摘要
路标识别方法是室外智能移动机器人机器视觉导航的关键技术之一。该文针对公路上的３种交通标志（箭头）提出了快速识别算法。算法通过搜索候选者、“第一次生长”、阈值自动选择、“第二次生长”４步从图象中分割出目标，而后进行分类。在基于轮廓结构分析的分类器设计中，利用“最短光栅距离”设计了直线判别准则，并用回溯法通过轮廓关键点的匹配来完成模式识别。整个算法结构简单，时间复杂度低，已成功地应用于ＴＨＭＲ－Ⅲ在校园道路上的实时导航，实验表明算法具有良好的实时性和鲁棒性。
The approach to the recognition of road labels is one of the key technology in vision based navigation of outdoor intelligent mobile robots. This article discusses a fast algorithm for recognition of three kinds of traffic road labels on highway. A road label is captured from an image and then classified by a four step process which is composed of goal search, “first time growth”, automatic threshold selection and “second time growth”. In the design of frame analysis based classifier, the standard of “shortest grid range” is used to find out straight line. The pattern recognition is realized by the backward frame key point matching. The algorithm is simple and near real time. It has been applied to the real time navigation of THMR Ⅲ (Tsinghua mobile robot model Ⅲ) successfully on the campus roads, which has demonstrated good performances such as on robust and real time.
出处
《清华大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
1998年第S1期107-110,共4页
Journal of Tsinghua University(Science and Technology)
基金
国家"八六三"高技术项目
