3D vision systems for industrial bin-picking applications

Abstract

This paper relates to a vision-based system for robotic grasping of various objects randomly organized in a bin. In other words, this work deals with the bin-picking problem where the objects that are about to recognize by the vision system are placed in a bin non-oriented, interlocked, jumbled and/or heavily occluding each other. Concerning the industrial applications, demands for a functional technology to cope with the problem are still rising. In this paper we present a short review regarding the bin-picking applications followed by selected results concerned with the system structure description, control system structure, bin-picking methodology and pilot testing projects solutions.