基于改进RefineDet的织物疵点检测

doi:10.19398/j.att.202107051

现代纺织技术 ›› 2022, Vol. 30 ›› Issue (5): 12-20.DOI: 10.19398/j.att.202107051

• 特约专栏:图像处理与数值模拟 • 上一篇下一篇

基于改进RefineDet的织物疵点检测

阮梦玉, 李敏, 何儒汉, 姚迅

武汉纺织大学计算机与人工智能学院,武汉 430200

收稿日期:2021-07-30 出版日期:2022-09-10 网络出版日期:2022-09-19
通讯作者:李敏,E-mail:2008031@wtu.edu.cn
作者简介:阮梦玉(1996-),女,湖北黄冈人,硕士研究生,主要从事图像处理方面的研究。
基金资助:
湖北省教育厅科技项目(D20161605)

Fabric defect detection based on improved RefineDet

RUAN Mengyu, LI Min, HE Ruhan, YAO Xun

School of Computer and Artificial Intelligence, Wuhan Textile University, Wuhan 430200

Received:2021-07-30 Published:2022-09-10 Online:2022-09-19

摘要/Abstract

摘要： 为了实现织物疵点的自动检测与分类,提出了一种基于改进RefineDet的疵点检测方法。首先,将VGG16改为全卷积网络对织物图像特征进行提取;其次,为了获取疵点重要的特征并抑制不必要的特征,在Anchor细化模块(Anchor refinement module,ARM)中加入了注意力机制;为了提高网络的分类性能,在传输连接块(Transfer connection block,TCB)中加入了SE模块(Squeeze and excitation,SE);最后,目标检测模块(Object detection module,ODM)将检测的结果回归到准确的目标位置,并预测疵点的类别,对疵点进行定位。结果表明:本文算法对孔、污渍、纱疵和线状4种类别织物图像的均值平均精度mAP达到了79.7%,比传统RefineDet检测方法均值平均精度提高了5.0%,具有良好的分类和定位效果。

关键词: 深度学习, 疵点检测, RefineDet, VGG16, 注意力机制

Abstract: To realize the automatic detection and classification of fabric defects, a defect detection method based on Refinedet is proposed. Firstly, VGG16 is changed into a full convolution network to extract fabric image features. Secondly, to obtain the important features of defects and suppress unnecessary features, an attention mechanism is added to the anchor refinement module (ARM). To improve the classification performance of the network, SE module (sequence and exception, SE)is added to the transfer connection block (TCB). Finally, the object detection module (ODM)returns the detection results to the accurate target position, predicts the category of defects and locates the defects. The experimental results show that the average accuracy of this algorithm is 79.7%,which is 5.0% higher than that of the traditional algorithm.

Key words: deep learning, defect detection, RefineDet, VGG16, attention mechanism

中图分类号:

TS101.9

阮梦玉, 李敏, 何儒汉, 姚迅. 基于改进RefineDet的织物疵点检测[J]. 现代纺织技术, 2022, 30(5): 12-20.

RUAN Mengyu, LI Min, HE Ruhan, YAO Xun. Fabric defect detection based on improved RefineDet[J]. Advanced Textile Technology, 2022, 30(5): 12-20.

