Range Image Registration via Probability Field
Abstract
This paper presents a powerful variant of the ICP (IterativeClosest Point) algorithm for registering range imagesusing a probability field. The probability field (p-field) representsthe probability distribution of the surface position.By capitalizing on the properties of the range image, fastconstruction, compact representation, and efficient query ofthe p-field can be achieved. Different sensor models aresupported by the p-field according to the properties of therange image. Range images can be precisely aligned bymaximizing the probability of overlapping surfaces via thep-field.
Images and movies
BibTex references
@Article {ZHK04, author = "Zhang, Haitao and Hall-Holt, Olaf and Kaufman, Arie E.", title = "Range Image Registration via Probability Field", booktitle = "Proceedings of the Computer Graphics International", journal = "na", pages = "546--552", year = "2004", publisher = "IEEE Computer Society Washington, DC, USA", keywords = "Probability Field, Range Image, ICP, Image Alignment, Registration", url = "http://cvc.cs.stonybrook.edu/Publications/2004/ZHK04" }