Froth Flotation Classification of Antimony Based on Histogram of Bubbles Perimeters

Abstract

The process of flotation is one of the most complex industrial processes for purifying minerals, and the control of flotation process is one of the most challenging issues in the mineral processing industry. This paper describes a method based on machine vision system to classify different grades of Antimony during the floatation process. It is proved that the size of bubbles provides valuable information about froth flotation process. The proposed machine vision system, after collecting froth flotation images of Antimony, segments each image bubbles with Extended-Maxima transform method and creates a descriptor based on bubbles perimeters. Based on different grades of Antimony, images are divided in to four classes. To classify Antimony froth images, the created descriptors are assigned to a classifier like support vector machine. The proposed method is used in an Antimony flotation cell, and results shows that it is able to classify froth images based on Antimony's concentrate grade with acceptable accuracy. The experimental results indicate that this method can classify froth flotation images better than some common methods like GLCM and CCM.