Filling holes in an elevation surface structured by a sparse 3D volume

Abstract

Geometric modeling is a field of research based on applied mathematics and computational geometry. It studies methods and algorithms on the mathematical description of shapes or realistic object for computer graphics and 3D simulation. Surface reconstruction in 3D is one of the important steps in the geometric modeling’s application to process a 3D surface. Reconstruction of a 3D surface consists of many steps such as simplification, triangulation, subdivision, hole filling and refining. As we known, the method of filling holes in a triangular surface has been researched and developed on 3D meshes, triangular meshes and polygon meshes. However, filling the holes of 3D point clouds is still a challenge for researchers. In this thesis, we propose a method for filling the holes of 3D point clouds. Our method includes three steps. In the first step, we determine the holes of 3D point clouds’ surface. The second one is filling these holes. Finally, we refine the holes in order to adapt and keep the local curvature of the surface. We present our method and some experimental results, respectively. Key words: Geometric models, reconstruction surface, 3D point clouds, filling the holes.