Title :
Attentional Landmark Selection for Visual SLAM
Author :
Frintrop, Simone ; Jensfelt, Patric ; Christensen, Henrik I.
Author_Institution :
Computational Vision & Active Perception Lab., Stockholm
Abstract :
In this paper, we introduce a new method to automatically detect useful landmarks for visual SLAM. A biologically motivated attention system detects regions of interest which "pop-out" automatically due to strong contrasts and the uniqueness of features. This property makes the regions easily redetectable and thus they are useful candidates for visual landmarks. Matching based on scene prediction and feature similarity allows not only short-term tracking of the regions, but also redetection in loop closing situations. The paper demonstrates how regions are determined and how they are matched reliably. Various experimental results on real-world data show that the landmarks are useful with respect to be tracked in consecutive frames and to enable closing loops
Keywords :
SLAM (robots); image matching; robot vision; attentional landmark selection; automatic landmark detection; biologically motivated attention system; feature similarity; scene prediction; short-term tracking; visual SLAM; visual landmarks; Biology computing; Cameras; Focusing; Intelligent robots; Layout; Mobile robots; Robot kinematics; Robot vision systems; Simultaneous localization and mapping; Tracking loops;
Conference_Titel :
Intelligent Robots and Systems, 2006 IEEE/RSJ International Conference on
Conference_Location :
Beijing
Print_ISBN :
1-4244-0258-1
Electronic_ISBN :
1-4244-0259-X
DOI :
10.1109/IROS.2006.281711