XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3 1491
at higher speeds could be beneficial (cataracting regime).
The generation of fine particles (–38 µm) while the remain-
ing of coarse particles (+100 µm) suggest that the mill was
operating in cascading regime. The optimal milling condi-
tions regarding the produced PSD was obtained for 5 kg
milled for 30 minutes. However, this set up produce the
higher number of fine particles (–38 µm), indicating that
a possible desliming stage would be required previous the
flotation in the industrial application.
Figure 3 shows an example of the digital image process-
ing using the ImageJ. It is possible to see that the software is
cable of recognize, identify, and label the mineral particles
present in the image. The Figure 3a shows why flotation is
necessary for this ore, as a gangue particle (white particle)
can be seen in the right-hand side of the image and a mixed
chromite particle can be seen in the left-hand side (dark
particle).
Figure 4 shows digital images of the products of 30
minutes milling before and after IPIA. As expect the process
is not perfect and particles close to each other can be seen
by the algorithm as one, a common problem also known as
clustering (see Figures 6b, 6d, and 6e). More work is needed
to increase the algorithm efficiency recognizing the shapes
and patterns of the particles, leading to a better segmenta-
tion of the images. At the moment, the clusters are going
to be seen as one particle. Therefore, a filter was added to
the algorithm to remove very small or large very particles,
always considering the sieve used for the classification of
the particles. Therefore, it is not recommended to apply
this methodology to particles without previous sieving.
The Figure 5 shows a visual chart for estimating par-
ticles morphological features as proposed by Krumbein and
Schloss (1963). Although the definitions of roundness and
sphericity has changed over the years this chart shows that a
lower Krumbein number indicates a more angular particle,
which is exactly the feature required to increase the favor-
able conditions for floatation to happen. Figure 6 presents
an example of the results obtained at the end of the digital
image processing for the chromite concentrate particles. To
scale the particles all three are presented with the same size
for the higher dimension (width or height). Three particles
are presenting have different values of roundness (R) and
circularity (c). The morphological figures for these particles
were C= 0.70 and R =0.70, C =0.80 and R =0.81, and C
=0.77 and R =0.75, for Figures 6a, 6b and 6c, respectively.
It is possible to notice that the particles are similar to the
chart proposed by Krumbein and Schloss (1963).
Figure 7 presents the morphological results obtained
after the milling tests. Although a slightly change had
occurred with the shape of the particles, most of the results
remained in the same category proposed by Krumbein and
Schloss (1963). However, it is possible to notice that the
milling of 10 kg for 30 min produced the better results
for the particles shape. On one hand the particles size
decreased less than other tests, but on the other hand the
morphology shifted to the lower part of the feed distribu-
tion. This may indicate that the particles broke producing
shapes more angular. The milling of 5 kg for 30 minutes
lead to an accentuated production of elongated fine par-
ticles and rounded coarse particles, corroborating the PSD
and in agreement with the hypothesis that the mill oper-
ated in cascading regime.
CONCLUSIONS
A chromite concentrate sample were submitted to experi-
ments in order to correlate the milling conditions and the
morphological features of the milled product. Although a
milling condition that produces a concentrate with d75 of
100 µm could not be found, it was possible to notice the
Figure 3. Digital image before (a) and after (b) processing
at higher speeds could be beneficial (cataracting regime).
The generation of fine particles (–38 µm) while the remain-
ing of coarse particles (+100 µm) suggest that the mill was
operating in cascading regime. The optimal milling condi-
tions regarding the produced PSD was obtained for 5 kg
milled for 30 minutes. However, this set up produce the
higher number of fine particles (–38 µm), indicating that
a possible desliming stage would be required previous the
flotation in the industrial application.
Figure 3 shows an example of the digital image process-
ing using the ImageJ. It is possible to see that the software is
cable of recognize, identify, and label the mineral particles
present in the image. The Figure 3a shows why flotation is
necessary for this ore, as a gangue particle (white particle)
can be seen in the right-hand side of the image and a mixed
chromite particle can be seen in the left-hand side (dark
particle).
Figure 4 shows digital images of the products of 30
minutes milling before and after IPIA. As expect the process
is not perfect and particles close to each other can be seen
by the algorithm as one, a common problem also known as
clustering (see Figures 6b, 6d, and 6e). More work is needed
to increase the algorithm efficiency recognizing the shapes
and patterns of the particles, leading to a better segmenta-
tion of the images. At the moment, the clusters are going
to be seen as one particle. Therefore, a filter was added to
the algorithm to remove very small or large very particles,
always considering the sieve used for the classification of
the particles. Therefore, it is not recommended to apply
this methodology to particles without previous sieving.
The Figure 5 shows a visual chart for estimating par-
ticles morphological features as proposed by Krumbein and
Schloss (1963). Although the definitions of roundness and
sphericity has changed over the years this chart shows that a
lower Krumbein number indicates a more angular particle,
which is exactly the feature required to increase the favor-
able conditions for floatation to happen. Figure 6 presents
an example of the results obtained at the end of the digital
image processing for the chromite concentrate particles. To
scale the particles all three are presented with the same size
for the higher dimension (width or height). Three particles
are presenting have different values of roundness (R) and
circularity (c). The morphological figures for these particles
were C= 0.70 and R =0.70, C =0.80 and R =0.81, and C
=0.77 and R =0.75, for Figures 6a, 6b and 6c, respectively.
It is possible to notice that the particles are similar to the
chart proposed by Krumbein and Schloss (1963).
Figure 7 presents the morphological results obtained
after the milling tests. Although a slightly change had
occurred with the shape of the particles, most of the results
remained in the same category proposed by Krumbein and
Schloss (1963). However, it is possible to notice that the
milling of 10 kg for 30 min produced the better results
for the particles shape. On one hand the particles size
decreased less than other tests, but on the other hand the
morphology shifted to the lower part of the feed distribu-
tion. This may indicate that the particles broke producing
shapes more angular. The milling of 5 kg for 30 minutes
lead to an accentuated production of elongated fine par-
ticles and rounded coarse particles, corroborating the PSD
and in agreement with the hypothesis that the mill oper-
ated in cascading regime.
CONCLUSIONS
A chromite concentrate sample were submitted to experi-
ments in order to correlate the milling conditions and the
morphological features of the milled product. Although a
milling condition that produces a concentrate with d75 of
100 µm could not be found, it was possible to notice the
Figure 3. Digital image before (a) and after (b) processing