摘要
RoboCup标准平台比赛(SPL),由于环境的动态性,需要NAO机器人根据场上情况快速稳定地确定自身位姿状态。介绍了一种基于KD树的最近邻搜索实现特征匹配建立色表、通过颜色识别进行场地建模、根据单目视觉测距来实现NAO机器人的自定位的方法。实验证明,方法在各种复杂的环境影响下,能够快速有效的识别目标,并准确的实现机器人的自定位。
On Soccer Standard Platform League, the NAO robot need determine the posture itself fast and stably according to the dynamic environment.This paper presents a self-localization approach of NAO robot basing the monocular distance measurement.To create the color table by implementing the feature matching which is according to the nearest neighbor search based on KD tree,and modeling the soccer field by color identification.Experiments have shown that this method could quickly and efficiently identify the target and accurately implement the robot self-localization under various complex conditions.
出处
《工业控制计算机》
2013年第12期35-36,39,共3页
Industrial Control Computer
关键词
最邻近搜索
目标识别
单目视觉
自定位
the nearest neighbor search
target identification
monocular vision
self-localization
