Abstract: A large number of froth images of laboratory flotation of chalcopyrite were collected and pre-processed. Digital image analysis technique was used to analyze these froth images and their gray histogram. The statistical texture feature parameters of the gray histogram of froth images were extracted and the model for the relation between the flotation indexes of chalcopyrite and the statistical texture feature parameters of those froth images was established by using RBF neural network. The simulation proves the high precision of this model.

Key words: Flotation froth, digital image, image gray histogram, RBF neural network, flotation index prediction model