THE USE OF ENVISAT ALTERNATING POLARIZATION SAR IMAGES IN
AGRICULTURAL MONITORING IN COMPARISON WITH RADARSAT-1 SAR IMAGES
Mika Karjalainen ", Harri Kaartinen *, Juha Hyyppà *, Heikki Laurila ^, Risto Kuittinen “
* Finnish Geodetic Institute, Geodeetinrinne 2, 02430 Masala, FINLAND (Mika.Karjalainen, Harri.Kaartinen,
Juha.Hyyppa or Risto.Kuittinen) @fg1.fi
b University of Helsinki, Faculty of Agriculture and Forestry, Heikki.Laurila@helsinki.fi
Commission VII, WG VII/2
KEY WORDS: Agriculture, SAR, polarization, satellite, monitoring
ABSTRACT:
In this paper Envisat alternating polarization SAR images were used in the agricultural monitoring and yield damage assessment. The
main advantage of using synthetic aperture radar (SAR) is that satellite images can be acquired frequently at user-specified times
from the target area, even through clouds. Consequently, agriculture has been considered as one of the most promising civilian
application areas of SAR imagery. Our test area is located in the Western Finland near the city of Seinäjoki. The test area is one of
the northernmost consistent agricultural areas in the world and it is situated approximately at the latitude of 63° north. The main food
crops are barley and oats, however, small areas of wheat and rye are also cultivated. A corresponding agricultural monitoring project
was carried out in summer 2001 in the same test area using Radarsat-1 Fine beam SAR images, therefore, a comparison of the
usability of Envisat ASAR and Radarsat-1 SAR images in the agricultural monitoring can be made. In order to carry out the research,
altogether 16 Envisat ASAR images were requested from the summer 2003. The image request was possible in the framework of
ESA’s Envisat announcement of opportunity (AOE-488). In order to collect reference data, ground survey campaigns were organized
for the selected set of test parcels simultaneously with each image acquisition. The ground surveys consisted of measurements of soil
roughness, soil moisture and crop height, as well as general notes about growing stage and possible yield damages caused by drought
or lodging. First results of the suitability of Envisat SAR images in the agricultural monitoring will be presented, as well as a
projection of the usability of SAR images in the yield estimation will be made.
1. INTRODUCTION
Synthetic aperture radar (SAR) is an active imaging instrument,
i.e. SAR sends a pulse of electromagnetic radiation and then
records the amplitude and phase of the radiation coming back
from the target. The backscattering coefficient, 0°, is a measure
describing the strength of the recorded radar signals from the
target per unit area. Advantage over the optical satellite images,
such as Landsat and SPOT, is that SAR uses cloud-penetrating
microwaves having wavelength of few centimetres to even
metres. Thus, it is possible to have satellite SAR images from
the target at user-specified times, which is important in
agricultural monitoring where satellite images are needed
regularly and the time window for image acquisition is narrow.
Although it is evident that cloud-penetrating SAR has great
potential in agricultural remote sensing, the exploitation of the
SAR backscattering, for example, in crop yield estimation is
still non-existent. Instead, optical satellite images have channels
revealing information of the photosynthetically active radiation
of the vegetation, and thus, there is a well-established
connection between satellite information and vegetation
biomass.
There have been several agricultural studies concerning the
temporal change of the SAR backscattering from agricultural
fields during a growing season (ESA, 1995), but the estimation
of the crop yield or vegetation biomass has proven to be a very
difficult and is still an unresolved problem. According to
previous studies the most useful frequency range for crop
biomass estimation would be the C-band, i.e. the wavelength of
132
about 5 cm, which is comparable with the size of the crop
leaves and stems. Skriver et al. (1999) found out that at the end
of the growing season C-band backscattering was dominated by
volume scattering from crop vegetation. Brown et al. (2003)
proposed that HH-VV amplitude difference of the
backscattering in C-band could be a good measure for
estimating the biomass of the crops. There have also been
promising results of the use of repeat-pass SAR interferometric
coherence with one day offset for the vegetation biomass
estimation, but at the moment there are no suitable SAR
satellite systems available for this purpose (Blaes et.al., 2003).
In general, crop yield estimation using remote sensing is an
inverse problem (Ulaby, 1998), which means that recorded
SAR backscattering is a function of several physical properties
such as soil moisture, soil surface roughness, vegetation
biomass, vegetation moisture, crop species, land slope and seed
row direction. In crop yield estimation one would like to
estimate the vegetation biomass, but its inversion from the
recorded SAR backscattering is very complicated since other
parameters are usually unknown. On the other hand, direct
problem solving (Ulaby, 1998), where the SAR backscattering
is modelled from the actual physical parameters (simulation), is
still needed to find optimal SAR parameters (wavelength.
polarization and look angle) for biomass estimation. Simulation
is also needed to gain understanding of complex scattering
mechanism of the microwaves from crop vegetation.
The Finnish Geodetic Institute (FGI) has conducted research in
the field of agricultural remote sensing since the beginning of
1990's. The early studies using CGMS (Crop Growth
Internation
ME GEI
Monitoring
based on w
low-resolu!
almost con
in Finland
vivid rese:
SAR image
the ERS-]
continuous
agricultura
species cla
assessment
In 2001 F
Seinäjoki
potential o
images in
yield estim
the growin
images We
Satellite S
agricultura
simultaneo
were prom
cultivation
caused onl
damages c
cases. Con
| is not st
crops, but
surface roi
detection €
April is ref
Figure 1.
In 2003 ıh
were obt
announcen
Envisat S/
which me
