Geoscience Reference
In-Depth Information
about lithology (including soils), vegetation and water on the earth's surface.
h ese bands are therefore usually combined into 24-bit RGB images. We i rst
read the data
I1 = hdfread(filename,'VNIR_Band3N','Fields','ImageData');
I2 = hdfread(filename,'VNIR_Band2','Fields','ImageData');
I3 = hdfread(filename,'VNIR_Band1','Fields','ImageData');
h ese commands generate three 8-bit image arrays, each representing the
intensity within a certain infrared (IR) frequency band of a 4200-by-4100
pixel image. We are not using the data for quantitative analyses and therefore
do not need to convert the digital number (DN) values into radiance and
rel ectance values. h e
ASTER User Handbook
provides the necessary
information on these conversions (Abrams and Hook 2002). Instead, we will
process the ASTER image to create a georeferenced RGB composite of bands
3N, 2 and 1, to be used in i eldwork. We i rst use a contrast-limited adaptive
histogram equalization method to enhance the contrast in the image by
typing
I1 = adapthisteq(I1);
I2 = adapthisteq(I2);
I3 = adapthisteq(I3);
and then concatenate the result to a 24-bit RGB image using
cat
.
naivasha_rgb = cat(3,I1,I2,I3);
As with the previous examples, the 4200-by-4100-by-3 array can now be
displayed using
imshow(naivasha_rgb,'InitialMagnification',10)
We set the initial magnii cation of this very large image to 10%. MATLAB
scales images to i t the computer screen. Exporting the processed image
from the Figure Window, we only save the image at the monitor's resolution.
To obtain an image at a higher resolution, we use the command
imwrite(naivasha_rgb,'naivasha.tif','tif')
h is command saves the RGB composite as a TIFF-i le
naivasha.tif
(ca. 52
MB) in the working directory, which can then be processed using other
sot ware such as Adobe Photoshop. h e processed ASTER image does not
yet have a coordinate system and therefore needs to be tied to a geographical
reference frame (
georeferencing
). h e HDF browser
hdftool('naivasha.hdf')
