptation requirements of image matching are used as model training constraints. The deep learning classification neural network model is trained to obtain the adaptation region screening model. The multiple image sub-blocks are input into the adaptation area screening model for feature recognition and adaptation degree determination to obtain the adaptation degree level of each image sub-block; Based on the adaptation level of each image sub-block and the geographic topology of the flight area, adjacent image sub-blocks are fused and divided to obtain the adaptation area division result.

[0008] According to one embodiment of this application, the step of planning flight paths based on the adaptation area division results to obtain the flight curve of the unmanned aerial vehicle includes: Based on the adaptation zone division results, the initial trajectory search space is constructed by taking the take-off point and the target point as the trajectory start and end nodes and combining the geographical obstacle information of the flight area. In the initial trajectory search space, with priority given to passing through the first adaptation zone and avoiding the third adaptation zone as trajectory constraints, several candidate polyline trajectories are searched and obtained. The candidate polyline track is evaluated by a combination of path smoothness and adaptation area coverage, and the candidate polyline track with the best score is selected. A continuous curve fitting algorithm is used to fit the candidate broken line segment trajectory with the optimal score to obtain the flight curve of the unmanned aerial vehicle without right-angle inflection points.

[0009] According to one embodiment of this application, the step of preparing and binding a reference base map of the flight curve of the unmanned aerial vehicle, obtaining the reference base map, and uploading it to the onboard operating system of the unmanned aerial vehicle includes: Sampling points are selected at preset intervals for the flight curve of the unmanned aerial vehicle. Multiple image acquisition areas are formed by expanding outward from each sampling point as the center, and adjacent image acquisition areas are set with a preset proportion of overlapping area. The image content corresponding to each image acquisition area is cropped as the base map source data. The base map source data is then subjected to coordinate calibration and resolution normalization to obtain the reference base map. The baseline maps are numbered and packaged according to the flight path sequence to complete the binding of the baseline maps, and then uploaded to the onboard operating system of the unmanned aerial vehicle.

[0010] According to one embodiment of this application, the step of matching a reference image and a real-time image based on an image matching algorithm to obtain the matching position of the center point of the real-time image on the reference image includes: Based on a deep learning feature extraction network, local feature points and global feature descriptions of the preprocessed benchmark image and real-time image are extracted respectively to construct a benchmark image feature set and a real-time image feature set. The preprocessing includes grayscale conversion, denoising and feature enhancement. Feature points are matched between the baseline graph feature set and the real-time graph feature set using a feature matching network to obtain initial feature matching pairs. Then, a random sampling consensus algorithm is used to remove mismatched points from the initial feature matching pairs to obtain accurate feature matching pairs. A coordinate transformation matrix is ​​constructed based on accurate feature matching pairs. The pixel coordinates of the center point of the real-time image are mapped to the pixel coordinate system of the reference image through the coordinate transformation matrix, so as to obtain the matching position of the center point of the real-time image on the reference image.

[0011] According to one embodiment of this application, obtaining the real-time spatial position coordinates of the unmanned aerial vehicle (UAV) based on the matched position, the current attitude of the UAV, and its relative ground altitude includes: The pixel coordinates of the matching position of the center point of the real-time map on the reference map are converted into geographic planar coordinates to obtain the planar geographic coordinates of the matching position. The current attitude parameters of the unmanned aerial vehicle are obtained, including roll angle, pitch angle and yaw angle. The attitude parameters are then transformed into coordinates based on the spatial attitude calculation model to obtain the attitude correction matrix. The original relative ground altitude of the unmanned aerial vehicle is obtained, and the original relative ground altitude is compensated and corrected for air pressure error and terrain error to obtain the corrected relative ground altitude. The planar geographic coordinates and the corrected relative ground height are input into the spatial coordinate calculation model, and the three-dimensional coordinates are calculated by combining the attitude correction matrix to obtain the real-time spatial position coordinates of the unmanned aerial vehicle.

[0012] According to one embodiment of this application, the process of cropping the reference base map based on the pose of the unmanned aerial vehicle to obtain the reference map includes: Based on the pose of the unmanned aerial vehicle and the intrinsic parameters of the airborne downward-looking camera, the geographic coordinate range corresponding to the camera's field of view is calculated using the camera imaging model. The clipping boundaries of the base map are determined based on the geographic coordinate range; Adaptive cropping is performed on the base image according to the cropping boundary, and resolution adaptation processing is performed on the cropped image to ensure that the resolution of the cropped image is consistent with the resolution of the real-time image captured by the airborne camera, thus obtaining the base image.

[0013] Secondly, this application provides a visual positioning device for unmanned aerial vehicles in a satellite-denied environment, the device comprising: The acquisition module is used to acquire digital orthophotos of the unmanned aerial vehicle's flight area, and to divide the digital orthophotos of the unmanned aerial vehicle's flight area into adaptation regions based on an image matching algorithm, thereby obtaining adaptation region division results. The adaptation regions include a first adaptation region, a second adaptation region, and a third adaptation region. The first processing module is used to plan the flight path based on the adaptation area division results to obtain the flight curve of the unmanned aerial vehicle. The second processing module is used to prepare and bind a reference base map of the flight curve of the unmanned aerial vehicle, obtain the reference base map and upload it to the onboard operating system of the unmanned aerial vehicle. The third processing module is used to crop the base map based on the pose of the unmanned aerial vehicle to obtain the base map and to obtain the real-time image captured by the camera directly below the unmanned aerial vehicle. The fourth processing module is used to match the reference image and the real-time image based on the image matching algorithm to obtain the matching position of the center point of the real-time image on the reference image. The fifth processing module is used to obtain the real-time spatial position coordinates of the unmanned aerial vehicle (UAV) based on the matched position, the current attitude of the UAV, and its relative ground altitude.

