一种球形机器人视觉定位系统研究  

Research on a Visual Positioning System for SphericalRobot

作　　者：赵鹏[1] 战强[1] ZHAO Peng;ZHAN Qiang(School of Mechanical Engineering and Automation,Beijing University of Aeronautics and Astronautics,Beijing 100191,China)

机构地区：[1]北京航空航天大学机械工程及其自动化学院,北京100191

出　　处：《机械制造与自动化》2018年第3期184-186,193,共4页Machine Building & Automation

摘　　要：针对球形机器人在地面运动时的定位问题,提出了一种基于sift特征以及霍夫圆变换的球形机器人视觉定位方案。采用对尺度、视角、旋转等具有鲁棒性的sift特征向量,利用由两个外部摄像机组成的视觉系统采集的左右图片,从中提取sift特征点,利用k-d tree和BBF算法进行特征匹配查找,利用RANSAC算法筛选匹配点并计算变换矩阵,通过变换矩阵将后续摄像机采集的图像融合,在融合的图像通过霍夫圆变换定位球形机器人。实验结果表明,该视觉定位系统视野开阔、定位稳定性良好。Aiming at the localization of spherical robot in ground motion,a visual positioning scheme for spherical robot based on sift feature and Hough transform is proposed. This paper uses the sift feature vector with robustness to scale,angle of view and rotation and two external cameras to collect the left and right images,which the sift feature points are extracted form and kd-tree and BBF algorithm is used to search the feature matching. The RANSAC algorithm is used to filter the matching points and calculate the transformation matrix. Then,the transformed images are fused by the transformation matrix,and the spherical robot is located by the Hough transform. The experimental results show that the visual positioning system has a wide field of view and good positioning stability.

关 键 词：球形机器人 视觉定位系统 图像拼接 霍夫圆变换 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]