FIR FILTER FOR SPECKLE NOISE REJECTION

A PROJECT REPORT
Submitted in partial fulfillment of the
requirement for the award of the
Degree of
BACHELOR OF TECHNOLOGY
in
ELECTRONICS AND COMMUNICATION ENGINEERING
By
DUNNA JASWANTH
11BEC1061
CHINTAKUNTA NANDA KISHORE REDDY
11BEC1068
Under the Guidance of
DR.T.VIGNESWARAN

SCHOOL OF ELECTRONICS ENGINEERING

VIT UNIVERSITY
CHENNAI (TN) 600127
MAY 2015

FIR FILTER FOR SPECKLE NOISE REJECTION

A PROJECT REPORT
Submitted in partial fulfillment of the
requirement for the award of the
Degree of
BACHELOR OF TECHNOLOGY
in
ELECTRONICS AND COMMUNICATION ENGINEERING
By
DUNNA JASWANTH
11BEC1061
CHINTAKUNTA NANDA KISHORE REDDY
11BEC1068
Under the Guidance of
DR.T.VIGNESWARAN

SCHOOL OF ELECTRONICS ENGINEERING

VIT UNIVERSITY
CHENNAI (TN) 600127
MAY 2015

CERTIFICATE
This is to certify that the Project work titled FIR filter for speckle noise rejection that is
being submitted by Dunna Jaswanth and Chintakunta Nanda Kishore Reddy is in
partial fulfillment of the requirements for the award of Bachelor of Technology, is a record
of bonafide work done under my guidance. The contents of this Project work, in full or in
parts, have neither been taken from any other source nor have been submitted to any other
Institute or University for award of any degree or diploma and the same is certified.

Guide
The thesis is satisfactory / unsatisfactory

Internal Examiner

External Examiner

Approved by

Program Chair

ii

ACKNOWLEDGEMENTS

I would like to thank my project guide Dr.T.Vigneswaran for his continuous guidance and
support. I would also like to thank my family and friends for supporting me through the
difficult times.

Reg. No. 11BEC1061

Reg. No. 11BEC1068

iii

TABLE OF CONTENTS

2
3

5
6
7
8

ABSTRACT

LIST OF ABBREVIATIONS

vi

LIST OF FIGURES

vii

LIST OF TABLES

viii

EXECUTIVE SUMMARY
Introduction
1.1 Objective

ix
1
2

1.2 Motivation
1.3 Background
Project Description and Goals
Technical Specification

3
4
20
22

3.1

22

Software module

3.2 Verilog

24

3.3

27

MATLAB

3.4 Hardware modules

Design Approach and Details
4.1 Hardware Design

28
32
32

4.2

33

Hardware Implementation

4.3 Software Implementation

4.4 Codes and Standards
4.5 Constraints, Alternatives and Tradeoffs
Schedule, Tasks and Milestones
Project Demonstration
Cost Analysis
Summary
References
Appendix

36
37
38
39
40
44
45
47
49

ABSTRACT

iv

Rejecting noise from the satellite image, medical image etc. is a challenge for the researchers
involved in digital image processing. The speckle noise is commonly found in Synthetic
Aperture Radar (SAR) images and satellite images.
The aim of this project is to remove the speckle noise from the satellite images using FIR
filter. The image filtering algorithms are applied on the images to remove the different types
of noises that are either present in the image during capturing, or injected in to the image
during the transmission.
In this project, an improved method is proposed that reduces the speckle noise in SAR images
and satellite images, and retains the original content of these images. This method enhances
the signal-to-noise ratio and help in perceiving the original features of the images.
The output de-noised image is produced by applying selective filter and it will also be
implemented in hardware using Field Programmable Gate Array (FPGA) board.

LIST OF ABBREVIATIONS
FFT

Fast Fourier Transform

v

FPGA

Field Programmable Gate Array

LCD

Liquid Crystal Display

LED

Light Emitting Diode

RADAR

Radio Detection and Ranging

SAR

Synthetic Aperture Radar

SDRAM

Synchronous Dynamic Random Access Memory

USB

Universal Serial Bus

LIST OF FIGURES
Figure No.

