杨连贺,张超

• 1:天津工业大学计算机科学与技术学院

摘要(Abstract)：

为了准确而高效地识别出布匹各种疵点的种类，采用改进的EfficientDet算法进行布匹疵点识别。首先采取改进的Ostu阈值分割算法进行特征边缘的检测，采用非极大值抑制方法对边缘进行筛选，确定候选区域；然后采用筛选器对候选区域的疵点进行识别和分类，其中筛选器采用改进的EfficientDet算法。改进的EfficientDet算法与其他优秀的目标检测算法以及原算法进行了比较。结果表明，改进的Ostu分割算法相较于传统算法不仅可以在更多的布匹图像中更准确地识别疵点区域，而且抑制了假边缘现象；该模型规模是几种算法中最小的，识别准确率达到94%，高于目前最优算法4个百分点。

关键词(KeyWords)： 疵点检测;迁移学习;Ostu;目标检测;EfficientDet

Abstract:

Keywords:

基金项目(Foundation): 国家自然科学基金资助项目（61771340）;; 天津市自然科学基金资助项目（18JCYBJC15300）

作者(Author): 杨连贺,张超

参考文献(References)：

• [1]刘茁梅，李鹏飞，景军锋.基于稀疏表示的印花织物疵点检测[J].西安工程大学学报，2018,32(2):197-202.LIU Z M,LI P F,JING J F. Patterned fabric defect detection based on sparse representation[J]. Journal of Xi'an Polytechnic University,2018,32(2):197-202(in Chinese).
• [2]张宏伟，张凌婕，李鹏飞.基于深度卷积神经网络的织物花型分类[J].纺织高校基础科学学报，2017,30(2):261-265,271.ZHANG H W,ZHANG L J,LI P F. Fabric pattern classification based on depth convolution neural network[J]. Basic Sciences Journal of Textile Universities,2017,30(2):261-265,271(in Chinese).
• [3]贺智明，彭亚楠.基于深度学习的织物疵点检测研究进展[J].毛纺科技，2019,47(8):83-88.HE Z M,PENG Y N. Progress in fabric defect detection based on deep learning[J]. Wool Textile Journal,2019,47(8):83-88(in Chinese).
• [4]王春妍，沈丹峰，杨国仲，等.应用Prewitt算子的织物疵点检测改进算法[J].纺织科技进展，2020(3):42-46.WANG C Y,SHEN D F,YANG G Z,et al. An improved fabric detect detection algorithm based on prewitt operator[J].Progress in Textile Science&Technology,2020(3):42-46(in Chinese).
• [5] LI Q,PENG Q,CHEN J Z,et al. Improving image classification accuracy with ELM and CSIFT[J]. Computing in Science&Engineering,2019,21(5):26-34.
• [6] JING J F,ZHANG H H,WANG J,et al. Fabric defect detection using Gabor filters and defect classification based on LBP and Tamura method[J]. The Journal of the Textile Institute,2013,104(1):18-27.
• [7]闫伟伟，孙宏昌，张廷，等.改进Canny算法在编织布疵点边缘检测中的应用[J].天津职业技术师范大学学报，2021,31(1):37-41.YAN W W,SUN H C,ZHANG T,et al. Application of improved Canny operator in edge detection of woven fabric defects[J]. Journal of Tianjin University of Technology and Education,2021,31(1):37-41(in Chinese).
• [8]龙涵彬，狄岚，梁久祯.基于畸变校正与视觉显著特征的纺织品瑕疵检测[J].模式识别与人工智能，2020,33(12):1122-1134.LONG H B,DI L,LIANG J Z. Fabric defect detection based on distortion correction and visual salient features[J]. Pattern Recognition and Artificial Intelligence,2020,33(12):1122-1134(in Chinese).
• [9]曹振军，景军锋，苏泽斌，等.基于树莓派的深度学习色织物疵点检测研究[J].棉纺织技术，2019,47(1):11-15.CAO Z J,JING J F,SU Z B,et al. Study on the defect detection of yarn-dyed fabric with deep learning based on raspberry pie[J]. Cotton Textile Technology,2019,47(1):11-15(in Chinese).
• [10]罗维平，徐洋，陈永恒，等.基于迁移学习和改进ResNet50网络的织物疵点检测算法[J].毛纺科技，2021,49(2):71-78.LUO W P,XU Y,CHEN Y H,et al. Fabric defect detection algorithm based on migration learning and improved ResNet50network[J]. Wool Textile Journal,2021,49(2):71-78(in Chinese).
• [11]李敏，杨珊，何儒汉，等.结合视觉显著性和卷积神经网络的提花织物疵点检测技术[J].现代纺织技术，2021(6):62-66.LI M,YANG S,HE R H,et al. Jacquard fabric defect detection combining visual saliency and convolutional neural network[J].Advanced Textile Technology,2021(6):62-66(in Chinese).
• [12]李东洁，郭帅，杨柳.基于改进图像阈值分割算法的纱线疵点检测[J].纺织学报，2021,42(3):82-88.LI D J,GUO S,YANG L. Yarn defect detection based on improved image threshold segmentation algorithm[J]. Journal of Textile Research,2021,42(3):82-88(in Chinese).
• [13]毕秀丽，邱雨檬，肖斌，等.基于统计特征的图像直方图均衡化检测方法[J].计算机学报，2021,44(2):292-303.BI X L,QIU Y M,XIAO B,et al. Histogram equalization detection based on statistical features in digital image[J]. Chinese Journal of Computers,2021,44(2):292-303(in Chinese).
• [14]刘露露，李波，何征，等.基于FS-YOLOv3及多尺度特征融合的棉布瑕疵检测[J].中南民族大学学报（自然科学版），2021,40(1):95-101.LIU L L,LI B,HE Z,et al. Cotton defect detection based on FS-YOLOv3 and multiscale feature fusion[J]. Journal of SouthCentral University for Nationalities(Natural Science Edition),2021,40(1):95-101(in Chinese).
• [15]袁小翠，黄志开，马永力，等. Otsu阈值分割法特点及其应用分析[J].南昌工程学院学报，2019,38(1):85-90,97.YUAN X C,HUANG Z K,MA Y L,et al. Analysis of characteristics and application of Otsu threshold method[J]. Journal of Nanchang Institute of Technology,2019,38(1):85-90,97(in Chinese).
• [16] TAN M X,PANG R M,LE Q V. EfficientDet:Scalable and efficient object detection[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR). Seattle,WA,USA:IEEE,2020:10778-10787.
• [17]陈志韬，殷恺铭，张洋，等.基于EfficientDet的安全帽佩戴检测研究[J].信息技术与标准化，2021(S1):19-23,29.CHEN Z T,YIN K M,ZHANG Y,et al. The research of safety helmet wearing detection based on EfficientDet[J]. Information Technology&Standardization,2021(S1):19-23,29(in Chinese).
• [18] TAN M X,LE Q V. EfficientNet:Rethinking model scaling for convolutional neural networks[EB/OL]. 2019. DOI:10.48550/arXiv. 1905.11946.
• [19] BAO J M,CHEN D,WEN F,et al. CVAE-GAN:Fine-grained image generation through asymmetric training[C]//2017IEEE International Conference on Computer Vision. Venice,Italy:IEEE,2017:2764-2773.
• [20]奚祥品，陈筱，朱向冰.生成对抗网络介绍及应用[J].无线电通信技术，2021(3):338-343.XI X P,CHEN X,ZHU X B. Introduction and application of generative adversarial networks[J]. Radio Communications Technology,2021(3):338-343(in Chinese).