Feature Descriptors Applied to Slag Characterization on Casting Process

Vision systems are increasingly entering the field of metallurgy, carrying out operations where a human operator is not possible due to the process conditions. The purpose of these systems is the monitoring and control of the process to improve the quality and manufacturing of the products. Nevertheless, the amount of slag, the presence of gases and high temperatures are the main problems that make this task difficult. In this proposal the characterization of the slag is treated, through the analysis of the light changes with the functions of Fourier and Gabor, which allow to identify or locate the location of the slag in the material, so that, in future works the slag It can be segmented, measured or used to detect the level of the metal in the refractory. In addition, results obtained when evaluating sensitivity and precision curves are presented, with which the information recovered by the algorithms is evaluated.