Remote Sensing Applications: Society and Environment 16 (2019) 100271

Contents lists available at ScienceDirect

Remote Sensing Applications: Society and Environment

journal homepage: http://www.elsevier.com/locate/rsase

Evaluation of SAR speckle filter technique for inundation mapping

Vikas Kumar Rana *, T.M.V Suryanarayana
Water Resources Engineering and Management Institute, Faculty of Technology & Engineering, The Maharaja Sayajirao University of Baroda, Gujarat, India

A R T I C L E I N F O A B S T R A C T

Keywords: The presence of speckle in visual images makes the automated digital image classification a challenging task.
Inundation Therefore, reduction of speckles is an important pre-processing step. The choice of speckle filter depends on the
SAR requirements of an application and the characteristics of the dataset. In this study, some most preferred speckle
Machine learning
filters are assessed for the data from Sentinel-1 to map flood extent. The Sentinel-1 (VV-vertical transmit, vertical
Sentinel-1
receive and VH- vertical transmit, horizontal receive) polarizing filter data were used to evaluate machine
learning algorithms, namely, random forest (RF) and support vector machine (SVM), to classify an inundated
area. The accuracies of the classifications were assessed by kappa coefficient, overall accuracies, and producer’s
and user’s accuracies. The present study suggests an approach to monitor damage and provide basic information
to help local communities manage water-related risk, land planning, water management, and flood control
programs.

1. Introduction spatial and temporal resolution images to track the rapidly retreating
flood process (Zhang et al., 2014).
In recent years, severe rainfall events have afflicted the state of Drastic variability in climate has accelerated the incidence of cata
Kerala in southern India causing damage to houses and infrastructures. strophic flood events in the last decade (Chunming et al., 2005). In any
Remotely sensed data can provide significant mapping capabilities flood-related study, identification of the flood extent and susceptible
during such severe rainfall events. However, obtaining remotely sensed areas is a prerequisite to assess the disaster impact. Flood mapping can
data with an ideal combination of fine spatial and temporal resolution best be achieved with the help of remote sensing due to the inaccessi
with the ability to see through clouds and discriminate flooding under bility to the flood-affected regions. However, cloudy conditions reduce
forest cover is a difficult task. The extent of inundation, caused by river flood mapping accuracy below the acceptable levels in optical remote
flooding and/or coastal storm surges, is required quickly to expedite sensing.
relief and repair services. The precipitation over Kerala amid June, July Synthetic Aperture Radar (SAR) imaging is an efficient remote
and August (1–19 August 2018) were 15%, 18% and 164% above sensing technique offering well-developed, consistent, efficient, and
normal, respectively. Due to intense rainfall, all the major reservoirs reliable means of collecting information to extract earth’s surface
were full by the end of July 2018 and had no buffer storage to accom dielectric properties (Lee and Pottier, 2009). The ability of SAR to
modate the inflows from 10th of August 2018 (Central Water Commis penetrate clouds is extremely useful in flood-related studies. Synthetic
sion, 2018). Serious spell of precipitation from the 14 August 2018 to 19 aperture radar uses microwave radiation to illuminate the earth’s sur
August 2018 brought appalling flood in 13 out of 14 districts. The face for recording the amplitude and phase of the back-scattered radi
perpetuated exceptional rainfall in August (170% above normal) in the ation, which makes the imaging process coherent. The active sensor of
catchment areas compelled the authorities to resort to hefty downstream Sentinel-1 forms a SAR image by coherently processing the returning
release into the rivers (India Meteorological Department, 2018). signals from successive radar pulses. Stronger or weaker final signals
Remote sensing promises exceptional capacity in catastrophe control (output) are generated by the out-of-the-phase waves by constructively
owing to its regular acquisition function over a large spatial extent or destructively interfering with each other. These interferences produce
(Serpico et al., 2012; Nirupama and Simonovic, 2007; Gitas et al., 2008; a seemingly random pattern of brighter and darker pixels giving the
Khan, 2005). Flooding is a complex phenomenon due to its heteroge radar images a distinctly grainy appearance known as ‘Speckle’
neity and spectral diversity. The analysis of flood mapping require high (Goodman, 1976; Lee et al., 1994). Speckle noise changes the spatial

* Corresponding author. WREMI, The Maharaja Sayajirao University of Baroda, India.

E-mail addresses: [email protected] (V.K. Rana), [email protected] (T.M.V Suryanarayana).

https://doi.org/10.1016/j.rsase.2019.100271
Received 16 June 2019; Received in revised form 3 October 2019; Accepted 22 October 2019
Available online 22 October 2019
2352-9385/© 2019 Elsevier B.V. All rights reserved.
V.K. Rana and T.M.V Suryanarayana Remote Sensing Applications: Society and Environment 16 (2019) 100271

retain the polarimetric characteristics of certain pixels. However, it over

filters the point targets, creates a combination of heterogeneous pixels
and degrades the overall spatial information. The filter is simple and
fast, however, it is not isotropic (i.e. circularly symmetric), but smooths
further along diagonals than along rows and columns. Also, disconti
nuities are found in the smoothed image due to an abrupt cut-off of
weights rather than decline gradually to null.

1.1.2. Gamma map filter

This filter is based on the Bayesian analysis of image statistics. The
scene reflectivity of the underlying image in Gamma-Map algorithm is
assumed to be Gamma distributed rather than normally distributed, and
Fig. 1. Selected Homogeneous area, linear feature and Edge.
speckle is noise within it. Thus this filter works best for geospatial im
ages containing homogenous areas such as oceans, forests, fields, etc.
statistics of the underlying scene backscatter making the classification of
(Lopes et al., 1990). It is given by following cubic equation (Frost et al.,
imageries a difficult task (Durand et al., 1987). A brief introduction of
1982).
some well-known despeckling methods is presented below.
bI 3 2
IbI þ σðbI DNÞ ¼ 0
1.1. DE-NOISING methods
bI ¼ required value
The presence of speckle is the major challenge in the SAR image
processing. A speckle reduces the resolution of an image and the I ¼ ​ local mean
detectability of the ground targets. It also distorts the spatial patterns of DN ¼ ​ input value
surface characteristics and reduces the accuracy of image classification σ ¼ ​ original ​ image ​ variance
(Wang and Ge, 2012). Speckles are signal-dependent and, therefore, act
like multiplicative noise (LeeSen, 1981). Gamma-Map approach has several advantages compared to the other
filters, as it can simultaneously take into account realistic first and
1.1.1. Boxcar filter second order statistical models for both speckle and underlying scene
A simple averaging filter that replaces the center pixel in a 3 � 3 or a reflectivity, and combine them through Bayesian inference. Thus this
larger sized moving kernel with the mean value of kernel pixels. It has filter works best for geospatial images containing homogenous areas
good performance in reducing speckles in a homogeneous area; how such as oceans, forests and fields.
ever, it degrades spatial resolution due to indiscriminately averaging
pixels from the inhomogeneous area and destroys the polarimetric
properties (Lee and Pottier, 2009). This easy operation can very well

Fig. 2. Study area Kerala.

