一种基于全景视觉系统的Robocop机器人定位方法  

A Localization Method for Robocup Robot Base on Omni-Vision System

作　　者：刘光宇[1] 刘国栋[1] 韩云生[1] 

机构地区：[1]江南大学通信与控制工程学院,江苏无锡214122

出　　处：《江南大学学报（自然科学版）》2009年第3期279-282,共4页Joural of Jiangnan University (Natural Science Edition)　

摘　　要：对全景视觉成像原理进行了讨论,提出了一种基于YUV色彩空间的多阈值图像分割和利用三角定位原理实现的机器人定位方法。通过识别全景视觉图像中坐标已知的特征信标,提取信标特征,进而应用几何原理计算出机器人的空间坐标,实现机器人的定位,为机器人导航和蔽障工作奠定了基础,有效地减少了彩色图像分割的计算量,避免了全景视觉图像柱面展开所带来的误差。通过实验证明,本方法能够较精确地实现机器人的定位。The paper presents a location method based on omni-vision system for mobile robot system, studies omni-vision system imaging principle, brings up a method using multi-threshold based on color space and making use of a triangle to location. By distinguishing the characteristic road signs, abstracting the road signs that the coordinate is already knew, the geometry principle can then be used to calculate the robotic space coordinate, which is the base of the robot guides and of avoiding the obstacles. The calculation having decreased by color image division avoids error brought by omni-vision image cylindrical surface development. The experimental results verified the real-time and reliability of the method.

关 键 词：全景视觉 移动机器人 自主定位 图像处理 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]