Fabric defect detection based on dynamic clustering arithmetic

The fabric image was processed with Gauss template smoothing pretreatment,and average grey value of columns and rows of the normal fabric image in small window was extracted.Mid grey value of the yarn evenness and the space between yarns were worked out through K-means clustering.Besides,average grey value and fluctuating range of mid grey value of the normal fabric image were also figured out.By comparing mid grey value of the image being checked and of normal fabric image,whether the fluctuating range of the former went beyond that of normal fabric image was judged,and whether there were flaws in the image being checked was distinguished and its location was lined out exactly.