In: Wagner W., Szekely, B. (eds.): ISPRS TC VII Symposium - 100 Years ISPRS, Vienna, Austria, July 5-7, 2010, IAPRS, Vol. XXXVIII, Part 7B
364
Figure 6: Colour-coded 70 models of two flight strips for the
overlapping region: (left) western flight strip and (right) eastern
flight strip.
Figure 7: Colour-coded difference models of two overlapping
flight strips for the original amplitude measurements and different
radiometric calibration values.
The colour-coded incidence angle corrected 70 model is displayed
in figure 6. It can be seen, that the surfaces of the western and the
eastern strip show similar values. However, within one strip roof
areas with surface normals pointing towards the sensor and those
with surface normals pointing away do not show similar values.
This is unexpected but seems to be a result of the unexpected
amplitude values as shown in figure 5.
In a next step, difference models of the different radiometric cali
bration values were calculated for the two overlapping flight strips
(see figure 7). The colour-coded difference image of the ampli
tude values shows the expected increase in difference towards the
borders of the overlapping swath. The differences in backscatter
cross section a already indicate similar values in the overlapping
flight strips for wide regions, especially for horizontal surfaces.
However, inclined roof surfaces still suffer from incidence angle
effects. This can particularly be seen at the huge casern building
in the centre of the difference image. The difference image of 7
shows differences of the same strength as in the difference im
age of a. The differences in cr° between the overlapping strips
on the other hand show again the amplification of the incidence
angle dependence at the inclined roof surfaces. The incidence an
gle corrected values and 70 strongly minimize the differences
for roof areas. Since the incidence angles’ estimation in vege
tated areas is uncertain and sometimes impossible, this drawback
can also be seen in the difference images of ere and 70 by strong
differences in either one or the other direction.
4 CONCLUSION
This paper presents a comparison of different radiometric calibra
tion values, the assets and drawbacks of each calibration value as
well as a quantitative comparison by analyzing difference models
of overlapping regions. The backscatter cross section a delivers
usable results especially for horizontal surfaces. However, area-
normalized values should be the aim in case measurements with
different resolution, e.g. acquired at significantly different flight
heights over ground, with significantly different beam divergence
and/or by different ALS sensors, shall be compared. Since <?o
proves to amplify the effect of the angle of incidence, 7 turns
out to be the preferred quantity of these so-called backscatter-
ing coefficients. Due to the fact that all these values still suf
fer from incidence angle dependencies, only incidence angle cor
rected values such as 07 and 70 are able to deliver homogeneous
values for a homogeneous surface. Such values can only be de
rived for echoes, where the estimation of the local surface normal
is successful, though. The analysis of the horizontal parade yard
shows that oe tends to overcorrect the incidence angle depen
dence, while 70 delivers homogeneous values for the homoge
neously reflecting parade yard. Further evaluation has to be done
especially concerning multi-temporal data, e.g. acquired by dif
ferent sensors (e.g. with a different beam divergence, laser wave
length, etc.) and/or from different flight heights. Additionally,
the unexpected behavior of the amplitude values that are visible
in the upper left part of figure 5 (c.f. section 3) will be studied in
the future.
ACKNOWLEDGEMENTS
We would like to thank Stadtvermessung Wien, Magistratsab-
teilung 41 (MA41, 2010), for providing the full-waveform ALS
data, which was used for this study. Furthermore, we would like
to thank the company Riegl Laser Measurement Systems GmbH
(RIEGL, 2010) for the development of the reflectometer, a tool
