基于机器视觉的异型卷烟分拣线误差自动控制方法  

Automatic error control method for sorting line of irregular cigarette based on machine vision

作　　者：吴德祥[1] 区剑平 肖波[3] 沈宇航[4] Wu Dexiang;Ou Jianping;Xiao Bo;Shen Yuhang(Lanzhou Cigarette Factory,Gansu Tobacco Industrial Co.,Ltd.,Gansu Lanzhou,730050,China;Hainan Hongta Cigarette Co.,Ltd.,Hainan Haikou,571137,China;Changchun Cigarette Factory,Jilin Tobacco Industry Co.,Ltd.,Jilin Changchun,130033,China;China Tobacco Zhejiang Industrial Co.,Ltd.,Zhejiang Hangzhou,310009,China)

机构地区：[1]甘肃烟草工业有限责任公司兰州卷烟厂,甘肃兰州730050 [2]海南红塔卷烟有限责任公司,海南海口571137 [3]吉林烟草工业有限责任公司长春卷烟厂,吉林长春130033 [4]浙江中烟工业有限责任公司,浙江杭州310009

出　　处：《机械设计与制造工程》2024年第6期39-42,共4页Machine Design and Manufacturing Engineering

摘　　要：为有效解决异型卷烟分拣困难的问题,提出一种基于机器视觉的异型卷烟分拣线误差自动控制方法。通过工业相机采集异型卷烟图像,采用数学形态学运算增加异型卷烟图像对比度,根据直方图对异型卷烟图像整体进行均衡化处理,有效保留图像信息,完成图像增强。分割目标图像,使用Gabor滤波器提取目标图像和样本图像特征,获取异型卷烟分拣线误差特征,同时标记连通域区域,根据提取的异常特征,进行分拣差错控制。实验测试结果表明,使用所提方法分拣线控制误差在0.4 mm内,控制时间在1.1 s内,可以更好地完成异型卷烟分拣。To effectively solve the problem of difficult sorting of irregular cigarettes,a machine vision based automatic error control method for irregular cigarette sorting lines is proposed.By using industrial cameras to capture images of irregular cigarettes,mathematical morphological operations are used to increase the contrast of the irregular cigarette images.Based on histograms,the overall image of irregular cigarettes is balanced to effectively preserve image information and complete image enhancement.By segmenting the target image and using Gabor filters to extract the features of the target image and sample image,it obtains the error features of the irregular cigarette sorting line,and marks the connected domain areas.Based on the extracted abnormal features,sorting error control is carried out.The experimental test results show that using the proposed method to control the sorting line error within 0.4 mm and the control time within 1.1 s can better complete the sorting of irregular cigarettes.

关 键 词：机器视觉 异型卷烟 分拣线误差 自动控制 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]