Computers & Geosciences 63 (2014) 3443

Contents lists available at ScienceDirect

Computers & Geosciences

journal homepage: www.elsevier.com/locate/cageo

An easy to use ArcMap based texture analysis program for extraction

of ooded areas from TerraSAR-X satellite image
Biswajeet Pradhan a,n, Ulrike Hagemann b, Mahyat Shafapour Tehrany a, Nikolas Prechtel b
a
b

Faculty of Engineering, Department of Civil Engineering, University Putra Malaysia, Serdang, Malaysia
Institute for Cartography, Faculty of Forestry, Geo and Hydro-Science, Dresden University of Technology, 01062 Dresden, Germany

art ic l e i nf o

a b s t r a c t

Article history:
Received 27 May 2013
Received in revised form
13 October 2013
Accepted 17 October 2013
Available online 6 November 2013

Extraction of the ooded areas from synthetic aperture radar (SAR) and especially TerraSAR-X data is one
of the most challenging tasks in the ood management and planning. SAR data due to its high spatial
resolution and its capability of all weather conditions makes a proper choice for tropical countries.
Texture is considered as an effective factor in distinguishing the classes especially in SAR imagery which
records the backscatters that carry information of kind, direction, heterogeneity and relationship of the
features. This paper put forward a computer program for texture analysis for high resolution radar data.
Texture analysis program is introduced and discussed using the gray-level co-occurrence matrix (GLCM).
To demonstrate the ability and correctness of this program, a test subset of TerraSAR-X imagery from
Terengganu area, Malaysia was analyzed and pixel-based and object-based classication were attempted.
The thematic maps derived by pixel-based method could not achieve acceptable visual interpretation
and for that reason no accuracy assessment was performed on them. The overall accuracy achieved by
object-based method was 83.63% with kappa coefcient of 0.8. Results on image texture classication
showed that the proposed program is capable for texture analysis in TerraSAR-X image and the obtained
textural analysis resulted in high classication accuracy. The proposed texture analysis program can be
used in many applications such as land use/cover (LULC) mapping, hazard studies and many other
applications.
& 2013 Elsevier Ltd. All rights reserved.

Keywords:
Texture analysis
Feature extraction
Remote sensing
TerraSAR-X
Pixel based
Malaysia

1. Introduction
Texture is considered as an important characteristic in image
processing which is useful in radar remote sensing and other elds
where it is necessary to interpret gray value images like in the
medical sector (Treitz et al., 1996; Mahmoud et al., 2011). Due to the
surface properties such as roughness, humidity and orientation
every object in SAR scene can have its own unique backscattering
properties (Haack and Bechdol, 2000). The recognition of texture
and object classication are the most challenging problems in the
remotely sensed data processing (Zhang et al., 2007; Hamedianfar
and Shafri, 2013). The main aim in image processing is to convert
the remote sensing (RS) imagery information into tangible information which can be understandable and possibly be used in combination with other data in Geographic Information System (GIS)
environment (Blaschke, 2010). Therefore precision of the data and
the texture analysis method are two main factors that have direct
impact on the level of accuracy and information that can be
achieved (Chica-Olmo and Abarca-Hernandez, 2000).

Corresponding author. Tel.: 60 3 89466383.

E-mail addresses: biswajeet24@gmail.com, biswajeet@lycos.com (B. Pradhan).

0098-3004/$ - see front matter & 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.cageo.2013.10.011

SAR images such as TerraSAR-X with high spatial resolution are

optimal solution for texture analysis from which meaningful
texture parameters can be deduced (DLR-Deutsches Zentrum fr
Luft- und Raumfahrt, 2009; Sousa et al., 2013). TerraSAR-X satellite
is Germany's rst active space borne remote sensing satellite which
has been in orbit since 15 June 2007 (Biro et al., 2012). SAR sensors
can provide its own illumination source and it can record data
independent of day and night time (Gibson, 2000; Buckreuss, et al.,
2006). Another advantage is the ability of this data to penetrate
even cloud cover, making the image recording independent of all
weather conditions (Christina Herzfeld and Zahner, 2001; Pradhan
et al., 2009; Pradhan and Shae, 2009). The penetration of radar
beams depends on the wavelength, humidity and roughness of the
surface (Haack and Bechdol, 2000). SAR has a wide range of
applications such as land use and regional/urban planning
(Fugura et al., 2011; Mahmoud et al., 2011), change detection
(Vidal and Moreno, 2011), disaster management (Elbialy et al.,
2013; Hassaballa et al., 2013; Tehrany et al., 2013b; Pradhan and
Shae, 2009; Pradhan et al., 2009; Pradhan and Youssef, 2011),
snow and glacier monitoring (Floricioiu et al., 2008) and sea as well
as sea ice and wind monitoring (Ren et al., 2012). Therefore, there is
a need to perform the classication and extract the environmental
properties of the ground (DellAcqua and Gamba, 2006).

