A robot recognizes a textured shopping item and its pose and grasps it. The system is based on the Textured Object Recognition stack (www.ros.org/wiki/tod) in the Robot Operating System (www.ros.org). The robot uses Oriented BRIEF features to describe the objects, locality-sensitive hashing (LSH) to find features that belong to the object, and the random sample consensus (RANSAC) algorithm to estimate the object's pose.
Link to this comment:
All Comments (0)