Geoscience Reference
In-Depth Information
Entity ID: EO1H1690582013197110KF_PF2_01
Acquisition Date: 16-JUL-13
Target Path: 169
Target Row: 58
taken on 16th July 2013. As before, we need to register with the USGS
website, log on, and then download the radiometrically corrected (but not
geometrically corrected) Level 1R (L1R) product (215.3 MB), which is then
stored on the hard drive in the i le
LEO1H1690582013197110KF_1R.ZIP
.
h e
.ZIP
archive consists of a metadata i le (
.MET
), a Federal Geographic
Data Committee (FGDC) metadata i le (
.fgdc
), an HDF data set i le (
.L1R
),
and multiple auxiliary i les. h e EO-1 User's Guides provides some useful
information on the data formats of these i les (Barry 2001, Beck 2003). We
can import the data from the
EO1H1690582013197110KF.L1R
i le using
clear
HYP = hdfread('EO1H1690582013197110KF.L1R',...
'/EO1H1690582013197110KF.L1R', ...
'Index', {[1 1 1],[1 1 1],[3189 242 256]});
h e parameter
Index
is a three-element cell array,
{start,stride,edge}
,
specifying the location, range, and values to be read from the data set.
h e value of
start
specii es the position in the i le to begin reading. In our
example it starts reading at the beginning of the i le, i.e., starting from the
[1
1 1]
element. h e value of
stride
dei nes the interval between the values to
be read, which in our example is
[1 1 1]
, i.e., every element of the data set
is to be read. h e value of
edge
specii es the size of the data in the i le; in our
example the dimensions of the data are 3,189-by-242-by-256. Typing
whos
shows how the hyperspectral image is stored in the workspace:
HYP 3189x242x256 395129856 int16
h e details indicate that the image is stored as a 3,189-by-242-by-256 array,
representing a 3,189-by-256 array for each of the 242 spectral bands. h e
listing of the current variables in the workspace also gives the information
int16
array, i.e., each array element representing one pixel contains signed
16-bit integers. We need to permute the array to move the bands to the third
dimension by typing
HYP = permute(HYP,[1 3 2]);
We next need to determine the radiance values from the digital number
