XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3 1161
channel, intending to capture the average color and the
color gradient for each grain. The extreme clusters (0 and
3) correspond respectively to the darkest and lightest end of
the color spectrum, with median pixel value for each color
channel (Red, Green, Blue) and color gradient very distinct.
Clusters 1 and 2 seem to be a mix of maroon, orange,
yellow and even dark brown grains. Cluster 2 has more
grains with “white spots” corresponding to light reflection,
creating a larger color gradient than grains in the cluster 1.
Since the classification criteria is an average of median pixel
value and color gradients this means that grains that are the
Figure 7. Results of the segmentation algorithm from the original photograph (A) points_per_side=15 (B) points_per_side=33
(C) and points_per_side=50
(a) (b)
(c) (d)
Figure 8. Classification using k-mean algorithm results for a blend of ilmenite and zircon products
channel, intending to capture the average color and the
color gradient for each grain. The extreme clusters (0 and
3) correspond respectively to the darkest and lightest end of
the color spectrum, with median pixel value for each color
channel (Red, Green, Blue) and color gradient very distinct.
Clusters 1 and 2 seem to be a mix of maroon, orange,
yellow and even dark brown grains. Cluster 2 has more
grains with “white spots” corresponding to light reflection,
creating a larger color gradient than grains in the cluster 1.
Since the classification criteria is an average of median pixel
value and color gradients this means that grains that are the
Figure 7. Results of the segmentation algorithm from the original photograph (A) points_per_side=15 (B) points_per_side=33
(C) and points_per_side=50
(a) (b)
(c) (d)
Figure 8. Classification using k-mean algorithm results for a blend of ilmenite and zircon products