Feature aware Digital Surface Model analysis and generalization based on the 3D Medial Axix Transform

Abstract

Modern Digital Surface Models (DSMs) are highly detailed and cover large areas. This brings great advantages for applications such as flood modeling, crisis management and 3D city modeling. Unfor- tunately, and despite recent developments on this subject, current methods are unable to fully take ad- vantage of modern DSMs. First, because of their huge data volume. And second, because conventional methods use only 2.5D data-structures and algorithms. As a result of the latter, valuable 3D information that is present in modern DSMs is ignored. This research aims to develop a methodology for the analysis and generalisation of modern DSMs that uses truly 3D data-structures and algorithms. It will be based on the Medial Axis Transform (MAT), a compact descriptor of the geometry and topology of shapes. It is the hypothesis of this research that the MAT can facilitate truly 3D analysis and generalization of modern 3D DSMs, through the definition of features (significant and application dependent surface characteristics) based on the geometrical and topological properties of the MAT. The appropiate algorithms and data-structures will be designed, prototyped and robustly imple- mented. To support this process, a number of case studies will be performed, that each focuses on a distinct and practical application for which conventional 2.5D data-structures and algorithms have proven to be unsatisfactory.