The Relationship between Concentrate Yield and Descriptors from a Machine Vision System in a Platinum Flotation Application

"Recent developments in image analysis techniques and the power of computer hardware have spurred international interest in the potential uses of image analysis techniques in a range of industrial applications. One of these applications is the control of the complex flotation process. Despite years of work, the flotation process is still poorly understood. Current attempts to control flotation cells focus on consistently duplicating the control achieved by human operators.In human control of flotation two of the dominant visual cues are the froth flow speed and the froth stability. These two factors can be measured and inferred respectively from image analysis of video of the froth surface. Two important hydrodynamic variables for control of the flotation process are the air addition rate and the level of the froth pulp interface.The current work focuses on the relationship between froth flow speed and the rate of recovery of concentrate from a flotation cell. Results are presented that show a strong relationship between flow speed and concentrate yield. These results include tests where the conditions in the cell were not changed, as well as tests involving changes to the cell level and air addition rate. Preliminary results are presented which show a correlation between froth stability and froth percent solids.Future developments of the system will include comprehensive froth texture typing, colour space and stability analyses."