Natural Landmark Detection for Visually-Guided Robot Navigation
Enric Celaya, Jose-Luis Albarral, Pablo Jimenez and Carme Torras
The 10th Congress of the Italian Association for Artificial Intelligence - Special Track on AI and Robotics
Roma, Italy, September 10-13, 2007
Abstract
Landmark-based navigation in unknown unstructured environments is still far from solved. The bottleneck nowadays seems to be the fast detection of reliable visual references in the image stream as the robot moves. To solve this problem, we developed a strategy to detect and track salient regions along image streams by just performing on-line pixel sampling. This strategy continuously updates the mean and covariances of the salient regions, as well as creates, deletes and merges regions according to the sample flow. Regions detected as salient for long enough periods of time can be considered as potential candidates to define a landmark that can be used in the navigation task. Once a salient region is found, a landmark associated with it is defined and characterized by a set of subregions with given color and normalized shape. Landmarks are stored in a database for posterior identification along the navigation process.