Efficient techniques for BGA solder joint identification in low resolution X-ray images

Abstract

The benefits of X-ray inspection are broad in scope due to the ability of X-rays to see through packages including: encapsulation, heat sinks, and metallic shielding to reveal obscured connections and identify potential quality nondestructive issues. X-ray inspection is particularly beneficial to applications that involve advanced packaging technologies such as Ball Grid Array (BGA). This paper proposes a method that relies on an image processing and computer vision technique that detects each ball of the X-ray image of BGA chip and flags the suspect balls for further evaluation and rework. It makes use of the basic geometric distinction between a circle (perfect solder ball) and an ellipse (balls with reshaped solder joint). Thus making it easier for the casual user to identify the suspect areas (solder joints) for further evaluation.