Satellite navigation-free unmanned aerial vehicle position estimation method capable of predicting flight path

Abstract

The invention provides a satellite navigation-free unmanned aerial vehicle position estimation method capable of predicting a flight path. The method comprises the following steps: S1, obtaining a detection image of a front downward view field of an unmanned aerial vehicle; s2, performing multi-source image fusion processing on the detection image to obtain a fused image; s3, perspective projection transformation is carried out on the fused image, feature matching is carried out on the fused image and DEM data, and feature point information is obtained; and S4, determining the current coordinates of the unmanned aerial vehicle based on the relative height of the unmanned aerial vehicle relative to the ground, the altitude of the unmanned aerial vehicle, the flight speed of the unmanned aerial vehicle and the feature point information, wherein the current coordinates are used for predicting the flight path of the unmanned aerial vehicle. Compared with the prior art, the non-satellite navigation unmanned aerial vehicle position estimation method capable of predicting the flight path can estimate the position coordinates of the unmanned aerial vehicle without satellite navigation, so that the current coordinates of the unmanned aerial vehicle are obtained, and flight path prediction and autonomous navigation are facilitated; a fault protection mechanism is provided for the unmanned aerial vehicle when the satellite navigation signal is lost, and smooth flight tasks of the unmanned aerial vehicle are guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field, and in particular to a method for estimating the position of a satellite-free navigation unmanned aerial vehicle capable of predicting a flight path. Background Art

[0002] In recent years, drones have been widely used in military, agriculture, forestry, mining engineering and other fields. More and more flight control and navigation of drones have become a hot topic in the field of drones in complex environments. The global navigation satellite system is widely used as the main positioning system or combined with inertial navigation to form a comprehensive navigation and positioning system for drone mission execution due to its high positioning accuracy, low hardware power consumption, small mass and volume.

[0003] Small drones use satellite navigation and inertial navigation to obtain attitude and position. When the satellite navigation signal is interfered with or lost, the inertial navigation used by small drones diverges in a very short time. The divergence of position information makes it impossible for drones to obtain position information, and attitude divergence affects the flight safety of drones. In wartime or certain specific application scenarios, it is very common for satellite navigation signals to be interfered with.

[0004] At present, in the actual flight process, if we only rely on satellite navigation as the only source of UAV geolocation, when the satellite navigation signal becomes unavailable due to various reasons (such as interference or deception), the UAV cannot determine the current coordinates, cannot navigate the UAV, and may not be able to continue to perform the original mission, which may cause safety hazards to the UAV. Summary of the invention

[0005] The purpose of the present invention is to overcome the shortcomings and deficiencies in the prior art and to provide a method for estimating the position of a satellite-free navigation drone with a predictable trajectory.

[0006] An embodiment of the present invention provides a method for estimating the position of a satellite-free navigation drone with a predictable trajectory, comprising:

[0007] S1: Acquire the detection image of the front downward field of view of the UAV;

[0008] S2: performing multi-source image fusion processing on the detection image to obtain a fused image;

[0009] S3: Perform perspective projection on the fused image and perform feature matching with DEM data to obtain feature point information;

[0010] S4: Determine the current coordinates of the drone based on the relative height of the drone relative to the ground, the altitude of the drone, the flight speed of the drone and the feature point information, and the current coordinates are used to predict the drone track.

[0011] In some optional embodiments, the detection image is acquired by a camera of the drone, the camera faces the front and lower field of view of the drone, and the angle of the shooting direction of the camera relative to the horizontal plane can be adjusted between 0°-90°.

[0012] In some optional implementations, the angle of the shooting direction of the camera relative to the horizontal plane is determined based on the relative height and flight speed of the drone.

[0013] In some optional embodiments, the relative height is measured by a laser rangefinder.

[0014] In some optional embodiments, the altitude is measured by a barometric altimeter.

[0015] In some optional implementations, the multi-source image fusion processing is implemented based on the SwinFusion general image fusion framework.

[0016] In some optional embodiments, the detected image includes at least a visible light image and an infrared image.

[0017] In some optional implementations, step S3 includes:

[0018] S31: performing perspective projection on the fused image to obtain three-dimensional point cloud data of the fused image;

[0019] S32: performing feature matching on the three-dimensional point cloud data of the fused image and the DEM data;

[0020] S33: Acquire feature point information, where the feature point information at least includes ground feature points.

