Journal of Textile Research ›› 2019, Vol. 40 ›› Issue (12): 50-56.doi: 10.13475/j.fzxb.20181200407

• Textile Engineering • Previous Articles     Next Articles

Fabric defect detection method based on improved fast weighted median filtering and K-means

ZHANG Huanhuan(), MA Jinxiu, JING Junfeng, LI Pengfei   

  1. School of Electronic and Information, Xi'an Polytechnic University, Xi'an, Shaanxi 710048, China
  • Received:2018-12-03 Revised:2019-06-02 Online:2019-12-15 Published:2019-12-18

RichHTML

14

PDF (PC)

137

Abstract:

In order to detect various defects in the production process of textured fabrics, a texture fabric defect detection method based on improved weighted median filtering and K-means clustering was proposed. Firstly, the fabric image was preprocessed by the improved weighted median filter to reduce the influence of texture information on the defect detection. At the same time, by assigning weights and pixels to the histogram dynamic data, the time to seek the median was effectively shortened to increase the execution speed. Then, the K-means algorithm was adopted to cluster the filtered fabric images, and the cluster centers of the fabric image defects and non-defects were calculated, thereby realizing the segmentation of the image defect regions. The experimental results show that the method can effectively detect the defects of various types of textured fabrics such as square, dot, star, plain, and twill and significantly increase the detection speed.

Key words: fabric defect detection, improved weighted median filtering, joint histogram, K-means clustering

CLC Number: 

  • TP391

Fig.1

Flowchart of fabric defect detection"

Fig.2

Joint-histogram illustration"

Fig.3

Illustration of balance and cut point"

Fig.4

Segment of fabric defect by K-means. (a)Texture part; (b)Defect part; (c)Fabric defect image"

Fig.5

Detection results of different r. (a)Grid beheaded; (b) r=10.0; (c) r=13.0; (d) r=15.0"

Fig.6

Detection results of different σ. (a) Box-patterned oil; (b) σ=100.0; (c) σ=135.0; (d) σ=500.0"

Tab.1

Parameters setting for modified faster weighted median filter"

织物类型 r σ
方格 [12.5~13.0] [135.0~245.0]
点形 [8.0~11.0] [90.0~200.0]
星形 [55.0~58.0] [250.0~390.0]
平纹 [5.0~10.0] [12.0~25.0]
斜纹 [9.0~10.0] [21.5~25.0]

Fig.7

Some results of square, dot, star fabric defect detection by method of this paper. (a) Fabric original image; (b) Result by using improved weighted median filtering; (c) Detection result"

Fig.8

Some results of plain and twill fabric defect detection by method of this paper. (a) Fabric original image; (b) Result by using improved weighted median filtering; (c) Detection result"

Fig.9

Comparison of results by different method. (a) Fabric original image; (b) Detection result by method of reference[3]; (c) Detection result by method of reference[17]; (d) Detection result by method of this paper"

Tab.2

Comparison of average time by three different methodss"

方法 方格
织物		 点形
织物		 星形
织物		 平纹
织物		 斜纹
织物
文献[3] 9.156 7 14.998 6 8.536 5 10.015 8 38.782 4
文献[17] 0.941 5 0.913 4 1.095 7 0.974 1 49.459 7
本文方法 0.193 7 0.233 4 0.312 2 0.049 5 0.507 7

Tab.3

Comparison of accuracy by three different methods"

