Automatic visual inspection of woven textiles using a two-stage defect detector

Jonathan G. Campbell, Fionn Murtagh

Research output: Contribution to journalArticle

47 Citations (Scopus)

Abstract

Automatic inspection of woven textile fabric is discussed. A two-stage detection process is adopted, with the second stage involving set of novel contextual decision fusion techniques. Three significant problems are addressed: (1) texture feature extraction: Fourier transform features are found to be well matched to the spatially periodic nature of the woven pattern; (2) detection of localized flaw patterns: since prior probabilities are impossible to estimate, and we cannot hope to enumerate all defect classes, a Neyman-Pearson approach is adopted, i.e., flaw detection is via measured deviation from nominal; and (3) detection of extended flaw patterns: the most common flaws are characterized by linear or other cluster shaped patterns; although these are weakly detectable by local detectors, they may be ignored when local detector sensitivity is set to achieve tolerably low false-alarm rates; a local-extended contextual decision fusion technique using morphological filtering enables us to achieve very low composite false-alarm rate. The performance of the system is evaluated on samples of denim fabric containing real defects. The predicted composite false-alarm rate is of the order 1 in 1013, or equivalent to 1 per 100 km of fabric roll. Experimental results demonstrate the compatibility of this favorable false-alarm rate with the reliable detection of flaws, which have been chosen for their Subtlety and detection difficulty.

LanguageEnglish
Pages2536-2542
Number of pages7
JournalOptical Engineering
Volume37
Issue number9
DOIs
Publication statusPublished - 1 Sep 1998
Externally publishedYes

Fingerprint

textiles
inspection
Textiles
Inspection
false alarms
Detectors
Defects
detectors
defects
fusion
Fusion reactions
nondestructive tests
composite materials
pattern recognition
Composite materials
compatibility
textures
Feature extraction
Fourier transforms
deviation

Cite this

@article{120ccb8e1da14857a48dd427433ddb77,
title = "Automatic visual inspection of woven textiles using a two-stage defect detector",
abstract = "Automatic inspection of woven textile fabric is discussed. A two-stage detection process is adopted, with the second stage involving set of novel contextual decision fusion techniques. Three significant problems are addressed: (1) texture feature extraction: Fourier transform features are found to be well matched to the spatially periodic nature of the woven pattern; (2) detection of localized flaw patterns: since prior probabilities are impossible to estimate, and we cannot hope to enumerate all defect classes, a Neyman-Pearson approach is adopted, i.e., flaw detection is via measured deviation from nominal; and (3) detection of extended flaw patterns: the most common flaws are characterized by linear or other cluster shaped patterns; although these are weakly detectable by local detectors, they may be ignored when local detector sensitivity is set to achieve tolerably low false-alarm rates; a local-extended contextual decision fusion technique using morphological filtering enables us to achieve very low composite false-alarm rate. The performance of the system is evaluated on samples of denim fabric containing real defects. The predicted composite false-alarm rate is of the order 1 in 1013, or equivalent to 1 per 100 km of fabric roll. Experimental results demonstrate the compatibility of this favorable false-alarm rate with the reliable detection of flaws, which have been chosen for their Subtlety and detection difficulty.",
keywords = "Decision fusion, Defect detection, Fourier transform, Morphological filtering, Neyman-Pearson, Textile, Texture",
author = "Campbell, {Jonathan G.} and Fionn Murtagh",
year = "1998",
month = "9",
day = "1",
doi = "10.1117/1.601692",
language = "English",
volume = "37",
pages = "2536--2542",
journal = "Optical Engineering",
issn = "0091-3286",
publisher = "SPIE",
number = "9",

}

Automatic visual inspection of woven textiles using a two-stage defect detector. / Campbell, Jonathan G.; Murtagh, Fionn.

In: Optical Engineering, Vol. 37, No. 9, 01.09.1998, p. 2536-2542.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Automatic visual inspection of woven textiles using a two-stage defect detector

AU - Campbell, Jonathan G.

AU - Murtagh, Fionn

PY - 1998/9/1

Y1 - 1998/9/1

N2 - Automatic inspection of woven textile fabric is discussed. A two-stage detection process is adopted, with the second stage involving set of novel contextual decision fusion techniques. Three significant problems are addressed: (1) texture feature extraction: Fourier transform features are found to be well matched to the spatially periodic nature of the woven pattern; (2) detection of localized flaw patterns: since prior probabilities are impossible to estimate, and we cannot hope to enumerate all defect classes, a Neyman-Pearson approach is adopted, i.e., flaw detection is via measured deviation from nominal; and (3) detection of extended flaw patterns: the most common flaws are characterized by linear or other cluster shaped patterns; although these are weakly detectable by local detectors, they may be ignored when local detector sensitivity is set to achieve tolerably low false-alarm rates; a local-extended contextual decision fusion technique using morphological filtering enables us to achieve very low composite false-alarm rate. The performance of the system is evaluated on samples of denim fabric containing real defects. The predicted composite false-alarm rate is of the order 1 in 1013, or equivalent to 1 per 100 km of fabric roll. Experimental results demonstrate the compatibility of this favorable false-alarm rate with the reliable detection of flaws, which have been chosen for their Subtlety and detection difficulty.

AB - Automatic inspection of woven textile fabric is discussed. A two-stage detection process is adopted, with the second stage involving set of novel contextual decision fusion techniques. Three significant problems are addressed: (1) texture feature extraction: Fourier transform features are found to be well matched to the spatially periodic nature of the woven pattern; (2) detection of localized flaw patterns: since prior probabilities are impossible to estimate, and we cannot hope to enumerate all defect classes, a Neyman-Pearson approach is adopted, i.e., flaw detection is via measured deviation from nominal; and (3) detection of extended flaw patterns: the most common flaws are characterized by linear or other cluster shaped patterns; although these are weakly detectable by local detectors, they may be ignored when local detector sensitivity is set to achieve tolerably low false-alarm rates; a local-extended contextual decision fusion technique using morphological filtering enables us to achieve very low composite false-alarm rate. The performance of the system is evaluated on samples of denim fabric containing real defects. The predicted composite false-alarm rate is of the order 1 in 1013, or equivalent to 1 per 100 km of fabric roll. Experimental results demonstrate the compatibility of this favorable false-alarm rate with the reliable detection of flaws, which have been chosen for their Subtlety and detection difficulty.

KW - Decision fusion

KW - Defect detection

KW - Fourier transform

KW - Morphological filtering

KW - Neyman-Pearson

KW - Textile

KW - Texture

UR - http://www.scopus.com/inward/record.url?scp=0001518030&partnerID=8YFLogxK

U2 - 10.1117/1.601692

DO - 10.1117/1.601692

M3 - Article

VL - 37

SP - 2536

EP - 2542

JO - Optical Engineering

T2 - Optical Engineering

JF - Optical Engineering

SN - 0091-3286

IS - 9

ER -