Reconocimiento automatizado de menas metálicas mediante análisis digital de imagen: Un apoyo al proceso mineralúrgico. II: Criterios metalogenéticos discriminantes

Abstract

Ore microscopy may furnish very important information for geometallurgists, but today's needs for automation are difficult to meet with the optical microscope, unless an adequate methodology is developed. Some limitations of the routine procedure, related to risks of misidentification caused by the spectral similarity of some ores, ask for complementary criteria. Defining ore deposit typologies and the corresponding assemblages guides the choice of species and limits their number. Comparison of the reflectance values of the ores in each mineral association defined shows that their automated identification is possible in most of the common occurrences. The number of species to be actually considered being greatly limited, performance is increased. The system is not intended to substitute for a mineralogist, but to enhance enormously his performance, while offering the industry an economic procedure to produce a wealth of information which would not be possible with traditional methods, as the point counter.