Premalatha Balan
2 STUDY AREA AND DATA USED
N
on the + The study area is part of Western Ghats of India and is shown in
shorter . am rw RUP, Figure 1. The area is covered between 75? 20' E to 75? 25' E
rogram D C longitude and 13? 11' N to 13? 15' N latitude. The area is covered
ope for so "4 by dense rain forest and the topography shows slopes of more
ariation — than 40%, together with some cliffs. Coffee is the common crop
as any — cultivated in plantations along the hill slopes.
at earth ‘ ; \
mn c J crown he Data acquired by the SIR-C instrument flown on the Shuttle
QE me mission during 1994 were acquired for interferometric analysis.
orbital meneur 77 ^ Optical data acquired by the PAN sensor mounted on the Indian
be the Figure 1 Location map of the study area Remote Sensing satellite IRS-1D was also used for this study. A
(ESA) 1:50,000 scale topographic map of the area was used to generate
ed with “the reference DEM from 20m interval contours. The data sets used are listed in Table 1.
Platform | Sensor Wavelength Date of acquisition
tS Rd SIR-C SAR | 23.5em(L band) | 08 October 1994
ve SIR-C | SAR |235em( band) | 09 October 1994
ation is IRS-1D PAN 0.4 um —0.7 um | 17 December 1997
NW. Asa
n along Table 1 Dataset used
r every
racking
ccurate 3 METHODOLOGY
s using
ove the 3.1 Baseline estimation using state vectors
yosition :
Interferometric processing consists of a series of steps to generate a DEM from a pair of Single Look Complex (SLC)
data. As a first step, the SLC images were co-registered to sub-pixel accuracy using one image as the reference (master)
image and the second as the slave image. A complex interferogram is generated from the co-registered SLC images by
multiplying the complex value of the master image with the complex conjugate of the slave image on a pixel by pixel
der file basis. The phase value of this complex interferogram (Figure 2) is proportional to the height of the terrain. The wrapped
stimate phase of the interferogram contains the phase due to flat earth in addition to phase component determined by the height
h using of the terrain. To remove the phase component due to flat earth, the baseline was estimated using orbital information
provided in the leader file of the data set in the form of state vectors. The resultant interferogram (the “flattened
interferogram”) is shown in figure 3(a). The flattened interferogram was then unwrapped and converted to terrain
heights by inputting a number of ground control points. This elevation image, which was in slant range geometry, was
then orthonormalised to ground geometry. The orthonormalised height image is the InSAR derived DEM.
1 initial 3.1.1 DEM comparison
vectors To evaluate the effect of any processing parameter, it is necessary
the co- to have a reference DEM, ideally the theoretical InSAR DEM. In
parallel the absence of the theoretical InSAR DEM, another DEM derived
in the from established methods could be used as the reference DEM. In
terrain this case the differences are not only due to the processing
| as the parameter but due to other factors as well. The other factors that
' if the affect the quality of the DEM are system parameters and the terrain
vindow parameters. The system parameters include wavelength, actual (not
o allow estimated) baseline separation, pixel resolution and temporal
ent flat baseline (time delay between the interferometric pair acquisition).
results. Terrain parameters include vegetation density and type,
imation meteorological conditions, and slope. If the processing parameter
iate the introduces systematic error then height differences due to system
he best parameters and the terrain parameters can be ignored. In this study,
a DEM generated from 20m contours in a 1:50,000-scale map was Figure 2 Interferogram with phase trend due
used as the reference DEM. These contours were digitised using the to flat earth
Arc/Info software package. Spot heights were also digitised. The
DEM was generated by triangulation method using the ERDAS Imagine software package.
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B1. Amsterdam 2000. 31
