XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3 421
Ideal pre-concentration means removing of particles con-
taining 0% of ore and providing this with high stability
in the severe mining conditions. However, there is always
a certain misclassification of ore particles in such pro-
cesses, which makes this sorting technology less attractive.
Therefore, the sorting process must be carried out with the
best possible precision to reduce these losses in the waste
fraction. Currently used XRT sorting technology (X-ray
transmission), often applied in coal sorting (Kolacz, 2015),
is generally based on dual energy analysis (XRT-DE), which
has a resolution of 0.4–0.8 mm. It means, potential ore par-
ticles, which have sizes under 400 µm, are not detected at all
or they are detected with much lower accuracy, providing
significant ore losses in the waste fraction. New technolo-
gies described in this paper, include other sensing principles
and give possibilities to detect fine ore particles. This pro-
vides a significant breakthrough in sorting efficiency.
SORTING TECHNOLOGY
The new sorting system, as shown in Figure 1, can be con-
figured with different sensors depending on material types
and analyzed parameters, to provide the best identification
and separation for various process capacities (Kolacz, 2016).
The image analysis is realized by a camera system
installed over the transport belt conveyor, which can
employ visible light (RGB-red green blue) or SWIR (Short
Wave Infra Red) hyperspectral cameras. The sorting system
includes the X-ray attenuation analysis carried out by the
X-ray source and the XRT sensors. The sorting system can
be used with optical RGB, SWIR and XRT analysis simul-
taneously or separately depending on an application.
The sensor data after XRT, SWIR and RGB analysis are
shown in form of images. Particle recognition used to sepa-
rate different materials is based on a shape and color analy-
sis of these images, where the objects can also be identified
by over 20 parameters used for shape description. Some
of them are: diameter in different orientations, perimeter,
center of mass, moment of inertia, particle elongation fac-
tor, edge sharpness, etc. Additional combinations of these
parameters can also be used for distinguishing particles of
interest. The surface of the particles, where different colors
or contours vary in intensity and frequency, can be ana-
lyzed by digital filters to recognize differences in texture
and structure of the particles. This method brings much
more information about the analyzed particles rather than
simple color recognition or XRT analysis. Finally, the XRT,
RGB and SWIR analysis can be combined in different
mathematical models (Kolacz, 2014). A lot of computa-
tion power is required to perform signal and data process-
ing based on combined data from these sensors especially
Figure 1. Comex CXR sorting system in multi-sensor configuration
Ideal pre-concentration means removing of particles con-
taining 0% of ore and providing this with high stability
in the severe mining conditions. However, there is always
a certain misclassification of ore particles in such pro-
cesses, which makes this sorting technology less attractive.
Therefore, the sorting process must be carried out with the
best possible precision to reduce these losses in the waste
fraction. Currently used XRT sorting technology (X-ray
transmission), often applied in coal sorting (Kolacz, 2015),
is generally based on dual energy analysis (XRT-DE), which
has a resolution of 0.4–0.8 mm. It means, potential ore par-
ticles, which have sizes under 400 µm, are not detected at all
or they are detected with much lower accuracy, providing
significant ore losses in the waste fraction. New technolo-
gies described in this paper, include other sensing principles
and give possibilities to detect fine ore particles. This pro-
vides a significant breakthrough in sorting efficiency.
SORTING TECHNOLOGY
The new sorting system, as shown in Figure 1, can be con-
figured with different sensors depending on material types
and analyzed parameters, to provide the best identification
and separation for various process capacities (Kolacz, 2016).
The image analysis is realized by a camera system
installed over the transport belt conveyor, which can
employ visible light (RGB-red green blue) or SWIR (Short
Wave Infra Red) hyperspectral cameras. The sorting system
includes the X-ray attenuation analysis carried out by the
X-ray source and the XRT sensors. The sorting system can
be used with optical RGB, SWIR and XRT analysis simul-
taneously or separately depending on an application.
The sensor data after XRT, SWIR and RGB analysis are
shown in form of images. Particle recognition used to sepa-
rate different materials is based on a shape and color analy-
sis of these images, where the objects can also be identified
by over 20 parameters used for shape description. Some
of them are: diameter in different orientations, perimeter,
center of mass, moment of inertia, particle elongation fac-
tor, edge sharpness, etc. Additional combinations of these
parameters can also be used for distinguishing particles of
interest. The surface of the particles, where different colors
or contours vary in intensity and frequency, can be ana-
lyzed by digital filters to recognize differences in texture
and structure of the particles. This method brings much
more information about the analyzed particles rather than
simple color recognition or XRT analysis. Finally, the XRT,
RGB and SWIR analysis can be combined in different
mathematical models (Kolacz, 2014). A lot of computa-
tion power is required to perform signal and data process-
ing based on combined data from these sensors especially
Figure 1. Comex CXR sorting system in multi-sensor configuration