B. Pradhan et al. / Computers & Geosciences 63 (2014) 3443

Extraction of the ooded areas, using SAR data through the

texture analysis is one of the highly demandable areas of research.
Usually in tropical countries during the rainy season the area is
mostly covered by clouds (Al Fugura et al., 2011; Kia et al., 2012;
Pradhan, 2009; Pradhan and Youssef, 2011; Tehrany et al., 2013b;
Youssef et al., 2011). In this situation optical RS cannot be used. As
texture analysis plays an important role in the visual interpretation
and recognition, it is necessary to develop efcient method to extract
information precisely (Haack and Bechdol, 2000). In the recent
development of object-based methods for classication of RS data,
texture is one of the most important factors in the process of
segmentation (Ouma et al., 2008). In hazard mapping using radar
imagery, the most difcult task is extracting the texture of the area
which is not easy to perform. Some methods have been developed
but still there exist no user-friendly and simple method for texture
analysis (Kiema, 2002; Luo et al., 2012).
The current study aimed to produce an easy to use ArcMap
9.2 application to solve the problem and complexity of texture
recognition in ood mapping using TerraSAR-X imagery. Also the
proposed program can be used in other applications such as
agriculture studies, landslide mapping, oil spill monitoring etc.
At the beginning of this article mathematical background of
texture analysis will be explained. A novel application for texture
analysis using the gray-level co-occurrence matrix (GLCM) will be
introduced and discussed as well. Finally, to demonstrate the
ability and correctness of this application, a test subset of a
TerraSAR-X image will be analyzed using simple pixel-based and
object-based classication approaches.

2. Texture analysis: a preview

Tamura et al. (1978) considered texture as a repetitive pattern
that denes small areas. Haralick (1979) described it as pixels of a
certain type and number that have a spatial organization either
randomly or with a dependency of two or more pixels. He also
suggested that the term texture should not stand alone as it is always
connected to the tone. Both require each other in a kind of
interrelationship with one always dominating the other. Thus an
area with little gray value variation is dominated by tone; a high
variation, however, refers to texture as the dominant property.
Haralick (1979) also stated that it is important to examine the
properties of single gray value pixels as well as their spatial relationship with each other. Therefore it is necessary to use rst and second
order statistics. The rst order describes the properties of each gray
value pixel separately, whereas the second order statistics examine
the relationship of pixels and thus their organization. Especially the
second order calculation is important because texture is a quality of
not a single pixel but of a whole area.
In a recent paper, Mahmoud et al. (2011) performed texture
analysis using TerraSAR-X data in order to enhance the land use/
cover (LULC) classication near Pirna, Saxony, Germany. They
applied both separability and threshold (SEaTH) method to extract
textural information from the TSX image in order to assess the
enhancement of the classication accuracy. Their results proved the
efciency of TSX imagery and texture-based analysis in LULC
mapping which leaded to acquire an overall accuracy of 95% with
kappa coefcient 93%. Another study by Lee et al. (2012a), utilized
GLCM to detect the landslides on levees using SAR data. The
proposed method was applied on L-band SAR data collected from
NASA's UAVSAR of the Mississippi River levee system between
Vicksburg, MS and Clarksdale, MS, USA. All known levee landslides
in their study area could be detected with a low number of false
positives. In a related paper by Wei et al. (2012), oil spill was
monitored from ERS-2 SAR image using GLCM texture analysis in
Bohai Sea, China. Through their analysis, they discovered that

35

variance, contrast, dissimilarity and correlation are four texture

characteristics suitable for classication of oil spill. So based on the
literature it can be said that texture analysis is a proper tool which
assist hazard mapping and other applications.
Various approaches are available to examine texture. Among
those there are different structural, statistic, model-based and
transform solutions (Pant et al., 2010). The structural approach
concentrates on the hierarchical structure of texture. Hereby, the
relationship between micro-texture elements (primitives) respective to their spatial arrangement is of prime importance. The
statistic approach uses non-deterministic properties. Image gray
values distribution and relationships which are governed by these
properties indirectly represent texture in this case. Often second
order statistics are of importance (Lee et al., 2012b).
Another possibility to calculate texture is the model based
option (Kim and Kang, 2007). Both generative image and stochastically based models are used in this approach. The image analyses
are done by estimating parameters for the respective model.
However, it has a high computational complexity due to the need
to estimate parameters of statistic models (Pradhan, 2010). The
last approach uses transform methods like Gabor, Fourier or
wavelet transforms. In this case, the image is transformed into
another space where the respective coordinate system represents
characteristics of the respective texture such as its frequency or
size (Materka and Strzelecki, 1998).
In the current research, a statistical approach was used for texture
analysis. The concentration will rely on the gray level co-occurrence
matrix (GLCM) as this method was used for the implementation of
the texture analysis tool. The co-occurrence matrix that was rst
introduced by Haralick (1979), is among the most widely used for
derivation of these statistic features. The following section will
introduce the program which can be used in texture analysis of
TerraSAR-X and the efciency of this program will be tested.

3. Methodology
The practical and theoretical aspects of texture analysis implemented in this study involve several steps as shown in Fig. 1.
3.1. General information on design and functionality
The texture analysis program was implemented using ArcGIS
ArcObjects 9.2 and Visual Basic for Application (VBA). Fig. 2 shows
the interface of proposed program.
The program can be divided into three parts: raster denition,
matrix framework creation and formula application. The graphic
interface of the program consists of one window containing three
frames. The rst frame of the starting window allows the user to
select a raster le for the texture analysis either by selecting an
already opened le in ArcMap 9.2 or by searching through the
directory. The image le will always be saved as an .img in the
folder of the original image. Depending on chosen direction and
type of analysis, an appendix will be added to the le name, thus
giving it an easy to recognize the description.
The second and third frame makes various texture options for four
different directions available to the user. The texture options which
can be found in the second frame are divided into the three groups
such as contrast group, orderliness group and statistics group. To
achieve these three groups, GLCM should be calculated. The GLCM,
also called gray tone spatial dependency matrix, was rst introduced
by Haralick et al. (1973). To dene this matrix, one has to picture a
rectangular image with Nr rows, Nc columns and with a quantization
of Ng gray levels from the tone of each single pixel. Lr {0,1,2,.,Nr}
and Lc {0,1,2,.,Nc} will be the number of rows or spatial domains
and dened as the set of Ng quantied gray levels. If we create the

36

B. Pradhan et al. / Computers & Geosciences 63 (2014) 3443

Fig. 1. Methodology ow chart.