[0021] In some optional embodiments, during the implementation of steps S1-S4, the drone performs a flight mission based on the original coordinates of the drone.

[0022] Compared with the prior art, the method for estimating the position of a satellite-free navigation drone with a predictable track of the present invention can estimate the position coordinates of the drone when there is no satellite navigation, thereby obtaining the current coordinates of the drone, facilitating track prediction and autonomous navigation, and providing a fault protection mechanism for the drone when the satellite navigation signal is lost, thereby ensuring the smooth progress of the drone's flight mission.

[0023] In order to more clearly understand the present invention, the specific implementation of the present invention will be described below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A flowchart of a method for estimating the position of a satellite-free navigation drone with a predictable trajectory according to an embodiment of the present invention;

[0025] Figure 2 FIG. 1 is a flow chart of step S3 of an embodiment of the present invention. DETAILED DESCRIPTION

[0026] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0027] See also Figure 1 An embodiment of the present invention provides a method for estimating the position of a satellite-free navigation drone with a predictable trajectory, comprising:

[0028] S1: Acquire the detection image of the front downward field of view of the UAV;

[0029] S2: performing multi-source image fusion processing on the detection image to obtain a fused image;

[0030] S3: Perform perspective projection on the fused image and perform feature matching with DEM data to obtain feature point information;

[0031] S4: Determine the current coordinates of the drone based on the relative height of the drone relative to the ground, the altitude of the drone, the flight speed of the drone and the feature point information, and the current coordinates are used to predict the drone track. Since it takes a certain amount of time to perform steps S1-S4, it is necessary to determine the offset according to the flight speed. The coordinates determined by the relative height of the drone relative to the ground, the altitude of the drone and the feature point information are corrected by the offset to obtain the current coordinates of the drone.

[0032] When the UAV is flying, it can detect satellite signals through the detection device. When it detects that the satellite signal is lost or the satellite signal is interfered with, it implements steps S1-S4, so that the current position of the UAV can be obtained without satellite navigation, which is convenient for track prediction and autonomous navigation. It provides a fault protection mechanism for the UAV when the satellite navigation signal is lost, ensuring the smooth progress of the UAV's flight mission.

[0033] It should be noted that the front downward field of view refers to the field of view in front of the drone and facing downward. This field of view can observe the terrain that the drone is about to fly to. Therefore, the detection image of the front downward field of view can be used to solve the current coordinates of the drone.

[0034] The principle of using the coordinates of the drone to predict the trajectory of the drone is a technology well known to those skilled in the art and will not be described in detail here.

[0035] In some optional embodiments, the detection image is acquired through a camera of the drone, which faces the front and lower field of view of the drone, and the angle of the camera's shooting direction relative to the horizontal plane can be adjusted between 0°-90°, so that the angle of the camera's shooting direction relative to the horizontal plane meets the requirements for drone coordinate positioning.

[0036] In some optional embodiments, the angle of the camera's shooting direction relative to the horizontal plane is determined based on the relative height and flight speed of the drone, wherein the faster the drone's flight speed, the smaller the angle of the camera's shooting direction relative to the horizontal plane, thereby facilitating the early acquisition of more image information in front of the drone; the slower the drone's flight speed, the larger the angle of the camera's shooting direction relative to the horizontal plane, thereby facilitating more accurate acquisition of image information near the drone; the higher the drone's relative height, the larger the angle of the camera's shooting direction relative to the horizontal plane, thereby acquiring image information in a wider range; the lower the drone's relative height, the smaller the angle of the camera's shooting direction relative to the horizontal plane, thereby avoiding the camera being too pointed downward to acquire sufficient image information.

[0037] In some optional embodiments, the relative height is measured by a laser rangefinder. Of course, the relative height can also be measured by other suitable devices, not limited to this example.

[0038] In some optional embodiments, the altitude is measured by a barometric altimeter. Of course, the altitude can also be measured by other suitable devices, not limited to this example.

[0039] Multi-source image fusion can improve the accuracy of feature extraction and optimize the accuracy of track prediction and positioning. In some optional embodiments, the detection image includes at least a visible light image and an infrared image, and the combination of multiple images can improve the feature accuracy of the fused image at night or in low light environments.