[0014] Thirdly, this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the visual positioning method for unmanned aerial vehicles in a satellite denial environment as described in the first aspect above.

[0015] Fourthly, this application provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the visual positioning method for unmanned aerial vehicles in a satellite-denied environment as described in the first aspect above.

[0016] Fifthly, this application provides a chip including a processor and a communication interface, the communication interface being coupled to the processor, the processor being used to run programs or instructions to implement the visual positioning method for unmanned aerial vehicles in a satellite-denied environment as described in the first aspect.

[0017] In a sixth aspect, this application provides a computer program product, including a computer program that, when executed by a processor, implements the visual positioning method for unmanned aerial vehicles in a satellite-denied environment as described in the first aspect above.

[0018] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application.

[0019] The present invention provides a visual positioning method for unmanned aerial vehicles in satellite-denied environments, which has the following advantages over existing technologies: (1) This invention divides the digital orthophoto of the flight area into adaptation zones, distinguishes different levels of areas suitable for image matching, and plans the flight path in combination with the adaptation zone results, so that the unmanned aerial vehicle can preferentially pass through areas with rich features and high adaptability, thereby avoiding matching errors in areas with single textures from the source, greatly improving the accuracy of visual positioning, while avoiding geographical obstacles and low adaptation zones, and improving flight safety.

[0020] (2) This invention prepares a base map customized for the planned flight curve, divides the image acquisition area by sampling points and sets overlapping areas, which reduces the amount of image data processing on the airborne end, ensures the real-time positioning, and ensures the accuracy of the geographic coordinates and image consistency of the base map by coordinate calibration and resolution normalization, laying the foundation for subsequent image matching. This improves the autonomy and reliability of unmanned aerial vehicles in complex environments.

[0021] (3) This invention achieves adaptive cropping of the reference base map based on the pose of the unmanned aerial vehicle, so that the reference map matches the field of view of the airborne camera. Combined with preprocessing such as grayscale and noise reduction and deep learning feature extraction, and with random sampling consistency algorithm to remove mismatched points, the invention achieves accurate matching between the reference map and the real-time map, effectively improving the accuracy of the matching position.

[0022] (4) This invention obtains the attitude correction matrix by performing coordinate transformation on the attitude parameters of the unmanned aerial vehicle when solving the spatial position coordinates, compensates for air pressure and terrain errors on the relative ground height, and performs three-dimensional coordinate calculation by combining planar geographic coordinates. This effectively eliminates the positioning deviation caused by attitude offset and altitude measurement error, further improves the positioning accuracy and stability, completely gets rid of the dependence on satellite positioning, and has strong adaptability in satellite denial environment. Attached Figure Description

[0023] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is one of the flowcharts illustrating the visual positioning method for unmanned aerial vehicles in a satellite-denied environment provided in the embodiments of this application; Figure 2 This is a schematic diagram of the process for preparing the reference base map provided in the embodiments of this application; Figure 3 This is a flowchart illustrating the calculation of the current coordinates of an aircraft according to an embodiment of this application; Figure 4 This is the second flowchart illustrating the visual positioning method for unmanned aerial vehicles in a satellite-denied environment provided in the embodiments of this application. Figure 5 This is one of the positioning maps of the unmanned aerial vehicle provided in the embodiments of this application; Figure 6 This is the second positioning diagram of the unmanned aerial vehicle provided in the embodiments of this application; Figure 7 This is a schematic diagram of the structure of the visual positioning device for unmanned aerial vehicles in a satellite-denied environment provided in the embodiments of this application; Figure 8 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0024] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0025] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and the number of objects is not limited; for example, a first object can be one or more. Furthermore, the use of "and" and / or to indicate at least one of the connected objects in the specification and claims, and the character " / ", generally indicates an "or" relationship between the preceding and following objects.

[0026] The following description, in conjunction with the accompanying drawings, details the unmanned aerial vehicle (UAV) visual positioning method, UAV visual positioning device, electronic equipment, and readable storage medium provided in this application embodiment through specific embodiments and application scenarios.

[0027] Among them, the visual positioning method for unmanned aerial vehicles in satellite-denied environments can be applied to terminals, specifically executed by hardware or software within the terminal.

[0028] The terminal includes, but is not limited to, portable communication devices such as mobile phones or tablets with touch-sensitive surfaces (e.g., touchscreen displays and / or touchpads). It should also be understood that, in some embodiments, the terminal may not be a portable communication device, but rather a desktop computer with touch-sensitive surfaces (e.g., touchscreen displays and / or touchpads).

[0029] The following embodiments describe a terminal including a display and a touch-sensitive surface. However, it should be understood that the terminal may include one or more other physical user interface devices such as a physical keyboard, mouse, and joystick.

