Described here i an extraction and recongition method of road shape models from static color images of road scene. As the basic process of vision-based envir onment recongiton, it is used extensively for the researc of autonomous vehicles and ITS (Intelligent Transport Systems). The usuall methd is depending on the recognition result of road lane marks and generally used on expressways or hihways only. In contrast, we propose a region sementation and recgnition algorithms using color and texture information, and a robust recognition method of road shape model for general roads. Experiment based on real roadscene images shows the feasibility o f our approach.
展开▼