利用视觉灭点检测室内机器人偏航角的方法  

Method for detecting yaw angle of indoor robot using visual vanishing point

作　　者：王成 崔希民[1] 蔡量力 刘欢[1] 张建 WANG Cheng;CUI Ximin;CAI Liangli;LIU Huan;ZHANG Jian(College of Geoscience and Surveying Engineering,China University of Mining&Technology-Beijing,Beijing 100083,China)

机构地区：[1]中国矿业大学(北京)地球科学与测绘工程学院,北京100083

出　　处：《测绘通报》2021年第1期13-17,23,共6页Bulletin of Surveying and Mapping

基　　金：国家自然科学基金(51474217);越崎杰出学者奖励计划(800015Z1181)。

摘　　要：室内障碍地图在呈现给用户端浏览的同时,也是自主运动视觉机器人路径规划和障碍规避的依据。由于自主运动视觉机器人的起始位姿选择具有随机性,当机器人沿随机方向行走时,往往导致所建室内障碍地图产生偏斜。本文利用单张已标定像片检测视觉正交灭点,结合相机几何透视原理和灭点几何特性,提出了一种室内机器人偏航角的简单快速检测方法,并成功应用于自主运动机器人的轨迹优化和所建障碍地图的方向校正。对比试验表明,基于视觉灭点的室内机器人偏航角检测方法精度高,在纠正所建障碍地图朝向的同时,保证了机器人运行轨迹短、耗时低,具有良好的推广应用前景。While the indoor obstacle map is presented to the user to browse,it is also the basis for path planning and obstacle avoidance of the autonomous motion visual robot. Due to the randomness of the initial pose selection of the autonomous motion vision robot,when the robot walks in a random direction,it often causes the indoor obstacle map to be skewed. In this paper,a single calibrated photo is used to detect the orthogonal vanishing point. Combined with the geometric perspective principle of the camera and the vanishing point geometric characteristics,a simple and fast method for detecting the yaw angle of the indoor robot is proposed. And it is successfully used in the trajectory optimization of the autonomous motion robot and direction correction of the created obstacle map.The comparison experiment indicates that the indoor robot yaw angle detection method based on visual vanishing point has high accuracy. While correcting the orientation of the obstacle map,the robot has a short running trajectory and low time consumption. It has good prospects for promotion and application.

关 键 词：灭点 偏航角 检测 障碍地图 纠偏 

分 类 号：P228[天文地球—大地测量学与测量工程]