[0030] The UAV visual positioning method under satellite denial environment provided in this application embodiment can be executed by an electronic device or a functional module or entity in an electronic device that can implement the UAV visual positioning method under satellite denial environment. The electronic devices mentioned in this application embodiment include, but are not limited to, mobile phones, tablets, computers, cameras and wearable devices. The following uses an electronic device as the execution subject to illustrate the UAV visual positioning method under satellite denial environment provided in this application embodiment.

[0031] Figure 1 This is one of the flowcharts illustrating the visual positioning method for unmanned aerial vehicles in a satellite-denied environment provided in this application embodiment, such as... Figure 1As shown, the visual positioning method for unmanned aerial vehicles in a satellite-denied environment includes steps 110, 120, 130, 140, 150, and 160.

[0032] Step 110: Obtain digital orthophotos of the unmanned aerial vehicle (UAV) flight area, and divide the digital orthophotos of the UAV flight area into adaptation regions based on an image matching algorithm to obtain adaptation region division results. The adaptation regions include a first adaptation region, a second adaptation region, and a third adaptation region. In some embodiments, the adaptation region division of the digital orthophoto of the unmanned aerial vehicle's flight area based on the image matching algorithm to obtain the adaptation region division result includes: The digital orthophoto of the unmanned aerial vehicle's flight area is preprocessed into blocks to obtain multiple image sub-blocks; A deep learning classification neural network model is constructed, and the feature adaptation requirements of image matching are used as model training constraints. The deep learning classification neural network model is trained to obtain the adaptation region screening model. The multiple image sub-blocks are input into the adaptation area screening model for feature recognition and adaptation degree determination to obtain the adaptation degree level of each image sub-block; Based on the adaptation level of each image sub-block and the geographic topology of the flight area, adjacent image sub-blocks are fused and divided to obtain the adaptation area division result.

[0033] Figure 2 This is a schematic diagram of the process for preparing the reference base map provided in the embodiments of this application, such as... Figure 2 As shown, a high-precision digital orthophoto of the unmanned aerial vehicle (UAV) mission area is acquired. This image contains the geographic coordinate information and terrain texture features of the mission area. Based on the image matching algorithm, the digital orthophoto is divided into adaptation regions, dividing the mission area into the first adaptation region (rich features, clear texture, and high image matching accuracy), the second adaptation region (general features, image matching can be achieved but the accuracy is slightly lower), and the third adaptation region (simple texture / no features, effective image matching cannot be achieved), thus obtaining the adaptation region division results.

[0034] Image block preprocessing: The digital orthophoto of the flight area is divided into blocks according to a preset pixel size (e.g., 512512 pixels) to obtain multiple non-overlapping image sub-blocks, while retaining the geographic coordinate information of each image sub-block.

[0035] Adaptation region selection model training: Construct a deep learning classification neural network model, select image samples of different feature types as training sets, and use the feature adaptation requirements of image matching (such as the number of feature points, the uniformity of feature point distribution, texture contrast, etc.) as model training constraints to train the model so that the model can accurately identify the matching degree of image regions. After training, the adaptation region selection model is obtained.

[0036] Fit determination: All image sub-blocks are input into the fit area selection model. The model performs feature recognition and fit quantification scoring on each image sub-block. The fit level is divided according to the scoring results (e.g., 90 points or above is level 1, 70-90 points is level 2, and below 70 points is level 3).

[0037] Fusion and division: Based on the adaptation level of each image sub-block, combined with the geographic topology of the flight area (such as terrain continuity, administrative boundaries, road and water system distribution), adjacent image sub-blocks of the same level are fused to form continuous areal adaptation areas, and the final adaptation area division result is obtained. The first-level adaptation area is the first adaptation area, the second-level is the second adaptation area, and the third-level is the third adaptation area.

[0038] Step 120: Based on the adaptation area division results, perform flight trajectory planning to obtain the flight curve of the unmanned aerial vehicle; In some embodiments, the step of planning flight paths based on the adaptation area division results to obtain the flight curve of the unmanned aerial vehicle includes: Based on the adaptation zone division results, the initial trajectory search space is constructed by taking the take-off point and the target point as the trajectory start and end nodes and combining the geographical obstacle information of the flight area. In the initial trajectory search space, with priority given to passing through the first adaptation zone and avoiding the third adaptation zone as trajectory constraints, several candidate polyline trajectories are searched and obtained. The candidate polyline track is evaluated by a combination of path smoothness and adaptation area coverage, and the candidate polyline track with the best score is selected. A continuous curve fitting algorithm is used to fit the candidate broken line segment trajectory with the optimal score to obtain the flight curve of the unmanned aerial vehicle without right-angle inflection points.

[0039] Establish the initial trajectory search space: Using the takeoff point and mission target point of the UAV as the start and end nodes of the trajectory, import the adaptation area division results and the geographical obstacle vector data of the flight area to build a three-dimensional initial trajectory search space, and exclude the spatial area occupied by geographical obstacles.

[0040] Searching for candidate polyline tracks: In the initial track search space, track constraints are set to prioritize passing through the first adaptation zone, try to pass through the second adaptation zone, and strictly avoid the third adaptation zone. Path search algorithms such as A* and Dijkstra are used to search for several candidate polyline tracks that meet the constraints.

[0041] Comprehensive scoring and selection: Establish a comprehensive scoring system to quantitatively score the path smoothness (such as the number of inflection points and the number of inflection angles) and adaptation area coverage (such as the proportion of the length traversed in the first adaptation area to the total length of the track) of each candidate polyline segment track. The scoring formula can be set as: Comprehensive score = 0.6 * adaptation area coverage score + 0.4 * path smoothness score. Select the candidate polyline segment track with the highest comprehensive score as the optimal polyline segment track.

