Proposal for a Platform Based on Artificial Vision for the Identification and Classification of Ceramic Tiles

It is proposed to develop a prototype platform based on artificial vision that allows us to identify the defects that ceramic tiles have, with which we can classify them according to their quality. The algorithms chosen to develop the platform are Support Vector Machine (SVM) and K-Nearest Neighbor (KNN). The SVM algorithm has the characteristic that, a priori, we know the classes to which our individuals belong, it is not a grouping by similarities, but we have well-defined classes, in the treatment of images there is a high percentage of effectiveness of this algorithm. On the other hand, because of its ease of implementation, the KNN algorithm is one of the most widely used non-parametric classifiers. Its theoretical properties guarantee that its error probability is bounded by twice the Bayesian error probability, in the treatment of images with this algorithm there are high percentages of effectiveness in the classification The necessary parameters are established for the proposal based on the study of related works and the same application methodology is presented, which contemplates the pre-processing of the images, the obtaining of the descriptors, the use of the algorithms and the results obtained.