基于HALCON的PCB图像拼接方法研究  

Research on Image Mosaic Method based on Halcon

作　　者：郭鹏霞 冯冲 薛严冰[1] 徐丹丹[1] GUO Pengxia;FENG Chong;XUE Yanbing;XU Dandan(School of Electrical and Information Engineering,Dalian Jiaotong University,Dallian 116028,China;School of Information and Communication Engineering,Dalian Minzu University,Dalian 116000,China;Post Doctoral Mobile Station of Electronic Science and Technology,Dalian University of Technology,Dalian 116000,China)

机构地区：[1]大连交通大学电气信息学院,辽宁大连116028 [2]大连民族大学信息与通信工程学院,辽宁大连116000 [3]大连理工大学电子科学与技术博士后流动站,辽宁大连116000

出　　处：《大连交通大学学报》2018年第2期113-117,共5页Journal of Dalian Jiaotong University

基　　金：国家自然科学基金资助项目(61201092;61471080);辽宁省教育厅优秀人才资助项目(LJQ2013047);辽宁省企业博士后项目

摘　　要：以印刷电路板(PCB)贴装生产中的视觉检测为背景研究图像拼接方法.方法包括图像预处理、特征点提取和拼接图像三个步骤,并且在拼接图像步骤中使用了三种技术.利用Halcon软件以两幅PCB图像为目标进行了三个实验测试,并用Do EM方法对实验结果进行评价,评价结果分别为:0.9260(实验1),0.8106(实验2),0.7635(实验3).实验1对应的图像拼接方法能以较少的匹配点数得到最优的匹配结果,可用于PCB贴装质量的视觉检测.The image mosaic method is studied taking the visual detection in the surface mount production of PCB as background.The methods include image preprocessing,feature point extraction and image mosaicing,and the image mosaicing uses three techniques.Three experiments were carried out using Halcon software and targeting two PCB images,the experimental results were evaluated by DoEM method.The results are 0.9260(Experiment 1),0.8106(Experiment 2),and 0.7635(Experiment 3).The image mosaic method corresponding with Experiment 1 can obtain optimal matching results with fewer matching points and can be used for the visual detection of PCB mounting quality.

关 键 词：图像拼接 HALCON 特征点提取 算子 

分 类 号：TN41[电子电信—微电子学与固体电子学] TP391.41[自动化与计算机技术—计算机应用技术]