[0042] Curve fitting: Continuous curve fitting algorithms (such as Bézier curve and spline curve fitting algorithms) are used to fit the optimal polyline segment trajectory, eliminating right-angle inflection points in the trajectory, and obtaining a smooth UAV flight curve that conforms to the flight dynamics characteristics of the UAV.

[0043] Step 130: Prepare and bind the reference base map for the flight curve of the unmanned aerial vehicle, obtain the reference base map, and upload it to the onboard operating system of the unmanned aerial vehicle; In some embodiments, the step of preparing and binding a reference base map of the flight curve of the unmanned aerial vehicle (UAV), obtaining the reference base map, and uploading it to the UAV's onboard operating system includes: Sampling points are selected at preset intervals for the flight curve of the unmanned aerial vehicle. Multiple image acquisition areas are formed by expanding outward from each sampling point as the center, and adjacent image acquisition areas are set with a preset proportion of overlapping area. The image content corresponding to each image acquisition area is cropped as the base map source data. The base map source data is then subjected to coordinate calibration and resolution normalization to obtain the reference base map. The baseline maps are numbered and packaged according to the flight path sequence to complete the binding of the baseline maps, and then uploaded to the onboard operating system of the unmanned aerial vehicle.

[0044] Divide the image acquisition area: Select sampling points for the flight curve at preset intervals (such as 50 meters or 100 meters, adjusted according to the flight altitude of the UAV and the field of view of the camera). With each sampling point as the center, expand the preset geographical range (such as a radius of 100 meters) to form a circular / rectangular image acquisition area. At the same time, set an overlap area of ​​10%-20% between adjacent image acquisition areas to avoid image discontinuities on the flight path.

[0045] Base map source data processing: Based on the geographic coordinates of the image acquisition area, the corresponding image content is cropped from the digital orthophoto of the flight area as the base map source data. The base map source data is then subjected to coordinate calibration (correcting the geographic coordinate offset of the image) and resolution normalization (unifying all base map source data to a preset resolution, such as 0.1 meters / pixel) to obtain a single base map.

[0046] Binding and Uploading: All base maps are numbered and packaged according to the flight path sequence of the flight curves to complete the binding of the base maps and form a base map library corresponding to the flight path; the base map library is uploaded to the onboard operating system of the UAV via wireless communication, and the onboard system stores it by number for easy access later.

[0047] Step 140: Based on the pose of the UAV, crop the reference base map to obtain the reference map, and obtain the real-time image captured by the camera directly below the UAV. Figure 3 This is a flowchart illustrating the calculation of the current coordinates of an aircraft according to an embodiment of this application, as shown below. Figure 3 As shown, in some embodiments, the process of cropping the reference base map based on the pose of the unmanned aerial vehicle to obtain the reference map includes: Based on the pose of the unmanned aerial vehicle and the intrinsic parameters of the airborne downward-looking camera, the geographic coordinate range corresponding to the camera's field of view is calculated using the camera imaging model. The clipping boundaries of the base map are determined based on the geographic coordinate range; Adaptive cropping is performed on the base image according to the cropping boundary, and resolution adaptation processing is performed on the cropped image to ensure that the resolution of the cropped image is consistent with the resolution of the real-time image captured by the airborne camera, thus obtaining the base image.

[0048] After takeoff, the onboard unit acquires its own pose information (including position prediction and attitude information) in real time. Based on the pose, it adaptively crops the base map to obtain a reference map. Simultaneously, it captures ground images of the flight area using the onboard downward-facing camera to obtain real-time images, specifically including: Calculate the geographic range of the camera's field of view: Based on the real-time pose (longitude, latitude, altitude, heading) of the UAV and the intrinsic parameters (focal length, pixel size, number of pixels) of the airborne downward-facing camera, combined with the pinhole camera imaging model, the ground geographic coordinate range (latitude and longitude range) corresponding to the current field of view of the camera is calculated through coordinate transformation.

[0049] Determine and crop the boundaries: Based on the calculated geographic coordinate range, retrieve the corresponding base map from the base map library to determine the precise cropping boundaries. Adaptively crop the base map according to the boundaries to obtain an initial base map that matches the camera's field of view.

[0050] Resolution adaptation: The initial reference image is scaled to make its resolution completely consistent with the resolution of the real-time image captured by the airborne camera, thus obtaining the final reference image.

[0051] Acquiring real-time images: Real-time images of the ground directly below the UAV are captured by an airborne downward-facing camera. At the same time, the real-time images are standardized to ensure that they are consistent with the image format of the reference image.

[0052] Step 150: Match the reference image and the real-time image based on the image matching algorithm to obtain the matching position of the center point of the real-time image on the reference image; In some embodiments, the step of matching the reference image and the real-time image based on the image matching algorithm to obtain the matching position of the center point of the real-time image on the reference image includes: Based on a deep learning feature extraction network, local feature points and global feature descriptions of the preprocessed benchmark image and real-time image are extracted respectively to construct a benchmark image feature set and a real-time image feature set. The preprocessing includes grayscale conversion, denoising and feature enhancement. Feature points are matched between the baseline graph feature set and the real-time graph feature set using a feature matching network to obtain initial feature matching pairs. Then, a random sampling consensus algorithm is used to remove mismatched points from the initial feature matching pairs to obtain accurate feature matching pairs. A coordinate transformation matrix is ​​constructed based on accurate feature matching pairs. The pixel coordinates of the center point of the real-time image are mapped to the pixel coordinate system of the reference image through the coordinate transformation matrix, so as to obtain the matching position of the center point of the real-time image on the reference image.

