基于视觉引导着陆的微小型四旋翼设计及实现

为了增强四旋翼的自主性,借用视觉导航方法要求在GPS丢失信号或信号精度不高的情况下进行辅助定位,使得四旋翼能够在某些特殊需求场合自动着陆。针对Harris角点检测算法计算量大的问题,设计了一种快速、准确的标志图像识别算法,将一个区域内的所有角点转换成一个单值角点,替代了Harris算法中的局部非极大值抑制算法,减小了计算量,并进一步设计了视觉模式下的姿态与位置转换算法和视觉着陆控制算法。为了验证所设计算法的可行性和有效性,基于STM32单片机搭建了微型四旋翼实验平台系统,并实时在线进行了性能实验,结果表明,标志图像识别算法能正确检测和识别出标志图像,视觉着陆算法也成功实现了四旋翼实验平台的着陆功能。

Design and implementation of a micro quadrotor based on visual guidance landing

In order to enhance the autonomy of the quadrotors, the visual navigation method is used to assist in the positioning of the GPS missing signal or the low signal accuracy, and automatic landing on special occasions. A fast and accurate image recognition algorithm is designed for the problem of large amount calculation in Harris corner detection. Through transformation of all corners in a region into a single value corner, which replaces the local non-maximal suppression algorithm in the Harris, reduces the amount of calculation, designs attitude and position conversion algorithm and visual landing control algorithm in visual mode. To verify the feasibility and the effectiveness of the proposed algorithm, a quadrotor experimental platform is set up based on the STM32 microcontroller. Real-time online experimental results show that the target image recognition algorithm correctly detects and identifies the target image, and eventually achieves quadrotor visual landing function.