Fig. 2. Texture analysis program interface.

matrix framework Lr  Lc, then a set of pixels can be received. These

are ordered according to row and column respectively.
2

0; 0

6 1; 0
6
6
Lr  Lc : 6
6 2; 0
6
4
N r ; 0
Lr f0; 1; 2; :; N r g

0; 1

0; 2

1; 1

1; 2

2; 1

2; 2

Nr ; 1

0; N c

N r ; 2

1; N c 7
7
7
2; N c 7
7
7

5
N r ; N c

Lc f0; 1; 2; :; N c g
Each pixel is dened as a pair of coordinates and has a gray level G
assigned to itself by a function that can be described as the image
Lr  Lc ; I : Lr  Lc -G. Now, the relative frequency of occurrence of a
pair of neighboring pixels or resolution cells that have a distance d
between them can be expressed as a matrix P ij . Hereby the variables i
and j describe respective gray levels. This matrix P ij or is called the
GLCM. It is symmetric and denes distance and direction, the so
called angular relationship, of two pixel neighbors as a function
(Haralick, 1979).

B. Pradhan et al. / Computers & Geosciences 63 (2014) 3443

37

Fig. 3. Development of GLCM (example): kernel (1), matrix framework Fij (2), symmetrical matrix Vij (3), normalized matrix Pij (4).

Here is the brief explanation of how the GLCM is calculated.

This will lead to the formulas implemented in the texture analysis
tool (Hall-Beyer, 2007). As mentioned previously, the GLCM has to
be calculated prior to the implementation of the formulas. This
requires the following steps as shown in Fig. 3. Matrix number
(1) in Fig. 3 will be a 3  3 subset of an image with the respective
gray values shown. The undertaken calculation will produce a
horizontal GLCM in a 3  3 window.
First, the matrix framework needs to be prepared. It will contain
the number of occurrences of each pair of gray value combination.
The pair itself consists of a neighbor pixel and a reference pixel
having both a certain distance d away from each other and a specic
spatial relationship. This can be either a horizontal, vertical or
diagonal relationship.
The offset has to be constant during the entire calculation. In the
example a distance of d 1 and a horizontal relationship was
selected, thus the relationship of pixel lying next to each other is
being examined. For instance, it can be noted that the pixel
combination (0, 1) occurs twice in the image, therefore the number
two will be placed at its corresponding position. After the number of
occurrences has been determined the matrix is made symmetrical.
The reason is, that both opposed directions are taken into account. It
does not matter whether a pair of pixels has the combination (1,0) or
(0,1) as this will change when the opposite direction is considered.
The symmetrical matrix V can be realized by adding the matrix
framework F to its transposed matrix F T :
2
3 2
3 2
3
0 2 0
0 1 1
0 3 1
6
7 6
7 6
7
T
F F 4 1 1 0 5 4 2 1 0 5 4 3 2 0 5 V
2
1 0 1
0 0 1
1 0 2
as a nal step, the symmetrical matrix V has to be normalized.
Therefore each value of this matrix is divided by the sum of all values
in V.
P i;j

V ij
1
N
i;j 0 V ij

Now that the GLCM has been calculated, the formulas can be
applied. As mentioned they can be divided into three groups:
contrast group, orderliness group and statistics group. The previously used symmetric normalized GLCM example P will be used
in the following paragraphs as a demonstration for calculation. The
second matrix will be containing the respective weights resulting
from the given formula. The scalar product of these two matrices
will be the sought for result.

Similar to contrast the dissimilarity is calculated (Eq. 5). The image

can display values from zero in areas of equal tone to x, where x is
a positive real number depending on the radiometric resolution.
For an 8 bit input image the highest value would thus be 255.
N1

P ij ji jj

The last part of the contrast group is homogeneity (Eq. 6).

Contrary to the two aforementioned methods it uses an inverse
difference approach. To avoid an error in calculations the enumerator
was increased by one so that no zero can exist under the fraction bar:
N1

N1

P ij i  j2

i;j 0

P ij

2
i;j 0 1 i  j

3.1.2. Orderliness group

The second group concentrates on the orderliness of values in a
kernel. This means that it looks at their regularity (Yenugu et al.,
2010). Likewise the contrast group weighted values are used.
Angular second moment (ASM), and energy are the two options
that are clubbed together. Hereby the ASM uses its own cooccurrence matrix as weight. Thus values from one to almost zero
are possible. When close to zero a value will depend on the size of
the searching window and direction of the GLCM. The formula 1/n
1
applies with n N
i;j 0 V ij . The value one can only be reached, if an
area displays a homogenous tone with only one gray value. The
same applies to the energy, also referred to as uniformity, is
calculated by applying the square root to the ASM result:
ASM