[0053] This method employs a combination of deep learning and traditional matching algorithms to accurately match the baseline image and the real-time image, obtaining the matching position of the center point of the real-time image on the baseline image. Specifically, this includes: Image preprocessing and feature extraction: The baseline image and the real-time image are preprocessed separately, and feature enhancement operations such as grayscale conversion, Gaussian denoising, and histogram equalization are performed in sequence. The preprocessed baseline image and real-time image are input into a deep learning feature extraction network to extract local feature points (such as corner points and edge points) and global feature descriptions (such as feature vectors) of the two images respectively, and construct the feature set of the baseline image and the feature set of the real-time image.

[0054] Feature matching and mismatch removal: Feature points are matched between the baseline map feature set and the real-time map feature set based on the feature matching network to obtain initial feature matching pairs; the random sampling consensus algorithm is used to filter the initial feature matching pairs and remove mismatch points caused by changes in lighting and terrain to obtain accurate feature matching pairs.

[0055] Coordinate mapping yields the matching position: Based on the pixel coordinates of the accurate feature matching pair, a coordinate transformation matrix (homography matrix) is constructed using the least squares method; the pixel coordinates of the center point of the real-time image are substituted into the coordinate transformation matrix and mapped to the pixel coordinate system of the reference image to obtain the pixel coordinates of the center point of the real-time image on the reference image, which is the matching position.

[0056] Step 160: Based on the matched position, the current attitude of the UAV and its relative ground altitude, obtain the real-time spatial position coordinates of the UAV.

[0057] In some embodiments, obtaining the real-time spatial position coordinates of the unmanned aerial vehicle (UAV) based on the matched position, the UAV's current attitude, and its relative ground altitude includes: The pixel coordinates of the matching position of the center point of the real-time map on the reference map are converted into geographic planar coordinates to obtain the planar geographic coordinates of the matching position. The current attitude parameters of the unmanned aerial vehicle are obtained, including roll angle, pitch angle and yaw angle. The attitude parameters are then transformed into coordinates based on the spatial attitude calculation model to obtain the attitude correction matrix. The original relative ground altitude of the unmanned aerial vehicle is obtained, and the original relative ground altitude is compensated and corrected for air pressure error and terrain error to obtain the corrected relative ground altitude. The planar geographic coordinates and the corrected relative ground height are input into the spatial coordinate calculation model, and the three-dimensional coordinates are calculated by combining the attitude correction matrix to obtain the real-time spatial position coordinates of the unmanned aerial vehicle.

[0058] Based on the matching position obtained in step 150, and combined with the current attitude of the UAV and the corrected relative ground altitude, the real-time three-dimensional spatial position coordinates of the UAV are obtained through coordinate calculation, specifically including: Pixel coordinates to geographic plane coordinates: Based on the geographic coordinate calibration parameters of the reference map (such as pixel latitude and longitude resolution), the pixel coordinates of the matching location are converted into the corresponding geographic plane coordinates (longitude and latitude) to obtain the projected plane coordinates of the unmanned aerial vehicle on the ground.

[0059] Attitude correction matrix calculation: The current attitude parameters of the unmanned aerial vehicle are obtained through airborne attitude sensors, including roll angle, pitch angle and heading angle; based on the spatial attitude calculation model (such as Euler angle coordinate transformation model), the attitude parameters are transformed to obtain the attitude correction matrix used to correct the spatial position.

[0060] Relative ground height correction: The original relative ground height of the unmanned aerial vehicle is obtained by airborne barometric altimeter and laser altimeter; the original height is compensated for by barometric error (correcting measurement deviations caused by atmospheric pressure and temperature) and terrain error (correcting deviations caused by terrain undulations by combining terrain elevation data from the base map) to obtain the corrected relative ground height.

[0061] Three-dimensional spatial coordinate calculation: The geographic plane coordinates and the corrected relative ground height are input into the spatial coordinate calculation model (such as the WGS-84 geodetic coordinate system calculation model), and the three-dimensional coordinates are calculated in combination with the attitude correction matrix to obtain the real-time spatial position coordinates of the unmanned aerial vehicle, thus completing the accurate visual positioning in the satellite-denied environment.

[0062] One prerequisite for the usability of visual positioning is that the scenery features in the flight path area are relatively obvious, meeting the requirements of information richness, saliency, stability, uniqueness, and accuracy. Therefore, the flight path area is divided into three levels of adaptation zones: high, medium, and low. High adaptation zones are suitable for matching, while low adaptation zones are not. The planned flight path must avoid low adaptation zones and, as far as possible, pass through high adaptation zones. Based on the planned flight path and high-precision, high-timeliness digital orthophotos, a reference base map to be attached to the aircraft is prepared. After the aircraft takes off, the real-time generated reference map and the real-time map are matched, and then the aircraft's current actual coordinate position is calculated based on the matched position and flight attitude. Figure 4 This is the second flowchart illustrating the visual positioning method for unmanned aerial vehicles in a satellite-denied environment provided in this application embodiment. Figure 4 As shown, it includes the following steps: (1) Adaptation zone selection: Combined with image matching algorithm, a classification neural network structure of deep learning is used to train the adaptation zone selection model file. The model file is used to predict the digital orthophoto of the flight area and divide it into three levels of adaptation zones: high, medium and low.

