摘要
目的对球栅阵列封装(Ball Grid Array,BGA)芯片中由于焊接过程中产生的气体未能及时逸出所致的空洞缺陷进行3D检测,以降低缺陷检测难度和提高准确率,为实现BGA产品的流水线检测奠定基础.方法利用X射线三维显微镜对BGA芯片进行扫描与重建得到3D模型,将模型等距切片后转为灰度图像,根据3D可视化结果和灰度直方图选择固定阈值进行全局阈值分割,将分割得到的二值图进行连通区域标记并计算各区域面积,最后采用积分法求取焊球和空洞体积并计算空洞率.结果与2D检测方法对比,该方法可以有效去除图像中的多元器件重叠的不利因素,可直接观察空洞缺陷的大小及位置;在BGA切片图像中标记分割阈值的等值面并测量焊球和空洞的直径,将测量结果与DR图像中的测量值对比,最大误差为3.726μm,表明该方法可以准确地分割焊球及空洞特征.结论该方法可以有效地检测出BGA中的空洞缺陷,并准确地计算出焊球和空洞体积及基于体积的空洞率.
In order to reduce the difficulty and improve the accuracy of defect detection,this paper studies the 3D detection method of void defects in Ball Grid Array(BGA)chips due to the gas failed to escape in time during welding.This method lays a foundation for the production line inspection of BGA products.The BGA chip is scanned and reconstructed by X-ray three-dimensional microscope to get a 3D model.The model is divided into several two-dimensional images at equal distances and then transformed into gray images.According to the results of 3D visualization and gray histogram,fixed thresholds are selected to segment the image globally.The binary image obtained by threshold segmentation is labeled with connected regions and the area of each region is calculated.Finally,the volume of welding balls and voids is calculated by integral method and the void rate is calculated.Compared with the 2D detection method,this method can effectively remove the disadvantageous factors of overlapping multiple elements in the image,and can directly observe the size and location of void defects.The equivalent surface of segmentation threshold is marked in BGA slice image,and the diameters of welding balls and voids are measured.The maximum error is 3.726 um by comparing the measured results with those in DR image.Experiments show that this method can accurately segment the characteristics of welding balls and voids,and accurately calculate the volume and void rate of welding balls and voids.This method can effectively detect void defects in BGA,and accurately calculate the volume of welding balls and voids and void rate based on volume.
作者
须颖
刘永斌
安冬
邵萌
XU Ying;LIU Yongbin;AN Dong;SHAO Meng(School of Mechanical Engineering,Shenyang Jianzhu University,Shenyang,China,110168;Micro-Nano Detection and Motion Control Institute,Shenyang Jianzhu University,Shenyang,China,110168;Sanying Precision Instruments Co.Ltd.,Tianjin,China,300399)
出处
《沈阳建筑大学学报（自然科学版）》
CAS
CSCD
北大核心
2020年第1期155-162,共8页
Journal of Shenyang Jianzhu University：Natural Science
基金
国家重点研发计划项目(2017YFC0703903)
辽宁省科学技术基金项目(20180550002)
辽宁省高等学校基本科研项目(LJZ2017035)
辽宁省重点研发计划项目(2017225016)
