论文部分内容阅读
为实现旋翼飞行机器人在室内环境中的精确定位及控制,设计了基于单目视觉的机载视觉定位方法.该方法通过矩特征提取、图形分类来识别地面标记图形,进而利用EPn P方法解算摄像机坐标.为验证机载视觉定位方法的有效性,利用Kinect传感器搭建了旋翼飞行器外部定位系统,设计了针对旋翼飞行器目标的视觉跟踪算法,定位精度达2 mm.实验结果表明,机载视觉定位方法能够很好地进行飞行器定位.
In order to realize the precise positioning and control of rotor-flying robots in indoor environment, an on-board vision positioning method based on monocular vision is designed.This method recognizes the ground marker images by moment feature extraction and graph classification, and then uses EPn P method to solve Camera coordinates.In order to verify the effectiveness of the on-board vision positioning method, a Kinect sensor was used to build an external positioning system of the rotorcraft and a vision tracking algorithm was designed for the target of the rotorcraft with a positioning accuracy of 2 mm. The experimental results show that, The method can well locate the aircraft.