In order to achieve the automatic monitoring and intelligent evaluation of the bauxite flotation process, a method
of analyzing and selecting the best production condition of the cleaner froth based on machine vision is presented. Firstly, the
roughness feature of froth surface texture is extracted based on the improved local binary pattern(LBP) operator. Then, the
best roughness interval of the cleaner froth texture feature in the flotation production process is obtained through analyzing
the relation between the flotation concentrate grade and the extracted surface texture features of the cleaner froth image.
This method can monitor the change of the texture feature of the cleaner froth image in real time and automatically identify
whether the cleaning process is under the best production condition, which lays the foundation of the optimal control of the
bauxite flotation process.