N1

P i;j 2 Energy

p
ASM

i;j 0

The maximum probability, short MAX is retrieved through the

simplest calculation. The highest value of the GLCM is searched
and then applied. Similar to the ASM, the result can lie between
1
one and 1/n, where n is n N
i;j 0 V ij . In the example the largest P ij
value in the window is one fourth or 0.25. This might indicate a
slightly higher occurrence of one or two combinations. However,
no dominance is visible.
Entropy, which is another measure for order, refers to the
amount of chaos or disorder in a window. Its value ranges from
zero to n, with n being dependent on the size of the searching
window and the direction of the neighbor pixel relationship. In
this case n 2.485. The smaller the value is, the higher the degree
of order in an image will be.
N1

3.1.1. Contrast group

The Eqs. (4)(6) belong to the contrast group. In this group the
contrast or difference in gray value between related pixels is
emphasized (Eq. 4). As it can be seen in the Fig. 2 the contrast
group has three subdivisions of contrast, similarity and homogeneity. This calculation emphasizes higher gray value differences:

i;j 0

P i;j  ln P i;j

i;j 0

3.1.3. Statistics group

The last group that should be measured is statistics group. The
mean value in this calculation can be derived by either using the
reference pixel i or j. In the case of the symmetric GLCM, this will
yield an equal result because values in this matrix will appear to be
mirrored on the diagonal. Therefore i and j need not be calculated
separately. As a result values from zero to n, where n is the number of

38

B. Pradhan et al. / Computers & Geosciences 63 (2014) 3443

different gray values possible in the original image, can be calculated.

The mean itself displays a value around which most likely the
dominant reference pixel value can be found. In areas of high contrast
this value, however, will be further away from the reality:
N1

i P i;j

i;j 0

The variance is closely connected to the mean that describes the

deviation from the previous statistic element. Similar to the mean
calculation, it is possible to determine both the s2i and the value s2i .
Due to the symmetry of the matrix, however, it will not be necessary,
as both results will be equal. Variance may have an advantage over
the mean as its results have a higher range of positive numbers, thus
a greater stretch might be visible on output images.
The standard deviation is similar to the variance. Its only
difference lies in another range of values as it is dened as the
square root of the variance:
s2

N 1

P i;j i  2

i;j 0

Standard deviation : s

p
s2

10

The last member of this statistics group is the correlation: it is an

indication of how much a gray value depends on its neighbors.
Results may range from zero to one, where zero points to an
absolutely uncorrelated image subset and one refers to fully correlated areas. It can be seen that an error may be received as a result, if
the variance of an area is zero because the variance can be found in
the enumerator of the formula and thus a division by zero can occur.
This might only happen if only one gray value exists in the respective
area. This subset would be fully correlated. Thus, when programming
the correlation the result can be set to one in this case:
2
3
N1
N 1
i  j
6i j j j 7
P i;j 4 q 5 -simplified : P i;j
11
s2
i;j 0
i;j 0
s2 i s2 j

3.1.4. Program usability

Certain steps have to be undertaken in a program, in order to
make it work smoothly and user friendly. Although they are not
necessary themselves for the main program, but they can be useful
for the user as they give further options and also they can give
warning of errors. Among those it can be mentioned to the
possibility of opening the les that have already been opened in
ArcMap 9.2, to be informed about existing les of the same name.
Also it can give the opportunity to overwrite these les and to
create multiband images for four directions of the same texture
option. Also using le selection and drop down menu the mistake
of accidentally selecting a non-raster le can be prevented. These
functions are also added to the main program.

followed. First, the speckle effect needs to be removed (Idrees

et al., 2013). ERDAS IMAGINE provides a few lters which will
suppress and smooth out this effect (Haack and Bechdol, 2000).
However, not all of them are equally suitable. Dong et al. (2000)
discussed this issue, comparing different lters regarding their
mean, edge and textural information preservation as well as their
reduction of the standard deviation. They also indicated that a
speckle reducing lter should not distort and degrade the inherent
texture if it is intended for texture preservation. After evaluating
the Lee, Kuan, Frost, mean, median and edge-sharpening lter they
concluded, that the median lter was not suitable as a speckle
lter for SAR data as it strongly distorts the texture.
Comparison has been done between the outputs of these lters
using signal to noise ratio (SNR). Visual interpretation concluded
that frost lter performed better than the rest in this study. The
Frost lter displays this area very well, with low noise levels. For
this purpose, 5  5 window Frost lter was used to remove the
speckle and the results showed features are still visible without
appearing blur. Moreover, line and point features are clearly visible
and the texture context seems to have been least altered.
In the next step the radiometric resolution has to be rescaled from
unsigned 16 bit to unsigned 8 bit which means that the number of
gray values available has to be downscaled from 65,536 to 256. The
reason therefore is an incompatibility of a method used with
ArcObjects. It does not matter which method is used, information will
always be lost when an image is rescaled (Zan et al., 2008). The
question that should now be answered is: which method does the
least damage to the original texture information of the image? Three
different methods using ERDAS IMAGINE were used and analyzed.
Finally the method which selected was the one that can be found as
rescale option through ERDAS Interpreter menu under the option
utilities. The reason for this selection is it has neither the problem of a
strongly altered contrast situation nor does it provide oating gray
values.
Finally, suitable subset has to be found. Fig. 4 shows the
proper scene. The Terengganu river can be recognized as well as
a mountainous area and different agricultural elds with different
types of crops. Also rectangular structure of different buildings and
the linear, dark structure of streets can be seen.

