An efficient method of occluded solder ball segmentation for automated BGA void defect inspection using X-ray images

摘要

One challenge concerning the reliability of ball grid array (BGA) packages assembly is to detect the void defects occurring inside solder balls. Additionally, for use in mass production, an automated inspection system has increasingly become an attractive solution. In practice, the first procedure of this system is required to segment the individual solder balls from the background. Here, we have proposed the efficient method to analyze the X-ray images obtained from arbitrary rotating angles of a printed circuit board (PCB). Specifically, we have succeeded in dealing with the general pattern of solder balls arrangement by applying the Delaunay triangulation technique and verifying the unit cell of periodic lattice. The proposed method is robust against occluded balls caused by some interference and outperforms the conventional method in term of yielding higher accuracy of solder ball segmentation and lower processing time. These conclusions were confirmed by our experiments.