基于机器视觉的BGA芯片缺陷检测及其MATLAB实现

Defects inspection and MATLAB realization of BGA chips based on machine vision

文章应用机器视觉技术,采集回流之前的球栅阵列(BGA)芯片图像,对图像进行预处理后,运用点分析方法对图像中各个焊球区域进行标记,通过计算获得焊球数目以及各个焊球位置、大小等特征信息,建立判断标准并依次序对这些信息进行比对,从而判断芯片合格与否,并对不合格芯片判断其缺陷类型;研究采用MATLAB完成图像预处理以及具体的检测程序编写;实验结果证明,此方法可以正确识别缺陷类型,在检测速度以及可靠性方面有很好的保证。

This study captures the images of ball grid array（BGA）chips before they are reflowed by applying machine vision technology. Then the blob analysis algorithm is used to sign each solder ball after the image pretreatment. By calculating the balls＇ features can be got, such as total number, their positions and sizes, etc. The distinctive standard is established and the obtained features are compared with the standard so as to check the regular chips and know the defects of other chips. In the study, MATLAB is used to carry out image pretreatment and detection programming. Experiment results prove that this algorithm can distinguish defects properly and can assure the detecting speed and reliability.