Premalatha Balan 
  
  
   
  
  
  
  
timates The DEM generated 
imates. == using the fringe rate 
nce the method to estimate 
om the the perpendicular 
d from 2 baseline component 
s and orbital 
2 information used to 
estimate the parallel 
component showed 
d DEM EN " vum Ade som Sue less tilt compared to 
on the E (2) a Pi ®) e the previous two 
d Using Figure 6 Profile along the tilt of the DEM generated using baseline estimate based on Cases. Also the datum 
ing sx image offset parameters. (a) Profile location and (b) the profile along the tilt of the InSAR DEM 4s 
ograms ; above the level of the 
gh they datum of the reference DEM which is evident from Figure 6(b) where the profile of the difference image starts with a 
nber of negative value. If we look at the DEM cross section shown in Figure 8(c) the height value of the InSAR DEM in the 
ns and west-most location is just 100m below the height value of the reference DEM whereas the east-most location has a 
grams, height value almost equal to the height value of the reference DEM. The difference image still shows a considerable tilt 
stimate (Figure 6(b)), and the profile exhibits an increasing trend from top to bottom of the image (Figure 5(a)) which is shown 
(holes) as left to right in the profile in Figure 6(b). In this case also the shape of the DEM profile agrees well with the shape of 
e to an the profile of the reference DEM, when the presence of noise and holes are ignored. 
eported 
DEMs The DEM generated 
tracted using the fringe rate 
ifferent method to estimate 
values the perpendicular 
nce the component and : 
and for image offset | == 3 
! Shown parameters to = 3 
estimate the parallel 
component of the | = : 
oM ee baseline showed less | iR $ 2800 "n m 8008 
    
  
  
  
tilt when compared == ss Distance (meters) 
to all other DEMs, (a) Hu (b) 
though there is still Figure 5 Profile along the tilt of the DEM generated using baseline estimate based on 
some tilt in the fringe rate method to estimate the perpendicular component and orbital information to 
difference image estimate the parallel component of the baseline. (a) Profile location and (b) the profile 
(Figure 7(b). The along the tilt 
  
  
cross section of the 
InSAR DEM shows a good agreement with the cross section of the reference DEM. The west-most height value in 
8000 Figure 8(d) is just in the level of the reference DEM height value in Figure 8(e), whereas the east-most height value 
little higher than the reference DEM height value (Figure 8(d) and Figure 8(e) respectively). The shape of the InSAR 
DEM shows as very good agreement with the shape of the reference DEM even for the smaller undulations in the DEM 
  
  
  
  
      
n cross section. The tilt still present in this DEM can be removed by many ways. One of them can be the iterative 
methods of locating the flat window for the fringe rate method. If the fringe estimation window is not perfectly flat, 
ist- west ur ce which is the case of 
own in = the most of the “real 
R DEM world” situations, the 
st-most estimated baseline 
| agrees 2 would produce either 
z 2 an under-estimated or 
= i an over-estimated 
thwest- value, depending 
rthwest ; em = ye mt upon the direction of 
d using Distance (meters) the slope within the 
jare the (b) fringe rate estimation 
er level Figure 7 Profile along the tilt of the DEM generated using baseline estimate based on window, with respect 
ference fringe rate method to estimate the perpendicular component and image offset parametersto | tO the look direction 
his case estimate the parallel component of the baseline. (a) Profile location and (b) the profile of the SAR systems. 
revious along the tilt Another method can 
  
  
  
  
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B1. Amsterdam 2000. 35