THE ROBUST SEQUENTIAL ESTIMATOR - A GENERAL-APPROACH AND ITS APPLICATION TO SURFACE ORGANIZATION IN RANGE DATA

We present an autonomous, statistically robust, sequential function ap proximation approach to simultaneous parameterization and organization of (possibly partially occluded) surfaces in noisy, outlier-ridden (n ot Gaussian), functional range data. At the core of this approach is t he Robust Sequential Estimator, a robust extension to the method of se quential least squares. Unlike most existing surface characterization techniques, our method generates complete surface hypotheses in parame ter space. Given a noisy depth map of an unknown 3-D scence, the algor ithm first selects appropriate seed points representing possible surfa ces. For each nonredundant seed it chooses the best approximating mode l from a given set of competing models using a modified Akaike Informa tion Criterion. With this best model, each surface is expanded from it s seed over the entire image, and this step is repeated for all seeds. Those points which appear to be outliers with respect to the model in growth are not included in the (possibly disconnected) surface. Point regions are deleted from each newly grown surface in the prune stage. Noise, outliers, or coincidental surface alignment may cause some poi nts to appear to belong to more than one surface. These ambiguities ar e resolved by a weighted voting scheme within a 5 x 5 decision window centered around the ambiguous point. The isolated point regions left a fter the resolve stage are removed and any missing points in the data are filled by the surface having a majority consensus in an 8-neighbor hood.