IEEE Geoscience and Remote Sensing Magazine - June 2014 - 17

(a)

(b)

(c)

(d)

FiguRE 11. TanDEM-X DEM of Newfoundland visualized in Google Earth: (a) DEM, (b) amplitude mosaic, (c) height error map (dark color

means low height errors), (d) water Indication mask (light blue: water detected once, dark blue: water detected multiple times).
set of suitable height references. For TanDEM-X the ICESat
(Ice, Cloud and land Elevation Satellite) data are used as
ground control points. The ICESat spaceborne laser altimeter provides precise height references at a sampling distance of 170 m in along-track and a maximum across track
distance of 80 km at the equator [12]. A small subset of
selected ICESat points is used for calibration, and a much
higher number is used for validation of the final DEM
heights. According to an accuracy study [31] the standard
deviation for GCPs is below 2 m for ICESat points on flat
bare land. Secondly, tie points in the more than 3 km wide
overlap between neighboring DEM swaths are used to
ensure the relative height calibration (see Fig. 7). Equidistantly spaced areas (approximately every 1 km in azimuth)
along the overlap center are compared to derive tie point
differences as input for the DEM calibration.
Estimates for each data take offset, tilt in range, and
slope in the azimuth direction are being derived by means
of a weighted least-squares adjustment [32]. The idea is
that the heights in overlapping areas should be identical,
apart from the random noise, after the error function is
applied. The advantage of this method is that the correction parameters can be found independently from terrain
types. As the data acquisition for the DEMs is following
the global scenario presented in Section II.E above, DEM
calibration of larger blocks is triggered when all acquisitions for that area are complete.
june 2014

ieee Geoscience and remote sensing magazine

In the mosaicking, the calibration corrections are
applied to each DEM and different coverages are mosaicked
by a weighted average of the K elevation values:
K

1
hk
v 2HEM, k
h=
/ 21
k = 1 v HEM, k

k =1
K

(3)

where the weights are derived from the relative height error
v HEM, k of one certain elevation h k . Relative height error
values are also fused together with the help of the relative
height error map (HEM). After the DEM mosaicking process
an interactive quality control is performed for each mosaic.
Fig. 11 presents the main layers of the final TanDEM-X
DEM product of Newfoundland. A by-product is the socalled Water Indication Mask (WAM). The intention of this
WAM is to provide a supplementary information layer to
support later DEM editing processes rather than to generate
a global water mask. As the TanDEM-X DEM is an unedited,
interferometric product, water surfaces will show a noisy
relief, because of low coherence and/or low backscatter.
V. DEM PERFORMANCE
The requirements for the global TanDEM-X DEM are specified by the following key parameters (see Table 1): a relative
vertical accuracy below 2 m and 4 m for flat and steep terrain, respectively, and an absolute vertical accuracy below
17

Table of Contents for the Digital Edition of IEEE Geoscience and Remote Sensing Magazine - June 2014