On Tuesday 5 April from 12 to 14 in lecture room R a number of our Geomatics students will tell about their MSc thesis research. You are cordially invited to attend this lunch meeting.
The topics are:
— P3 Presentation: Florian W. Fichtner —
Automatic space subdivision for multi-story pathfinding on a 3D point cloud using an octree
Supervisor: Abdoulaye A. Diakité, Sisi Zlatanova, Robert Voûte (CGI)
Under supervision of CGI Nederland
Point clouds can be acquired rapidly and relatively cheaply, but they lack a structure and semantics which are necessary for pathfinding. This thesis aims to create a workflow to subdivide the 3D space of an indoor point cloud of a building and to create a semantically enriched model for pathfinding between different floor levels.
During this presentation I will show how far the research question can currently be answered: “To what extent can an octree data structure be used to subdivide 3D space and to create a model for multi-story pathfinding?”. Furthermore, intermediate results will be shown and an estimate about what can be achieved until P4 will be given.
— P3 Presentation: Merwin Rook —
Automatic semantic and thematic labelling of 3D city models
Supervisors: Abdoulaye A. Diakité, Filip Biljecki
Current geographic information systems (GIS) use 2D information, while a wide field of GIS applications, like flood modelling or real estate registration and valuation, require 3D information. Evolving technologies in remote sensing bring new possibilities in capturing the human environment in 3D. In order to effectively use this 3D information, the data must be enriched with semantic information; in other words, does a surface represent a roof, a wall or terrain? My research aims at finding ways to automate the process of semantic and thematic classification of geographical data about the built environment.
— P3 Presentation: Ivo de Liefde —
Exploring the Use of the Semantic Web for discovering and retrieving data from Sensor Observation Services
Supervisors: Marian de Vries, Martijn Meijers
Sensor Observation Services (SOS) are sources of open sensor data on the internet based on OGC and ISO standards. The data in these services are free to be used by anyone. However, compared to their commercial competitors they are hard to find and cover only a limited area. Therefore, this thesis aims to explore the use of the semantic web to automatically discover, retrieve, integrate and aggregate data from multiple sensor observation services, allowing easier access to these sources of open data.