基于邻域去噪及图像差分的碳纤维丝束展宽表面缺陷检测  

Detection of surface defects in carbon fiber tow spreading based on neighborhood denoising and image difference

作　　者：钱小敏 甘学辉[1,2] 刘香玉 卢瑶瑶 张晓晓 巨安奇 马晓建 QIAN Xiaomin;GAN Xuehui;LIU Xiangyu;LU Yaoyao;ZHANG Xiaoxiao;JU Anqi;MA Xiaojian(College of Mechanical Engineering,Donghua University,Shanghai,China;Center for Civil Aviation Composites,Donghua University,Shanghai,China;College of Materials Science and Engineering,Donghua University,Shanghai,China)

机构地区：[1]东华大学机械工程学院,上海 [2]东华大学民用航空复合材料协同创新中心,上海 [3]东华大学材料科学与工程学院,上海

出　　处：《东华大学学报（自然科学版）》2024年第6期120-125,共6页Journal of Donghua University(Natural Science)

基　　金：上海市高性能纤维复合材料协同创新中心开放课题基金(X12812201/021);中国纺织工业联合会应用基础研究项目(J202301)。

摘　　要：碳纤维展宽质量是影响预浸料质量的一个关键因素,人工检测存在误差大、效率低等问题。为提高碳纤维展宽质量的检测效率和准确度,提出基于邻域去噪及图像差分的碳纤维丝束展宽表面缺陷检测方法。针对碳纤维丝束表面光照不均导致的缺陷细节表征问题,采用一种筛选邻域孤点填充的降噪方法。该方法通过筛选出可适应大小的邻域孤点后进行背景填充,实现图像降噪,并将待检测图像与滤波处理后的背景图像进行差分运算,以提取缺陷区域。通过搭建的碳纤维丝束展宽表面缺陷检测系统进行实验,结果表明,该方法可对大丝束碳纤维展宽后表面劈裂、明缝、起毛等缺陷进行较为准确的检测,缺陷准确检出率为96.9%,满足工程实际生产需要。As a key factor affecting the quality of prepreg,the quality of carbon fiber spreading has the problems of large error and low efficiency of manual inspection.In order to improve the detection efficiency and accuracy of carbon fiber spreading quality,a surface defect detection method based on neighborhood denoising and image difference for carbon fiber tow spreading is proposed.Aiming at the problem of defect detail characterization caused by uneven illumination on the surface of carbon fiber tows,a noise reduction method is proposed to screen the neighborhood lone point filling.By filtering out the adaptable size of the neighborhood solitary point after the background filling,the image noise reduction is realized,and the difference operation between the image to be detected and the filtered background image is carried out in order to extract the defective region.The experiment was carried out by the carbon fiber tow broadening surface defect detection system,and the results show that this method can carry out more accurate detection of defects such as splits,open seams,and hairiness on the surface of large carbon fiber tow after spreading,and the accurate detection rate of defects is 96.9%,which meets the needs of actual production of the project.

关 键 词：表面缺陷检测 降噪算法 碳纤维丝束 图像处理 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]