A Machine Vision System of Ball Grid Array Inspection on RT-Linux OS

摘要

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