Spatial Analysis of Two Different Urban
Landscapes Using Satellite Images and
Landscape Metrics
Duygu Şimşek and Elif Sertel
Abstract
This research aims to compare landscape metrics of two dif-
ferent cities having distinct landscape characteristics by using
up-to-date, very high resolution
SPOT
6/7 images-derived and
thematically extensive urban land cover/use (
LCLU
) maps
produced from object based classification approach. Object
based classification method was applied to 1.5 m resolution
SPOT
images in conjunction with some opensource geoinfor-
mation in order to create accurate urban
LCLU
maps. Land-
scape metrics were calculated using detailed land cover/use
classes and comparisons were conducted for two different city
landscapes; one having huge industrialized areas and another
with urban sites and archeological city residences which
was declared as World Cultural Heritage. Patch Density (
PD
),
Edge Density (
ED
), Largest Patch Index (
LPI
), Euclidean Near-
est Neighbor Distance Distribution (
ENN_MN
), Area-Weighted
Mean Fractal Dimension Index (
FRAC_AM
) and Contagion
(
CONTAG
) metrics were calculated for two cities using these
land cover/use maps. Analyzes of urban areas with landscape
metrics allowed objective evaluations to define the structure
of urban areas and representation of differences or similari-
ties in the spatial structure of different urban regions.
Introduction
The main objective of urban sustainability is to carefully de-
sign urban environments having minimal impact on ecosys-
tems and maintain comfortable life for citizen. Development
and implementation of socio-economic and environmental
policies that will minimize the impact of cities on local and
global environment are important for urban sustainability
(Dempsey and Jenks, 2006). Rational and efficient land use
and management are critical to ensure urban sustainability.
For the planning models to be prepared in this direction, it is
necessary to know the related regions’ past and current land
cover and land use information and their spatio-temporal
changes on the basis of urban settlements and resources
(Iyyer, 2009). Remotely sensed data has been widely used for
variety of urban applications, such as mapping and monitor-
ing of urban sprawl and urban growth trends, urban change
detection, urban utility and infrastructure planning, urban
land zoning, urban environment and impact assessment,
urban management and modeling, urban hydrology (Iyyer,
2009). Need for remote sensing data to obtain information on
built environment especially for cities in developed countries
might not be that significant because detailed information is
already available from regulatory, governmental or commer-
cial sources. However, in developing countries, remote sens-
ing applications becomes crucial due to the lack of efficient,
periodic and timely geospatial information (Miller and Small,
2003; Sertel and Akay, 2015). New developments in satellite
technology with the capability of providing very high spatial
resolution and advances in image processing methods allow
the precise determination of urban and suburban features to
characterize urban
LCLU
resulted in effective usage of these
data sets not only for urban environmental studies but also for
urban planning purposes (Xian, 2015).
There have been wide-ranging studies and projects con-
ducted on the creation of
LCLU
maps from satellite images
from local to continental scales. There are several initiatives
at European level as well; such as Coordination of Informa-
tion on the Environment (
CORINE
) and Urban Atlas Projects.
The
CORINE
project is part of the Copernicus services coor-
dinated and managed by the European Commission creating
continuous
LCLU
maps of
EU
countries for every six year pe-
riod and the most recent one is created for 2012.
CORINE
LCLU
maps contain a total of 44
LCLU
classes; 11 of these classes
contain artificial surfaces and the minimum mapping unit
(
MMU
) is 25 hectares with a 1:100 000 scale. Although
CORINE
database is very useful for country wide planning and differ-
ent country inter-comparisons of landscape characteristics, its
scale, thematic and geometric details are not sufficient for ur-
ban specific large scale applications. The Urban Atlas Project
is part of the “European World Observation Program-Coperni-
cus” funded by the European Commission with the support of
the European Regional Development Fund (ERDF) and jointly
produced by the Member States with the support of the Euro-
pean Environment Agency. Urban Atlas is available for years
2006 and 2012. 2012 data includes 0.25 ha minimum map-
ping unit for seventeen different urban related classes and
1 hectare
MMU
for ten rural classes. For the year 2012 refer-
ence year, urban atlas maps of 697 large urban areas and the
surrounding area (population more than 100,000) providing
land information were created (Copernicus, n.d). The aims of
the Urban Atlas project are to fill the information gap on land
use in European cities, to produce comparable land use data,
to facilitate the creation of more evidence-based policies, to
compare land use models among the major cities in Europe,
and thus, to compare cities in Europe and the use of satellite
imagery to cost-effectively create reliable and comparable
high-resolution maps of urban areas (European Environment
Agency, 2010). The Methodology of Urban Atlas contains,
object based classifications based on high resolution satellite
Elif Sertel is with the Civil Engineering Faculty-Geomatics
Engineering Division, Istanbul Technical University, ITU
Ayazaga Campus 34669 Maslak, Istanbul, Turkey
(
).
Duygu
Ş
im
ş
ek is with the Center for Satellite Communication
and Remote Sensing (CSCRS), Istanbul Technical University,
ITU Ayazaga Campus 34669 Maslak, Istanbul, Turkey.
Photogrammetric Engineering & Remote Sensing
Vol. 84, No. 11, November 2018, pp. 711–721.
0099-1112/18/711–721
© 2018 American Society for Photogrammetry
and Remote Sensing
doi: 10.14358/PERS.84.11.711
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
November 2018
711
