Contour reconstruction for multiple 2D regions based on adaptive boundary samples

Abstract

There exist a lot of algorithms for 2D contour reconstruction from sampling points which guarantee a correct result if certain sampling criteria are fulfilled. Nevertheless nearly none of these algorithms can deal with non-manifold boundaries of multiple regions. We discuss, which problems occur in this case and present a boundary reconstruction algorithm, which can deal with partitions of multiple regions, and nonsmooth boundaries (e.g. corners or edges). In comparison to well-known contour reconstruction algorithms, our method requires a lower sampling density and the sampling points can be noisy. © Springer-Verlag Berlin Heidelberg 2009.