摘要
文中提出了一种基于数学形态学与种子填充相结合的分层提取匹配算法,检测印刷电路板(PCB)中的贴片电阻和电容丢失与立碑、集成电路(IC)芯片丢失与极性错误等缺陷。算法分为PCB图像预处理、分层提取、模板分类匹配3部分。算法的关键在于消除无关信息干扰并通过种子填充算法对各类元器件进行分层提取,然后将不同元器件分别进行基于单链表的模板分类匹配。本算法在检测长度为19 cm、宽度为13.5 cm的PCB样本时,在确保精度为0.1 mm以上的情况下最快检测速度可以达到2 s/块。
A hierarchical extraction classification matching algorithm based on mathematical morphology and seed filling is proposed,to detect the defects of printed circuit board(PCB),include the missing and cocked of chip resistors and capacitors,al?so include the missing and polarity errors of IC chip.The algorithm is divided into three parts:PCB image preprocessing,hierar?chical extraction and template classification.The key to the algorithm is to eliminate irrelevant information interference and the va?rious components of PCB are extracted by seed filling algorithm.And then different components were based on single?linked list of template classification and matching.This algorithm is used to detect PCB samples with a length of 19 cm and a width of 13.5 cm,in the case of ensuring a precision of 0.1 mm or more,the fastest detection speed can be up to 2 sec/block
作者
赵翔宇
周亚同
何峰
王帅
张忠伟
ZHAO Xiang-yu;ZHOU Ya-tong;HE Feng;WANG Shuai;ZHANG Zhong-wei(School of Electronics and Information Engineering,Hebei University of Technology,Tianjin Key Laboratory of Electronic Materials and Devices,Tianjin 300401,China;Beijing An shi zhong dian Technology Co.,Ltd.,Beijing 100871,China)
出处
《仪表技术与传感器》
CSCD
北大核心
2018年第8期84-89,共6页
Instrument Technique and Sensor
基金
中国博士后基金(F2013202254)
河北省自然科学基金(E2016202341)
教育部人文社会科学研究规划基金(15YJA630108)
2017年省级研究生创新资助项目(CXZZSS2017035)
关键词
机器视觉
电路板元器件检测
数学形态学
种子填充
分层提取
模板分类匹配
machine vision
circuit board components detection
mathematical morphology
seed filling
layered extraction
template classification match
