Surface creation and segmentation for 3D point clouds of an urban environment

Abstract

This paper presents a fast and robust approach to surface creation and feature extraction. The methodology is based on segmentation of point clouds iteratively till a set bound is reached. This paper concentrates on developing the methodology for developing planar surfaces. To achieve this goal vegetation is filtered and planar surfaces are created using the Delaunay triangulation. Surface creation process uses segmented point clouds based on fluctuation of normal of the surfaces in the segmented cubes. Results produced using this technique show the effect of imposing geometric constraints on the reconstruction to generate realistic surfaces.