Blue tab
Green tab
Brown Tab

Preliminary results (examples)

 

 

Test site 2. Witbank - WV2 Iron oxide and jarosite mapping

 

The Witbank WV2 11Dec07084502 image was atmospheric corrected using ACORN. Henceforth are preliminary results of mapping the iron oxides (Goethite & Hematite) and Jarosite concentrations in the soil using some ASD spectra from the 02.2011 Witbank field campaign data (Fig. 1).

ASD spectra data graphs
Fig. 1- (A) The ASD spectra from the 02.2011 Witbank field campaign data, (B) the ASD spectra after resampling to WV2 spectral configuration, (C) The continuum removal values of the ASD spectra (resampled to WV2 spectral configuration).

Iron oxides

Using the WV2 spectral configuration Goethite and Hematite show a clear absorption in the continuum removal spectra, at 478nm for Goethite and at 545nm for Hematite (Fig. 2).

ASD spectrometer graph
Fig. 2- The continuum removal spectra of Goethite (purple line) and Hematite (black line) as measured with the ASD spectrometer and resampled to WV2 spectral configuration.

The Continuum removal method was applied to the WV2 reflectance image resulting with a continuum removal image. The relative concentration of Goethite and Hematite in each pixel in the image was derived from the relative intensity of the continuum removal value at 478 and 545nm, respectively. Some examples are shown in Figs. 3-6.

WV2 reflectance image results
Fig. 3- (A) The location of two pixels (marked with X) in an image subset, (B) the concentration of Goethite in the soil within the subset, (C) the concentration of Hematite in the soil, (D) the continuum removal spectra of the two pixels (purple of Goethite and black of Hematite).
 
WV2 reflectance image results
Fig. 4- (A) The location of two pixels (marked with X) in an image subset, (B) the concentration of Goethite in the soil within the subset, (C) the concentration of Hematite in the soil, (D) the continuum removal spectra of the two pixels (purple of Goethite and black of Hematite).
 
WV2 reflectance image results
Fig. 5- (A) The location of a pixel (marked with X) in an image subset, (B) the concentration of Goethite in the soil within the subset, (C) the continuum removal spectrum of the Goethite pixel.
 
WV2 reflectance image results
Fig. 6- (A) The location of a pixel (marked with X) in an image subset, (B) the concentration of Hematite in the soil within the subset, (C) the continuum removal spectrum of the Hematite pixel.

Using another spectral manipulation approach, a spectral first derivative of the WV2 spectral configuration, was able to distinguish Jarosite from other minerals, as seen in Fig. 7.

WV2 reflectance image results
Fig. 7- The first derivative spectra of the ASD spectra from Fig. 1.

One can see that the two TDBA Jarosite first derivative spectra are distinguishable from the other spectra. Thus, we used the first derivative method to identify Jarosite in the image.

TOP

Jarosite

A first derivative algorithm (IDL) was applied to the reflectance image resulting with a first derivative image. Then the Spectral Angle Mapping algorithm was applied to this first derivative image, using the first derivative ASD spectrum of Jarosite (TDBA2 in Fig. 7) as the endmember (reference) spectrum. A major Jarosite coverage area was identify (Fig. 8).

WV2 reflectance image results
Fig. 8- (A) The location of a pixel (marked with X) in an image subset, (B) the concentration of Jarosite in the soil within the subset, (C) the reflectance spectrum of Jarosite (green as measured with the ASD spectrometer and magenta for the pixel in the reflectance image), (D) the first derivative spectrum (green for the ASD's spectrum and magenta for the pixel in the first derivative image).


Conclusions so far

It seems that the WorldView2 data can be used to map iron oxide minerals. This however needs validation, either from experts who recognize the area or from detailed georeferenced ground control data. Geomorphological, soil and geological maps could be of help to validate and\or improve the atmospheric correction and to analyze the data and mapping other minerals. For example, many pixels in the image have a spectrum similar to the one shown in Fig. 9. Is it possible that this is due to biogenic content in the soil as demonstrated in Fig. 10? Again, this needs a detailed validation mission. Maybe the next RSA campaign can focus on this issue….

WV2 reflectance image results
Fig. 9- (A) The location of a pixel (marked with X) in an image subset, (B) the reflectance spectrum of the pixel.
 
WV2 reflectance image results
Fig. 10- The ASD spectra of Kaolinite (cyan), Biogenic crust (black), spectral linear mixing of Kaolinite and Biogenic crust (blue) and an image pixel's spectrum (red).

TOP