[0063] (2) Flight trajectory planning: avoid low adaptation zone and try to pass through high adaptation zone. Manually draw the broken line segment from the take-off point to the target point, automatically fit it into a curve, and then specify parameters such as flight altitude and speed to complete the trajectory planning work of the aircraft flight mission.

[0064] (3) Preparation and binding of the reference base map: Points are taken at equal intervals on the planned flight path, and a square is generated by expanding a certain distance outward from each sampling point. Note that adjacent squares have a certain overlap in area. The digital orthophoto area within the range of each square is cropped as the reference base map and uploaded to the designated location of the aircraft's onboard operating system.

[0065] (4) Generate a reference image / real-time image. Based on the aircraft's pose, a reference image is obtained by cropping from the reference base image. The real-time image is generated by rotating, scaling and cropping the real-time image taken by the camera directly below the aircraft.

[0066] (5) Baseline / real-time image matching: Using a deep learning image matching algorithm, the baseline image and the real-time image are matched, and the matching position of the center point of the real-time image on the baseline image is calculated.

[0067] (6) Solve the spatial position of the aircraft. Using the matching position result from the previous step, the current attitude of the aircraft, and the corrected relative ground height, solve the spatial position coordinates of the aircraft.

[0068] Figure 5 This is one of the positioning maps of the unmanned aerial vehicle provided in the embodiments of this application. Figure 6 This is the second positioning diagram of the unmanned aerial vehicle provided in the embodiments of this application, such as... Figure 5 and Figure 6 As shown, when both pitch and roll angles change dynamically within a range of ±25 degrees, the average positioning error can be stabilized at around 5 meters, and the maximum positioning error is generally around 8 meters.

[0069] It should be noted that before takeoff, the ground-based mission planning and data processing program completes the selection of the adaptation area, flight path planning, and preparation and binding of the base map. After takeoff, the airborne visual image matching and positioning program specifically performs real-time map / base map generation, real-time map / base map matching, and aircraft spatial position calculation.

[0070] The task planning and data processing flow is shown in the following diagram. Flight path planning avoids low-fitness zones and tries to pass through high-fitness zones. The process involves manually drawing a broken line segment from the takeoff point to the target point, automatically fitting it into a curve, and then specifying parameters such as flight altitude and speed to complete the flight path planning for the aircraft mission.

[0071] Fitting curves: Reduce errors. The maximum error is likely to occur near this right-angle turn. The speed and angular velocity change smoothly. Severe fluctuations in angular velocity will cause the captured image and the position recorded when the image was captured to not correspond, with a delay of 200ms.

[0072] Image data is acquired using the aircraft's downward-facing camera, processed into a real-time image based on parameters such as heading angle and relative ground altitude, and the range of a reference image is calculated based on flight attitude and camera parameters and cropped from the digital orthophoto to obtain a reference image. Based on the algorithm performance of the matching module in the visual positioning function used by the aircraft's airborne platform, the image pair sequence composed of the reference image and the real-time image is divided into three levels: high, medium, and low.

[0073] According to the satellite-denied environment visual positioning method for unmanned aerial vehicles provided in the embodiments of this application, by dividing the digital orthophoto of the flight area into adaptation zones, different levels of areas suitable for image matching are distinguished. The trajectory planning is combined with the adaptation zone results, so that the unmanned aerial vehicle prioritizes passing through areas with rich features and high adaptability, thereby avoiding matching errors in areas with simple textures from the source, greatly improving the accuracy of visual positioning, while avoiding geographical obstacles and low adaptation zones, and improving flight safety.

[0074] The UAV visual positioning method in a satellite-denied environment provided in this application embodiment can be executed by a UAV visual positioning device in a satellite-denied environment. This application embodiment uses the execution of the UAV visual positioning method in a satellite-denied environment by a UAV visual positioning device as an example to illustrate the UAV visual positioning device in a satellite-denied environment provided in this application embodiment.

[0075] This application also provides a visual positioning device for unmanned aerial vehicles in satellite-denied environments, such as... Figure 7 As shown, the visual positioning device for unmanned aerial vehicles in a satellite-denied environment includes: an acquisition module 310, a first processing module 720, a second processing module 730, a third processing module 740, a fourth processing module 750, and a fifth processing module 760.

[0076] The acquisition module 710 is used to acquire digital orthophotos of the unmanned aerial vehicle's flight area, and to divide the digital orthophotos of the unmanned aerial vehicle's flight area into adaptation regions based on an image matching algorithm, thereby obtaining adaptation region division results. The adaptation regions include a first adaptation region, a second adaptation region, and a third adaptation region. The first processing module 720 is used to perform flight trajectory planning based on the adaptation area division results to obtain the flight curve of the unmanned aerial vehicle. The second processing module 730 is used to prepare and bind a reference base map of the flight curve of the unmanned aerial vehicle, obtain the reference base map and upload it to the airborne operating system of the unmanned aerial vehicle. The third processing module 740 is used to crop the reference base map based on the pose of the unmanned aerial vehicle to obtain the reference map and to obtain the real-time image captured by the camera directly below the unmanned aerial vehicle. The fourth processing module 750 is used to match the reference image and the real-time image based on the image matching algorithm to obtain the matching position of the center point of the real-time image on the reference image. The fifth processing module 760 is used to obtain the real-time spatial position coordinates of the unmanned aerial vehicle based on the matching position, the current attitude of the unmanned aerial vehicle, and its relative ground altitude.

