基于直线检测的室内移动机器人视觉定位方法  

Visual location method of indoor mobile robot based on beeline detection

作　　者：周绍磊[1] 吴修振[1] 祁亚辉[1] 公维思 

机构地区：[1]海军航空工程学院控制工程系,山东烟台264001

出　　处：《华中科技大学学报（自然科学版）》2016年第10期93-97,共5页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(61004002)

摘　　要：针对室内移动机器人定位问题,提出了基于直线检测的室内移动机器人视觉定位方法.该算法通过Canny算子对图像进行边缘检测,设计自适应霍夫变换检测边缘中的直线,对直线进行平移和旋转变换,求取视场中心点到直线的距离,使用证据推理判断视场中心点是否穿越直线,从而获得机器人的实时坐标.实验结果表明:该算法切实可行,实时性好,定位精度高,具有很强的鲁棒性,很好地解决了室内移动机器人的定位问题.In order to solve the problem of localization of indoor mobile robot, the visual location method of indoor mobile robot based on beeline detectoion was put forward. In this algorithm, image edge was detected by Canny operator, and adaptive Hough transform was used to check the beelines in the image edge. After translational transformation and rotational transformation, by calculating the distance of the center of the visual field to the beelines, the real-time coordinate was obtained by tracking the beelines crossing of the center of the visual field through evidence reasoning. Experiments show that the algorithm with high positioning accuracy, is feasible and has a good real-time perform- ance and strong robustness, which gives a good solution to the problem of localization of indoor mobile robot.

关 键 词：视觉定位 自适应霍夫变换 平移变换 旋转变换 证据推理 

分 类 号：V249.325[航空宇航科学与技术—飞行器设计]