基于二维码图像识别的无人机定位方法研究

针对物流无人机(Unmanned Aerial Vehicle,UAV)定点投放定位精度不高的问题,提出一种无人机物流精确定位解决方案。该方案将二维码图像识别应用于飞控系统,建立空中二维码图像位置和三维空间坐标模型。运用基于局部特征的CamShift跟踪算法引导无人机飞行、降落,实现无人机低空精准着陆。户外实验测试结果表明,用该方案设计的物流飞控系统跟踪速度快,平均每帧图像处理时间为0.017 ms,5 m内高空着地误差不超过20 cm,能够为物流无人机物资投放提供较好的技术参考。

POSITIONING METHOD OF UAV BASED ON TWO-DIMENSIONAL CODE IMAGE RECOGNITION

To solve the problem of low positioning accuracy of fixed-point delivery of logistics UAV, we proposed a solution for precise positioning of logistics UAV. In this scheme, the two-dimensional code image recognition was applied to the flight control system, and the position and three-dimensional coordinate model of the two-dimensional image in the air were established. CamShift tracking algorithm based on local features was used to guide UAV to fly and land so as to achieve precise landing at low altitude. The outdoor test results show that the tracking speed of the logistics flight control system designed by this scheme is fast, the average image processing time per frame is 0.017 ms, and the landing error is less than 20 cm within 5 m. It can provide better technical reference for the material delivery of logistics UAV.