Regions segmentation using multiple cues for robot navigation on outdoor environment

This paper proposes a method segments the object from an image taken by moving robot in outdoor environ- ment. We show the method of segmentation based on low-level features using multiple cues. Multiple local cues are color, straight line, edge, context information, Hue Co-occurrence Matrix (HCM), Principal Components (PCs) and Vanishing Point (VP). Model the objects of outdoor environment that define their characteristics individually. We classify the object into natural and artificial ones. We detect sky and tree of natural object and building of artificial object. We segment the region as mixture using the proposed features and methods. Objects can be detected when we combine predefined multi- ple cues. In this paper, we segment region of sky, tree and building. We use features of color, edge, context information to extract of sky. Tree region use the features of color, edge, context information of HCM. We use to extract building using color, straight line, PCs, edge and vanishing point. Finally, we confirm the result of region segmentation using multiple cues on outdoor environment.