3.2. Analysis
3.2.1. Data and study area
The study is carried out in Terengganu which is situated in
north-eastern part of Peninsular Malaysia, and is bordered in the
northwest by Kelantan, the southwest by Pahang, and the east by
the South China Sea (Zaleha et al., 2006). The data used in this
work was recorded by TerraSAR-X satellite, short TSX-1, from 27th
November 2009. Data was single look, stripmap modus, with three
meters spatial resolution and HH polarization. TerraSAR-X data
used with radiometric resolution of 16 bit thus providing 65,536
different gray values.
3.2.2. Pre-processing
In order to guarantee good results, the images have to be preprocessed (Albinet et al., 2012). Therefore the following steps are

Fig. 4. Subset of area of the Terengganu River, Malaysia.

B. Pradhan et al. / Computers & Geosciences 63 (2014) 3443

Fig. 5. Original image (upper left) and contrast group results: contrast (upper right), dissimilarity (lower left), homogeneity (lower right).

Fig. 6. Orderliness group results: ASM (upper left), energy (upper right), entropy (lower left), MAX (lower right).

39

40

B. Pradhan et al. / Computers & Geosciences 63 (2014) 3443

3.2.3. Texture analysis and classication

After all these steps have been undertaken, the texture analysis
could be applied to the subset image using the proposed program. In
recent years, the object-based classication methods proved to be
more efcient than pixel based due to the spatial information that is
used in these methods (Al Fugara et al., 2009; Blaschke et al., 2008;
Mahmoud et al., 2011). Results of the texture analysis were used in
pixel-based classication and four classes could be recognized by the
ERDAS program: water body, settlement, agriculture/elds and
forest. Also advanced object-based classication method was applied
on the outputs of the texture analysis and nally compared with the

pixel-based result. As we know segmentation is the basis of the

object-based classication that divides the image into the homogeneous objects and classies these objects based on spectral, spatial,
textural, relational and contextual information (Petropoulos et al.,
2012). These regions are homogenous in some way but also differ
from their adjacent regions (Morris et al., 1986). Thus this segmentation corresponds in a way with the human perception of areas.
For object-based classication Deniens eCognition 7.0 was
used which is very popular in optical and radar remote sensing
(Tehrany et al., 2013a). However, before the actual segmentation
process it was necessary to combine all texture results which

Fig. 7. Statistics group result: mean (upper left), variance (upper right), Standard deviation (lower left), correlation (lower right).

Fig. 8. Exemplary result of a pixel based classication with ERDAS Imagine.

B. Pradhan et al. / Computers & Geosciences 63 (2014) 3443

achieved from texture analysis into one image. This was done
using Composite Bands of ArcMaps Data Management Tools. This
created an image le with 44 bands. After a recalculation of its
statistics and the assignment of the no data value 9999, the
image could be used for segmentation. Scale and color values were
selected, depending on the properties of the image used. Hereby
the resolution of the image object level is dened by the scale
parameter. Thus low values indicate that smaller objects will be
created. The second parameter, color, signies the importance of
image color homogeneity in the segmentation process. For this
work, scale received a parameter value of ten and 0.9 was assigned
as the color parameter. After the execution of the segmentation
process an image containing segmentation polygons was calculated and returned. Nearest neighbor was chosen as the classication method. Color and mutual embedding were both taken into
consideration. Training areas could be dened by assigning certain
segments to either a positive or negative predened output class.
This pre-denition had to be done by the user and consisted of a
name and a color for the later class. For example, the rst
classication cycle contained the positive output class water (blue)
and the negative output class land (green). This way an evaluation
of the texture analysis program and its results can be conducted.

4. Results and discussion

The results of the structural texture analysis are as follows. The
image contains four bands to display the four possible directions.
Small part of the study area has been chosen to represent here, in

41

order to show the impact of each group of texture analysis on data

more clearly. As can be seen in Fig. 5 results of the contrast group
produced concordant results. Settlement areas had high contrast
values, and thus contained low homogeneity values, whereas
water bodies and ooded areas displayed a low contrast and thus
high homogeneity.
The surrounding area displays lower contrast values. Also the
results of orderliness group and statistics group can be seen in
Figs. 6 and 7 respectively. Classication scheme was applied using
the texture analysis results to examine the efciency of two
classiers and to assess the impacts of the texture information
on the precision of the results of each classication. The result of
the pixel-based classication is shown in Fig. 8.
Visual interpretation of the thematic map derived by ERDAS
imagery has illustrated the weakness of pixel-based to extract the
feature classes. Due to the high oat value variation in the
different images a pixel based approach is not advisable. This
would simply yield inhomogeneous areas with a high amount of
noise as well as misclassication. As can be seen in Fig. 8, water
was not recognized by the program. Although training areas were
used and a supervised classication was conducted, it was interpreted as forest or agriculture area. Also clearly visible are the
inhomogeneous areas where forest or agricultural areas would be
suspected. Therefore a pattern recognition approach using segmentation was examined. Fig. 9 shows the image before applying
the segmentation. It was possible to differentiate water from the
land. Also a settlement can be recognized. Fig. 10 shows a part of
the segmented image containing the biggest settlement as well as
elds and ooded area.

Fig. 9. Six layer false color representation of the texture band composite in eCognition with possible elements for later classication (rotated 901 clockwise). (For
interpretation of the references to color in this gure legend, the reader is referred to the web version of this article.).

Fig. 10. Result of the multiresolution segmentation process.

