         
|
GRASP Lab Seminar 2003-2004
June 25, 11:00 AM, Levine Hall 307, hosted by Kostas Daniilidis.
Toon Goedeme
PSI-VISICS, Katholieke Universiteit Leuven, Belgium
Visual Topological Navigation using Omnidirectional Images
Abstract: Autonomous robots still often use special markers.This is especially the case in cluttered scenes, where the walls are difficult to pick up from laser scanner data. In this project we have developed algorithms for picking up novel kind of landmarks that are available in most scenes. Hence, No special markers are required, as these landmarks can be extracted from the scene itself. Moreover, the characterization of these landmarks is based on the joint invariance under viewpoint and illumination changes.
The system is albe to build up a topological model of the environment, solely based on information from omnidirectional images that are taken during a training tour through the environment. Unlike popular metrical world models such as occupancy maps and 3D models, it concerns a graph-like representation which reflects the topological structure of the environment. Streets, for instance, become edges that connect the crossroads and squares in a town. An indoor environment is modeled at the level of corridors connecting rooms. As a consequence, the robot can navigate through the environment robustly at a high cognitive level. It will be possible for the robot to repeat the learnt trajectories by comparing its current position against those from where the example images have been taken. The robot can find its position even when it is off course and under different illumination conditions, due to the invariance in the description of the natural landmarks. The robot will automatically adapt its position to match it against the most similar example image. In this project we aim at the application of a self-steering electronic wheel chair. The topological navigation technique can implement the global navigation tasks of such a moving vehicle.Biography: Toon Goedeme obtained a Masters in Electrotechnical Engineering (Multimedia and Signal Processing) at the Katholieke Universiteit Leuven in Belgium in June 2002. Since September 2002, he has been working as a research assistant at the Computer Vision Lab (VISICS) under the sypervision of Prof. Luc Van Gool.
