Abstract:Fabric defect detection is a crucial quality control step in the textile manufacturing industry. In this article, machine vision system based on the Sylvester Matrix Based Similarity Method (SMBSM) is proposed to automate the defect detection process. The algorithm involves six phases, namely resolution matching, image enhancement using Histogram Specification and Median-Mean Based Sub-Image-Clipped Histogram Equalization, image registration through alignment and hysteresis process, image subtraction, edge detection, and fault detection by means of the rank of the Sylvester matrix. The experimental results demonstrate that the proposed method is robust and yields an accuracy of 93.4%, precision of 95.8%, with 2275 ms computational speed.