89
on September 28, 2007, was used in this study. ASTER sensor provides seasonal coverage of
the overall weight to a spatial resolution of 15 m for the VNIR, of 30 m for the SWIR (Short
Waves InfraRed), and 90 m for the TIR (Thermal InfraRed). All images were georeferenced
to UTM zone 32 North projection with WGS84 datum. An image subset was derived after
that for VNIR and SWIR images and was radiometrically normalized using the Internal Average Relative Reflection (IARR) to estimate surface spectral reflectance. This methodology was
previously well presented and introduced as an efficient method for similar areas (Hosseinjani
et al., 2005; Hosseinjani et al., 2011; Guha et al., 2018). The TIR image was calibrated using
in-scene-atmospheric correction (ISAC) method (Guha et al., 2015) and an emissivity normalization method was after that implemented for deriving ASTER based emissivity (Guha, 2014).
Matched Filtering MF algorithms were implemented on the normalized three VNIR,
SWIR and TIR data using image derived spectra in order to map phosphate outcrop for ore
exploration. The principle of the MF algorithm consists on the maximization of the response
of the known end member and the suppression of the response of the background materials by mathematically projecting each pixel vector onto a subspace, which is orthogonal to
the background spectra (Liu, 2005). All endmembers were selected using PCA (principal
component analysis) statistical analysis (Theseira, 2003). Resulting maps were compared to
delineate phosphate outcrop and their abundances.
4 RESULTS AND DISCUSSION
4.1 Phosphate ore characterization
The depiction of X-ray diffractograms, resulting from XRD analysis, revealed the mineralogy
of the different samples. All samples have shown the presence of both calcite (CaCO 3 ) and
dolomite (Ca Mg (CO 3 )
2 ) carbonates, as dominant minerals, with important concentrations.
Figure 3. Methodological flowchart.
1111
Atmospherically
c
Quartering
& geometrically
;;;
...
corrections
"
u
0
..
a.
olJ
1111
c
XRD analysis
;;;
Phosphate
...
" u
mapping:
0
..
ASD
Mfmethod
a.
.,
reflectance
..
a.
measurements
c
olJ 0
VNIR, SWIR & TIR
... ·Mineralogical
.. ..
-
ell
based phosphate
::s.,
composition
... -
.,_
maps
0::: ell
>
on September 28, 2007, was used in this study. ASTER sensor provides seasonal coverage of
the overall weight to a spatial resolution of 15 m for the VNIR, of 30 m for the SWIR (Short
Waves InfraRed), and 90 m for the TIR (Thermal InfraRed). All images were georeferenced
to UTM zone 32 North projection with WGS84 datum. An image subset was derived after
that for VNIR and SWIR images and was radiometrically normalized using the Internal Average Relative Reflection (IARR) to estimate surface spectral reflectance. This methodology was
previously well presented and introduced as an efficient method for similar areas (Hosseinjani
et al., 2005; Hosseinjani et al., 2011; Guha et al., 2018). The TIR image was calibrated using
in-scene-atmospheric correction (ISAC) method (Guha et al., 2015) and an emissivity normalization method was after that implemented for deriving ASTER based emissivity (Guha, 2014).
Matched Filtering MF algorithms were implemented on the normalized three VNIR,
SWIR and TIR data using image derived spectra in order to map phosphate outcrop for ore
exploration. The principle of the MF algorithm consists on the maximization of the response
of the known end member and the suppression of the response of the background materials by mathematically projecting each pixel vector onto a subspace, which is orthogonal to
the background spectra (Liu, 2005). All endmembers were selected using PCA (principal
component analysis) statistical analysis (Theseira, 2003). Resulting maps were compared to
delineate phosphate outcrop and their abundances.
4 RESULTS AND DISCUSSION
4.1 Phosphate ore characterization
The depiction of X-ray diffractograms, resulting from XRD analysis, revealed the mineralogy
of the different samples. All samples have shown the presence of both calcite (CaCO 3 ) and
dolomite (Ca Mg (CO 3 )
2 ) carbonates, as dominant minerals, with important concentrations.
Figure 3. Methodological flowchart.
1111
Atmospherically
c
Quartering
& geometrically
;;;
...
corrections
"
u
0
..
a.
olJ
1111
c
XRD analysis
;;;
Phosphate
...
" u
mapping:
0
..
ASD
Mfmethod
a.
.,
reflectance
..
a.
measurements
c
olJ 0
VNIR, SWIR & TIR
... ·Mineralogical
.. ..
-
ell
based phosphate
::s.,
composition
... -
.,_
maps
0::: ell
>