Recognition and Bin-Picking of Coil Springs by Stereo Vision

Abstract

It is difficult to recognize each of coil springs randomly placed in a pile by conventional machine vision techniques because of their shape characteristics such as a succession of identical shapes and a complicated outline. In this paper, we propose a method of recognition and pose estimation of coil springs using their highlights made by illumination with stereo vision. In this method, we extract and discriminate their highlights. They are grouped into highlight groups in left and right images so that a highlight group includes highlights that belong to a coil spring. Then, we find correspondence between left and right highlight groups to estimate the pose of coil springs by stereo vision. We implemented this method as a bin-picking system with an industrial robot. Bin-picking of coil springs was almost successful on the system. A main reason for picking failure was collisions between the fingers of the hand and the part box, and those between the fingers and other coil springs. Therefore, implementation of collision avoidance would make bin-picking more reliable.