42

B. Pradhan et al. / Computers & Geosciences 63 (2014) 3443

As can be seen in the Fig. 10, different classes could be

separated very well through the segmentation and it shows the
efciency of the object-based classication and the strangeness of
eCognition software in precise segmentation. Fig. 11 shows the
thematic map acquired by object-based method and ve land
cover classication results were obtained.
As can be seen in Fig. 11, the thematic map has acceptable and
representable appearance which proves the efciency of the
object-based classication. In some areas some misclassications
are obvious but accuracy assessment should be done because it is
not proper to do any judgments without considering and evaluating the statistics. For that reason, the level of correctness for the
results of classications was evaluated by accuracy assessment
through confusion matrix (Foody, 2002; Magee, 2011). Table 1
shows the results of the accuracy assessment. The results indicated
that the proposed object-based method produced an acceptable
producer and user accuracies for all the classes except the class of
settlement and wood. The reason could be related to the mixture
of the spectral information of two classes of settlement and wood,

so the probability of having errors is very high. Also due to spectral

similarity of the class of settlement and wood and the class of
wood, their accuracies are lower than other classes. This implies
that it was difcult for the program to differentiate between these
two classes.

5. Conclusion and recommendation

In this paper, we have developed an easy to use ArcMap
program which is made in visual basic environment and it is able
to perform the texture analysis of features using TerraSAR-X
imagery. Secondly, ERDAS and eCognition software were used for
pixel-based and object-based classication respectively in order to
examine the precision of both methods in extracting the features.
Results showed that pixel-based approach was not that successful,
whereas object-based approach showed very good results. Objectbased by 83.63% overall accuracy and 0.8 kappa coefcient proved
its efciency in detecting the features over traditional pixel-based
method. Results on image texture classication showed that the
proposed program is capable in the analysis of TerraSAR-X
imagery. Hence the developed program can be used in many earth
observation applications such as LULC mapping, change detection,
and hazard studies. The program could differentiate between the
objects which lead to enhance their results. The program itself
performed well on the task but the computation process took little
longer time. If the algorithm is enhanced that might improve the
processing time. However, the most time consuming aspect of this
program is the creation of pixel blocks, therefore improvement of
algorithm does not have signicant impact. Thus to really reduce
the runtime it is advisable to use smaller signicant subsets of a
scene. The currently available version is a robustly running beta
version. However, some features have not yet been implemented.
The two most important features are, are the use of different sized
kernels in order to assess a greater area of texture and the
question of the no-data value handling. Further, improvements
in usability and user options might be implemented. Also using
data fusion can improve the accuracy of the texture analysis.

Acknowledgments
The German Aerospace Center (DLR) provided Terra-SAR-X data
under the Science proposal ID: HYD0326. Thanks to Thomas Hahmann for his valuable inputs on the processing of Terra-SAR-X data.
References
Al Fugura, A., Billa, L., Pradhan, B., Mohamed, T.A., Rawashdeh, S., 2011. Coupling of
hydrodynamic model and aerial photogrammetry-derived digital surface model
for ood simulation scenarios using GIS: Kuala Lumpur ood. Malaysia. Disaster
Adv. 4 (4), 2028.

Fig. 11. Result derived by object-based method.

Table 1
Object-based classication accuracy assessment.

