International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004 
  
  
  
  
  
  
  
  
  
  
9 02 August 2003 UK-PAC 6 
10 11 August 2003 D-PAC 3 
11 24 August 2003 I-PAC 4 
12 15 September 2003 | UK-PAC 3 
  
Unfortunately, the image delivery from ESA's Processing and 
Archiving Facilities took several weeks, thus it was not possible 
to make near real-time monitoring as in the case of Radarsat-1 
SAR images in 2001. The Envisat SAR images were ordered in 
the precision image format, which means that the pixel's digital 
number corresponds to the amplitude of the recorded SAR 
backscattering. Firstly, all images were orthorectified into the 
Finnish uniform coordinate system using ground control points 
and image-to-image tie points. PCI geomatics software was 
used in the orthorectification. The residual error for the ground 
control points was 0.53 pixels in west-east direction and 0.82 
pixels in south-north direction (the pixel size is 12.5 meters). 
The ground control points were originally digitized from the 
Finnish 1:20 000 basic maps. The rectification accuracy was 
also visually excellent, which guaranteed that parcel specific 
information could be derived from the time series of SAR 
images. In Figure 2 there is an example of multitemporal dual- 
polarization Envisat SAR image of our test area. 
  
Figure 2. A multitemporal dual-polarization Envisat SAR image 
from the Seinäjoki test area in summer 2003. 
Original data © European Space Agency. Processed 
by FGI. 
After orthorectification, a pixel value corresponds to the 
amplitude of SAR backscattering. In order to be able to 
compare images from different dates, amplitude must be 
converted into the backscattering coefficient, which is usually 
given in decibels (dB). Backscattering coefficient was 
calculated using following formula (ESA, 2002): 
PixelValue)? . |. 
o o e vate i) (1) 
where ois backscattering coefficient (dB) 
PixelValue is a digital number of a pixel 
K is calibration constant 
i is incidence angle 
Calibration constant K varies from image to image and it can be 
found from the header information of the each image file. 
Incidence angle also varies from image to image, but also 
between each parcel, so incidence angle values were calculated 
for each parcel in every SAR image. 
The true parcel boundaries for calculating SAR backscattering 
time series were obtained from the Land Parcel Identification 
System of Finland. Totally about 40000 base parcels were 
acquired from the Ministry of Agriculture and Forestry. Some 
of the base parcels can have one or more smaller parcels, so 
firstly we selected only those parcels that contained one base 
parcel i.e. one crop species and that had an area over | ha. This 
selection had 5571 parcels and for these an average SAR 
backscattering in VV and VH polarization was calculated. 
Averaging of the SAR image pixels inside a parcel was needed 
because of the SAR speckle. Finally, a time series of average 
SAR backscattering was acquired for major crop species in 
Seinüjoki area in Finland in 2003. The backscattering time 
series for test parcels are represented in Figure 3. 
VV (Envisat 2003) 
  
  
  
  
  
4 
8-Jun-03 3-Jul-03 28-Jul-03 22-Aug-03 16-Sep-03 
6 
-8 ay N a I 
ART | f 
a. Mh\ / Z [ | 
2 V i X zy i | 
> VV zi i I 
= HN 
© 2 lt X TT T, P | 
5712 ED | | T7 
o Y l^ oon 
S M T NP 
S 14 | x 
L 
-16 
-18 
-20 : 
Date 
VH (Envisat 2003) Spring w heat 
4 a Autumn rye 
8-Jun-03 3-Jul-03 28-Ju-03 22-Aug-03 Bar 
- Barley 
-6 
-- Barley (malt) 
gi a | Oats 
Potato 
© ; 
5 -10 Turnip rape 
T Grass 
= -12 Grass (silage) 
2 | CAP fallow 
8 A fe es } zt MM EEE 
WA ALA Ped 
8 | fi LAS t > 
. y 1 
-20 i. = 
Date 
Figure 3. Time series of SAR backscattering in VV and VH 
polarizations for major crop species in 2003 in 
Finland. 
The time series of SAR backscattering were then compared to 
the information acquired in field surveys in order to have more 
detailed information about temporal variation of backscattering 
against cultivation practises and crop growth. 
Internationc 
PU 
2.2 Field s 
In 2003 fic 
surveys in | 
surveys is 1 
SAR back: 
window, wl 
image acqui 
more detail 
following ir 
= 
u 
= 
=» 
= 
Soil surfac 
distometer 
and after th 
growing se 
ThetaKit T 
measures v 
height was 
damages W 
surveys. Tii 
parcels are 
Volumetric soil moisture % 
> 
© 
  
  
8-Jun-03 
  
Figure 4. \ 
3.1 Envis: 
In Figure 3 
series in \ 
plants in t 
obtained fc 
2001 (Kar 
2001 and 2 
weather co 
the usabil 
agricultura 
In the be 
ploughed 
consequent