Geoscience Reference

In-Depth Information

2 The First Calculation Concept

Since 2012 LiDAR-data with a point density between 1 and 4 points per square

meter are available in Bavaria. This data are suitable for calculating a precise

digital elevation model (DEM) as well as a surface model. Currently the DEM is

available in the minimum mesh size of 1 m for approximately 90 % of Bavaria's

territory. The overall coverage will be achieved in the middle of 2014.

Together with the building ground plans from the cadastre, LiDAR-data are

suited for the first acquisition of a 3D Building Model (Schilcher et al.
1998
,
1999
;

Schilcher and Roschlaub
1999
). The intersection of the building ground plans with

the DEM provides the building root points. The DSM is used as data basis for

the roof recognition. In the following, the task of recognizing roofs automatically

from LiDAR-data is described. Subsequently the Bavarian method for the initial

acquisition of a nationwide 3D Building Model is described.

2.1 General Problem Definition

The usage of randomly spread point clouds for the derivation of 3D Building

Models causes difficulties for the interpretation of basic and unstructured eleva-

tion models and the subsequent modeling of the complex vector geometries. For

all point clouds (steadily or unsteadily arranged) the data information content is

limited to the elevation. Additional structures are missing. Nevertheless, by link-

ing the point clouds with building ground plans (Fig.
1
) which are for example

available up to date in the cadastre, the point clouds get a first semantic and spatial

attribution.

The assignment of the building ground plan complies a selection of relevant

points from the point cloud. The elevation information of the selected points can

now be interpreted as representatives of the roof. First roof structures like ridge-

lines can plastically be visualized with a Delaunay Triangulation. A relevant geo-

metrical attribution as vectorial surface is missing.

One possible method of automatic roof reconstruction proceeds as follows:

Each triangle surface, which was calculated from the relevant points by an ele-

vation-independent (2.5-dimensional) Delaunay Triangulation, is dedicated to a

standardized surface normal with a length of 1 m.

Subsequently it will be analyzed to which side of the building ground plan each

surface normal belongs. As presented in Fig.
2
any of the surface normals will be

selected. It can then be projected in the xy-plane due to its tilt and a respective

direction angel of the xy-plane can be calculated. Considering a clockwise build-

ing ground plan direction it is first tested, whether the direction angel of the pro-

jected surface normal is orthogonal to the side of the building. A certain variance

between the direction angels is considered. Secondly the minimal orthogonal dis-

tance is calculated from the base point of the surface normal to one of the sides of