Fig. 1. System design

Fig. 2. Original images of ceramic tile. (a) Crack in the tile head; (b) crack in the texture region

Fig. 3. Block diagram of detection algorithm. (a) Image preprocessing; (b) crack detection; (c) crack determination

Fig. 4. Image preprocessing of tile head area and texture area. (a) Original images; (b) red channel images; (c) median filtered images

Fig. 5. Defect detection for tile head area. (a) Filter template; (b) filtered image; (c) filter detection result

Fig. 6. Defect detection for texture area. (a) Seed point selection; (b) automatic region growth

Fig. 7. Detection results. (a) Morphological processing image of the tile head; (b) morphological processing image of the texture region

Fig. 8. Detection effects of different algorithms for tile head area. (a) Original image of tile head area; (b) maximum entropy; (c) sobel; (d) discrete wavelet transform; (e) proposed method

Fig. 9. Detection effects of different algorithms for texture regions. (a) Original image of texture region; (b) maximum entropy; (c) sobel; (d) discrete wavelet transform; (e) proposed method