Building Footprint Extraction from LiDAR Data and Imagery Information

Abstract

This study presents an automatic method for regularisation of building outlines. Initially, building segments are extracted using a new fusion method. Data- and model-driven approaches are then combined to generate approximate building polygons. The core part of the method includes a novel data-driven algorithm based on likelihood equation derived from the geometrical properties of a building. Finally, the Gauss-Helmert and Gauss-Markov models adjustment are implemented and modified for regularisation of building outlines considering orthogonality constraints.