| Abstract | This article introduces an approach, based on Bayesian Networks, for the grouping of 3-D surfaces extracted from data obtained by a laser ranging sensor. A methodology based on the decomposition of an object into its sub-parts is used for specifying the structure of the network. Conditional probabilities are computed using a set of compatibility functions that measure the are a measure in the quality of fit of the data to a model that the features may have come from. These compatibility functions are akin to measures that are used for the perceptual organization of features in the computer vision domain except that they have been developed for 3-D range data. An approach is presented for the mapping of the compatibility functions to conditional probabilities that are required by the Bayesian network. An example of a Bayesian network is presented that models the detection of corners and continuity among planar surfaces and uses both range and intensity values as features sets. The Bayesian network is used to compute a belief value in the formation of corners and continuity among the surfaces which in turn can be used to decide if surfaces should be joined. Results and analysis are presented for an actual set of intensity and range images taken of a typical indoor scene of a robotics laboratory. |
|---|
