A computer vision system for identification of granite-forming minerals based on RGB data and artificial neural networks

[Abstract]: Granitic stones are widely used in the field of Cultural Heritage in the north-western Iberian Peninsula. In some activities regarding conservation, such as the laser cleaning, it is of great interest the identification of the minerals on the granitic stone surface in order to improve the treatment and to avoid damages by means of the adaption of the fluence to each different forming mineral. The aim of this work is the optimization of a back propagation artificial neural network (ANN) in order to obtain the rapid and reliable identification of forming minerals in granitic rocks by means of RGB images. Our goal is, eventually, in situ monitoring the laser cleaning of granitic stonework. The results obtained, though preliminary, led a high degree of correct identification of the forming minerals for three different granitic types.This work has been supported by the Spanish Research Project BIA2014-54186-R funded by Ministerio de Economía y Competitividad. The authors would like to thank INGEMARGA S.L., especially to Miguel Martinez for his kind attention. J.S. Pozo-Antonio was supported by a postdoctoral contract with the University of Vigo within the framework of the 2011–2015 Galician Plan for Research, Innovation and Growth (Plan I2C) for 2014