文档介绍：沈阳理工大学
硕士学位论文
基于边缘检测的移动机器人视觉导航方法研究
姓名:刘微
申请学位级别:硕士
专业:信号与信息处理
指导教师:胡玉兰
20100301
沈阳理工大学硕士学位论文

摘要

近年来,随着计算设备运算速度的提高、图像技术的进步,视觉导航成为移
动机器人导航的一个发展趋势。在移动机器人的视觉导航技术中,路径识别与跟
踪是其关键技术,能够快速、准确地识别出机器人导航路径并进行跟踪对机器人
视觉导航具有重要意义,为此,本文研究了基于边缘检测的移动机器人视觉导航
方法,主要研究内容如下:
(1) 分析了传统的边缘检测算法、给出了一种改进的形态学边缘检测方法,并
进行了仿真实验,对仿真结果进行了对比分析,表明了改进的形态学边缘检测方
法具有较好的抗噪性能,得到的边缘细节丰富、连续性较好,本文采用此算法对
视觉图像进行处理,辅以 Hough 变换得到导航路径,同时,为了提高 Hough 变换
检测直线的速度和准确性,本文在 Hough 变换检测直线之前先对边缘检测图进行
细化处理。
(2) 分析了图像坐标和世界坐标的变换关系,并进行了标定实验;将移动机器
人前进方向与路经的角度偏差和距离偏差作为控制器的输入,设计了双输入单输
出的模糊控制器以实现移动机器人的路径跟踪功能。
本文将理论研究进行了仿真实验与分析,均获得较满意的实验结果,验证了本
文提出的基于边缘检测的移动机器人视觉导航方法的有效性。

关键词:视觉导航,边缘检测,数学形态学,Hough 变换,路径跟踪
沈阳理工大学硕士学位论文

Abstract

In recent years, with the increasing putational speed puting devices,
advancing in imaging technology, visual navigation is ing a trend. In the visual
navigation technology, the path identification and follow is the key. It is meaningful for
the mobile robot vision navigation to recognize and follow the robot navigation path
accurately and fast. Therefore, this article has studied method of mobile robot vision
navigation based on edge detection, main research contents are as follows:
(1) This article analyse the traditional edge detection algorithm, gave improved
morphological edge detection method, carried out simulation experiments and analyzed,
proved that the improved morphology edge detection method has the good anti-noise
performance, detail of obtained edge is rich and the continuity is good. This article has
used this algorithm to process the visual image, auxiliary obtained the navigation path
by the Hough transformation. In order to enhance the speed and the accuracy of
detection straight line, this article has firstly carried on refinement processing to the
edge image before Hough transformation detect straight line.
(2) This article has analyzed the transformation relations between the image
coordinate and