摘要
根据晶闸管模块的封装工艺要求,提出利用机器视觉实现晶闸管模块自动粘片系统,解决目前手工封装存在的问题,实现自动粘片过程。自动粘片系统利用CCD智能工业相机拍摄晶圆芯片的图像,采用基于几何特征的模板匹配定位算法,确定有效芯片的位置,然后控制执行机构,把有效芯片从晶圆蓝膜上取下,并自动放置在铜板底座的固定位置。通过在机器视觉实验平台上验证,实现的晶闸管模块自动粘片系统能有效识别晶圆芯片的墨点、崩角、裂纹等缺陷,且定位快速准确,满足实际生产要求。
With the requirements of the thyristor module package process,the thyristor module automatic die-bonding system using machine vision can realize automatic die-bonding process and solve the problem of the manual package.The thyristor module automatic die-bonding system captures the image of the wafer chip through the CCD industrial camera,then determines the valid chip location using a template matching location algorithm based on geometric characteristics.In control of the actuator,the valid chip is removed from the wafer blue film,and it is automatically placed in the fixed position of the copper plate base.Tested and verified on the experimental platform of machine vision,the thyristor module automatic die-bonding system can automatically identify the ink dot,chip die and crack die,etc.It achieves positions accuracy and fast,meets the practical production requirements.
出处
《组合机床与自动化加工技术》
北大核心
2013年第12期101-104,共4页
Modular Machine Tool & Automatic Manufacturing Technique
基金
浙江省自然科学基金(Y1100743)
关键词
机器视觉
晶闸管模块
自动粘片系统
模板匹配
几何特征
machine vision
thyristor module
automatic die-bonding system
template matching
geometric characteristics