参考文献

[1] XIA D, JIANG G M, LI Y Y, et al. Warp-knitted fabric defect segmentation based on non-subsampled Contourlet transform[J]. Journal of the Textile Institute,2017,108(2): 239-245.
[2] 袁小军,陈晓东,邱莉,等.基于最大熵阈值算法的织物疵点检测与疵点织物应用[J].毛纺科技,2019,47(3):66-70.
YUAN Xiaojun, CHEN Xiaodong, QIU Li, et al. Fabric defect detection and application of defect fabric based on maximum entropy threshold algorithm[J]. Wool Textile Journal,2019, 47(3): 66-70.
[3] 俞新星,任勇,支佳雯.织物表面疵点检测方法的设计与实现[J].现代纺织技术,2021,29(1):62-67.
YU Xinxing, REN Yong, ZHI Jiawen. Design and implementation of defect detection method for fabric surface[J]. Advanced Textile Technology,2021, 29(1): 62-67.
[4] 赵艳,左保齐.机器视觉在织物疵点检测上的应用研究综述[J].计算机工程与应用,2020,56(2):11-17.
ZHAO Yan, ZUO Baoqi. Analysis on application of machine vision in fabric defect detection[J]. Computer Engineering and Application, 2020, 56 (2): 11-17.
[5] JUN X, WANG J G, ZHOU J, et al. Fabric defect detection based on a deep convolutional neural network using a two-stage strategy[J]. Textile Research Journal, 2021, 91(1/2): 130-142.
[6] 周文明,周建,潘如如.应用上下文视觉显著性的色织物疵点检测[J].纺织学报,2020,41(8):39-44.
ZHOU Wenming, ZHOU Jian, PAN Ruru. Yarn-dyed fabric defect detection based on context visual saliency[J]. Journal of Textile Research, 2020, 41 (8): 39-44.
[7] 景军锋,范晓婷,李鹏飞,等.应用深度卷积神经网络的色织物缺陷检测[J].纺织学报,2017,38(2):68-74.
JING Junfeng, FAN Xiaoting, LI Pengfei, et al. Yarn-dyed fabric defect detection based on deep-convolutional neural network[J]. Journal of Textile Research, 2017, 38 (2): 68-74.
[8] 刘纪,张团善,徐杰.应用对抗网络和霍夫变换的织物点状缺陷检测研究[J].毛纺科技,2020,48(12):62-66.
LIU Ji, ZHANG Tuanshan, Xu Jie. Research on fabric point defect detection using generative adversarial network and Hough Transform[J]. Wool Textile Journal, 2020, 48 (12): 62-66.
[9] 李敏,崔树芹,陈佳.应用视觉显著性的小提花织物疵点检测[J].纺织学报,2016,37(12):38-42,48.
LI Min, CUI Shuqin, CHEN Jia. Defect detection for mini-jacquard fabric based on visual saliency[J]. Journal of Textile Research, 2016, 37 (12): 38-42,48.
[10] ZHANG S F, WEN L Y, BIAN X, et al. Single-shot refinement neural network for object detection[C]. IEEE Conference on Computer Vision and Pattern Recognition. Salt Lake City, UT, USA. IEEE,2018:4203-4212.
[11] 荣亮,高清维,李笑语,等.RefineDet网络与注意力机制结合的目标检测算法[J].传感器与微系统,2021,40(3):130-133.
RONG Liang, GAO Qingwei, LI Xiaoyu, et al. Target detection algorithm combining RefineDet network and attention mechanism[J]. Transducer and Microsystem Technologies, 2021, 40 (3): 130-133.
[12] JIE H, LI S, GANG S, et al. Squeeze-and-excitation networks[C]. IEEE Conference on Computer Vision and Pattern Recognition. Salt Lake City, UT, USA. IEEE,2018:7132-7141.
[13] 丁淑敏,李春雷,霍昭辰,等.基于弱监督学习的三分支卷积网络织物疵点检测方法:CN111882546A[P].2020-11-03.
DING Shumin, LI Chunlei, HUO Zhaochen, et al. Three branch convolution network fabric defect detection method based on weak supervised learning: CN111882546A[P]. 2020-11-03.
[14] 肖振远,王逸涵,罗建桥,等.基于部分加权损失函数的RefineDet[J].计算机应用,2021,41(7):1928-1932.
XIAO Zhenyuan, WANG Yihan, LUO Jianqiao, et al. Refinedet based on partial weighted loss function[J]. Computer Applications, 2021, 41 (7): 1928-1932.

基于改进RefineDet的织物疵点检测

Fabric defect detection based on improved RefineDet

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 7

编辑推荐

Metrics

[1]	从明芳, 李子印, 卢鸯, 韩高锋, 谢凌佳, 王启真. 基于Mask R-CNN深度学习的羊绒羊毛纤维识别技术[J]. 现代纺织技术, 2022, 30(2): 36-40.
[2]	李敏，杨珊，何儒汉，姚迅，崔树芹. 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, 0(6): 62-66.
[3]	汪坤,史伟民,李建强,彭来湖. 基于深度学习的织物印花分割算法研究[J]. 现代纺织技术, 2021, 0(3): 45-50.
[4]	俞新星，任勇，支佳雯. 织物表面疵点检测方法的设计与实现[J]. 现代纺织技术, 2021, 0(1): 62-67.
[5]	吴军. 织物表面疵点自动检测方法研究[J]. 现代纺织技术, 2020, 0(1): 47-51.
[6]	王孟涛，李岳阳，杜帅. 基于机器视觉的疵点检测方法的研究进展[J]. 现代纺织技术, 2019, 0(05): 57-61.
[7]	纪惠军1，张小栋2，杨建民1. 基于机器视觉的织物疵点检测技术探讨[J]. 现代纺织技术, 2011, 0(02): 11-15.