Ill-defined processes such as the froth flotation of minerals are mostly controlled in an empirical way by using rules of thumb. These processes involve so many variables that the plant operator finds it difficult to visualise or even observe a change in process conditions. In froth flotation the operator is supposed to visually observe process changes from the appearance of the froth, which is an unreasonable demand under industrial conditions. An on-line computer vision system based on a textural analysis of the froth phase has been developed in South Africa and has been in operation on two industrial plants since early 1995. Textural parameters are determined online, and disturbances in process conditions, such as a change in reagent addition or froth depth, are visualised via a Self-Organizing Map (SOM) neural net.