NimbRo Logistics -- Project KittingBot
This work addresses a specific challenge in robotics and computer vision for logistics applications, but it appears incremental as it builds on existing methods for pose estimation.
The paper tackles the problem of object pose estimation from point clouds by exploiting object symmetry and static camera images, achieving detection of nuts in a cluttered tabletop scene with screws, washers, and placeholders.
Recovering the pose of an object from mere point clouds is often hindered by the lack of the information that they provide. In this lab, we address this problem by proposing a method that exploits the symmetry of objects as well as using pictures taken from a static camera of the same scene. We apply this approach to detects nuts in a table top scene that includes screws, nuts, washers and several placeholders for grasp planning.