Model reconstruction from cloud data

Abstract

Model reconstruction provides a powerful paradigm for modeling shapes from samples. For cloud data with only geometric coordinates as input, the surface reconstruction algorithms that utilizing marching cubes have been shown to be quite effective both in theory and practice. However, the marching-cubes-based algorithms are not sufficiently efficient for handling large data set. In this paper, a data-filtering algorithm is first presented to automatically reduce the number of cloud points under given tolerance. Then the triangle mesh surface is reconstructed from the simplified data set by using the marching-cubes algorithm. The proposed data-filtering procedure and surface reconstruction approach are highly integrated and the efficiency of the solution is greatly improved. An approach is also put forward to create padding triangles with optimized shapes for covering the unwanted gaps in the reconstructed triangle meshes. A practical example is included to demonstrate our method.