An Acquiring and texture comparing method of shaped floor ceramic tiles is proposed

After the image of tile collected by CCD has been segmented by gray threshold

the search of contours judged by feature of rectangle patterns is used to acquire the normalizing affine transformed tile region.Based on co-occurrence matrix feature vectors

the minimum distance classifying principle is adopted to carry out texture classifying decision

and adaptive threshold parameter is implemented to improve the robustness of algorithm.Shaped floor ceramic tiles system run successfully in actual scene and show the classifying is effective.