April and May). Figure 4 shows the development of the
radar backscatter coefficient throughout the growing
season for the 4 most important croptypes. Although
the digital radar images have known intensity scales,
an absolute calibration lacks in these measurements.
Therefore the average backscatter coefficient of the
sugarbeet fields was determined in the images and
compared to calibrated ground based measurements,
which were always taken at the same date and in the
same area. The resulting correction factor was applied
to the whole image. The data in figure 4 is for
horizontal polarization and 15° grazing angle. The
frequency is 9.4 GHz (X-band).
Nowadays, a more direct approach to the calibration
problem of the SLAR system is available. The radar is
fitted with an internal delay line calibration. The
radar data is calibrated in the PARES preprocessing
phase (ref. 2). This algorithm takes care of the
complete geometric and radiometric correction and
calibration of the radar images. The calibration
accuracy is in the order of 1 dB or better, as was
tested with the aid of corner reflectors.
From figure 4 it can be seen that a large contrast
Figure 4. E
the growing
horizontal
exists bet
April and
while all
July a goo
types, whe
backscatte
the one fo
The larg
croptypes
low grazin
the winter
winter and
other crop
their biom
ground covi
backscatte
because tb
amount of '
leafs Stic'
ably to tb
angles say
is much im
is smaller
between tb
Thus we :
winter- am
May, and s:
wintercrop
identify a
histogram t
coefficienl
figure it :
be comp let«
by applyinj
Now that
to classify
This demons
contrast wi
(Ref. 1) wl
scatter thr
discriminat
Sofar the
■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■I
