Vision Methods for Outdoor Mobile Robot Navigation

摘要

This study deals with a vision system for outdoor mobile robot navigation. Theconcentrated aspects include edge extraction, pattern recognition, positioning, and image sensing. In this study, several edge detection or extraction methods are described and developed. The vision system for vehicle navigation is used for real-time control because of the fast processing and reliability of our new edge definition and searching strategy. In this study, the proposed fast tactics for detecting and recognizing signs are based on a sign's color information and shapes. On the basis of the simplified formulas of three-reference-points, a new mode is presented which uses only two reference points together with a hardware device which can measure the heading angle automatically.