Building hierarchical solid models from sensor data

Abstract

The authors consider the problem of aggregating the geometrical information provided by sensors such as range finders, and its applications to scene modeling. The information consists of collections of three-dimensional surface points that form a discrete subspace of the objects-to-free-space boundaries within the world to be modeled. The ultimate goal is to obtain a valid surface model which can in turn be transformed into an efficient volumetric representation for solid-interference-detection algorithms. This representation should achieve search efficiency, compactness, and parallelism. The authors first introduce such a volumetric representation which uses a tetrahedrization of space. Then they show that surface connectivity is a nontrivial and essential element of a valid representation. To that end, they introduce a formal graph-theoretic definition for model validity which they use to guide the process of aggregating the different views. Finally, the topological problems posed by objects nonhomeomorphic to spheres, such as multiholed tori, are introduced.< >