Fabric defect detection based on texture enhancement

Abstract

Fabric defect detection plays an important role in automated inspection of fabric quality based on computer vision, which aims at locating defect regions accurately. In this paper, a method for defect detection based on texture enhancement is proposed. Firstly texture properties of the defect areas and the background can be discriminated efficiently as well as the boundary of defect regions can be located precisely by using texture enhancement method. Then combined with gray level co-occurrence matrices algorithm, the defect texture features are extracted and a classifier can be designed for texture defect detection. The experiment results demonstrate that the detection accuracy of fabric defect is substantially improved by using the proposed method.