Vision calibration method and system, and unmanned aerial vehicle

Abstract

The invention discloses a vision calibration method. The method comprises the steps of collecting a vision calibration pattern arranged on the ground through a camera on an unmanned aerial vehicle; identifying a concentric annulus, a concentric square and a direction calibration image in sequence to obtain identification results; according to the identification results, adjusting flight attitude of the unmanned aerial vehicle to enable the direction of the head of the unmanned aerial vehicle to be consistent with that of the direction calibration image; and when the direction of the head of the unmanned aerial vehicle is consistent with that of the direction calibration image, reducing the flight speed of the unmanned aerial vehicle to a preset speed, thereby enabling the unmanned aerial vehicle to land at the preset speed. According to the method, the unmanned aerial vehicle identifies the provided vision calibration pattern richer in geometric characteristic and easier to be subjected to image identification, so that the autonomous landing precision of the unmanned aerial vehicle is improved and the occurrence frequency of autonomous landing exception events is reduced. The invention furthermore discloses a vision calibration system and the unmanned aerial vehicle, which have the abovementioned beneficial effects.

Description

Technical field [0001] This application relates to the field of image recognition technology, and in particular to a method, system and unmanned aerial vehicle for visual calibration. Background technique [0002] With the development of science and technology, various UAVs are used more and more widely. Among them, intelligent UAVs rely on various devices installed on them to achieve autonomous landing and recovery, which will greatly reduce the long-distance Controlling the workload of personnel and improving work efficiency, but at the same time there will be interference and influence caused by various factors, resulting in the poor realization of the autonomous landing function of the UAV. [0003] At present, the prior art recognizes ground icons composed of multiple concentric squares, and recognizes each concentric square from large to small, and after the recognition is passed, it lands in the square with the most center and the smallest side length. Because only the geom...

AI Technical Summary

Problems solved by technology

Because only the geometric feature of the existing square is recognized and detected, it is easy to make mistakes, and other patterns similar to concentric squares are included in the recognition range, so as to land at the wrong place. In addition, there is an inconsistency in the orientation of each drone landing. The problem may affect the next drone take-off in a certain environment

Images