Signature search method for 3-D pose refinement with range data

摘要

In many applications in robotics, the geometry of the tasknenvironment is uncertain and so the pose of the target object may benknown only approximately. For the object to be grasped successfully itsnactual pose must be determined using some form of vision sensing. Thisnpaper presents a novel method of processing 3-D range data for posenrefinement. Given the model of the object and its approximate pose, thenalgorithm adjusts the pose of the object model for a best fit to thenmeasured 3-D data. The main attributes of this algorithm are that itsnperformance is largely unaffected by background clutter including somentolerance to occlusion and that it imposes no restrictions on thensurface shape or representation scheme used for the model