Flotation froth characterisation by optica-digital sectioning techniques

The recognition and extraction of the pictorial features characterizing froth surface is an essential step to design a machine vision system for monitoring industrial flotation cells. It is well known, in fact, as such characteristics are strongly related to the froth structure. In this perspective, segmentation techniques based on watershed and on erosion and dilation algorithms were applied to isolate the domains representative of the froth bubbles. To determine morphological and morphometrical features of the bubbles, digital froth images were initially enhanced in order to improve the possibility to extract useful data from them. Several enhancement techniques, ranging from simple operations such as brightness and contrast adjustment, to the more complex spatial and morphological filtering operations, were applied. The developed segmentation procedures were thus compared and critically evaluated with respect to different froths resulting from different plant performance condition.