基于机器视觉的PCB图像焊点数据自动识别技术  

Automatic recognition technology of PCB image solder joint data based on machine vision

作　　者：张玉登 刘新妹[1] 殷俊龄[1] Zhang Yudeng;Liu Xinmei;Yin Junling(State Key Laboratory of Electr onic Testing Technology,North University of China,Taiyuan 030051,China)

机构地区：[1]中北大学电子测试技术国家重点实验室,太原030051

出　　处：《国外电子测量技术》2020年第11期12-16,共5页Foreign Electronic Measurement Technology

基　　金：山西省回国留学人员科研项目(2017-090);山西省重点研发项目(201903D121058)资助。

摘　　要：针对传统的自动测试系统在被测电路无法提供原始计算机辅助设计(computer aided design,CAD)文件的情况下,需要以人工手动方式输入电路板焊盘坐标,工作量大且效率低,提出一种基于机器视觉的印制电路板(printed circuit board,PCB)图像焊点坐标自动提取的方法。通过最小二乘法法对焊点定位,完成连通区域的标记以及像素中心参数的标定,实现PCB图像焊点的识别及坐标提取。实验分析表明,通过机器视觉提取的焊点坐标与使用计算机软件导出的真实坐标对比,证明该方法提取的焊点坐标最大误差低于0.1mm,为自动化测试系统提供了坐标位置支持。In view of the traditional automatic test system,when the circuit under test can not provide the original CAD design file,it needs to input the PCB pad coordinates manually,which is heavy workload and low efficiency.presents a method for automatic extraction of solder joint coordinates in PCB image based on machine vision.The solder joint is located by interpolation roundness detection method,the connected area is marked and the parameter of pixel center is calibrated.eventually,the solder joint recognition and coordinate extraction of PCB image are realized.The experimental results show that the maximum error of solder joint coordinates extracted by this method is less than 0.1 mm,compared with the real coordinates derived by PCB softw are I.t will provide coordinate support for automatic test system.

关 键 词：机器视觉 焊点 最小二乘法 连通区域标记 标定 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]