Visual information model based predictor for froth speed control in flotation process Felipe Núñez, Aldo Cipriano * College of Engineering, Pontificia Universidad Católica de Chile, Santiago, Chile article info Article history: Received 29 May 2008 Accepted 17 October 2008 Available online 9 December 2008 Keywords: Dynamic textures Froth flotation Expert systems Modelling abstract Image processing sensors are emerging as an important measurement option in mineral processing, mainly due to their non-intrusive characteristics. Their principal application areas have been the deter- mination of ore size distributions in grinding and froth features in flotation. The incorporation of visual information in control loops is the logical step. However, the excessive processing required brings a new problem that must be solve: to count with a strategy able to provide a measured value for each visual sensor in the plant. A first approach is to assign one computer to each sensor yielding a distributed archi- tecture, but this means the implementation of a huge computer network. A more efficient alternative is alternated sampling, but the succeed of this option is limited to the existence of virtual sensors capable of give accurate values that must be use during the unsampled period. In this paper we begin by reviewing classical image processing algorithms used in flotation froth feature extraction. Then a new method is introduced for the characterization and recognition of visual information using dynamic texture tech- niques. Finally we developed a dynamic texture based virtual sensor for the prediction of froth speed in the unsampled period, tested with industrial data. Ó 2008 Elsevier Ltd. All rights reserved. 1. Introduction The aim of mineral processing is to concentrate raw ore in prep- aration for the subsequent metal-extraction stage. Usually, the valuable minerals are first liberated from the ore matrix by commi- nution and size-separation processes and then separated from the gangue using processes capable of selecting the particles according to their physical or chemical properties. These include characteris- tics such as surface hydrophobicity, specific gravity, magnetic sus- ceptibility and color (Wills, 1992). Over the last two decades, image-processing sensors have emerged as an interesting measurement option in mineral process- ing, mainly due to their non-intrusive characteristics. The principal application fields of these sensors are the grinding and flotation stages. In the grinding process, the most successful application of visual sensors is fragmentation measurement in traditional mining oper- ations with large crushing and milling circuits that require effi- ciency in comminution to maximize profit on the final product. With the help of on-line image analysis systems, operators can ac- tively monitor the various stages in the comminution circuit to en- sure their operation is running within specification. Regarding the flotation process, it is widely known that the color, morphology and speed of the froth are closely related to mineral concentrations, process status and recovery, respec- tively. As a consequence, operators often make operating changes based on the visual appearance of the froth as well as past experience. Image processing sensors provide quantita- tive measures of some representative froth characteristics and also offer the option of making control decisions independently of the operator, resulting in a control law that depends exclu- sively on these measures. Most popular automatic control strat- egies that incorporate image processing sensors measures have been expert systems. 2. Visual information in froth flotation control In most flotation plants it is customary for plant operators to monitor the surface froth visually and make adjustments to pro- cess parameters based on their interpretation of the surface appearance. To this end, operators categorize froths mentally into a number of distinct types and develop appropriate operating strategies for each one, looking for the ideal froth appearance. Visual information on surface froth from operating flotation plants has not been systematically catalogued. The interpretation of this information is based on intuitive experience and tends to vary from one operator to another. As a result, achieving optimal control with human operators is often not possible. 2.1. Froth features and industrial sensors Operator control decisions are based on surface properties such as froth stability, froth speed, bubble size, bubble shape, bubble 0892-6875/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.mineng.2008.10.005 * Corresponding author. E-mail addresses: fenunez@ing.puc.cl (F. Núñez), aciprian@ing.puc.cl (A. Cipriano). Minerals Engineering 22 (2009) 366–371 Contents lists available at ScienceDirect Minerals Engineering journal homepage: www.elsevier.com/locate/mineng