基于视觉与IMU的跟踪机器人研究  

Research of tracking robot based on vision and IMU

作　　者：邴志刚[1,2,3] 王永霞[1,2] 陈弘达[3] 

机构地区：[1]天津职业技术师范大学自动化与电气工程学院,天津300222 [2]天津信息感知与智能控制重点实验室,天津300222 [3]中国科学院半导体研究所,北京100083

出　　处：《天津工程师范学院学报》2010年第3期1-4,共4页Journal of Tianji University of Technology and Education

基　　金：国家"863"计划项目(2007AA04Z254)

摘　　要：针对普通轮式机器人在运动和控制等方面的限制,设计并实现了轮式滑动转向移动机器人,该机器人具有结构简化、运动灵活、可靠性高、可控性好等特点。基于EKF实现单目视觉和IMU数据融合,运用SURF实现被跟踪对象特征的检测与匹配,结合垂直线等环境特征获取Range/Depth信息以及被跟踪对象及道路转弯信息。实验表明:该系统及相关算法有较强的实时性和鲁棒性。Considering the vulnerability of wheeled robot for the control and movement,a skid-steered mobile robot is designed and implemented in this paper,which has the characteristics of structural simplification,movement agility,high reliability and good controllability.Data fusion of monocular vision and IMU is implemented by EKF and feature detection and match of tracking object are fulfilled by SURF.The Range/Depth information as well as the information of object turning is determined combining with vertical line feature of circumstance.Experiments show that the system and the algorithm are real-time and robust.

关 键 词：SURF 垂直线 单目视觉 IMU 跟踪机器人 

分 类 号：TP262.6[自动化与计算机技术—检测技术与自动化装置]