2
­
­
V.K. Rana and T.M.V Suryanarayana Remote Sensing Applications: Society and Environment 16 (2019) 100271

VH polarization)
DV (dual VV þ
Polarization

–
23337
Orbit

7624
Track

165

–
Descending

Descending
Pass

21-Aug-

22-Aug-
Sensing
date

18

18
Acquisition mode

Wide swath (IW)

Interferometric

INS-NOBS

Fig. 3. Study area Asssam

1.1.3. Frost filter

This filter uses local image statistics to remove high-frequency noise
S2MSI1C
Detected

(speckles) while preserving features (edges) by averaging less in the

Product

Ground
Range
type

edge areas. It replaces the pixel of interest with a weighted sum of the
values within an n � n moving kernel (Qiu et al., 2004). The despeckled
pixel value is estimated using a sub-window of the processing window.
Instrument

Instrument

The size of the sub-window varies as a function of the target local het
Spectral

erogeneity measured with a coefficient of variation.

C-SAR

Multi-

X
Digital number ðDNÞ ¼ kαe αjtj
nXn
12 days

10 days
Repeat
cycle

� �� �
α¼ 4 σ2
nσ 2 I
2
S1A_IW_GRDH_1SDV_20180821T004109_20180821T004134_023337_0289D5_B2B2

k ¼ normalized constant
I ¼ local mean
S2B_MSIL1C_20180822T050649_N0206_R019_T43PFL_20180822T085140

σ ¼ local variance
σ ¼ image coefficient of variation value
jtj ¼ |X X0| þ | Y Y0|are present.
n ¼ moving kernel size

When uniform regions are filtered, the Frost filter acts as a mean
filter. When high contrast regions are filtered, the filter acts as a high-
pass filter with rapid decay of elements away from the filter center.
Thus, large uniform areas will tend to be smoothed out and speckle
Specifications of Sentinel-1 and Sentinel-2 products.

removed, whilst high contrast edges and other objects will retain their
signal values and not be smoothed. After application of the Frost filter,
the denoised images show better sharpness at the edges.

1.1.4. Lee filter

Lee filter is based on the assumption that the filtered or output pixel
value is a weighted sum of the reference pixel value and the mean of the
Product name

values within the kernel (LeeSen, 1981). The filter removes the noise by
minimizing either the mean square error or the weighted least square
estimation (Qiu et al., 2004). The Lee filter utilizes the statistical dis
tribution of the digital number values within the moving kernel to es
timate the value of the pixel of interest. This filter assumes the normal
SENTINEL-

SENTINEL-

distribution for the noise in image data.

Table 1

1A

2B

Iout ​ ¼ ½mean� þ K½Uin mean�

3
V.K. Rana and T.M.V Suryanarayana Remote Sensing Applications: Society and Environment 16 (2019) 100271

VH polarization)
Iout ¼ filtered output

DV (dual VV þ
Uin ¼ unfiltered input

Polarization
mean ¼ average of pixels in a moving kernel
VarðxÞ

–
K¼
mean2 σ 2 þ VarðxÞ

28113

12314
Orbit
Variance of � is defined as:

½variance ​ within ​ kernel� þ ½mean ​ within ​ kernel�2

Track

VarðxÞ ¼
41 σ2 þ 1

–
½mean ​ within ​ kernel�2
Descending
Ascending

Lee’s smoothing filter is adaptive to the local statistics in an image,

however, it is an isotropic adaptive filter which cannot remove noise in
Pass

the edge region effectively. Lee filter is reportedly superior in its ability
to preserve prominent edges, linear features, point target, and texture
information.
Sensing

14-Jul-

16-Jul-
date

19

19

1.1.5. Lee sigma filter

This filter is based on sigma probability of the Normal distribution.
Acquisition mode

Wide swath (IW)

The sigma (Standard Deviation) of the entire scene is first computed and
Interferometric

then each central pixel in a moving window is replaced with the average
INS-NOBS

of only those neighborhood pixels that have intensities within a fixed

sigma range of the center pixel. It is well known that, in the normal
distribution, the two-sigma likelihood is 0.955. The pixels outside the
two-sigma range are considered outliers and ignored.
S2MSI1C
Detected

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Product

Ground

variation
Range

S tan dard ​ deviation ​ of ​ an ​ image ​ ¼

type

mean
¼ ​ Coefficient of Variation ¼ Sigma ðσÞ

Due to the use of a fixed sigma computed for the entire scene de
Instrument

Instrument

Leeuw and de Carvalho, 2009, found that the Lee sigma filter blurred
Spectral
C-SAR

Multi-

some of the low-contrast edges and linear features.

1.1.6. Median
12 days

10 days
Repeat

This filter is not an adaptive filter as it does not account for the
cycle

particular speckle properties of the image. Destructive and constructive

interferences in SAR information are represented by extreme values
S1A_IW_GRDH_1SDV_20190714T115653_20190714T115718_028113_032CCD_F972

(low-value and high-value pixels), which are efficiently suppressed by

the Median filter (Sheng and Xia, 1996; Qiu et al., 2004). The median
filter is successful at removing pulse and spike noise while retaining step
S2B_MSIL2A_20190716T042709_N0213_R133_T46RDQ_20190716T083530

and ramp functions. Therefore, the median filter is better than the mean
filter in terms of preserving the edges between two different features,
but it does not preserve single pixel-wide features, which will be altered
if speckle noise is present. Median filter preserves the texture informa
tion very well for small window size (3 � 3) but does not retain the mean
value at an acceptable level. Since the median is less sensitive than the
mean to extreme values (outliers), those extreme values are more
effectively removed.
Specifications of Sentinel-1 and Sentinel-2 products.

1.2. Measuring performance efficiency of SAR speckle filters

A speckle suppression filter is expected to filter the homogeneous

areas with reasonable speckle reduction. A good SAR despeckling
technique should have the following characteristics (i) scene feature
preservation (such as texture, linear features, and point features) (ii)
radiometric preservation (iii) speckle-noise reduction, smoothing, blur
Product name

reduction, and edge preservation. The evaluation of the performance of

the filters in de-speckling the SAR image is, therefore, necessary.
Selected Homogeneous area, linear feature and Edge from Kerala SAR
data is shown in Fig. 1.
Following are the parameters to evaluate the performances of a
SENTINEL-

SENTINEL-

despeckling filter:
Table 2

1A

2B

4
V.K. Rana and T.M.V Suryanarayana Remote Sensing Applications: Society and Environment 16 (2019) 100271

Fig. 4. Methodology.

Fig. 5. Mean square error. Fig. 6. Speckle suppression index.

