面向QFN封装视觉定位系统  

The Vision Positioning System for QFN Packaging

作　　者：高金龙 胡海燕[1] 高阳 杨云龙 GAO Jinlong;HU Haiyan;GAO Yang;YANG Yunlong(Mechanical and Electrical Engineering,Soochow University Mechanical and Electrical Engineering,Suzhou 215123,China;Jiangsu Jing Chuang advanced electronic technology Co.,Ltd.,Suzhou 215123,China)

机构地区：[1]苏州大学机电工程学院、江苏省先进机器人技术重点实验室机器人与微系统研究中心,江苏苏州215123 [2]江苏京创先进电子科技有限公司,江苏苏州215123

出　　处：《电子工业专用设备》2019年第6期26-30,44,共6页Equipment for Electronic Products Manufacturing

摘　　要：以全自动划片机针对QFN封装图形的自动对准为研究对象,为了适应划片机生产效率和工艺的要求,提出了一种快速的切割位置检测定位方法;并根据QFN的表面特征,利用图像相似度匹配算法获取图像位置,结合QFN封装图形的分布特征,制定相应的搜索策略可以快速的实现切割位置的找寻;同时,为了保证对大尺寸芯片翘曲造成的切割道不是直线的情况,采用最小二乘法,对切割道的图形间隔采样对准,根据位置进行最小二乘法拟合,不仅可以减小切割造成的位置偏差,同时也提高了芯片质量成品率。这种QFN快速对准定位检测方法对芯片检测设备软件系统开发与应用具有重要的参考价值。Take automatic image matching in the QFN(Quad Flat No-lead Package)of full-automatic dicing saw as research object in this paper. In order to meet the requirements of the production efficiency and technical process of dicing saw,a rapid cutting position detection and location method is proposed. According to the surface features of QFN,the image location is obtained by using the image similarity matching algorithm. In combination with the distribution characteristics of QFN package graphics,the corresponding search strategy is formulated to quickly find the cutting location.Meanwhile,adopting least square method to align sampling image interval of cutting street in case that big-size warping chip creates non-straight cutting street. Fit least square method based on position not only can reduce position deviation caused by cutting but also improve qualified chip yield. The QFN alignment and rapid positioning detection method have important reference value for the development and application of chip detection equipment software system.

关 键 词：全自动划片机 相似度 搜索策略 最小二乘法 

分 类 号：TN305.94[电子电信—物理电子学]