Wim van Rossum 
  
(less than approximately 20 m apart, so as to avoid baseline decorrelation) during passes spaced in time by several 
months ("repeat pass interferometry"), the subsidence of the area will be mapped.. 
In order to obtain coherence between the two passes, the azimuth aspect angle of the two passes must be nearly the 
same. In the case of PHARUS, a difference of no more than 0.6? (30 % coherence loss) is allowed. This necessitates 
active, real-time beam steering, to compensate for aircraft attitude. This geographically locked beam steering mode 
differs from the other steering modes for PHARUS; the beam is kept pointing in a specific direction, while the usual 
beam steering mode keeps the beam at zero Doppler and the spotlight steering mode keeps the beam directed to a 
specific point on the ground. 
The experiments are on-going. Due to the relatively narrow antenna beam of PHARUS maintaining angular overlap 
between the passes proved to be difficult. For this reason a sliding spot technique is used as an alternative. In this mode 
the beam is pointed to an elongated spot (instead of the normal smaller spot). Thus, the effective observation angle is 
increased w.r.t. the normal RPI mode and thereby the chances of achieving overlap. 
4 | ASARDEMONSTRATOR 
4.1 ENVISAT's Advanced Synthetic Aperture Radar (ASAR) 
Compared to the single-channel (VV) SAR onboard of ERS-1 and ERS-2, ENVISAT's ASAR employs a number of 
new developments, which have resulted in some attractive improvements. Most importantly, instead of a passive 
radiator array, the ASAR antenna uses an active phased array which results in a large coverage of incidence angles. 
Also, the possibility to switch between polarisation channels allows scenes to be imaged with a combination of 
horizontal and vertical polarized beams. Because of the active array and the possibility to switch between channels, the 
ASAR instrument can be operated in five modes, which are mutually exclusive and can be operated on request. The five 
modes are described in Table 7. 
Table 7. ASAR operating mode parameters [ESA, 1998]. 
  
  
Image wide swath alternating Wave global 
polarisation monitoring 
Polarization HH or VV HH or VV HH/HV,HH/VV HHor VV HH or VV 
or HV/VV 
Spatial «30m « 150m «30m « 0m « 1000 m 
resolution (4 looks) (12 looks) (2 looks) (single look) (7 looks) 
Swath width < 100 km >400 km <100 km 5 x5 km 2400 km 
Looking angle — 15?-45? 17°-43° 15°-45° 15°-45° 17°-43° 
  
  
  
Because PHARUS is fully polarimetric and has a better resolution than ASAR these five modes can be simulated using 
PHARUS data. Obviously, because of its resolution of 3 m together with a swath width of about 10 km, PHARUS is 
best suited to demonstrate ASAR applications that require medium resolution. These applications include geocoding, 
stereoscopy, land-use classification and the determination of directional wave spectra. Geocoding and land-use 
classification will be discussed in more detail. 
With regard to land applications, several studies have shown that land-use classification with one single channel, e.g. 
the VV channel of the ERS satellites, is very restricted. Fully polarimetric SAR systems like PHARUS have been quite 
successful in providing detailed land-use classification results (see Figure 2). Therefore, PHARUS offers an unique 
opportunity to evaluate the possibilities of ASAR's dual-channel alternating polarization mode for land-use 
classification. Images with the characteristics of ENVISAT's ASAR operating in image mode, i.e. with a resolution of 
30 m, a number of looks of 4, have been simulated using PHARUS data. An elevated area near Freiburg (Germany) and 
an area without elevation, the Heerde test site, were classified using a Bayes type classifier. It was concluded that in 
elevated areas only fully polarimetric (three channels) images could be classified. The classification using only two 
channels gave disappointing results. For areas without relief the extension from one to two channels is profitable and 
the classifier works well with only two channels. 
  
344 International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B1. Amsterdam 2000. 
4,2 € 
With | 
Becau 
that th 
proces 
betwe 
contal 
means 
from : 
both t 
UTM 
identi 
  
The al 
8. Th 
PHAF 
geoco