Image Analysis Technique as a Tool for Extracting Features from the Copper Surface Froth in the Flotation Process

Abstract

The froth can be adopted as an indicator of the performance of flotation processes. The study of froth image structure would enable us to establish a number of parameters from which could convey the froth characteristics. To monitor the operating performance of the floatation cell by machine vision system, it is crucial to identify and extract those features that are descriptive of the surface froth. Consequently, it can provide interdependency between the froth characteristics with the operating conditions on one hand (e.g., aeration rate, froth depth, chemical compound and pH variation) and the cell parameters performance on the other hand, as well (e.g., copper grade, recovery and solid contents). The aim of the present study is to examine the copper froth characteristics, by adopting an image analysis technique and hence evaluating froth features such as the average bubble size, bubbles distribution, bubble shape features, bubble elongation factor, image average color and the color distribution. Owing to the intricacy aspect of the froth structure and in order to match properly between the real froth and the segmentation images, this algorithm adopts features similar to proper filters in the pre-processing stage, edge detection functions, threshold functions and different mathematical morphology models. The findings of this work reveal that the size and shape of the froth bubbles plays an important role in classifying the froth. Hence, it is possible to incorporate such features for either evaluating the flotation cell performance or adopting it for the automatic on-line control of the flotation process. The findings of this research could also be implemented towards the training of the operators.