Chi-square histogram distance is one of the distance measures that can beused to find dissimilarity between two histograms. Motivated by the fact thattexture discrimination by human vision system is based on second-orderstatistics, we make use of histogram of gray-level co-occurrence matrix (GLCM)that is based on second-order statistics and propose a new machine visionalgorithm for automatic defect detection on patterned textures. Input defectiveimages are split into several periodic blocks and GLCMs are computed afterquantizing the gray levels from 0-255 to 0-63 to keep the size of GLCM compactand to reduce computation time. Dissimilarity matrix derived from chi-squaredistances of the GLCMs is subjected to hierarchical clustering to automaticallyidentify defective and defect-free blocks. Effectiveness of the proposed methodis demonstrated through experiments on defective real-fabric images of 2 majorwallpaper groups (pmm and p4m groups).
展开▼