全区域覆盖移动机器人的边界识别研究

【作者】 闫肖肖；

【导师】 王华坤；

【作者基本信息】 南京理工大学 ， 机械电子工程， 2008， 硕士

【摘要】 本文以全区域覆盖移动机器人为研究对象,主要研究其工作区域的边界识别问题。文中,机器人采用视觉传感器和光电编码器与电子罗盘构建“环境地图信息”:在工作过程中,机器人通过传感器实时获得当前的位姿信息,并将其与存储的“环境地图信息”进行对比,判断是否到达边界。在边界识别过程中,机器人采用视觉导航方法识别、跟踪边界直线,通过边界直线检测自身的位置,并修正定位误差。本文首先提出了边界识别方法,研究了机器人如何利用视觉传感器识别、跟踪边界;然后构建了机器人的视觉模型,通过视觉模型得到了机器人位置与图像坐标的对应关系;接着研究了图像处理方法:滤波、增强、分割、形态学运算等方法,图像经过处理后,利用Hough变换方法拟合到边界直线,然后再根据坐标变换关系得到边界直线与机器人位置的对应关系;最后设计了所需的图像采集处理系统以及边界识别系统。本方法利用直线检测机器人的位置,提高了机器人定位的精度,使机器人一次覆盖下来的漏割区域大为减少,节约了机器人进行全区域覆盖时所需的时间,具有一定的应用价值。更多还原

【Abstract】 This paper based on the region-wide coverage robots, mainly research how do they identify the boundary of the work area when they work. In this paper,the robot uses visual sensors and photoelectric encoder and electronic compass to establish "map information";when it works,it gets position fom sensors to compare with "map information" to detect whether the robot is reaching the boundary. In the course of its boundary identification,the robot uses visual navigation to identify and track the boundary lines,uses lines to detect its own positions and corrects wrong positions.Fist, say the method about the identification of boundary,research how the robot identifies doundary with the visiual sensors;then establish visual model,and get the relationship between the robot’s positions and the image coordinates;then research methods of image processing:filtering、enhancing、threshold segmentation、morphological operations and so on,after these finishing these methods,get boundary lines in the image through Hough transform, and get the boundary lines with the corresponding robot’s positions;at last,design the image acquisition and processing system, as well as the boundary identification system.This method uses lines to detect the robot’s positions, improves the accuracy of the robot’s positions, reduce the missing cut area after once covering, saves region-wide coverage’s time, and has a certain value.更多还原