Fabric Surface Defect Detection Based on GMRF Model

Fabric defect detection is a challenging task due to unpredictable forms of the fabric defect. This paper presents a method based on Gaussian Markov Random Field (GMRF) model which automatically inspects fabric surface defects. This paper presents a method for detecting unpredictable fabric defects based on machine vision through two steps. First, the parameter distribution of GMRF model is achieved from defect-free fabric images. Second, the image of the fabric patch to be inspected is divided into fixed-size non-overlapping windows, and each window is classified as defective or non-defective according to the GMRF model parameters of each window.