基于欠完备字典重构的无监督织物疵点检测方法

doi:10.16183/j.cnki.jsjtu.2023.205

摘要/Abstract

摘要：

针对当前自动织物检测方法大多仍需人工挑选训练集而无法实现无监督学习的问题,提出使用中值稳健扩展局部二值模式(MRELBP)特征的无疵图像筛选方法与欠完备字典重构疵点检测方法,实现自动无监督疵点检测,并采用自适应字典大小搜索算法,自动选取合适字典大小.算法首先对织物样本进行无疵图像的自动筛选,然后使用K-SVD算法将筛选后的正常图像块作为训练集获取欠完备字典,最后将通过计算重构后的结构相似性指标(SSIM)作为阈值进行疵点检测.在334张含有经向、纬向、块状疵点的平纹白坯布上进行实验,与使用残差分割疵点的K-SVD方法相比,正检率平均提升21.81%,误检率平均降低0.72%,每张图像的检测速率平均提升50%,在AITEX数据集上取得了平均83.29%的正检率,证明了本算法的有效性.

关键词: 织物疵点检测, 中值稳健扩展局部二值模式, 字典学习, 无监督检测

Abstract:

To address the issue that most current automatic fabric defect detection methods still require manual selection of training sets and cannot achieve unsupervised learning, an automatic unsupervised defect detection method using the median robust extended local binary pattern (MRELBP) feature for flawless image screening and an under-complete dictionary reconstruction method for defect point detection are proposed. An adaptive dictionary size search algorithm is also proposed to automatically select a suitable dictionary size. First, the algorithm selects the flawless images from fabric images. Then, K-SVD is applied to obtain an under-complete dictionary from the normal image blocks as the training set. Finally, the reconstruction-base scheme is used to identify defects with the structural similarity index measure (SSIM) threshold. Experimental results of 334 plain fabrics with warp, weft, and block defects show that compared to the K-SVD method that uses residual segmentation for defect detection, the proposed method increases the correct detection rate up to 21.81%, reduces the false detection rate up to 0.72%, and a 50% increase in detection speed per image on average. The proposed algorithm achieved an average correct detection rate of 83.29% on the AITEX dataset, demonstrating its effectiveness.

Key words: fabric defect detection, median robust extended local binary pattern (MRELBP), dictionary learning, unsupervised detection

中图分类号:

刘建欣, 潘如如, 周建. 基于欠完备字典重构的无监督织物疵点检测方法[J]. 上海交通大学学报, 2025, 59(2): 283-292.

LIU Jianxin, PAN Ruru, ZHOU Jian. Unsupervised Fabric Defect Detection Based on Under-Complete Dictionary Reconstruction[J]. Journal of Shanghai Jiao Tong University, 2025, 59(2): 283-292.