1.2.1. Mean square error (MSE)

Mean square error (MSE) is the measurement of the difference be Iu ¼ unfiltered image
tween the output image and the input image. Higher the value of MSE, If ¼ filtered image
higher is the dissimilarity between the unfiltered image and the filtered K ¼ total number of pixels
image. A lower MSE value represents better image quality of the filtered
image (Senthilnath et al., 2013). MSE based measurements, however, 1.2.2. Speckle suppression index (SSI)
yield little information about the preservation of specific features as it The ability of a filter to suppress speckles is measured in terms of the
assesses the whole image. standard deviation of the image to its mean intensity. For homogeneous
� �
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi�
1 hX
n
�2 i areas, the ratio ​ VarðIu Þ meanðI Þ is regarded as the measurement
MSE ¼ Iu If u
K i¼1 of speckle strength. The speckle suppression index (SSI) is the coefficient
of variance of the filtered image standardized by that of the unfiltered

5
V.K. Rana and T.M.V Suryanarayana Remote Sensing Applications: Society and Environment 16 (2019) 100271

Fig. 7. Speckle mean preservation index. Fig. 10. Percent change in standard deviation.

1.2.3. Speckle mean preservation index (SMPI)

SSI is not accurate when the mean value is overestimated due to the
existence of extreme values in a relatively lower region of the image. In
addition to SSI, therefore, SMPI (Speckle Suppression and Mean Pres
ervation Index) is used to evaluate the filter efficiency (Wang and Ge,
2012). In terms of mean conservation and noise removal, lower SMPI
values show better filter efficiency (Shamsoddini et al., 2010).
!
qffiffiffiffiffiffiffiffiffiffiffiffiffi�ffiffi.
SMPI ¼ ​ Q � Var If p ffi ffi ffi ffiffi ffi ffi ffi ffi ffiffi ffi ffi ffi
VarðIu Þ

where Q is calculated as under:

� ��
�Q ¼ 1 þ meanðIu Þ mean If �

1.2.4. Equivalent number of looks (ENL)

Fig. 8. Equivalent number of looks. Another commonly used evaluation criterion is the equivalent
number of looks (ENL), also known as measure of the signal-to-noise
ratio. This index is calculated using the following equation:
� � �2
mean If
ENL ¼ �
standard deviation If

Higher ENL value for a filter represents higher efficiency in

smoothing speckle-noise over homogeneous areas (Bruniquel and Lopes,
1997).
The performance of a filter method is evaluated by considering
changes in mean and standard deviation. Ideally, the implementation of
filters should not result in any change in the mean of the target image,
while it should reduce the standard deviation.
To date, different classification algorithms, including the Support
Vector Machine (SVM), Maximum Likelihood (ML), Classification and
Regression Tree (CART), K-Nearest Neighbor (KNN), and Random Forest
(RF), have been applied in various studies. The RF classifier is one of the
Fig. 9. Percent change in mean. most effective approaches for classification (Breiman, 2001). Various
studies have been conducted using pixel-based RF algorithm for wetland
image, which is defined as: vegetation mapping using high spatial resolution SAR data (Amani et al.,
! 2017; Fu et al., 2017; Mahdianpari et al., 2017). Dumitru et al. (2015)
qffiffiffiffiffiffiffiffiffiffiffiffiffi�ffiffi � �
� pffiffiffiffiffiffiffiffiffiffiffiffiffiffi. applied the SVM classifier for the rapid mapping of damage assessment
SSI ¼ Var If mean I � � VarðIu Þ meanðI Þ
f u for flood in Germany in 2013 and the tsunami in Japan in 2011 using
TerraSAR-X pre- and post-flood data. The key element of a quantitative
accuracy assessment is the creation of a confusion matrix (Congalton,
VarðIf Þ ¼ variance of filtered image 2001; Janssen and Vanderwel, 1994; Story and Congalton, 1986). The
VarðIu Þ ¼ variance of the unfiltered image confusion matrix is represented by a table that shows correspondence
between the classification result and a reference image assigned to a
SSI has an inverse relationship with the suppression ability of the particular category, which is relative to the actual category as indicated
filter. The filtered image has lower variance because of speckle sup by the reference data. Senthilnath et al. (2013) used quantitative accu
pression SSI smaller than 1.0 indicates efficient speckle suppression racy assessment for flood mapping, which was estimated from the error
(Sheng and Xia, 1996). or confusion matrix. The overall accuracy was calculated by summing
the number of correctly classified values and dividing it by the total
­
6
V.K. Rana and T.M.V Suryanarayana Remote Sensing Applications: Society and Environment 16 (2019) 100271

Table 3
Quantitative evaluation of the filters over Kerala.
Kerala

MSE SSI SMPI ENL Percent Change – Mean Percent Change – Standard Deviation

Filters 3� 3 5� 5 3� 3 5� 5 3� 3 5� 5 3� 3 5� 5 3� 3 5� 5 3� 3 5� 5

Frostrowhead 0.005 0.004 0.937 0.983 0.930 0.979 0.120 0.109 0.825 0.396 7.05 2.138
Gamma map 0.007 0.006 0.503 0.361 0.503 0.359 0.417 0.812 0.100 0.609 49.71 64.14
Lee sigma 0.007 0.113 0.986 0.979 0.963 0.974 0.109 0.110 2.636 0.554 3.99 2.651
Lee 0.094 0.124 0.505 0.386 0.505 0.386 0.414 0.708 0.055 0.142 49.51 61.41
Boxcar 0.095 0.129 0.507 0.350 0.507 0.351 0.411 0.860 0.004 0.009 49.31 64.95
Median 0.124 0.148 0.248 0.165 0.221 0.142 1.712 3.883 12.58 15.77 78.28 86.11

Table 4
Quantitative evaluation of the filters over Assam.
Assam

MSE SSI SMPI ENL Percent Change – Mean Percent Change – Standard Deviation

Filters 3� 3 5� 5 3� 3 5� 5 3� 3 5� 5 3� 3 5� 5 3� 3 5� 5 3� 3 5� 5

Frost 0.007 0.004 0.769 0.886 0.763 0.874 0.297 0.224 0.924 1.579 23.82 12.75
Gamma map 0.011 0.030 0.757 0.593 0.756 0.590 0.306 0.499 0.122 0.539 24.38 41.00
Lee sigma 0.004 0.003 0.964 0.953 0.968 0.976 0.189 0.193 0.292 2.155 3.28 2.63
Lee 0.011 0.030 0.757 0.594 0.757 0.592 0.306 0.498 0.062 0.270 24.34 40.79
Boxcar 0.011 0.031 0.757 0.593 0.757 0.593 0.307 0.499 0.000 0.001 24.30 40.70
Median 0.015 0.043 0.668 0.431 0.625 0.379 0.394 0.945 7.085 13.26 37.95 62.60

