Robotic Vision for Bin-Picking Applications of Various Objects

Abstract

In this paper we present a new experimental work and results regarding the bin-picking problem that has been still a challenging topic in robotic so far. Although the number of potential applications for bin-picking is growing in industry representing a huge market potential, in many situations it is done manually. We present a functional solution for the bin-picking problem that has been verified on a variety of objects. Initially, there is presented a pilot project concerned with a picking of standard nuts from a bin. Further, based on interests from automotive industry, we dealt with objects complex in shape (sheet-metal parts). However, the solution is not universal; the whole bin-picking system has to be adjusted according to various object specifications. The bin-picking system, presented in this work, for objects manipulation (grasping and placement) is mainly concerned with an industrial robot, 3D vision system mounted on a movable linear axis frame and a proper end effector - standard gripper or vacuum gripper depending on the variety of objects to manipulate.