Surface Reconstruction from Multi-resolution Sample Points

Abstract

Robust surface reconstruction from sample points is a challenging problem, especially for real-world input data. We significantly improve on a recent method by Hornung and Kobbelt [HK06b] by implementing three major extensions. First, we exploit the footprint information inherent to each sample point, that describes the underlying surface region represented by that sample. We interpret each sample as a vote for a region in space where the size of the region depends on the footprint size. In our method, sample points with large footprints do not destroy the fine detail captured by sample points with small footprints. Second, we propose a new crust computation making the method applicable to a substantially broader range of input data. This includes data from objects that were only partially sampled, a common case for data generated by multi-view stereo applied to Internet images. Third, we adapt the volumetric resolution locally to the footprint size of the sample points which allows to extract fine detail even in large-scale scenes. The effectiveness of our extensions is shown on challenging outdoor data sets as well as on a standard benchmark.