Mineral granulometric analysis using mathematical morphology

Abstract

The most accurate granulometric analysis techniques present some disadvantages, mainly related with high cost of acquisition and maintenance of sophisticated devices. On the other hand, cheaper systems are generally characterized by low precision or long time to conclude the analysis. The only techniques capable of conducting a morphological analysis, today considered fundamental to mineral characterization, are image based, generally by microscopy. The main objectives of this paper consist on comparing different image acquisition devices and evaluating the feasibility of using mathematical morphology analysis to obtain granulometric profiles. Each configuration for image acquisition devices proved useful in specific ranges of particle sizes. Computational sieving by mathematical morphology also achieved promising results, and was even capable of discriminating grain sizes of particles trapped in a given physical sieve, thus enhancing granulometric precision.