摘要
由于当前绝缘体上硅(SOI)压阻传感器芯片的封装质量仍依赖人工检测,本文提出了一种自动实现该项检测的视觉检测方法。分析了压阻传感器的工作原理,研究了芯片定位精度和结合面质量对传感器性能的影响。以传感器性能和质量为导向,提出了一种以中心定位偏差和键合面结合度为检测点的封装结合面检测方法。该方法通过对Hough圆检测效果和实际图像的分析完成定位精度的检测;基于对传感器质量影响因素的分析和气泡面积的统计实现结合面质量的检测。在传感器实际制造封装过程中对该视觉检测算法进行了实验验证。结果表明:该方法能识别的结合面上的最小气泡直径为6μm;玻璃内孔半径检测误差约为0.015mm.。本文提出的基于视觉检测的方法基本满足了压阻传感器封装对结合面检测的要求,有助于实现封装质量的自动化检测。
As the packaging quality of a SOI（Silicon On Insulator）piezoresistive sensor is still dependent on manual detection,this paper proposes a method to implement its automatic visual inspection.The principle of piezoresistive sensor was analyzed,and the influence of the positioning accuracy of chips and the quality of bonding joint surface on sensor performance was investigated.By taking sensor performance and quality for a research target,a method to inspect the center deviation and bonding area was proposed for joint surface detection.Through analyzing the Hough circle detection and the actual image,the method implements the positioning accuracy detection.By analysis of the effect factors on sensor quality and statistics of the bubble area,the quality of bonding joint surface was measured.The proposed visual inspection method was verified in the practical packaging processing for sensor fabrication.Experimental results indicate that the method identifies the bubble with a diameter over 6μm and the error of glass inner hole radius is about 0.015 mm.The method proposed satisfies the requirement of piezoresistive sensor package joint surface inspection by detecting center deviation and contributes to the automated inspection of packaging quality.
出处
《光学精密工程》
EI
CAS
CSCD
北大核心
2016年第6期1382-1388,共7页
Optics and Precision Engineering
基金
国家863高技术研究发展计划资助项目(No.2012AA041201)
高等学校博士学科点专项科研基金资助项目(No.20133201130003)
