Detection of Fabric Defect Based on Optimal Tree Structure of Wavelet Decomposition

Abstract

This paper adopts optimal tree structure of wavelet transform algorithm. And extraction of sub-image characterization—frequency bands of intermediate frequency of fabric image texture. Calculating the value of entropy and energy is as characteristic quantity of defect detection. And they were also compared to characteristic value of the normal image for judging whether defect exists. Meanwhile, some typical defect forms of plain weave white cotton fabric has been detection in this paper. The experiment results show that the methods are effective and feasible.