织物类型 检测方法 ACC/% TPR/% FPR/%
文献[3] 93.93 95.07 12.87
方格 文献[17] 92.50 94.29 20.00
本文方法 97.50 98.09 6.67
文献[3] 92.49 96.01 29.97
点形 文献[17] 95.09 97.82 20.00
本文方法 94.17 95.24 13.57
文献[3] 93.62 97.14 25.88
星形 文献[17] 91.67 97.73 13.29
本文方法 94.36 95.89 19.73
文献[3] 95.83 98.09 20.05
平纹 文献[17] 90.94 95.23 30.48
本文方法 94.23 96.84 25.75
文献[3] 91.58 98.56 10.98
斜纹 文献[17] 92.79 96.76 17.79
本文方法 94.64 93.42 5.96
[1] CONCI A, PROENÇA C B. A fractal image analysis system for fabric inspection based on a box-counting method[J]. Computer Networks & Isdn Systems, 1998,30(20/21):1887-1895.
[2] 朱丹丹, 潘如如, 高卫东. 基于傅里叶特征谱和相关系数的织物瑕疵检测[J]. 计算机工程与应用, 2014,50(19):182-186.
ZHU Dandan, PAN Ruru, GAO Weidong. Fabric defect detection using characteristic spectrum of Fourier transform and correlation coefficient[J]. Computer Engineering and Applications, 2014,50(19):182-186.
[3] JING J F, YANG P, LI P F. Supervised defect detection on textile fabrics via optimal Gabor filter[J]. Journal of Industrial Textile, 2013,44(1):40-57.
[4] LI W, XUE W, CHENG L. Intelligent detection of defects of yarn-dyed fabrics by energy-based local binary patterns[J]. Textile Research Journal, 2012,82(19):1960-1972.
[5] AMET A L, ERTÜZÜN A, ERÇIL A. An efficient method for texture defect detection: sub-band domain co-occurrence matrices[J]. Image and Vision Computing, 2000,18(6/7):543-553.
doi: 10.1016/S0262-8856(99)00062-1
[6] KUMBHAR P Y, MATHPATI T. Textile fabric defects detection and sorting using image processing[J]. International Journal for Research in Emerging Science and Technology, 2016,3(3):19-24.
[7] KUMAR A, PANG G K H. Defect detection in textured materials using Gabor filters[J]. IEEE Transactions on Industry Applications, 2002,38(2):425-440.
[8] TSAI I S, HU M C. Automatic inspection of fabric defects using an artificial neural network technique[J]. Textile Research Journal, 1996,66(7):474-482.
[9] HAJIMOWLANA S H, MUSCEDERE R, JULLIEN G A. et al. 1D autoregressive modeling for defect detection in web inspection systems[C]// Midwest Symposium on Systems & Circuits. [S. l.]: IEEE Computer Society, 1998: 318-321.
[10] 李敏, 崔树芹, 谢治平. 高斯混合模型在印花织物瑕疵检测中的应用[J]. 纺织学报, 2015,36(8):94-98.
LI Min, CUI Shuqin, XIE Zhiping. Application of Gaussian mixture model on defect detection of print fabric[J]. Textile Research Journal, 2015,36(8):94-98.
[11] MEI S, WANG Y, WEN G. Automatic fabric defect detection with a multi-scale convolutional denoising autoencoder network model[J]. Sensors, 2018,18(4):1064.
[12] YAPI D, MEJRI M, ALLILI M S. A Learning-based approach for automatic defect detection in textile images[J]. Biochemical & Biophysical Research Communications, 2015,48(3):2423-2428.
[13] TSANG C S C, NGAN H Y T, PANG G K H. Fabric inspection based on the Elo rating method[J]. Pattern Recognition, 2016,51(4):378-394.
[14] LI Y, ZHAO W, PAN J. Deformable patterned fabric defect detection with fisher criterion-based deep lear-ning[J]. IEEE Transactions on Automation Science and Engineering, 2017,14(2):1256-1264.
[15] ZHANG Q, XU L, JIA J. 100+ times faster weighted median filter[C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Columbus: CUPR, 2014: 2830-2837.
[16] NG M K, NGAN H Y T, YUAN X, et al. Patterned Fabric Inspection and Visualization by the Method of Image Decomposition[J]. IEEE Transactions on Automation Science & Engineering, 2014,11(3):943-947.
[17] 张缓缓, 李仁忠, 景军锋. Frangi滤波器和模糊C均值算法相结合的织物瑕疵检测[J]. 纺织学报, 2015,36(9):120-124.
ZHANG Huanhuan, LI Renzhong, JING Junfeng. Fabric defect detection based on Frangi filter and fuzzy C-means algorithm in combination[J]. Journal of Textile Research, 2015,36(9):120-124.
[1] ZHU Lei, REN Mengfan, PAN Yang, LI Botao. Fabric defect detection based on similarity location and superpixel segmentation [J]. Journal of Textile Research, 2020, 41(10): 58-66.
[2] . Detection of fabric defects based on Gabor filters and Isomap [J]. JOURNAL OF TEXTILE RESEARCH, 2017, 38(03): 162-167.
[3] . Warp knit fabric defect detection method based on optimal Gabor filters [J]. JOURNAL OF TEXTILE RESEARCH, 2016, 37(11): 48-54.
[4] . Fast fabric defect detection algorithm based on integral image [J]. JOURNAL OF TEXTILE RESEARCH, 2016, 37(11): 141-147.
[5] . Fabric defects detection method based on texture saliency features [J]. JOURNAL OF TEXTILE RESEARCH, 2016, 37(10): 42-049.
[6] . Fabric defect detection using monogenic wavelet analysis [J]. JOURNAL OF TEXTILE RESEARCH, 2016, 37(09): 59-64.
[7] . Research of young female somatotype based on longitudinal profile curves [J]. JOURNAL OF TEXTILE RESEARCH, 2014, 35(8): 87-0.
[8] . Research on detection of defects in fabrics using improved singular value decomposition [J]. JOURNAL OF TEXTILE RESEARCH, 2014, 35(6): 62-0.
[9] TIAN Chengtai;BU Honggang;WANG Jun;CHEN Xia;. Fabric defect detection based on fractal feature of time series [J]. JOURNAL OF TEXTILE RESEARCH, 2010, 31(5): 44-47.
[10] PAN Ruru;GAO Weidong;ZHANG Xingye. Fabric defect detection based on dynamic clustering arithmetic [J]. JOURNAL OF TEXTILE RESEARCH, 2008, 29(10): 122-126.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
京ICP备14006648号
Copyright 2021 © Editorial office of Journal of Textile Research
Supported by Beijing Magtech Co.Ltd