摘要
提出了一种用混合编码路标的方法对移动机器人精确定位。根据路标像与像面上平均能量之间的关系,选取出合适的动态阈值,在大多数环境中能从天花板像中提取出路标像;在特殊情况下,阈值自动调整。采用竖直向下、水平向左相邻栅格查询的办法对阴影区分类,从而确定路标像的区域。依据路标像上的编码信息和重心坐标换算出机器人在环境中的位置和航向。实验表明,机器人航向角误差不超过3,°机器人位置误差不超过1%。
Mixed coding landmarks are used in the precise localization of a mobile robot. The appropriate dynamic threshold is selected to pick up a landmark image from a ceiling image in the most environments according to the relation between the landmark image and average energy on the image. The threshold is automatically adjusted under special conditions. The shadow regions are classified by inquiring neighbor lower and left grids, and then the region of the landmark image is confirmed. The position and the direction of the mobile robot are calculated on the basis of the centre of gravity and coding information of the landmark image. Experimental results show that errors of navigation angles are no more than 3^. and errors of positions are no more than 1 %.
出处
《数据采集与处理》
CSCD
北大核心
2007年第2期196-200,共5页
Journal of Data Acquisition and Processing
基金
国家自然科学基金(60475034)资助项目
霍英东教育基金(91055)资助项目
关键词
混合编码路标
视觉定位
动态闽值
栅格分类
mixed coding landmark
visual localization
dynamic threshold
grid classifying
