Automatic flotation froth bubble size distribution estimation using mean shift and watershed transforms

Abstract

Flotation is a process used in the mineral industry to concentrate valuable minerals. The efficiency of the process is influenced by among other the bubble size distribution of the flotation froth. In this investigation, we propose a new method for automatic estimation of the bubble size distribution using the mean shift algorithm and watershed transform. The mean shift algorithm is used to find pixel clusters of particular modes of the probability density function of the image data. Pixel clusters of which the modes are above a given threshold are used as markers for the watershed transform. Experimental results show that the proposed method was fast and robust for determining the size distribution of different classes of flotation froth.