[0040] In some optional embodiments, the multi-source image fusion processing is implemented based on the SwinFusion general image fusion framework. The steps of performing multi-source image fusion processing on the detection image can be roughly divided into three steps: feature extraction, feature fusion, and image reconstruction. The principles of feature extraction, feature fusion, and image reconstruction are well known to those skilled in the art and will not be described in detail here. It should be noted that for the feature extraction step, when the detection image includes a visible light image and an infrared image, the local features of the visible light image and the local features of the infrared image are extracted based on CNN, and the global features of the visible light image and the global features of the infrared image are extracted based on Swin Transformer.

[0041] See also Figure 2In some optional implementations, step S3 includes:

[0042] S31: performing perspective projection on the fused image to obtain three-dimensional point cloud data of the fused image;

[0043] S32: performing feature matching on the three-dimensional point cloud data of the fused image and the DEM data; the feature matching algorithm adopts the nearest neighbor search or RANSAC algorithm to improve the accuracy and robustness of the feature matching.

[0044] S33: Acquire feature point information, where the feature point information at least includes ground feature points, and the ground feature points can facilitate determination of the horizontal plane coordinates of the UAV.

[0045] DEM data is a pre-stored three-dimensional model containing ground elevation information. The current map location of the drone is determined through feature matching, and then the current coordinates of the drone can be accurately calculated by combining the altitude and relative height.

[0046] In some optional embodiments, during the implementation of the steps S1-S4, the drone performs a flight mission based on the original coordinates of the drone. The original coordinates are the latest drone coordinates obtained through the satellite signal before the loss of the satellite signal or the interference of the satellite signal is detected. Since it takes a certain amount of time to implement the steps S1-S4, especially the feature matching takes a long time, the drone cannot update the coordinates in real time at this time. For this reason, the drone flies according to the original flight mission to avoid affecting the efficiency of mission execution. After obtaining the current coordinates of the drone through the steps S1-S4, the drone continues to perform the flight mission based on the current coordinates. It should be noted that before the detection device detects that the satellite signal is restored, the steps S1-S4 need to be cyclically performed to continuously update the current coordinates of the drone, thereby meeting the drone track prediction and autonomous navigation requirements, so that the drone can accurately perform the flight mission.

[0047] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for estimating the position of a satellite-free unmanned aerial vehicle with a predictable trajectory, characterized in that: include: S1: Acquire the detection image of the front downward field of view of the UAV; S2: performing multi-source image fusion processing on the detection image to obtain a fused image; S3: Perform perspective projection on the fused image and perform feature matching with DEM data to obtain feature point information; S4: Determine the current coordinates of the drone based on the relative height of the drone relative to the ground, the altitude of the drone, the flight speed of the drone and the feature point information, and the current coordinates are used to predict the drone track.

2. The method for estimating the position of a satellite-free unmanned aerial vehicle with a predictable trajectory according to claim 1, characterized in that: The detection image is acquired through a camera of the drone, the camera faces the front and lower field of view of the drone, and the shooting direction of the camera can be adjusted at an angle of 0°-90° relative to the horizontal plane.

3. The method for estimating the position of a satellite-free unmanned aerial vehicle with a predictable trajectory according to claim 2, characterized in that: The angle of the shooting direction of the camera relative to the horizontal plane is determined based on the relative height and flight speed of the drone.

4. The method for estimating the position of a satellite-free unmanned aerial vehicle with a predictable trajectory according to claim 1, characterized in that: The relative height is measured by a laser rangefinder.

5. The method for estimating the position of a satellite-free unmanned aerial vehicle with a predictable trajectory according to claim 1, characterized in that: The altitude is measured by a barometric altimeter.

6. A method for estimating the position of a satellite-free unmanned aerial vehicle with a predictable track according to any one of claims 1 to 5, characterized in that: The multi-source image fusion processing is implemented based on the SwinFusion general image fusion framework.

7. A method for estimating the position of a satellite-free unmanned aerial vehicle with a predictable track according to any one of claims 1 to 5, characterized in that: The detected image includes at least a visible light image and an infrared image.

8. A method for estimating the position of a satellite-free unmanned aerial vehicle with a predictable track according to any one of claims 1 to 5, characterized in that: Step S3 includes: S31: performing perspective projection on the fused image to obtain three-dimensional point cloud data of the fused image; S32: performing feature matching on the three-dimensional point cloud data of the fused image and the DEM data; S33: Acquire feature point information, where the feature point information at least includes ground feature points.

9. A method for estimating the position of a satellite-free unmanned aerial vehicle with a predictable track according to any one of claims 1 to 5, characterized in that: During the implementation of steps S1-S4, the drone performs a flight mission based on the original coordinates of the drone.

Images