ie mor. 
e prob. 
| works 
essing, 
remote 
‘€ to be 
y Short 
Am. soil 
ufficient 
nts are 
ping of 
be very 
ing and 
tions is 
- crops, 
conser- 
of soil 
infrared 
tion for 
rch, 68 
of soil, 
31, 
fects of 
sensor 
Remote 
_amber- 
etry En 
189. 
in solid, 
applica- 
is. Pro- 
phy vol. 
o meas 
il water 
No 21. 
ssing of 
ping. In 
Remote 
ISPRS, 
0-24. 
ion pro- 
. Photo- 
No 4 
WAVELENGTII, POLARIZATION, AND INCIDENT ANGLE AS VARIABLES IN SETTLEMENT DETECTION 
WITH SATELLITE SAR IMAGERY 
Floyd M. Henderson 
GIS Remote Sensing Laboratory 
Department of Gcography and Planning 
State University of New York at Albany 
Albany New York 12222 USA 
Klaus Herrig 
Department of Geography 
Julius-Maximillians Universitact Wuerzburg 
Am Hubland 
977074 Wuerzburg 
Germany 
Commission VII, Working Group 9 
KEY WORDS: SAR, Human Settlement, Accuracy, Polarization, Wavelength, Multispectral 
ABSTRACT 
Knowledge of the current distribution, growth, and population of human settlements is of vital importance not only for census 
estimates but also for proper resource management. This study is part of a larger project to examine the use of multispectral 
SAR, merged SAR, and merged SAR and MSS sensor systems for settlement, urban growth, and population-related data 
collection. The purpose of this paper is to report the influence of wavelength, polarization, and incient angle as variables in 
settlement visibility. SIR-C SAR imagery of an area southwest of Munich, Germany was acquired at two incident angles, three 
wavelengths, and four polarizations. In addition three multispectral SAR composite scenes were examined. Findings indicate 
that visibility improved with wavelength regardless of incident angle. The preferred polarization varied but like-polarized 
imagery was the poorest. Multispectral composites produced the highest accuracy. Overall, accuracies for the 17 data sets 
ranged from 17 to over 96 per cent. 
INTRODUCTION 
Knowledge of the current distribution, growth, and population 
of human settlements is of vital importance not only for census 
estimates but also for proper resource management. Census 
data are often out-of-date, lacking, and at times inaccurate 
using current ficld and sampling procedures. This situation is 
particularly truc for developing nations of the world where 
populations are growing thc fastest, existing data collection 
techniques are the least advanced, and where the extent, 
frequency, and impacts of migration (especially rural to urban) 
the most evident. However, the fact that developed nations are 
seeking new, faster, more frequent, and better methods of 
acquiring population data is evidence that a demand also exists 
here. The need for current information on population and 
human settlement exists at all levels of acquisition. 
Information on the patterns of settlement growth at local, 
regional, national, and global scales is also needed. For 
example, in what specific geographic areas are settlements 
being established, growing, or declining? Which settlements 
arc expanding and what existing land cover and land use are 
287 
being converted to urban activity? Answers to such questions 
are essential in assessing the current distribution of renewable 
and non-renewable resources and for proper planning for the 
future. 
Remote sensing systems are being explored to supply such 
information. "The. advantages and characteristics of radar 
imagery compared to other remote sensing systems are well 
known to the geoscience remote sensing community. The 
recent increased availability of SAR data, especially operational 
satellite systems, has been accompanied by a rise in the interest 
in SAR as a remote sensing system and its potential geoscience 
applications. This study is part of a larger project to examine 
the use of multi-spectral SAR, merged SAR, and merged SAR 
and MSS sensor systems for settlement, urban growth, and 
population-related data collection. Visibility and detection of 
settlements are the first steps in delimiting built-up areas for 
population and growth estimates and algorithm development. 
The purpose of this paper is to report on the influence of 
wavelength, polarization, and incident angle as variables in 
settlement visibility. 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996