Vision-based multi-stage precise landing method for unmanned aerial vehicle hangar

Abstract

The invention discloses a vision-based multi-stage precise landing method for an unmanned aerial vehicle hangar. The unmanned aerial vehicle flies to the airspace above the hangar, a landing mark is arranged on a parking apron of the hangar, and the unmanned aerial vehicle overlooks downwards to collect images through an airborne camera and controls autonomous landing in real time: in the first stage, a square in the landing mark is obtained through image recognition, the horizontal position is adjusted, and then the unmanned aerial vehicle lands; if a square cannot be completely captured in the image during landing, entering a second stage; in the second stage, a concentric circle and a course angle reference line in the landing mark are recognized, the horizontal position is adjusted through the concentric circle, the course is adjusted through the course angle reference line, and landing continues; and when the unmanned aerial vehicle lands below the preset height threshold value, direct landing is not adjusted any more until the unmanned aerial vehicle lands on the parking apron of the hangar. According to the method, the landing point position of the unmanned aerial vehicle hangar can be accurately positioned in the whole landing process of the unmanned aerial vehicle, and the unmanned aerial vehicle is guided to realize fixed-point landing of the hangar.

Description

technical field [0001] The invention belongs to a drone landing method in the technical field of drone control, and in particular relates to a vision-based multi-stage precise landing method for a drone hangar. Background technique [0002] At present, most UAV hangar landings use RTK positioning or an H-shaped marker image for assisted landing, which has the following limitations: [0003] 1. Although the RTK positioning accuracy error is small (10cm), it is easy to be interfered and the RTK information is easy to be lost, which will lead to a large landing error in the drone hangar, which will cause the drone to land outside the hangar; [0004] 2. There is only one marked image, and the camera needs to capture the overall image of the marked, and identify the center point of the image to achieve precise landing, that is, the camera needs to be aligned with the marked to achieve precise landing. However, when the flying height of the UAV is high, the area of ​​the marked ...

AI Technical Summary

Problems solved by technology

However, when the flying height of the UAV is high, the area of ​​the marked image is generally large so that the camera can collect the overall image of the mark; when the UAV descends to a certain height, the camera cannot collect the overall image of the mark, resulting in Unable to identify the center point of the image, the drone will drift in scale relative to the landing point in the air, etc.

Images