基于改进FAST算法的移动机器人导航  

Research on the navigation of mobile robots based on the improved FAST algorithm

作　　者：蒲兴成[1] 谭少峰 张毅[2] 

机构地区：[1]重庆邮电大学数理学院,重庆400065 [2]重庆邮电大学自动化学院,重庆400065

出　　处：《智能系统学报》2014年第4期419-424,共6页CAAI Transactions on Intelligent Systems

基　　金：国家自然科学基金资助项目(51075420);重庆市教委科学技术研究资助项目(KJ100516;KJ1400432)

摘　　要：针对移动机器人在单目视觉导航方面实时性与鲁棒性较差的问题,提出一种基于FAST算法和一点RANSAC算法的移动机器人导航新方法,该方法主要由3个步骤组成:首先,机器人通过自身摄像头采用改进FAST角点提取算法采集周围的环境信息;其次,在匹配过程中采用结合扩展卡尔曼滤波的一点RANSAC算法进行误匹配剔除;最后,用得到的匹配点进行三维环境重建和地图创建。实验结果表明,改进方法能有效提高移动机器人定位与导航的实时性与鲁棒性。In order to solve the problem of poor real-time performance and weak robustness of mobile robots during monocular vision navigation , a new improved navigational algorithm is put forward on the basis of the FAST algo-rithm and the one point RANSAC algorithm for mobile robots .This improved new algorithm is made up of three main steps.Firstly, the mobile robots use the FAST algorithm to collect the information by camera .Secondly,in the process of matching , we use the improved RANSAC algorithm to reject the mismatching .Finally, a 3-D map of the environment is built .Experiments show that this method can improve the real-time performance and robustness ef-fectively regarding the aspects of positioning and navigation of the mobile robots .

关 键 词：移动机器人 单目视觉 FAST算法 扩展卡尔曼滤波 RANSAC算法 机器人导航 鲁棒性 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]