引线键合匹配定位算法研究  

Study of Grey Correlation Matching Location Algorithm for Wire Bonding

作　　者：孔凡芝[1,2] 王以忠[2] 李君兰[3] 赵兴玉[3] 张大卫[3] 

机构地区：[1]哈尔滨工程大学信息与通信工程学院,黑龙江哈尔滨150000 [2]天津科技大学电子信息与自动化学院,天津300222 [3]天津大学机械工程学院,天津300222

出　　处：《微计算机信息》2008年第30期232-233,228,共3页Control & Automation

基　　金：国家自然科学基金资助项目(50505032)

摘　　要：为提高全自动引线键合机视觉定位系统的检测速度和定位精度,在对定位系统进行了深入分析的基础上,分别对系统的示教和键合过程进行了参数优化。为了降低归一化灰度相关匹配算法的计算量,将算法的有效搜索范围的设定与眼点相联系,提高了算法的效率。最后,利用MATLAB软件编程实现了算法。仿真实验结果表明,针对678×576的引线框架实际图像,该算法定位的平均误差可控制在0.4像素以内,计算时间约为300ms,基本能满足自动键合机视觉系统的要求。For the detecting speed and accuracy of pattern recognition system（PRS） have much effect on the performance of the automatic wire bonder, much work was done in this paper for optimal design of the teaching and bonding process. A new method based on associating the searching area with the selected eye point was presented in this paper, which could decrease the computation com- plexity of the normalized cross correlation greatly, and improve the efficiency of the location algorithm. The detection and location system was implemented at last using MATLAB language. The results of the simulation experiments show that the algorithms can im- prove the speed and accuracy of the PRS and satisfy the requirement of automatic wire bonder.

关 键 词：引线键合 模板匹配 归一化互相关 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]