A Classification of Flotation Froth Based on Geometry

In order to get the exact information of the bubble distribution which is extracted by the watershed transformation algorithm, because of the abnormity of the bubble shape, using the ellipsoidal shape to calculate the major and minor axes is the most appropriate method to describe the flotation froth. At the same time, it was found that weighting by cross sectional area was also the most appropriate method to determine average ellipsoidal diameter. It not only depresses the noise, but also makes for the classification of the froth image. The algorithm of image classification for flotation froth in this paper is based on the flotation froth classification that is proposed by Moolman and classifies the froth image into four classes. And the relationship between the bubble size distribution and the performance of flotation is discussed. It is the basis to the control the flotation by image processing.