Texture feature extraction for a visual inspection of ceramic tiles

The aim of this research was to determine an acceptable method of texture feature extraction that could be used for a visual inspection of ceramic tiles quality. Traditional extraction of statistical features has been carried out as well as feature extraction based on local binary pattern operators. For the purpose of testing the proposed methods good and defective ceramic tiles have been identified. The obtained results indicate a possibility for developing a system of a visual inspection of ceramic tiles quality and point out local binary pattern operators as a powerful tool applicable in this particular case.