3D Scene Modeling for Robot Navigation Using Region Based Stereo

In this paper we describe an approach to efficient and robust depth extraction for 3D scene modeling for an autonomous mobile robot navigation system operating in static indoor environments. The depth extraction is performed with region based stereo aided by an a priori world model. Furthermore, a Multi Primitive Hierarchical matching strategy is employed to ensure fast and robust matching of lower level features (lines and corners) that are used to refine the depth estimates delivered by the regions alone. Experimental results proved that region based stereo used with an a priori scene model has a potential in 3D scene modeling for mobile robot navigation.