In: Stilla U, Rottensteiner F, Paparoditis N (Eds) CMRT09. IAPRS, Vol. XXXVIII, Part 3/W4 — Paris, France, 3-4 September, 2009 
3D BUILDING RECONSTRUCTION FROM LIDAR 
BASED ON A CELL DECOMPOSITION APPROACH 
Martin Kada a , Laurence McKinley b 
a Institute for Photogrammetry, University of Stuttgart, Geschwister-Scholl-Str. 24D, 70174 Stuttgart, Germany 
martin.kada@ifp.uni-stuttgart.de 
b Virtual City Systems, Zellescher Weg 3, 01069 Dresden, Germany 
lmckinley@virtualcitysystems.de 
Commission III, WG III/4 
KEY WORDS: LIDAR, Reconstruction, Building, Automation, Algorithms 
ABSTRACT: 
The reconstruction of 3D city models has matured in recent years from a research topic and niche market to commercial products and 
services. When constructing models on a large scale, it is inevitable to have reconstruction tools available that offer a high level of 
automation and reliably produce valid models within the required accuracy. In this paper, we present a 3D building reconstruction 
approach, which produces LOD2 models from existing ground plans and airborne LIDAR data. As well-formed roof structures are of 
high priority to us, we developed an approach that constructs models by assembling building blocks from a library of parameterized 
standard shapes. The basis of our work is a 2D partitioning algorithm that splits a building's footprint into nonintersecting, mostly 
quadrangular sections. A particular challenge thereby is to generate a partitioning of the footprint that approximates the general shape 
of the outline with as few pieces as possible. Once at hand, each piece is given a roof shape that best fits the LIDAR points in its area 
and integrates well with the neighbouring pieces. An implementation of the approach is used now for quite some time in a production 
environment and many commercial projects have been successfully completed. The second part of this paper reflects the experiences 
that we have made with this approach working on the 3D reconstruction of the entire cities of East Berlin and Cologne. 
1. INTRODUCTION 
3D building reconstruction has been a topic for quite some time 
now. Many research papers have been published; commercial 
services and software are available. (Brenner, 2005), e.g., gives 
a good overview of reconstruction methods and points out that 
“research is still far from the goal of the initially envisioned 
fully automatic reconstruction systems”. This situation has not 
yet changed much, although a lot of research is still devoted to 
this topic, as can be seen in the multitude of recent publications 
(e.g. (Arefi et al., 2008), (Moser et al., 2009), (Sohn et al., 
2008)). 
The subject of this paper is on the generation of realistic 3D city 
models in LOD2 as it is defined in the official OGC standard 
CityGML (see e.g. (Kolbe, 2009)). At this LOD, buildings have 
distinctive roof structures and flat facades that are textured from 
terrestrial or oblique aerial images. 
As the data basis, we rely on existing ground plans and airborne 
LIDAR data. A frequent requirement, especially from customers 
within the mainland Europe, is that the provided building 
outlines are to be preserved with only little tolerance and that 
ridge and eaves heights must be very accurate. This is especially 
important so that the facades and roofs can be properly mapped 
from oblique aerial images. 
The presented reconstruction approach is motivated from our 
research on the simplification of 3D building models for map 
like representations (Kada, 2007). An integral part of this work 
lies on a new method to decompose a 2D building footprint into 
a small set of nonintersecting primitives. Although the resulting 
partitioning only approximates the original outline, it is still 
accurate enough for reconstruction purposes. The benefit is, 
however, that the algorithm separates the sections nicely, 
especially for residential houses with gabled or hipped roofs. 
This eases the task of determining and assembling a valid roof 
structure from parameterized, standard shapes. 
In the second part of the paper, we give insight into two large- 
area projects that we have completed using the described 3D 
reconstruction system: East Berlin and Cologne. Figure 1 shows 
the reconstructed 3D city model of Berlin with textures mapped 
from oblique imagery. 
Figure 1. Real-time visualization of the 3D city model of 
Berlin. 
47