《Metric_localization_with_scale-invariant_visual_features_using_a_single_camera》.pdf

Metric Localization with Scale-Invariant Visual Features using a Single Perspective Camera Maren Bennewitz, Cyrill Stachniss, Wolfram Burgard, and Sven Behnke University of Freiburg, Computer Science Institute, D-791 10 Freiburg, Germany Abstract. The Scale Invariant Feature Transform (SIFT) has become a popular fea- ture extractor for vision-based applications. It has been successfully applied to met- ric localization and mapping using stereo vision and omnivision. In this paper, we present an approach to Monte-Carlo localization using SIFT features for mobile robots equipped with a single perspective camera. First, we acquire a 2D grid map of the environment that contains the visual features. To come up with a compact envi- ronmental model, we appropriately down-sample the number of features in the final map. During localization, we cluster close-by particles and estimate for each cluster the set of potentially visible features in the map using ray-casting. These relevant map features are then compared to the features extracted from the current image. The observation model used to evaluate the individual particles considers the differ- ence between the measured and the expected angle of similar features. In real-world experiments, we demonstrate that our technique is able to accurately track the po- sition of a mobile robot. Moreover, we present experiments illustrating that a robot equipped with a different type of camera can use the same map of SIFT features for localization. 1 Introduction Self-localization is one of the fundamental problems in mobile robotics. The topic was studied intensively in the past. Many approaches exist that use distance infor- mation provided by a proximity sensor for localizing a robot in the environment. However, for some types of robots, proximity sensors are not the appropriate choice because they do not agree with their design principle. Humanoid robots, for example, which are constructed to