number of values. the subtle structures of the image is essential, the efficiency of noise
Producer’s accuracy is the probability that value in a given class was suppression must be balanced with the effectiveness of the filter in order
correctly classified. to keep fine detail. The most preferred speckle filters are, therefore,
assessed in the current study over the data from Sentinel-1, intended for
Flooded area properly identified in a classification method
Producer’s ​ accuracy ¼ flood mapping applications. The Sentinel-1 (VV-vertical transmit, ver­
Flooded area in the reference ground truth
tical receive and VH-vertical transmit, horizontal receive) polarised
User’s accuracy is the probability that a value predicted to be in a filtered data were later used for performance evaluation of machine
certain class is really in that class. learning algorithms, namely, random forest (RF) and support vector
machine (SVM), to classify an inundated area. The accuracies of the
User’s accuracy ¼
Flooded area properly identified in a classification method classifications were assessed by the confusion matrix parameters, which
Total flooded area calculated from the method include kappa coefficient, overall accuracies, producer’s and user’s
accuracies.
The kappa coefficient measures the agreement between classification
and truth-values. A kappa value of 1 represents perfect agreement, while 2. Study area and datasets
a value of 0 represents no agreement
Observed accuracy Expected agreement 2.1. a. Kerala
Kappa coefficient ¼
1 Expected agreement
Kerala is a small, elongated coastal state in peninsular India’s south-
where western tip. It is surrounded by the Western Ghats in the east and the
Arabian Sea in the west. A part of the state of Kerala was considered in
Ct * ft þ nct * nft
Expected agreement ¼ this study. The state faces severe and varied damages due to floods and
ðAÞ2 heavy rainfall. Monsoon circulation dominates the climate of India and
Kerala in particular. The wind blows from the oceans to the south of the
Ct ¼ total flooded area from classified method Asian land masses during the half of the year, while a seasonal wind
ft ¼ actual total flooded area blows from the Asian land masses to the oceans in the south during the
nct ¼ total non-flooded area from a classified method other half of the year causing a spectacular reversal of pressure and wind
nft ¼ actual non-flooded total area patterns between the two six-month periods. South-west monsoon
A ¼ total area under the study (June–September) and post-monsoon (October–November) are the main
rainy seasons in Kerala. The state witnessed heavy floods in the year
The objective of this study was to develop an approach for opera­ 1924 and 1961. The IMD recorded rainfalls for 15 to 17 August 2018
tional flood extent mapping. Despeckling is an important preprocessing were found to be comparable to the rigorous storm that occurred in 1924
step as speckle complicates the interpretation of visual images making (Central Water Commission, 2018). Heavy rainfall resulted in high
automated digital image classification a challenging task. The choice of surface runoff in Kerala’s major river basins, filling all dams and sub­
filter relies on the particular application requirements and the data set sequent opening of these dams, causing widespread flooding in down­
characteristics used. Despeckle filters with excellent noise extraction stream areas, low-lying coastal areas, and Kerala’s backwaters. Fig. 2
capacities often appear to degrade the spatial and radiometric accuracy shows the area covered under the study.
of the actual image and trigger image reflection deterioration. This may The National Aeronautics and Space Administration Alaska Satellite
be acceptable for applications involving large-scale interpretation or Facility (NASA/ASF) houses a complete archive of Sentinel-1 SAR data
mapping of images. However, in many cases where the preservation of processed by the European Space Agency (ESA). The Sentinel-1 Level-1

7
V.K. Rana and T.M.V Suryanarayana Remote Sensing Applications: Society and Environment 16 (2019) 100271

Fig. 11. Visual comparison of de-noising methods on VV polarization over Kerala test data (a) Boxcar 3 � 3 (b) Frost 3 � 3 (c) Gamma map 3 � 3 (d) Lee 3 � 3 (e) Lee
sigma 3 � 3 (f) Median 3 � 3 (g) Boxcar 5 � 5 (h) Frost 5 � 5 (i) Gamma map 5 � 5 (j) Lee 5 � 5 (k) Lee sigma 5 � 5 (l) Median 5 � 5 and (m) Unfiltered test image in
VV polarization.

ground range detected (GRD) data acquired in interferometric wide owing to the increase in water concentrations of the Brahmaputra river
swath (IW) mode, which is the predefined mode over land with VV and and its related tributaries likely led from high continuous rainfall in the
VH polarizations, were downloaded via the ASF application program upper catchment regions of the Brahmaputra Basin. Fig. 3 shows the
ming interface (API). Sentinel-1 and Sentinel-2 data that are available area covered under the study. Sentinel-1 and Sentinel-2 data that are
closest to event date were acquired on 21 August 2018 at 00:40:44 and available closest to event date were acquired on 14 July 2019 at
22 August 2018 at 05:06:49, respectively. The specific parameters of the 11:57:18 and 16 July 2019 at 04:27:09, respectively. The specific pa
Sentinel-1 and Sentinel-2 products are given in Table 1. rameters of the Sentinel-1 and Sentinel-2 products are given in Table 2.

3. Methodology
2.2. b. Assam
3.1. Pre-processing
Assam is a state in northeast India, situated south of the eastern
Himalayas along the Brahmaputra and Barak River valley. The state has
A schematic of the Sentinel-1-based processing chain is outlined in
recently witnessed heavy flood in July 2019. The Brahmaputra basin
Fig. 4. The downloaded Sentinel-1 Leve-1 GRD data acquired in IW with
falls within the monsoon rainfall regime, getting an average rainfall of
VV and VH polarizations were loaded onto Sentinel Application
about 230 cm. The heavy floods in the Brahmaputra river in Assam

8
V.K. Rana and T.M.V Suryanarayana Remote Sensing Applications: Society and Environment 16 (2019) 100271

Fig. 12. Visual comparison of de-noising methods on VV polarization over Assam test data (a) Boxcar 3 � 3 (b) Frost 3 � 3 (c) Gamma map 3 � 3 (d) Lee 3 � 3 (e) Lee
sigma 3 � 3 (f) Median 3 � 3 (g) Boxcar 5 � 5 (h) Frost 5 � 5 (i) Gamma map 5 � 5 (j) Lee 5 � 5 (k) Lee sigma 5 � 5 (l) Median 5 � 5 and (m) Unfiltered test image in
VV polarization.

Fig. 13. Point target scattering on (a) Boxcar 3 � 3 (b) Lee 3 � 3 (c) Boxcar 5 � 5 and (d) Lee 5 � 5.

