A rotation-invariant and non-referential approach for ball grid array (BGA) substrate conducting path inspection

The aim of this paper is to locate and classify boundary defects such as op en and short circuits, mousebites, and spurs on ball grid array (BGA) subst rate conducting paths using machine vision. Boundary defects are detected b y a boundary-based corner detection method using covariance matrix eigen-va lues. Detected defects are then classified by discrimination rules derived from variation patterns of eigenvalues and the geometrical shape of each de fect type. Real BGA substrates with both synthetic and real boundary defect s are used as test samples to evaluate the performance of the proposed meth od. Experimental results show that the proposed method achieves 100% correc t identification for BGA substrate boundary defects under a sufficient imag e resolution. The proposed method is invariant with respect to the orientat ion of the BGA substrates, and it does not require prestored templates for matching. This method is suitable for various types of BGA substrate in sma ll-batch production because precise positioning of BGA substrates and the p restored templates are not necessary.