Design and development of a vision-based micro-assembly system

Abstract

Cost- and time-effective automated assembly of micro-sized objects within small tolerances is still a challenge to the present-day manufacturing industry. In this article, the details of the development of a low-cost, two-dimensional system capable of rapid micro-assembly are presented. The use of a basic webcamera without any external optics makes the system light and compact. In addition, an innovative thresholding technique for binarization and a novel algorithm for fast autofocusing with defocused images have also been discussed. Peg-in-hole insertion tasks using micro-objects of 100 μm have been successfully demonstrated using the system. The system has an accuracy of about 20 μm along the X- and Y-axes and a repeatability of less than 4 μm in performing peg-in-hole tasks.