Platform (SNAP). SNAP offers a wide range of tools and features for the study sites to reduce the processing time. Multi-looking was then
Sentinel-1 imagery processing and analysis. Due to the large swath performed to reduce the standard deviation of the noise. The number of
width of the Sentinel-1 data, the image was first divided into a subset for Azimuth looks and the number range of looks (2 � 2) with mean GR

9
V.K. Rana and T.M.V Suryanarayana Remote Sensing Applications: Society and Environment 16 (2019) 100271

Fig. 14. Kerala (a) Raw VH amplitude data; (b) Raw VV amplitude data; (c) Multi-looked, calibrated, Filtered (Lee) VH data in dB; (d) Multi-looked, calibrated,
Filtered (Lee) VV data in dB; (e) The Range-Doppler terrain corrected VH data; (f) The Range-Doppler terrain corrected VV data; (g) Histogram for Sigma0 VH in dB,
and (h) Histogram for Sigma0 VV in dB.

mean pixel of 20 m were applied to a 1 m � 5 m (single look). The multi-

DNi ¼ pixel’s digital number
looked data were then calibrated to transform the pixel values from the
Ai ¼ absolute ​ calibration constant
digital values recorded by the sensor into backscatter coefficient values
or Sigma0 (σ0). This was achieved using the following equation:

jDNi j
2 3.2. Application of filters
σ0 ¼
Ai 2
Speckles inherently corrupt all radar images, degrading the image
quality, and making it more difficult to interpret features. Thus, it is

10
V.K. Rana and T.M.V Suryanarayana Remote Sensing Applications: Society and Environment 16 (2019) 100271

Fig. 15. Assam (a) Raw VH amplitude data; (b) Raw VV amplitude data; (c) Multi-looked, calibrated, Filtered (Lee) VH data in dB; (d) Multi-looked, calibrated,
Filtered (Lee) VV data in dB; (e) The Range-Doppler terrain corrected VH data; (f) The Range-Doppler terrain corrected VV data; (g) Histogram for Sigma0 VH in dB,
and (h) Histogram for Sigma0 VV in dB.

Table 5 Table 6
Comparison of user’s accuracy (UA), producer’s accuracy (PA), overall accuracy Inundated area statistics of RF and SVM over cloud-free optical data for Kerala
(%), and kappa coefficient using random forest tree and support vector machine region.
algorithms for VV and VH polarization over Kerala region. VV Polarization VH Polarization NDWI
VV Polarization VH Polarization
RF SVM RF SVM –
RF SVM RF SVM
Inundated area (Km2) 40.25 34.82 34.17 35.19 41.78
PA UA PA UA PA UA PA UA Rest (Km2) 16.29 21.72 22.38 21.36 14.77

Inundation 0.89 0.96 0.79 0.99 0.78 0.99 0.79 0.99

Rest 0.88 0.72 0.98 0.60 0.98 0.58 0.98 0.60
kernel size, used in the study were available in SNAP and applied using
Kappa 0.72 0.64 0.61 0.63
coefficient default system parameters.
Overall Accuracy 88.80% 83.80% 82.60% 83.60%
(%)
3.3. Evaluating the performance efficiency of filters
Note: UA - User’s accuracy, PA - Producer’s accuracy VV - Vertical-Vertical, VH -
Vertical-Horizontal, RF - Random Forest, SVM -Support Vector Machine. Several techniques are available to quantitatively assess the effi
ciency of a speckle filter in distinct ways, for example, edge conservation
often necessary to enhance the image by filtering speckles before data and conservation of features. The findings of the various measurements
can be used in different applications. All of the filters, namely, Boxcar, may be contradictory. Therefore, distinct techniques of evaluation
Median, Frost, Gamma map, Lee and Lee sigma with 3 � 3 and 5 � 5 should be used to discover the optimum tradeoff between the various

11
­
V.K. Rana and T.M.V Suryanarayana Remote Sensing Applications: Society and Environment 16 (2019) 100271

Table 7 producer’s accuracy, user’s accuracy, and the overall accuracy of the
Comparison of user’s accuracy (UA), producer’s accuracy (PA), overall accuracy classifiers. The overall accuracy gives the correctly classified regions for
(%), and kappa coefficient using random forest tree and support vector machine the image and is calculated by the proportion of the correctly classified
algorithms for VV and VH polarization over Assam region. pixels to the total number of pixels in the confusion matrix.
VV Polarization VH Polarization To calculate inundation for entire scene, thresholding was done in
RF SVM RF SVM SNAP by using the conditional function given below after carefully
analysing the histogram:
PA UA PA UA PA UA PA UA

Inundation 0.77 1.0 0.82 1.0 0.81 0.99 0.89 0.99 If Sigma0 VV/VH db < X then 1 else 0
Rest 1.0 0.62 1 0.69 0.99 0.67 0.99 0.77
Kappa 0.65 0.72 0.70 0.81 X ¼ threshold value
coefficient
1 ¼ inundated pixels
Overall Accuracy 83.60% 87.60% 86.60% 92.00%
(%) 0 ¼ Rest

Note: UA - User’s accuracy, PA - Producer’s accuracy VV - Vertical-Vertical, VH -