[0077] The visual positioning method for unmanned aerial vehicles (UAVs) in satellite-denied environments provided in this application embodiment prepares a customized reference base map based on the planned flight curve. By dividing the image acquisition area into sampling points and setting overlapping areas, the amount of image data processing on the airborne end is reduced, ensuring real-time positioning. Furthermore, coordinate calibration and resolution normalization ensure the accuracy of the geographic coordinates and image consistency of the reference base map, laying the foundation for subsequent image matching. This improves the autonomy and reliability of UAVs in complex environments. The visual positioning device for unmanned aerial vehicles in satellite-denied environments provided in this application embodiment can achieve… Figures 1 to 6The various processes implemented in the embodiment of the visual positioning method for unmanned aerial vehicles in satellite-denied environments will not be described in detail here to avoid repetition.

[0078] In some embodiments, such as Figure 8 As shown, this application embodiment also provides an electronic device 800, including a processor 801, a memory 802, and a computer program stored in the memory 802 and executable on the processor 801. When the program is executed by the processor 801, it implements the various processes of the above-described embodiment of the visual positioning method for unmanned aerial vehicles in a satellite denial environment and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0079] It should be noted that the electronic devices in the embodiments of this application include the mobile electronic devices and non-mobile electronic devices described above.

[0080] This application also provides a non-transitory computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the various processes of the above-described embodiment of the visual positioning method for unmanned aerial vehicles under satellite denial environment and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0081] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0082] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described visual positioning method for unmanned aerial vehicles in a satellite denial environment.

[0083] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0084] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface and the processor are coupled. The processor is used to run programs or instructions to implement the various processes of the above-described embodiment of the visual positioning method for unmanned aerial vehicles in a satellite denial environment, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0085] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a device-level chip, device chip, chip device, or on-chip device chip, etc.

[0086] It should be noted that, in this document, the terms include, encompass, or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, including an element by a statement does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0087] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the visual positioning method of unmanned aerial vehicles in a satellite denial environment of the various embodiments of this application.

[0088] In the description of this application, the first feature and the second feature may include one or more of the features.

[0089] In the description of this application, "multiple" means two or more.

[0090] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

[0091] In the description of this specification, the references to the terms "an embodiment," "some embodiments," "illustrative embodiments," "examples," "specific examples," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0092] Although embodiments of this application have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of this application, the scope of which is defined by the claims and their equivalents.

Claims

1. A visual positioning method for unmanned aerial vehicles in a satellite-denied environment, characterized in that, The method includes: A digital orthophoto of the unmanned aerial vehicle (UAV) flight area is acquired. An adaptation region is then divided into the digital orthophoto of the UAV flight area based on an image matching algorithm to obtain the adaptation region division result. The adaptation region includes a first adaptation region, a second adaptation region, and a third adaptation region. Based on the adaptation zone division results, flight trajectory planning is performed to obtain the flight curve of the unmanned aerial vehicle; A baseline map is prepared and bound for the flight curve of the unmanned aerial vehicle, and the baseline map is then uploaded to the onboard operating system of the unmanned aerial vehicle. The baseline image is obtained by cropping the baseline image based on the pose of the UAV, and the real-time image captured by the camera directly below the UAV is obtained. The baseline image and the real-time image are matched based on the image matching algorithm to obtain the matching position of the center point of the real-time image on the baseline image. Based on the matched location, the current attitude of the UAV, and its relative altitude to the ground, the real-time spatial coordinates of the UAV are obtained.

2. The visual positioning method for unmanned aerial vehicles in a satellite-denied environment according to claim 1, characterized in that, The method of dividing the digital orthophoto of the unmanned aerial vehicle's flight area into adaptation regions based on the image matching algorithm yields the following adaptation region division results: The digital orthophoto of the unmanned aerial vehicle's flight area is preprocessed into blocks to obtain multiple image sub-blocks; A deep learning classification neural network model is constructed, and the feature adaptation requirements of image matching are used as model training constraints. The deep learning classification neural network model is trained to obtain the adaptation region screening model. The multiple image sub-blocks are input into the adaptation area screening model for feature recognition and adaptation degree determination to obtain the adaptation degree level of each image sub-block; Based on the adaptation level of each image sub-block and the geographic topology of the flight area, adjacent image sub-blocks are fused and divided to obtain the adaptation area division result.

3. The visual positioning method for unmanned aerial vehicles in a satellite-denied environment according to claim 2, characterized in that, The flight trajectory planning based on the adaptation area division results to obtain the flight curve of the unmanned aerial vehicle includes: Based on the adaptation zone division results, the initial trajectory search space is constructed by taking the take-off point and the target point as the trajectory start and end nodes and combining the geographical obstacle information of the flight area. In the initial trajectory search space, with priority given to passing through the first adaptation zone and avoiding the third adaptation zone as trajectory constraints, several candidate polyline trajectories are searched and obtained. The candidate polyline track is evaluated by a combination of path smoothness and adaptation area coverage, and the candidate polyline track with the best score is selected. A continuous curve fitting algorithm is used to fit the candidate broken line segment trajectory with the optimal score to obtain the flight curve of the unmanned aerial vehicle without right-angle inflection points.

