Post Facto Registration Tools for Urban Modelling

Abstract

Urban modelling applications require high-precision geometric models both for graphical rendering and for engineering purposes. While geometric models, photographic images and laser-scanned point clouds are ideally co-registered to national coordinate grids at the time of acquisition, the quantity and diversity of data sources means that registration must often be performed post facto. Moreover, the sheer size of urban point clouds prevents automated conversion and registration of the entire data set at once. We describe an interactive tool that manages the workflow for converting urban-scale point clouds to grid-registered geometric models. Our 'user in the loop' approach lets us leverage natural human understanding of the data to bypass issues of scale.