The resulted output was later used to calculate the difference in
Vertical-Horizontal, RF - Random Forest, SVM -Support Vector Machine.
inundated areas calculated by thresholding technique and images clas
sified using the random forest and support vector machine algorithms
elements of the image. having highest accuracy for both the study areas.
The sigma0 values of VV and VH polarised data with the applied filter
was terrain corrected using SNAP’s ‘Range Doppler Terrain Correction’ 4. Results and discussion
algorithm with an SRTM 1 ArcSecond digital elevation model. The
bilinear interpolation method was used for DEM and Image resampling 4.1. Speckle filtering
with a pixel spacing of 20 � 20 m. Terrain correction helps in improving
the geometric representation of the real-world surface. This is needed The performance efficiency of filters in speckle suppression, feature
because, during image capture, topographical variations and off-nadir preservation, and preventing the loss of meaningful data was evaluated
distortion unsettle the image. using MSE, ENL, SSI, SMPI, examination of mean, standard deviation,
and also by close visual assessment.
The MSE value indicates the amount of error present in a filtered
3.4. Machine learning algorithms for classification
image. It allows the comparison of the pixel values of a filtered image to
the degraded image before filtering. It is observed that for study area
The terrain corrected images were classified using the random forest
Kerala (window size 3 � 3) Frost, Gamma map, and Lee sigma filters
and support vector machine algorithms as a next step. For both the
showed very low MSE, indicating their effectiveness in feature preser
classifiers, the same number of training samples was used. The training
vation. Median filter showed the highest MSE values indicating poor
inundated pixels covered 5.2 Km2 and the rest of the training pixels
performance in terms of feature preservation. Boxcar and Lee filters
covered 3.1 Km2 of the study area Kerala. Similarly, for the study area
performed moderately. For Kerala (window size 5 � 5) and Assam
Assam, the training inundated pixels covered 11 Km2 and the rest of the
(window size 3 � 3 and 5 � 5), Frost and Lee sigma showed low MSE,
training pixels covered 54 Km2.
Median filter showed the highest MSE and moderate performance was
During the southwest monsoon season, it is nearly impossible to
observed for Boxcar, Gamma map and Lee (Fig. 5). The high value of
obtain 100% cloud-free data, however, a small extent of the cloud-free
MSE depicts a greater difference between the original test image and
data can be used for validation. The normalized difference water
despeckled image which concludes the significant speckle reduction but
index (NDWI) is defined for Sentinel–2 data as ((B03) (B08)/(B03) þ
at the cost of feature loss.
(B08)), where B03 is a green band and B08 is the near-infrared band
Lower SSI and SMPI values were obtained for Median filter (3 � 3
(Mcfeeters, 1996). When NDWI is applied over a multispectral image,
and 5 � 5) in both the study areas indicating high efficiency in speckle
the water feature has positive values, while soil and terrestrial vegeta
suppression. Moderate values were obtained for Gamma map, Lee and
tion features have zero or negative values. This is because NIR is
Boxcar indicating medium proficiency in the reduction of speckle. None
absorbed strongly by water but reflected strongly by terrestrial vegeta
of the filters produced value 0 and 1. A value of 0, indicates complete
tion and dry soil, while in a green light, water has high reflectance than
mean preservation and 100% noise reduction. Whereas, a value of 1
terrestrial vegetation and soil. Therefore, the NDWI was applied to
indicates 0% speckle reduction. It was observed that as the window size
extract water from the optical data. A cloud-free part of satellite optical
increased from 3 � 3 to 5 � 5, the SSI values reduced for all the filters
image was collected by Sentinel-2 at 05:06:49 on 22 August 2018, 28 h
(Figs. 6 and 7), indicating superior speckle suppression with an increase
after the Sentinel-1 pass over the study area Kerala. Similarly, a
in the size of the moving window. The lowest SSI values, as well as SMPI
cloud-free part of satellite optical image was collected by Sentinel-2 at
values, corresponds to Median filter (3 � 3 and 5 � 5) indicating high
04:27:09 on 15 July 2019, 40 h after the Sentinel-1 pass over the study
efficiency in speckle suppression.
area Assam. Sentinel-2 data were converted to reflectance and dark
For both the study areas Median filter (3 � 3 and 5 � 5) had high ENL
object subtraction atmospheric correction (DOS1) was applied. The
values indicating a higher efficiency in smoothing speckle noise over
corrected Sentinel-2 image was used to validate the extent of the flood.
homogeneous areas, which shows enhanced capacity to distinguish the
The normalized difference water index (NDWI), established earlier
distinct gray levels within the image. For constant flat areas where the
(Mcfeeters, 1996) to extract the water from the optical data, was
sample variance is null, ENL become ∞, this will repute highly blurred
calculated as:
data as excellent. According to ENL, SSI and SMPI values, Gamma map
ρGreen ​ ​ ρNIR (3 � 3 and 5 � 5), Lee (3 � 3 and 5 � 5) and Boxcar (3 � 3 and 5 � 5)
NDWI ¼
ρGreen þ ρNIR filters showed moderate performance for both the study areas (Fig. 8).
The application of filters should ideally not bring about any change
Moreover, stratified random sampling was used with 500 points for
in the mean of target image while it should reduce the standard devia
the accuracy assessment. The classification maps were evaluated in
tion. The filter, Median (3 � 3 and 5 � 5) for both the study areas was
terms of their overall accuracy (OA), producer’s accuracy (PA), user’s
most effective in increasing the standard deviation but also changed the
accuracy (UA), and the kappa index of agreement (k) or kappa coeffi
mean value considerably thereby implying that the filters reduced
cient. Confusion matrix was created to compare the kappa coefficient,

12
V.K. Rana and T.M.V Suryanarayana Remote Sensing Applications: Society and Environment 16 (2019) 100271

Fig. 16. Kerala (a) Random forest tree classification on filtered VH; (b) Support vector machine classification on VH; (c) Random forest tree classification on filtered
VV; (d) Support vector machine classification on VV.

13
 V.K. Rana and T.M.V Suryanarayana Remote Sensing Applications: Society and Environment 16 (2019) 100271

Fig. 17. Assam (a) Random forest tree classification on filtered VH; (b) Support vector machine classification on VH; (c) Random forest tree classification on filtered
VV; (d) Support vector machine classification on VV.

speckle but also caused considerable loss of meaningful data. The me performance of Boxcar and Lee filters was far better in feature preser
dian filter is better than the Boxcar filter in terms of preserving the edges vation in the filtered images followed by Gamma map. Although quan
between two different features, but it does not preserve single pixel-wide titative measures are often employed to compare different speckle
features, which will be altered if speckle noise is present. Median filter suppression filters, it has been noted by Raouf and Lichtenegger (1997)
preserves the texture information very well for small window size and others that visual inspection probably provides the best assessment
(3 � 3) but does not retain the mean value at an acceptable level. Since of the performance of a speckle filter. Visual assessment is an easy and
the median is less sensitive than the mean to extreme values (outliers), efficient way to investigate both the capability of a filter to suppress
those extreme values are more effectively removed. On the contrary, speckles and its effectiveness in preserving image details. Lee et al.
Boxcar (3 � 3 and 5 � 5) for both the study areas, made the least change (1994) stated that, in general, filters using small windows (such as 3 � 3)
in the mean, while reducing the standard deviation moderately. How preserve texture information better. Visual examination was, therefore,
ever, Lee filter (3 � 3 and 5 � 5) provided a fair balance by reducing the carried out and it was observed that the filters which reduced speckles
standard deviation without drastically affecting the mean (Figs. 9 and effectively also resulted in considerable loss of meaningful data (Figs. 11
10, and Tables 3 and 4). (a)-(m) and 12 (a)-(m)). Lee sigma, and Gamma map clearly resulted in
The Median filter performed well in terms of ENL, SSI, and SMPI the loss of edges and details. It was difficult to grade Boxcar, Frost and
values; however, its performance in terms of speckle reduction and Lee filter visually, since the variation was not perceivable. However, it
feature preservation was far inferior compared to the Boxcar and Lee was observed that point scatters were over filtered, transformed to
filter. However, since the objective of the speckle filtering was to use spread targets and sharp edges were generally blurred in Boxcar filter
these SAR images for inundation mapping, the performance of the filters (Fig. 13). We chose Lee filter for further analysis as it had low MSE, SSI
on a water body in terms of reduction of standard deviation while pre and SMPI values, and a higher percentage change in standard deviation
serving the mean of the original image was the most important. The compared to the Boxcar filter for most of the cases.

14
­
­
V.K. Rana and T.M.V Suryanarayana Remote Sensing Applications: Society and Environment 16 (2019) 100271

Fig. 18. (a) Natural colour composite R–B04 G-B03 B–B02, (b) Green band (c) Near-infrared (d) Calculated NDWI over cloud-free extent for Kerala region.

