Automatic soldering system based on computer vision

Abstract

Due to the fine pitch of today's micro-chips, accurate alignment is vital to acquire high quality joints with less detrimental effects on other components. We developed an integrated soldering system which combines an X-Y stage, a computer vision and a laser head. Based on the analysis of experimental data, we find that lens distortions and assembly angular errors of the X-Y stage play a distinct role on the positioning errors. To improve the positioning accuracy of this system, we proposed an algorithm based on straight line method to reduce image distortions, and a compensation algorithm on the non-perpendicular X-Y positioning stage. Finally, experimental results and comparative data are presented to evaluate the performance of these algorithms. © 2009 Springer-Verlag Berlin Heidelberg.