AVENUE: Automated site modeling in urban environments

摘要

This paper is an overview of the AVENUE project at Columbia University. AVENUE's main goal is to automate the site modeling process in urban environments. The first component of AVENUE is a 3-D modeling system which constructs complete 3-D geometric models with photometric texture mapping acquired from different viewpoints. The second component is a planning system that plans the Next-Best-View for acquiring a model of the site. The third component is a mobile robot we have built that contains an integrated sensor suite for automatically performing the site modeling task. We present results for modeling buildings in New York City.