Classification of Travertine Tiles with Supervised and Unsupervised Classifiers and Quality Control
Classification of Travertine Tiles with Supervised and Unsupervised Classifiers and Quality Control is a scholarly work, published in 2018 in ''International Journal of Engineering and Technology''. The main subjects of the publication include quality, artificial intelligence, self-reconfiguring modular robot, control in organization management, machine learning, pattern recognition, and computer science. Natural stone tiles, which are used in decoration field, should be in a specific visual standard.In order to meet this expectation, it is very important that the tiles be classified before being packed.Especially in operation where exports of processed natural stones are increasing, all steps from cutting to packing must be done with the least mistakes.In the natural stone sector, selection workers are mostly used for classification and packaging operations.This situation; changing environmental factors such as light, temperature, can result in errors due to the fact that the human eye tends to get tired and lose perception over time.Therefore, there is a need for an automation system that can be used correctly for classification.Along with developing technologies, it is seen that the errors that can occur in classification can be reduced to a minimum.In this study both of supervised and unsupervised algorithms were used and it was tried to determine which classifier and dataset gave the most accurate result.Besides this, the corner fractures of the travertine tiles were also detected and quality control was made.