A machine vision system of ball grid array inspection on RT-Linux OS

摘要

In this paper, a machine vision system for ball grid array (BGA) inspection is proposed. The binocular vision system can provide enough geometrical information by an optical approach for monitoring the quality of solder balls. A new optical method for BGA lead coplanarity measurement has been proposed, whose feasibility was verified in theory. This study is focused on algorithms for evaluating the critical characteristics, including the height of solder balls, the size of solder balls, coplanarity, and pitch, which are used to detect solder ball defects. These algorithms are implemented on an RT-Linux operating system. It provides a reference for the development of higher-speed and real-time inspection methods.