Sharp corner/edge recognition in domestic environments using RGB-D camera systems

This brief proposes a sharp corner/edge detection and evaluation scheme for RGB-D camera-based domestic applications. The proposed approach first preprocesses the original depth map of objects in domestic environments captured by an RGB-D camera system and then uses the recovered depth map to recognize sharp corners/edges on objects and calculate their sharpness. Furthermore, the approaching speed of a mobile manipulator equipped with the system is derived, and the threat of detected sharp corners/edges is evaluated. To the best of our knowledge, this is the first work that explores sharp corner/edge detection and evaluation in the literature. Simulation results demonstrated that the proposed algorithm can effectively identify dangerous sharp corners/edges in domestic environments