Settlement
Water
Settlement and wood
Wood
Woodless
Sum
Producer accuracy %

Settlement

Water

Settlement and wood

Wood

Woodless

Sum

User accuracy %

46
0
0
1
0
50
98

1
51
2
4
2
37
85

2
0
29
4
0
51
83

0
0
7
33
1
50
80
Overall accuracy
Kappa coefcient

0
0
12
1
30
50
70

49
0
50
43
33
238

94
100
58
77
91

83.6%
0.8

B. Pradhan et al. / Computers & Geosciences 63 (2014) 3443

Al Fugara, A.M., Pradhan, B., Mohamed, T.A., 2009. Improvement of land-use

classication using object-oriented and fuzzy logic approach. Appl. Geomat. 1
(4), 111120.
Albinet, C., Borderies, P., Fabre, S., 2012. Radar and optical parallel modelling of
forest remote sensing data. Int. J. Remote Sens. Appl. 2, 4.
Biro, K., Pradhan, B., Sulieman, H., Buchroithner, M., 2012. Evaluation of TerraSAR-X
data for land use/land cover analysis using object-oriented classication
approach in the African Sahel Area, Sudan. J. Indian Soc. Remote Sens. , http:
//dx.doi.org/10.1007/s12524-012-0230-7.
Blaschke, T., Lang, S., Hay, G.J. (Eds.), 2008. Object-Based Image Analysis Spatial
Concepts for Knowledge-Driven Remote Sensing Applications, rst ed. Springer,
New York.
Blaschke, T., 2010. Object based image analysis for remote sensing. ISPRS J.
Photogramm. Remote Sens. 65 (1), 216.
Buckreuss, S., Mhlbauer, P., Mittermayer, J., Balzer, W., Werninghaus, R., 2006. The
TerraSAR-X Ground Segment. In: Proceedings of the European Conference on
Synthetic Aperture Radar (EUSAR). Dresden, Germany.
Chica-Olmo, M., Abarca-Hernandez, F., 2000. Computing geostatistical image
texture for remotely sensed data classication. Comput. Geosci. 26 (4),
373383.
Christina Herzfeld, U., Zahner, O., 2001. A connectionist-geostatistical approach to
automated image classication, applied to the analysis of crevasse patterns in
surging ice. Comput. Geosci. 27 (5), 499512.
DellAcqua, F., Gamba, P., 2006. Discriminating urban environments using multiscale texture and multiple SAR images. Int. J. Remote. Sens. 27 (18), 37973812.
DLR-Deutsches Zentrum fr Luft- und Raumfahrt, 2009. DLR Portal- Anwendungen
und Nutzung. DLR Portal. [Online] Deutsches Zentrum fr Luft- und Raumfahrt,
20. March 2009. (Zitat vom: 21. February 2010) http://www.dlr.de/desktopde
fault.aspx/tabid-4219/8885_read-16087/.
Dong, Y., Milne, A.K., Forster, B.C., 2000. A review of SAR speckle lters: texture
restoration and preservation. In: Proceedings of the Geoscience and Remote
Sensing Symposium 2000 (IGARSS 2000), IEEE 2000 International, vol. 2,
pp. 633635.
Elbialy, S., Mahmoud, A., Pradhan, B., Buchroithner, M., 2013. Application of
spaceborne synthetic aperture radar data for extraction of soil moisture and
its use in hydrological modelling at Gottleuba Catchment, Saxony, Germany. J.
Flood Risk Manag. , http://dx.doi.org/10.1111/jfr3.12037.
Floricioiu, D., Eineder, M., Rott, H., Nagler, T., 2008. Velocities of major outlet
glaciers of the Patagonia Iceeld observed by TerraSAR-X. In: Proceedings of the
Geoscience and Remote Sensing Symposium 2008 (IGARSS 2008). IEEE International, vol. 4, pp. IV347.
Foody, G.M., 2002. Status of land cover classication accuracy assessment. Remote
Sens. Environ. 80 (1), 185201.
Fugura, A., Billa, L., Pradhan, B., 2011. Semi-automated procedures for shoreline
extraction using single RADARSAT-1 SAR image. Estuar., Coast. Shelf Sci. 95 (4),
395400.
Gibson, P., 2000. Introductory remote sensing principals and concepts. Taylor and
Francis, New York.
Haack, B., Bechdol, M., 2000. Integrating multisensor data and RADAR texture
measures for land cover mapping. Comput. Geosci. 26 (4), 411421.
Hall-Beyer, M. GLCM Texture Tutorial, 2007. University of Calgary. (Online)
University of Calgary, 21. February 2007. (Zitat vom: 2. February 2010)
http://www.fp.ucalgary.ca/mhallbey/tutorial.htm.
Hamedianfar, A., Shafri, H.Z., 2013. Development of fuzzy rule-based parameters for
urban object-oriented classication using very high resolution imagery. Geocarto. Int. , http://dx.doi.org/10.1080/10106049.2012.760006.
Haralick, R.M., Shanmugam, K., Dinstein, I.H., 1973. Textural features for image
classication. IEEE Trans. Syst. Man. Cybern. 6, 610621.
Haralick, R.M., 1979. Statistical and structural approaches to texture. Proc. IEEE 67
(5), 786804.
Hassaballa, A.A., Althuwaynee, O.F., Pradhan, B., 2013. Extraction of soil moisture
from RADARSAT-1 and its role in the formation of the 6 December 2008
landslide at Bukit Antarabangsa, Kuala Lumpur. Arab. J. Geosci., 110, http://dx.
doi.org/10.1007/s12517-013-0990-6.
Idrees, M.O., Saeidi, V., Yusuf, Y.A., Pradhan, B., 2013. Advanced differential
interferometry SAR techniques for deformation monitoring: A review on
sensors and recent research development. Geocarto. Int. , http://dx.doi.org/
10.1080/10106049.2013.807305.
Kia M.B., Pirasteh S., Pradhan B., Mahmud A.M., Sulaiman W.N.A., Moraddi A., 2012.
An articial neural network model for ood simulation using GIS: Johor river
basin, Malaysia. Environ. Earth Sci. 67 (1), 251264, http://dx.doi.org/10.1007/
s12665-011-1504-z.
Kiema, J., 2002. Texture analysis and data fusion in the extraction of topographic
objects from satellite imagery. Int. J. Remote Sens. 23 (4), 767776.
Kim, S.C., Kang, T.J., 2007. Texture classication and segmentation using wavelet
packet frame and Gaussian mixture model. Pattern Recognit. 40 (4), 12071221.
Lee, M.A., Aanstoos, J.V., Bruce, L.M., Prasad, S., 2012a. Application of omnidirectional texture analysis to SAR images for levee landslide detection. In:
Proceedings of 2012 IEEE International Geoscience and Remote Sensing
Symposium (IGARSS), IEEE, pp. 18051808.

