摘要
BGA封装是是一种高集成的封装方式,其焊点缺陷会影响封装器件性能。为克服全局阈值分割、边缘检测方法对BGA焊点缺陷检测错误率较高的问题,本文提出采用Otsu阈值分割、轮廓提取、边界跟踪方法提取焊点轮廓,并用灰度形态学顶帽操作、直方图拉伸、Blob分析提取焊点气泡轮廓。通过分析BGA焊点缺陷类型及特征,提出基于焊点轮廓、气泡轮廓特征参数的焊点检测与分类算法。试验研究表明该算法较全局阈值、Canny算子焊点缺陷检测的准确率高,能够准确的完成焊点缺陷分类。
BGA package is one of the highly integrated packages, and its solder joint defects may significantly affect the performance of packaged parts. To overcome the solder joint’s low extraction accuracy problem of global threshold segmentation method and edge detection method, we present an integrated method of solder joint’s contour extraction based on threshold segmentation, edge detection and edge tracking, as well as an integrated method of solder joint’s void extraction based on top-hat transformation, histogram stretching and Blob analysis. Basing on the analysis of BGA solder joint defects’ types and characteristics, defect detection process is proposed under Joints characteristic parameters which are obtained by Joints contours and void contours. The comparison experiment shows that this method possesses higher accuracy in solder joint defection than global threshold segmentation and Canny algorithm, and improves joint defection classification accurately.
出处
《电子设计工程》
2014年第12期164-166,170,共4页
Electronic Design Engineering
基金
广东省科技计划项目(2012B011300016)
关键词
BGA封装
焊点缺陷
缺陷检测
缺陷分类
BGA package
solder joint defects
defect detection
defect classification
