Image analysis techniques combined with fractal measurements were used to quantify textural patterns (size and shape) of minerals in base metal ores. The zinc mineral sphalerite is the mineral to be liberated through grinding and floatation. Pyrite (FeS2), however, because of its hardness, abundance, and control on grinding efficiencies was the mineral of interest. Pyrite textural patterns vary from isolated cubes (ten micrometres across) to clustered cubes, masses of pyrite separated by small veinlets of other minerals, to irregular or framboidal shapes. Quantitative analysis of these textural patterns by means of image analysis could prove useful in furthering the development of liberation models.
The images used in this study were generated by an image analyzer with input from a backscattered image. The image analysis routines were sequential erosion, dilation and perimeter measurements. Fractal analysis entailed generating a Richardson plot of the total field perimeter versus steplength (the pixel size of one erosion step). Studies with sequential erosion depicted a size entity of the ore that reflected the smallest size unit to which the pyrite could be ground. Sequential dilation revealed a spacing of the pyrite grains and, hence, the size of the more valuable interstitial sphalerite. Plots of the field perimeters with sequential erosion/dilation cycles revealed different “slopes” for the textural patterns of pyrite. Recasting the data to show the amount of roughness lost for each erosion/dilation lead to the “liberationgram” which indicates the amount of material lost to tailings when ground to a certain size.