Detection of Defects on Warp-knit Fabric Surfaces Using Self Organizing Map

Abstract

Warp-Knit fabric surface is a patterned surface with repetitive pattern units on its structure. Only few researches have been reported for Detection of defects on patterned fabric surfaces. Since Warp-Knit structures develop with different patterns the defect detection problem becomes even more complicated. Here, we have proposed Self-Organizing Map based non-segmenting approach for Detection of defects on Warp-Knit surfaces. The approach includes simplification of the Defect detection problem by clustering followed by abnormality detection using Self-Organizing Map. Gray Level Co-occurrence Matrix and Local Binary Pattern were used to represent the texture of fabric surfaces. The results show over 90% accuracy in detection of defects with respect to 13 different fabric structures.