The raw SAR data in VH and VV polarization acquired during the classifiers produced comparable overall accuracies in VH polarization
crisis events on 21 August 2018 and 14 Jul 2019 for study areas Kerala (SVM achieved higher classification accuracy than RF by about 1%). The
and Assam are shown in Fig. 14 (a) and (b), and Fig. 15 (a) and (b), NDWI calculated for the cloud-free extent is shown in Fig. 18 (d). The
respectively. The multi-looked, calibrated, filtered (Lee) data of the VH inundated area in the calculated NDWI over the cloud-free extent is
and VV are shown in Fig. 14 (c) and (d), and Fig. 15 (c) and (d), 73.88%, which is 41.78 km2. However, it has also been observed
respectively. Multi-looked, calibrated, filtered (Lee) SAR data were not (Table 6) that the inundated area using random forest classification on
projected on the map coordinates of each pixel. The pixel was in the filtered VV data over the cloud-free extent is 71.18%, which is
original coordinate position of data (rows/columns) in the field of 40.25 Km2. For the study area Assam, the SVM classifier exhibited
ground range. In the orthorectified imageries, each of the pixels that maximum overall accuracy of 92% with the kappa coefficient value of
were corrected and projected using the Range-Doppler terrain correc 0.81. Both the classifiers obtained better accuracy results in VH polari
tion appeared at the actual position. The Range-Doppler terrain cor zation compared to the VV polarization. The least overall accuracy of
rected pixels and their respective histograms for Kerala and Assam 83.60% and a kappa coefficient value of 0.65 were observed with
regions are shown in Fig. 14 (e)–(h), and Fig. 15 (e)- (h), respectively. random forest in VV polarization, which was followed by the RF in VH
Tables 5 and 7, and Figs. 16 and 17 show the comparison of classi polarization. SVM achieved higher classification accuracy than RF by
fication results of random forest classifier and support vector machine about 5.38% in VH polarization. The NDWI calculated for the cloud-free
classifier for VV and VH polarization. The training data were kept the extent is shown in Fig. 19 (d). The inundated area in the calculated
same for both the classifiers to avoid optimistic bias in the classification. NDWI over the cloud-free extent is 74.09%, which is 491.47 km2.
For the study area Kerala, the random forest classifier exhibited However, it has also been observed (Table 8) that the inundated area
maximum overall accuracy of 88.80% with the kappa coefficient value using SVM classification on filtered VH data over the cloud-free extent is
of 0.72. Both the classifiers obtained better accuracy results in VV po 62.76%, which is 416.99 Km2. To calculate inundation for the entire
larization compared to the VH polarization. The least overall accuracy of scene, threshold value of 10.96 and 19.58 was selected for Kerala
82.60% and a kappa coefficient value of 0.63 were observed with and Assam region respectively after analysing the histograms. The
random forest in VH polarization, which was followed by the support calculated inundated area (Fig. 20) with thresholding technique for
vector machine in VV polarization. RF achieved higher classification Kerala was found 204 km2 (2% more than classified VV polarised data
accuracy than SVM by about 5% in VV polarization. However, both the using RF algorithm). Similarly, the calculated inundated area with

15
V.K. Rana and T.M.V Suryanarayana Remote Sensing Applications: Society and Environment 16 (2019) 100271

Fig. 19. (a) Natural colour composite R–B04 G-B03 B–B02, (b) Green band (c) Near-infrared (d) Calculated NDWI over cloud-free extent for Assam region.

balance in feature preservation as well in despeckling compared to the

Table 8 other filters used in the study. The accuracy assessment of machine
Inundated area statistics of RF and SVM over cloud-free optical data for Assam learning algorithms for flood classification over Kerala shows that
region. random forest classifier has higher classification accuracy than the
VV Polarization VH Polarization NDWI support vector machine (about 5% higher in VV polarization). However,
RF SVM RF SVM – both the classifiers performed better in VV polarization than VH polar
ization. For study area Assam, SVM in VH polarization achieved higher
Inundated area (Km2) 363.92 389.21 387.62 416.99 491.47
Rest (Km2) 300.41 275.12 276.70 247.34 171.86
accuracy and least performance was observed by RF in VV polarization.
The proposed approach shows the potential for monitoring damages
caused by floods, providing basic information that can help local com
thresholding technique for Assam was found 3368.90 km2 (23.46% munities manage water-related risk, planning land and water manage
more than classified VH polarised data using SVM algorithm). For Assam ment as well as other flood control programs.
region the variation is very large, a single threshold did not hold well as
large swath of a SAR image suffers from environment heterogeneity Ethical approval
caused by wind-roughening and satellite framework parameters.
The presented research has not involved human subjects, human
5. Conclusion material, or human data and therefore ethical approval has not been
required.
Effective and quick response is required during disasters like flood
ing. Rapid mapping of such events will be beneficial to urban and Compliance with ethical standards
infrastructure planners, risk managers and disaster responses during
extreme and intense rainfall events. The study shows a simple and This study is in full compliance with all applicable ethical standards.
efficient method for mapping inundation extent with only the C-band
S1A, with coarser geometric resolution and fixed polarizations (VV-VH) Informed consent
by considering the case of Kerala and Assam. Backscattering coefficient
values become high as the water roughness causes high signal return, This study does not involve any subject participation and therefore
decreasing the contrast and making the separation of the land-water informed consent is not applicable.
covers difficult. Despeckle filters with good noise removal capabilities
often tend to degrade the spatial and radiometric resolution of an Declaration of competing interest
original image and cause the loss of image details. This may be
acceptable for applications involving large scale image interpretation or On behalf of all authors, the corresponding author states that there is
mapping. However, the retention of the subtle structures of an image is no conflict of interest.
important and, therefore, the performance of speckle noise suppression
technique must be balanced with the filter’s effectiveness to preserve the
fine details. The performance evaluation of de-noising methods in this
study showed that Lee filter with 3 � 3 kernel size provided a good

16
­
­
V.K. Rana and T.M.V Suryanarayana Remote Sensing Applications: Society and Environment 16 (2019) 100271

Fig. 20. (a) Random forest tree classification on filtered VV data (Kerala) (b) Classified filtered VV data with threshold value of 10.96 (Kerala) (c) Support vector
machine classification on filtered VH data (Assam) (d) Classified filtered VH data with threshold value of 19.58 (Assam).

