Interpretation Of Traffic Scenes Using A Hierarchical Data Structure

Abstract

This paper describes a hierarchical vision system for the detection and classification of traffic signs on freeways. The structure and color information of an image is used for detection of traffic signs in the image and coarse classification. Regions-of-interest (RoI) each of them containing only one hypothetical traffic sign, are investigated using pixel classification methods. In this paper, the use of structure information is emphasized and some aspects of its integration into a vision system are demonstrated.