基于Faster rcnn的棉麻纱混纺比自动检测
2020-07-07 11:36:49   来源:   评论:0 点击:

为了解决人工检测棉麻纱混纺比存在主观性强且要求检测人员经验丰富的问题,提出利用Faster rcnn目标检测网络进行棉麻纱混纺比自动检测;通过制作数据集,训练得到Faster rcnn模型,并对模型
肖登辉a,李立轻a,b,汪  军a,b
(东华大学 a.纺织学院;b.纺织面料技术教育部重点实验室:上海  201620)
 
摘要:为了解决人工检测棉麻纱混纺比存在主观性强且要求检测人员经验丰富的问题,提出利用Faster rcnn目标检测网络进行棉麻纱混纺比自动检测;通过制作数据集,训练得到Faster rcnn模型,并对模型进行评估;通过Faster rcnn模型在棉麻混纺纱的试验,在测试集上的平均精度的均值(mAP)为0.905,在实验中验证检测的棉麻混纺比与实际标准值的误差基本吻合。指出:利用Faster rcnn网络模型作为自动化检测棉麻纤维核心算法具有可行性和可靠性;该检测方法精度高,在制样以及图片采集过程中存在耗时、耗力的问题,需持续改进。
关键词:Faster rcnn;目标检测;棉纤维;麻纤维;混纺比;图像;模型
中图分类号:TS131.9    文献标志码:A    文章编号:1001-9634(2020)03-0001-04
 
收稿日期:2019-11-07
基金项目:国家自然科学基金项目(61271006)
作者简介:肖登辉(1995—),男,浙江宁波人,硕士研究生,主要从事织物图像的模式识别和机器学习等方面的研究。
 
Automatic Detection of Blending Ratio of Cotton and Hemp Blended Yarn Based on Faster Rcnn
XIAO Denghuia,LI Liqinga,b,WANG Juna,b
(Donghua University a.College of Textile;b.Key Laboratory of Textile Science & Technology Ministry of Education:Shanghai 201620,China)
 
Abstract:In order to solve the problem of manual testing the blending ratio of cotton and hemp blended yarn,which has strong subjectivity and requires rich experience of detection personnel,Faster rcnn target detection network is proposed for automatic detection of the blending ratio of cotton and hemp blended yarn.By making data sets,Faster rcnn model is received by test and evaluated.The average precision (mAP) on the test sets is 0.905 in the test of cotton and hemp blended yarn using Faster rcnn model,and the error between the detected cotton and hemp blended ratio and the actual standard value is basically consistent in the experiment.It is pointed out that using Faster rcnn network model as the core algorithm for automatic detection of cotton and hemp fibers is feasible and reliable.The detection method has high precision,but there are problems of time consuming and force consuming in the process of sample making and picture collection,which need continuous improvement.
Key Words:Faster rcnn;target detection;cotton fiber;hemp fiber;blending ratio;image;model
 
参考文献:
[1] 贾立锋,饶高昶,孟会娟.基于着色法的棉与苎麻混纺含量测定[J].纺织学报,2011,32(7):28-34.
[2] 杨元,李永贵,丁志强.棉/麻纤维混纺纱的定量分析方法探讨[J].上海纺织科技,2009(7):48-51.
[3] 高山.溶解—吸光法测定亚麻/棉混纺比[J].山东纺织科技,2001,42(2):46-48.
[4] 俞凌云,温演庆,朱谱新,等.三原色染色法测定棉麻纤维混纺比初探[J].中国测试,2013,39(5):62-64.
[5] 杨欣卉.近红外光谱在纤维成分含量定量分析中的应用研究进展[J].现代纺织技术,2017(2):37-42.
[6] Cortes C,Vapnik V.Support Vector Networks[J].Machine Learning,1995,20(5):273-297.
[7] REN S,HE K,GIRSHICK R,et al.Faster R-CNN:Towards Real-time Object Detection with Region Proposal Networks[J].IEEE Transaction on Pattern Analysis and Machine Intelligence,2017(6):1137-1149.
[8] REDMON J,DIVVALA S,GIRSHICK R,et al.You Only Look Once:Unified,Real-time Object Detection[C]//IEEE Conference on Computer Vision and Pattern Recognition,2016:779-788.
[9] LIU W,ANGUELOV D,ERHAN D,et al.SSD:Single Shot Multibox Detector[C]//European Conference on Computer Vision,2016:21-37.
[10] Tian Z,Shen C,Chen H,et al.FCOS:Fully Convolutional One-Stage Object Detection[OL].[2019-08-20].https://arxiv.org/pdf/1904.01355v2.pdf.
[11] 于伟东.纺织材料学[M].北京:中国纺织出版社,2006.
[12] 崔美琪,徐广标.木棉/棉混纺纱混纺比定量分析方法研究[J].上海纺织科技,2015(9):70-72.
 

相关热词搜索:

上一篇:碳纤维角联织机开口系统张力网络化控制研究
下一篇:单纤维测试仪在纺纱生产中的应用

分享到: 收藏