References Fu, B., Wang, Y., Campbell, A., Li, Y., Zhang, B., Yin, S., et al., 2017. Comparison of
object-based and pixel-based Random Forest algorithm for wetland vegetation
mapping using high spatial resolution GF-1 and SAR data. Ecol. Indicat. 73,
India Meteorological Department, 2018. Rainfall over Kerala during Monsoon Season-
105–117. https://doi.org/10.1016/j.ecolind.2016.09.029.
2018 and forecast for next 5 days. Retrieved from. https://www.imdtvm.gov.
Gitas, I.Z., Polychronaki, A., Katagis, T., Mallinis, G., 2008. Contribution of remote
in/images/rainfall. over kerala during monsoon season-2018 and forecast for next 5
sensing to disaster management activities: a case study of the large fires in the
days.pdf.
Peloponnese, Greece. Int. J. Remote Sens. 29 (6), 1847–1853. https://doi.org/
Amani, M., Salehi, B., Mahdavi, S., Granger, J., Brisco, B., 2017. Wetland classification in
10.1080/01431160701874553.
Newfoundland and Labrador using multi-source SAR and optical data integration.
Goodman, J.W., 1976. Some fundamental properties of speckle. JOSA 66 (11),
GIScience Remote Sens. 54 (6), 779–796. https://doi.org/10.1080/
1145–1150.
15481603.2017.1331510.
Janssen, L.L., Vanderwel, F.J., 1994. Accuracy assessment of satellite derived land-cover
Breiman, L., 2001. Random forests. Mach. Learn. 45 (1), 5–32. https://doi.org/
data: a review. Photogramm. Eng. Remote Sens. 60 (4).
10.14923/transinfj.2015IUP0008.
Khan, S.D., 2005. Urban development and flooding in Houston Texas, inferences from
Bruniquel, J., Lopes, A., 1997. Multi-variate optimal speckle reduction in SAR imagery.
remote sensing data using neural network technique. Environ. Geol. 47 (8),
Int. J. Remote Sens. 18 (3), 603–627. https://doi.org/10.1080/014311697218962.
1120–1127. https://doi.org/10.1007/s00254-005-1246-x.
Central Water Commission, 2018. Kerala Floods August 2018. Retrieved from. https://
Lee, J.S., Pottier, E., 2009. Polarimetric Radar Imaging: from Basics to Applications. CRC
reliefweb.int/sites/reliefweb.int/files/resources/Rev-0.pdf.
press.
Chunming, H., Huadong, G., Yun, S., Jingjuan, L., 2005, July. Detection of the flood
Lee, J.S., Jurkevich, I., Dewaele, P., Wambacq, P., Oosterlinck, A., 1994. Speckle filtering
boundary in SAR image using texture. In: Proceedings. 2005 IEEE International
of synthetic aperture radar images: a review. Remote Sens. Rev. 8 (4), 313–340.
Geoscience and Remote Sensing Symposium, vol. 5. IGARSS’05, pp. 3697–3699,
https://doi.org/10.1080/02757259409532206.
2005, (Ieee).
Lee, J. Sen, 1981. Speckle analysis and smoothing of synthetic aperture radar images.
Congalton, R.G., 2001. Accuracy assessment of wildland fire and other spatial
Comput. Graph. Image Process. 17 (1), 24–32. https://doi.org/10.1016/S0146-664X
information. Int. J. Wildland Fire 10, 321–328.
(81)80005-6.
de Leeuw, M.R., de Carvalho, L.M.T., 2009. Performance evaluation of several adaptive
Lopes, A., Nezry, E., Touzi, R., Laur, H., 1990. May). Maximum a posteriori speckle
speckle filters for SAR imaging. Anais XIV Simp� osio Brasileiro de Sensoriamento
filtering and first order texture models in SAR images. In: 10th Annual International
Remoto 7299–7305.
Symposium on Geoscience and Remote Sensing, pp. 2409–2412 (Ieee).
Dumitru, C.O., Cui, S., Faur, D., Datcu, M., 2015. Data analytics for rapid mapping: case
Mahdianpari, M., Salehi, B., Mohammadimanesh, F., Brisco, B., 2017. An assessment of
study of a flooding event in Germany and the tsunami in Japan using very high
simulated compact polarimetric SAR data for wetland classification using random
resolution SAR images. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 8 (1),
forest algorithm. Can. J. Remote Sens. 43 (5), 468–484. https://doi.org/10.1080/
114–129. https://doi.org/10.1109/JSTARS.2014.2320777.
07038992.2017.1381550.
Durand, J.M., Gimonet, B.J., Perbos, J.R., 1987. SAR data filtering for classification. IEEE
Mcfeeters, S.K., 1996. The use of the Normalized Difference Water Index (NDWI) in the
Trans. Geosci. Remote Sens. (5), 629–637. https://doi.org/10.1109/
delineation of open water features. Int. J. Remote Sens. https://doi.org/10.1080/
TGRS.1987.289842. GE-25.
01431169608948714.
Frost, V.S., Stiles, J.A., Shanmugam, K.S., Holtzman, J.C., 1982. A model for radar
images and its application to adaptive digital filtering of multiplicative noise. IEEE
Trans. Pattern Anal. Mach. Intell. 4 (2), 157–166.

17
V.K. Rana and T.M.V Suryanarayana Remote Sensing Applications: Society and Environment 16 (2019) 100271

Nirupama, N., Simonovic, S.P., 2007. Increase of flood risk due to urbanisation: a Serpico, S.B., Dellepiane, S., Boni, G., Moser, G., Angiati, E., Rudari, R., 2012.
Canadian example. Nat. Hazards 40 (1), 25–41. https://doi.org/10.1007/s11069- Information extraction from remote sensing images for flood monitoring and damage
006-0003-0. evaluation. Proc. IEEE 100 (10), 2946–2970.
Qiu, F., Berglund, J., Jensen, J.R., Thakkar, P., Ren, D., 2004. Speckle noise reduction in Shamsoddini, A., Trinder, J.C., Wagner, W., Sz� ekely, B., 2010. Image Texture
SAR imagery using a local adaptive median filter. GIScience Remote Sens. 41 (3), Preservation in Speckle Noise Suppression (na).
244–266. https://doi.org/10.2747/1548-1603.41.3.244. Sheng, Y., Xia, Z.G., 1996. May). A comprehensive evaluation of filters for radar speckle
Raouf, A., Lichtenegger, J., 1997. Integrated use of SAR and optical data for coastal zone suppression. In: IGARSS’96. 1996 International Geoscience and Remote Sensing
management. In: Third ERS Symposium on Space at the Service of Our Environment, Symposium, vol. 3. IEEE, pp. 1559–1561.
vol. 414, p. 1089. Story, M., Congalton, R.G., 1986. Accuracy assessment: a user’s perspective.
Senthilnath, J., Shenoy, H.V., Rajendra, R., Omkar, S.N., Mani, V., Diwakar, P.G., 2013. Photogramm. Eng. Remote Sens. 52 (3), 397–399.
Integration of speckle de-noising and image segmentation using Synthetic Aperture Wang, X., Ge, L., 2012. Evaluation of Filters for ENVISAT ASAR Speckle Suppression in
Radar image for flood extent extraction. J. Earth Syst. Sci. 122 (3), 559. https://doi. Pasture Area.
org/10.1007/s12040-013-0305-z. Zhang, F., Zhu, X., Liu, D., 2014. Blending MODIS and Landsat images for urban flood
mapping. Int. J. Remote Sens. 35 (9), 3237–3253. https://doi.org/10.1080/
01431161.2014.903351.

18