In this work we present the Object Labeling Toolkit(OLT), a set of software components publicly available forhelping in the management and labeling of sequential RGB-Dobservations collected by a mobile robot. Such a robot can beequipped with an arbitrary number of RGB-D devices, possiblyintegrating other sensors (e.g. odometry, 2D laser scanners,etc.). OLT first merges the robot observations to generate a3D reconstruction of the scene from which object segmentationand labeling is conveniently accomplished. The annotated labelsare automatically propagated by the toolkit to each RGB-Dobservation in the collected sequence, providing a dense labelingof both intensity and depth images. The resulting objects’ labelscan be exploited for many robotic oriented applications, includinghigh-level decision making, semantic mapping, or contextualobject recognition. Software components within OLT are highlycustomizable and expandable, facilitating the integration ofalready-developed algorithms. To illustrate the toolkit suitability,we describe its application to robotic RGB-D sequences taken ina home environment.
展开▼
