Defect Detection in Fabrics Using Local Binary Patterns - Advances in Image and Graphics Technologies - page 274

Defect Detection in Fabrics Using Local Binary Patterns
Pengfei Li, Xuan Lin, Junfeng Jing, and Lei Zhang
Xi'an Polytechnic University, College of Electronical and
Information, 710048 Xi'an, Shaanxi, China
{Li6208,linxuan0122}@163.com,
{413066458,11795503}@qq.com
Abstract . To detect defects in fabrics more efficiently, easily and accurately, a
method based on Local Binary Pattern (LBP) is proposed in this paper. The main
purpose of this algorithm is to extract the feature value of fabric images. Firstly
the feature of the whole defect-free fabric image is got with LBP algorithm. Then
the image is divided into small detection windows, and the feature of each
window can be obtained. Compare their similarity calculated by Chi-square
function to get the threshold. Then process the defective images according to the
same procedure. At last compare the similarity with the threshold to obtain defect
regions. The defects are detected at the same time. Experimental results demon-
strate that, LBP algorithm is effective in the area of detecting defects of fabrics.
Keywords : Local Binary Pattern, feature value, similarity, defect detection.
1 Introduction
The development of global fabrics is very rapid in recent years. Simultaneously textile
quality requirements for people are increasing. Consequently the quality of textiles
needs to be controlled strictly. Defects have a serious impact on the identification of
quality levels for fabrics [1]. According to statistics, if there is a defect, it may cause the
value of fabrics decreased by 45%-60% [2], [3]. Therefore, defect detection becomes
an essential step during fabric quality assessment.
With the continuous development of digital image processing technology, many
domestic and foreign scholarships have conducted extensive researches on defect de-
tection. To sum up, all the defecting methods can be divided into statistics based, model
based and spectrum based. Statistical methods are based on the gray properties of both
the pixels and their neighborhoods. Defects can be detected through studying the sta-
tistical characteristic in the texture area and determining the difference between the
defect region and the normal region. I-Shou TSai et al. [4] have proposed a method for
fabric defect detection using GLCM. With this method, threshold is not needed to
determine the defect area of the detected fabric image, but a very large amount of
calculation would exist. Model-based approaches describe the texture characteristics of
fabrics by the parameters of particular models. B. S. Manjunath [5] has conducted a