图/表 10

图1

图2

图3

图4

表1

表2

图5

表3

图6

图7

参考文献 19

[1]	田宸玮, 王雪纯, 杨嘉能, 等. 织物瑕疵检测方法研究进展[J]. 计算机工程与应用, 2020, 56(12): 8-18. doi: 10.3778/j.issn.1002-8331.2002-0169
	TIAN Chenwei, WANG Xuechun, YANG Jianeng, et al. Research progress on fabric defect detection methods[J]. Computer Engineering and Applications, 2020, 56(12): 8-18. doi: 10.3778/j.issn.1002-8331.2002-0169
[2]	NGAN H Y T, PANG G K H, YUNG N H C. Automated fabric defect detection—A review[J]. Image and Vision Computing, 2011, 29(7): 442-458.
[3]	王斌, 李敏, 雷承霖, 等. 基于深度学习的织物疵点检测研究进展[J]. 纺织学报, 2023, 44(1): 219-227.
	WANG Bin, LI Min, LEI Chenglin, et al. Research progress in fabric defect detection based on deep learning[J]. Journal of Textile Research, 2023, 44(1): 219-227.
[4]	吕文涛, 林琪琪, 钟佳莹, 等. 面向织物疵点检测的图像处理技术研究进展[J]. 纺织学报, 2021, 42(11): 197-206. doi: 10.13475/j.fzxb.20200702710
	LÜ Wentao, LIN Qiqi, ZHONG Jiaying, et al. Research progress of image processing technology for fabric defect detection[J]. Journal of Textile Research, 2021, 42(11): 197-206. doi: 10.13475/j.fzxb.20200702710
[5]	韩彦芳, 施鹏飞. 基于多层小波和共生矩阵的纹理表面缺损检测[J]. 上海交通大学学报, 2006, 40(3): 425-430.
	HAN Yanfang, SHI Pengfei. The texture defect detection based on multi-level wavelet transform and co-occurrence matrix[J]. Journal of Shanghai Jiao Tong University, 2006, 40(3): 425-430.
[6]	HU G H, WANG Q H, ZHANG G H. Unsupervised defect detection in textiles based on Fourier analysis and wavelet shrinkage[J]. Applied Optics, 2015, 54(10): 2963-2980.
[7]	HAMDI A A, SAYED M S, FOUAD M M, et al. Unsupervised patterned fabric defect detection using texture filtering and K-means clustering[C]// 2018 International Conference on Innovative Trends in Computer Engineering. Aswan, Egypt: IEEE, 2018: 130-144.
[8]	MEI S, WANG Y D, WEN G J. Automatic fabric defect detection with a multi-scale convolutional denoising autoencoder network model[J]. Sensors, 2018, 18(4): 1064.
[9]	WU Y, LOU L, WANG J. Global fabric defect detection based on unsupervised characterization[J]. Journal of Shanghai Jiaotong University (Science), 2021, 26(2): 231-238.
[10]	PENG Z R, GONG X Y, WEI B G, et al. Automatic unsupervised fabric defect detection based on self-feature comparison[J]. Electronics, 2021, 10(21): 2652.
[11]	屠恩美, 杨杰. 半监督学习理论及其研究进展概述[J]. 上海交通大学学报, 2018, 52(10): 1280-1291. doi: 10.16183/j.cnki.jsjtu.2018.10.017
	TU Enmei, YANG Jie. A review of semi-supervised learning theories and recent advances[J]. Journal of Shanghai Jiao Tong University, 2018, 52(10): 1280-1291.
[12]	OJALA T, PIETIKAINEN M, MAENPAA T. Multiresolution gray-scale and rotation invariant texture classification with local binary patterns[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2002, 24(7): 971-987.
[13]	LIU L, LAO S Y, FIEGUTH P W, et al. Median robust extended local binary pattern for texture classification[J]. IEEE Transactions on Image Processing, 2016, 25(3): 1368-1381.
[14]	王红雨, 尹午荣, 汪梁, 等. 基于HSV颜色空间的快速边缘提取算法[J]. 上海交通大学学报, 2019, 53(7): 765-772.
	WANG Hongyu, YIN Wurong, WANG Liang, et al. Fast edge extraction algorithm based on HSV color space[J]. Journal of Shanghai Jiao Tong University, 2019, 53(7): 765-772.
[15]	GUO Z H, ZHANG L, ZHANG D. Rotation invariant texture classification using LBP variance (LBPV) with global matching[J]. Pattern Recognition, 2010, 43(3): 706-719.
[16]	LIU L, FIEGUTH P, GUO Y L, et al. Local binary features for texture classification: Taxonomy and experimental study[J]. Pattern Recognition, 2017, 62: 135-160.
[17]	SILVESTRE-BLANES J, ALBERO-ALBERO T, MIRALLES I, et al. A public fabric database for defect detection methods and results[J]. Autex Research Journal, 2019, 19(4): 363-374.
[18]	AHARON M, ELAD M, BRUCKSTEIN A. K-SVD: An algorithm for designing overcomplete dictionaries for sparse representation[J]. IEEE Transactions on Signal Processing, 2006, 54(11): 4311-4322.
[19]	吴莹, 汪军. 基于K-SVD学习字典的机织物纹理表征及应用[J]. 纺织学报, 2018, 39(2): 165-170.
	WU Ying, WANG Jun. Woven fabric texture representation and application based on K-SVD dictionary[J]. Journal of Textile Research, 2018, 39(2): 165-170.

编辑推荐 0

Metrics

阅读次数

全文

222

HTML			PDF

最新录用	在线预览	正式出版	最新录用	在线预览	正式出版
0	0	5	189	0	28

来源	本网站	其他网站

次数	54	168
比例	24%	76%

摘要

147

最新录用	在线预览	正式出版

90	0	57

来源	本网站	其他网站

次数	60	87
比例	41%	59%

参数类别	参数值
光照方式	透射光
图像大小/像素	512×512
对应实际尺寸/cm	6.1×6.1
相机分辨率(PPI)	213
MRELBP半径	2
组织结构	1上1下
经密/(根·10 cm^-1)	202
纬密/(根·10 cm^-1)	202
织物种类	平纹白坯布
经/纬纱支	20^S

训练集	疵点部分		整体
训练集	MSSIM	PSNR	MSSIM	PSNR
正常图像	0.958 8	42.630 3	0.985 8	47.869 4
筛选图像	0.959 4	42.677 6	0.987 8	48.385 4
疵点图像	0.969 1	44.696 8	0.988 1	48.854 7

算法	疵点类型	r_C/%	r_F/%	检测时间/ms
K-SVD	经向疵点	69.14	1.23	70
	纬向疵点	70.31	1.15
	块状疵点	77.19	1.90
	平均综合检测率	72.21	1.43
本文	经向疵点	91.45	0.84	35
	纬向疵点	90.63	0.72
	块状疵点	100	0.57
	平均综合检测率	94.02	0.71