摘要
随着光缆应用需求增加在保证产量前提下针对光缆表面缺陷人工检测效率低、漏检率高的问题本文提出一种基于机器视觉在线高速检测算法方案为了达到高速和精确的要求将在线检测分为实时与准实时的多线程并行处理在实时处理中采集图像经图像增强和去噪的预处理同时运用改进的边缘检测算法进行光缆边界搜索通过统计过程控制法来进行缺陷预判产生可疑图像准实时处理则是对可疑图像进行强、弱边缘检测重点是采用改进的双向微分算法对弱边缘缺陷分割研究最后采用分水岭分割算法实现整体缺陷区域分割通过仿真以及在光缆生产线上实验测试达到最高4m/s的检测速度满足多型号光缆检测能够检测出表皮破损、划痕等常见缺陷为高速线缆生产线自动化提供质量监控手段。
With the increasing demand for optical cable applications, under the premise of ensuring the output, an online high-speed detection algorithm based on machine vision is proposed to solve the problems of low manual detection efficiency and high missed detection rate of optical cable surface defects. For high speed and high accuracy, online detection is divided into real-time and quasi-real-time multi-thread parallel processing. In the real-time processing, the acquired image is preprocessed by image enhancement and denoising. At the same time, the improved edge detection algorithm is used to search the cable boundary, and the defect prediction is performed by the statistical process control method to generate the suspicious image. Quasi-real-time processing is to detect strong and weak edges of suspicious images. The focus is on the improved bidirectional differential algorithm for segmentation of weak edge defects. Finally, the watershed segmentation algorithm is used to realize the segmentation of the whole defect region. Through simulation and experiment testing on the cable production line, the maximum detection speed reaches 4 m/s, which meets the requirements of multi-model cable inspection. The algorithm can detect common defects such as skin damage and scratches, and provide quality monitoring means for high-speed cable production line automation.
作者
吉咸阳
徐宁
杨庚
刘逢清
金仙力
Ji Xianyang;Xu Ning;Yang Geng;Liu Fengqing;Jin Xianli(College of Electronic and Optical Engineering, Nanjing University of Postsand Telecommunications, Nanjing 210023;School of Computer Science, Nanjing University of Posts and Telecommunications, Nanjing 210023)
出处
《高技术通讯》
EI
CAS
北大核心
2019年第8期808-815,共8页
Chinese High Technology Letters
基金
江苏省政策引导类计划(BY2016010-01)
江苏省研究生科研与实践创新计划(KYCX18＿0854)资助项目
关键词
光缆表面
视觉检测
高速
缺陷分割
cable surface
visual inspection
high speed
defect segmentation
作者简介
联系人:吉咸阳,男,1993年生,硕士生;研究方向:数字图像处理;E-mail:18305161877@163.com.