4. The visual positioning method for unmanned aerial vehicles in a satellite-denied environment according to claim 1, characterized in that, The process of preparing and binding a reference base map for the flight curve of the unmanned aerial vehicle (UAV), obtaining the reference base map, and uploading it to the UAV's onboard operating system includes: Sampling points are selected at preset intervals for the flight curve of the unmanned aerial vehicle. Multiple image acquisition areas are formed by expanding outward from each sampling point as the center, and adjacent image acquisition areas are set with a preset proportion of overlapping area. The image content corresponding to each image acquisition area is cropped as the base map source data. The base map source data is then subjected to coordinate calibration and resolution normalization to obtain the reference base map. The baseline maps are numbered and packaged according to the flight path sequence to complete the binding of the baseline maps, and then uploaded to the onboard operating system of the unmanned aerial vehicle.

5. The visual positioning method for unmanned aerial vehicles in a satellite-denied environment according to claim 4, characterized in that, The process of matching the reference image and the real-time image using an image matching algorithm to obtain the matching position of the center point of the real-time image on the reference image includes: Based on a deep learning feature extraction network, local feature points and global feature descriptions of the preprocessed benchmark image and real-time image are extracted respectively to construct a benchmark image feature set and a real-time image feature set. The preprocessing includes grayscale conversion, denoising and feature enhancement. Feature points are matched between the baseline graph feature set and the real-time graph feature set using a feature matching network to obtain initial feature matching pairs. Then, a random sampling consensus algorithm is used to remove mismatched points from the initial feature matching pairs to obtain accurate feature matching pairs. A coordinate transformation matrix is ​​constructed based on accurate feature matching pairs. The pixel coordinates of the center point of the real-time image are mapped to the pixel coordinate system of the reference image through the coordinate transformation matrix, so as to obtain the matching position of the center point of the real-time image on the reference image.

6. The visual positioning method for unmanned aerial vehicles in a satellite-denied environment according to claim 1, characterized in that, The process of obtaining the real-time spatial position coordinates of the unmanned aerial vehicle (UAV) based on the matched position, the UAV's current attitude, and its relative ground altitude includes: The pixel coordinates of the matching position of the center point of the real-time map on the reference map are converted into geographic planar coordinates to obtain the planar geographic coordinates of the matching position. The current attitude parameters of the unmanned aerial vehicle are obtained, including roll angle, pitch angle and yaw angle. The attitude parameters are then transformed into coordinates based on the spatial attitude calculation model to obtain the attitude correction matrix. The original relative ground altitude of the unmanned aerial vehicle is obtained, and the original relative ground altitude is compensated and corrected for air pressure error and terrain error to obtain the corrected relative ground altitude. The planar geographic coordinates and the corrected relative ground height are input into the spatial coordinate calculation model, and the three-dimensional coordinates are calculated by combining the attitude correction matrix to obtain the real-time spatial position coordinates of the unmanned aerial vehicle.

7. The visual positioning method for unmanned aerial vehicles in a satellite-denied environment according to claim 1, characterized in that, The process of cropping the base map based on the pose of the unmanned aerial vehicle to obtain the base map includes: Based on the pose of the unmanned aerial vehicle and the intrinsic parameters of the airborne downward-looking camera, the geographic coordinate range corresponding to the camera's field of view is calculated using the camera imaging model. The clipping boundaries of the base map are determined based on the geographic coordinate range; Adaptive cropping is performed on the base image according to the cropping boundary, and resolution adaptation processing is performed on the cropped image to ensure that the resolution of the cropped image is consistent with the resolution of the real-time image captured by the airborne camera, thus obtaining the base image.

8. A visual positioning device for unmanned aerial vehicles (UAVs) in a satellite-denied environment, implemented using the visual positioning method for UAVs in a satellite-denied environment as described in any one of claims 1 to 7, characterized in that, The device includes: The acquisition module is used to acquire digital orthophotos of the unmanned aerial vehicle's flight area, and to divide the digital orthophotos of the unmanned aerial vehicle's flight area into adaptation regions based on an image matching algorithm, thereby obtaining adaptation region division results. The adaptation regions include a first adaptation region, a second adaptation region, and a third adaptation region. The first processing module is used to plan the flight path based on the adaptation area division results to obtain the flight curve of the unmanned aerial vehicle. The second processing module is used to prepare and bind a reference base map of the flight curve of the unmanned aerial vehicle, obtain the reference base map and upload it to the onboard operating system of the unmanned aerial vehicle. The third processing module is used to crop the base map based on the pose of the UAV to obtain the base map and to obtain the real-time image captured by the camera directly below the UAV. The fourth processing module is used to match the reference image and the real-time image based on the image matching algorithm to obtain the matching position of the center point of the real-time image on the reference image. The fifth processing module is used to obtain the real-time spatial coordinates of the unmanned aerial vehicle (UAV) based on the matched position, the current attitude of the UAV, and its relative altitude to the ground.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the visual positioning method for unmanned aerial vehicles in a satellite-denied environment as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the visual positioning method for unmanned aerial vehicles in a satellite-denied environment as described in any one of claims 1 to 7.

Images