基于机器视觉的Mini LED背光板PCB焊盘检测系统  

Inspection System for PCB Pad of Mini LED Backlight Board Based on Machine Vision

作　　者：罗文 邹湘军[1,2] 陈建明 梁添贵 丁惠英 倪沁心 LUO Wen;ZOU Xiangjun;CHEN Jianming;LIANG Tiangui;DING Huiying;NI Qinxin(School of Engineering,South China Agricultural University,Guangzhou 510642,China;Guangdong Huaheng Intelligent Technology Co.,Ltd.,Dongguan 523575,China)

机构地区：[1]华南农业大学工程学院,广东广州510642 [2]广东华恒智能科技有限公司,广东东莞523575

出　　处：《自动化与信息工程》2022年第6期20-26,40,共8页Automation & Information Engineering

基　　金：广东省东莞大专项(20211800400092);广东省科技厅佛山大专项(广东省科技计划项目)(2120001008424)。

摘　　要：针对Mini LED背光板PCB焊盘检测精度和效率低的问题,提出一种基于机器视觉的Mini LED背光板PCB焊盘检测系统。首先,对采集的Mini LED背光板PCB焊盘图像进行预处理;然后,通过阈值分割法进行PCB焊盘粗分割,采用模糊C均值聚类算法进行PCB焊盘细分割;接着,利用Canny算子进行像素精度边缘提取;最后,采用基于Zernike矩的亚像素边缘检测方法提高边缘检测精度。实验结果表明,该系统具有较高的检测精度和效率,PCB焊盘尺寸测量误差在±8μm以内,位置测量误差在12μm以内。Aiming at the problem of low detection accuracy and efficiency of Mini LED backlight PCB pad,a machine vision based Mini LED backlight PCB pad detection system was proposed.First,preprocess the acquired PCB pad image of Mini LED backlight board;Then,the threshold segmentation method is used to roughly segment PCB pads,and the fuzzy C-means clustering algorithm is used to finely segment PCB pads;Secondly,Canny operator is used to extract pixel precision edge;Finally,the sub-pixel edge detection method based on Zernike moment is adopted to improve the edge detection accuracy.The experimental results show that the system has high detection accuracy and efficiency,and the PCB pad size measurement error is within±8μm,the position measurement error is within 12μm.

关 键 词：机器视觉 Mini LED背光板 PCB焊盘 图像处理 边缘提取 亚像素边缘检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]