International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B1, 2012 
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia 
KNOWLEDGE-BASED OBJECT DETECTION IN LASER SCANNING POINT CLOUDS 
F. Boochs *, A. Karmacharya, A. Marbs 
i3mainz, Insitute for Spatial Information and Surveying Technology, 
University of Applied Sciences Mainz, 
Lucy-Hillebrand-Str. 2, 55128 Mainz, Germany, 
(boochs, ashish, marbs) 9 geoinform.fh-mainz.de 
Commission III, Working Group III/2 
KEY WORDS: Point Cloud, Ontology, Modeling, Understanding, Processing, Classification, Algorithms 
ABSTRACT: 
Object identification and object processing in 3D point clouds have always posed challenges in terms of effectiveness and efficiency. In 
practice, this process is highly dependent on human interpretation of the scene represented by the point cloud data, as well as the set of 
modeling tools available for use. Such modeling algorithms are data-driven and concentrate on specific features of the objects, being 
accessible to numerical models. We present an approach that brings the human expert knowledge about the scene, the objects inside, and 
their representation by the data and the behavior of algorithms to the machine. This “understanding” enables the machine to assist human 
interpretation of the scene inside the point cloud. Furthermore, it allows the machine to understand possibilities and limitations of 
algorithms and to take this into account within the processing chain. This not only assists the researchers in defining optimal processing 
steps, but also provides suggestions when certain changes or new details emerge from the point cloud. Our approach benefits from the 
advancement in knowledge technologies within the Semantic Web framework. This advancement has provided a strong base for 
applications based on knowledge management. In the article we will present and describe the knowledge technologies used for our 
approach such as Web Ontology Language (OWL), used for formulating the knowledge base and the Semantic Web Rule Language 
(SWRL) with 3D processing and topologic built-ins, aiming to combine geometrical analysis of 3D point clouds, and specialists’ 
knowledge of the scene and algorithmic processing. 
also detected complex features using a knowledge base. Both 
  
1. INTRODUCTION works separate detection and qualification into two independent 
steps. Detection is based on predefined algorithmic sets, while 
A recent development in scanning technology is the ability to qualification uses knowledge to classify objects according to 
provide very precise data through the generation of highly their nature. Our work avoids such a separation and provides a 
dense 3D point clouds. Such high-density point clouds provide semantic bridge between scene knowledge and algorithmic 
a digital replica of the scanned scene. From the early days in 3D knowledge. Knowledge is part of an Algorithm Selection 
point cloud processing, the research has been focused on Module (ASM), which guides the processing independently of a 
investigating the reconstruction and recognition of geometrical particular scene. As a generic solution, it is extendable to any 
shapes (Wessel, Wahl, Klein, & Schnabel, 2008), (Golovinskiy, scene or algorithm. 
Kim, & Funkhouser, 2009). More complex strategies try to 
reconstruct complete sites. They can be broadly categorized The project uses environments from the Deutsche Bahn 
into two categories (Pu, 2010): data driven and model driven. (German Rail ) and Fraport (operator of Frankfurt Airport) 
Data driven methods (Beker, 2009), (Frueh, Jain, & Zakhor, to demonstrate effectiveness and versatility. Terrestrial laser 
2002) extract selected geometries from the point cloud and scanning technology is used to capture point cloud data. These 
combine them into a final model. With these methods, the huge datasets are then used on demand to create models of the 
redundancy of point cloud data produces difficulties due to objects inside of the installations. To date, the tasks of creating 
corresponding high ambiguity. Model driven methods try to and evaluating the 3D object models are solely manual, and 
take this into account. They use predefined primitive templates hence are costly in terms of both time and resources. The 
and information (as detected geometries) from the data to map existing tools do not provide significant assistance either, as 
them against the most likely templates (Ripperda, 2008). they are mostly data driven and concentrate on specific features 
of the objects to be used for numeric models. Algorithms have 
We present a knowledge driven method in which knowledge limited flexibility and can provide adverse effects when 
about the scene and the algorithmic processing is formalized deviated. Knowledge of algorithms and their limitations during 
logically for the generation of algorithmic sequences that detect implementation could limit such adverse effects. In the 
objects automatically. The developments are part of the project meantime, it provides flexibility to algorithmic manipulation 
WiDOP (Knowledge based detection of objects in point clouds for different scenarios. 
for engineering applications), which implements this method in = 
order to detect and identify objects in 3D point clouds. http://www.bahn.de 
Knowledge driven methods for object detection are relatively oh 
new. Pu (2010) detected exterior structures (mainly facades) of http://www.fraport.com/content/fraport-ag/en.html 
buildings through knowledge, and Maillot & Thonnat (2008) 
90 
Inte: 
neec 
diffe 
amo 
proc 
and 
scen 
incr 
fron 
with 
disti 
2.1. 
2.1.1 
Det: 
obje 
(“KI 
2.1. 
This 
scen 
that 
kno 
obje 
2.1. 
This 
kno 
reco 
thro 
2.2. 
This 
gene 
Met 
met] 
Fig 
The 
scen 
kno 
obje 
refir 
the | 
kno 
kno: