Distance Based Feature Detection on 3D Point Sets

Abstract

We propose a distance based algorithm for implicit feature detection on 3D point sets. Instead of directly determining whether a point belongs to a feature of the 3D point set or not, we first compute the distance between the point and its nearest feature. The obtained distance function is filtered, removing noise and outliers, and the features of the point set are computed as the zero set of the filtered function. Initial tests show that the proposed method is robust and can deal with amount of noise usually expected in a point set.