Fabric defects detection using adaptive wavelets

Abstract

Purpose – Fabric defects detection is vital in the automation of textile industry. The purpose of this paper is to develop and implement a new fabric defects detection method based on adaptive wavelet. Design/methodology/approach – Fabric defects can be regarded as the abrupt features of textile images with uniform background textures. Wavelets have compact support and can represent these textures. When there is an abrupt feature existed, the response is totally different with the response of the background textures, so wavelets can detect these abrupt features. This method designs the appropriate wavelet bases for different fabric images adaptively. The defects can be detected accurately. Findings – The proposed method achieves accurate detection of fabric defects. The experimental results suggest that the approach is effective. Originality/value – This paper develops an appropriate method to design wavelet filter coefficients for detecting fabric defects, which is called adaptive wavelet. And it is helpful to realize the automation of textile industry.