基于图像拼接的PCB基板质量检测算法  

PCB SUBSTRATE QUALITY INSPECTION ALGORITHM BASED ON IMAGE STITCHING

作　　者：崔家礼[1] 王鹏[1] Cui Jiali;Wang Peng(College of Information Science and Technology,North China University of Technology,Beijing 100144,China)

机构地区：[1]北方工业大学信息学院,北京100144

出　　处：《计算机应用与软件》2020年第6期191-198,共8页Computer Applications and Software

基　　金：国家重点研发计划项目(2017YFB0802300)。

摘　　要：印刷电路板(PCB)基板是PCB在制作过程中的重要载体,其合格与否直接影响产品的质量和性能。由于基板尺寸较大且孔洞繁多、密集,PCB基板质量检验成为了一件繁重而复杂的工作。针对此问题,提出一种基于图像拼接的基板质量自动检测方案,解决了大视野和高分辨率的矛盾。算法包括在子图上进行图像增强和目标分割;通过对应关键点进行图像拼接,获得高分辨的大视野图像;提出评估判据实现基板质量自动检测。实验结果表明,该算法可以准确地提取基板相关数据,检出错误并给出有关质量检测误差评估。Printed circuit board(PCB)substrate is an important carrier of PCB in the manufacturing process.Its qualification directly affects the quality and performance of products.Because of the large size and the numerous,dense holes,the quality inspection of PCB substrate has become a heavy and complicated work.To solve this problem,we propose an automatic inspection scheme of substrate quality based on image stitching,which solves the contradiction between the large field of view and high image resolution.The algorithm includes the following steps:image enhancement and target segmentation were performed on the sub-images;through the image stitching of corresponding key points,the high-resolution wide-field images were obtain;the evaluation criteria was proposed to achieve automatic substrate quality inspection.The experimental results show that this algorithm can accurately extract the data related to the substrate,detect the errors and give the evaluation of substrate quality.

关 键 词：PCB基板 质量检测 图像拼接 形状识别 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]