43

Lee, S.J., Serre, M.L., van Donkelaar, A., Martin, R.V., Burnett, R.T., Jerrett, M., 2012b.
Comparison of geostatistical interpolation and remote sensing techniques for
estimating long-term exposure to ambient PM2. 5 Concentrations across the
Continental United States. Environ. Health Perspect. 120 (12), 17271732.
Luo, Y., Zhao, S., Zhou, H., Wang, A., He, K., Tan, L., 2012. A novel classication
method based on texture analysis using high-resolution SAR and optical data.
In: Proceedings of the 2012 Second International Workshop on the Earth
Observation and Remote Sensing Applications (EORSA), IEEE, pp. 186190.
Magee, K.S., 2011. Segmentation, object-oriented applications for remote sensing
land cover and land use classication (Ph.D. dissertation). University of
Cincinnati, USA. (131pp).
Mahmoud, A., Elbialy, S., Pradhan, B., Buchroithner, M., 2011. Field-based landcover
classication using TerraSAR-X texture analysis. Adv. Space Res. 48 (5),
799805.
Materka, A. Strzelecki, M., 1998. Texture Analyisis Methods A Review.
Morris, O.J., Lee, M.D.J., Constantinides, A.G., 1986. Graph theory for image analysis:
an approach based on the shortest spanning tree. Communications, Radar and
Signal Processing. IEE Proceedings, F 133 (2), 146152.
Ouma, Y.O., Josaphat, S.S., Tateishi, R., 2008. Multiscale remote sensing data
segmentation and post-segmentation change detection based on logical modeling: theoretical exposition and experimental results for forestland cover
change analysis. Comput. Geosci. 34 (7), 715737.
Pant, T., Singh, D., Srivastava, T., 2010. Advanced fractal approach for unsupervised
classication of SAR images. Adv. Space Res. 45 (11), 13381349.
Petropoulos, G.P., Kalaitzidis, C., Prasad Vadrevu, K., 2012. Support vector machines
and object-based classication for obtaining land-use/cover cartography from
Hyperion hyperspectral imagery. Comput. Geosci. 41, 99107.
Pradhan, B., 2009. Flood susceptible mapping and risk area estimation using logistic
regression, GIS and remote sensing. J. Spat. Hydrol. 9 (2), 118.
Pradhan, B., Shae, M., 2009. Flood hazard assessment for cloud prone rainy areas
in a typical tropical environment. Disaster Adv. 2 (2), 715.
Pradhan, B., Shae, M., Pirasteh, S., 2009. Maximum ood prone area mapping
using RADARSAT images and GIS: Kelantan river basin. Int. J. Geoinf. 5 (2),
1123.
Pradhan, B., 2010. Remote sensing and GIS-based landslide hazard analysis and
cross-validation using multivariate logistic regression model on three test areas
in Malaysia. Adv. Space Res. 45 (10), 12441256.
Pradhan, B., Youssef, A.M., 2011. A100-year maximum oood susceptibility mapping using integrated hydrological and hydrodynamic models: Kelantan River
Corridor. Malaysia. J. Flood Risk Manag. 4 (3), 189202, http://dx.doi.org/
10.1111/j.1753-318X.2011.01103.x.
Ren, Y., Lehner, S., Brusch, S., Li, X., He, M., 2012. An algorithm for the retrieval of sea
surface wind elds using X-band TerraSAR-X data. Int. J. Remote Sens. 33 (23),
73107336.
Sousa, J.J., Magalhes, L.M., Ruiz, A.M., A.M.R., Cardoso, G., 2013. The viStaMPS tool
for visualization and manipulation of time series interferometric results, series
interferometric results. Comput. Geosci. 52, 409421, http://dx.doi.org/10.1016/
j.cageo.2012.11.012.
Tamura, H., Mori, S., Yamawaki, T., 1978. Textural features corresponding to visual
perception. IEEE Trans. Syst. Man Cybern. 8 (6), 460473.
Tehrany, M.S., Pradhan, B., Jebu, M.N., 2013a. A comparative assessment between
object and pixel-based classication approaches for land use/land cover
mapping using SPOT 5 imagery. Geocarto. Int., http://dx.doi.org/10.1080/
10106049.2013.768300.
Tehrany, M.S., Pradhan, B., Jebur, M.N., 2013b. Spatial prediction of ood susceptible
areas using rule based decision tree (DT) and ensemble bivariate and
multivariate statistical models. J. Hydrol. 504, 6979, http://dx.doi.org/
10.1016/j.jhydrol.2013.09.034.
Treitz, P.M., Soulis, E.D., Howarth, P.J., Kouwen, N., 1996. Texture processing of
synthetic aperture radar data using second-order spatial statistics. Comput.
Geosci. 22 (1), 2734.
Vidal, A., Moreno, M.R., 2011. Change detection of isolated housing using a new
hybrid approach based on object classication with optical and TerraSAR-X
data. Int. J. Remote Sens. 32 (24), 96219635.
Wei, L., Hu, Z., Guo, M., Jiang, M., Zhang, S., 2012. Texture feature analysis in oil spill
monitoring by SAR image. In: Proceedings of IEEE 20th International Conference on Geoinformatics (GEOINFORMATICS), 2012, pp. 16.
Yenugu, M.M., Marfurt, K.J., Matson, S., 2010. Seismic texture analysis for reservoir
prediction and characterization. Lead. Edge 29 (9), 11161121.
Youssef, A., Pradhan, B., Hassan, A.M., 2011. Flash ood risk estimation along the St.
Katherine road, southern Sinai, Egypt using GIS based morphometry and
satellite imagery. Environ. Earth Sci. 62, 611623.
Zaleha, K., Roswati, M.A., Iwasaki, N., 2006. Distribution of some species of
harpacticoid copepods in east coast of Peninsular Malaysia. Coast. Mar. Sci.
30 (1), 141145.
Zan, F.D., Guarnieri, A.M., Rocca, F., Tebaldini, S., 2008. Sentinel-1 radar interferometry applications. In: Proceedings of 2008 7th European Conference on
Synthetic Aperture Radar (EUSAR), VDE, 2008, pp. 14.
Zhang, J., Marszaek, M., Lazebnik, S., Schmid, C., 2007. Local features and kernels
for classication of texture and object categories: a comprehensive study. Int. J.
Comput. Vis. 73 (2), 213238.