Figure name

Page no.
vi

1.1

Original SAR image

1.2

SAR image corrupted with speckle noise

1.3

Wavelet Packet decomposition over 3 levels

10

3.1

n-bit Magnitude Comparator

23

3.2

8-bit comparator

24

3.3

Altera DE2-115 (Education and Development board)

29

3.4

Block Diagram of the DE2-115 Board

30

4.1

Hardware design

32

4.2

Connections between the slide switches and Cyclone IV E FPGA

33

4.3

Connections between the LEDs and Cyclone IV E FPGA

34

4.4

Design Procedure of Speckle Noise Reduction in MATLAB

36

6.1

Input image

40

6.2

Speckle noise added image

40

6.3

SavitzkyGolay filtered image

40

6.4

Median filtered image

41

6.5

Comparison of Speckled and De-speckled output images

41

6.6

Output of Verilog software code of 8-bit

42

6.7

Output of Verilog software code

42

6.8

Hardware output on Altera DE2-115 FPGA kit

43

LIST OF TABLES
Table No.
3.1

Table name
1-bit comparator truth table

Page no.
22
vii

7.1

Cost analysis

44

EXECUTIVE SUMMARY
SAR images are used to detect ground targets and to recognise spatial patterns. But
generally SAR image is corrupted by speckle noise. Reduction of speckle noise is one of the
viii

most important processes to increase the quality of radar coherent images. Speckle is a
granular noise that inherently exists in and degrades the quality of the active radar and SAR
images. Before using active radar and SAR images, the very first step is to reduce the effect
of Speckle noise.
The presence of speckle not only complicates visual image interpretation, but also
makes automated digital image classification a difficult problem. Therefore, speckle noise in
radar data must often be reduced before the data can be used for further analysis or
information extraction. Dozens of de-speckle filters are in order to remove multiplicative
speckle noise. In general, no filter consistently outperforms other filters. Each filter has its
unique strengths and limitations.
Thus, the choice of which filter to use is dependent on the requirements of the specific
application and the characteristics of the dataset employed. De-speckle filters with good noise
removal capabilities often tend to degrade the spatial and radiometric resolution of an original
image and cause the loss of image detail. This may be acceptable for applications involving
large scale image interpretation or mapping. However, in many cases where the retention of
the subtle structures of the image is important, the performance of noise suppression must be
balanced with the filters effectiveness in order to preserve fine detail.
It is necessary to have knowledge about the noise present in the image so as to select
the appropriate de-noising algorithm. The proposed method reduces speckle noise in SAR
images by retaining the original content of these images. This method enhances the signal to
noise ratio and perceives the original features of the images. The output de-noised image is
produced by applying selective filter.

ix

CHAPTER 1
INTRODUCTION
Visual information transmitted in the form of digital images is becoming a major
method of communication in the modern age. In image processing, image obtained after
transmission is often corrupted by different type of noises. But Synthetic Aperture Radar
(SAR) image is corrupted by speckle noise [1]. Reduction of speckle noise is one of the most
important processes to increase the quality of radar coherent images [2]. Image variances or
speckle is a granular noise that inherently exists in and degrades the quality of the active
radar and SAR images. Before using active radar and SAR images, the very first step is to
reduce the effect of Speckle noise.
SAR uses microwave radiation to illuminate the Earths surface. An image is formed
after the SAR system receives the coherent sum of reflected radiation at the antenna that is
synthesized by the large motion of the sensor system [3]. Since SAR provides its own
illumination, it overcomes some of the fundamental problems associated with conventional
passive remote sensing. For example, SAR is not affected by cloud cover or variation in solar
illumination, and thus can operate day or night [4]. Under certain circumstances, radar can
partially penetrate arid and hyper-arid surfaces, revealing subsurface features of the Earth.
Although radar does not penetrate standing water, it can reflect the surface action of oceans,
lakes, and other bodies of water and may sometimes provide information about the bottom
features of the water body [5].
Electromagnetic waves emitted by active sensors travel in phase and interact
minimally on their way to the target area. After interaction with the target, these waves are no
longer in phase even though they are coherent in frequency [6]. This is caused by several
factors, such as the difference in distance the waves travel back from different targets, or the
single versus multiple bounce scattering due to the variance in surface roughness. When the
synthesized antenna moves, the signals may go out of phase. The out-of-phase waves
interfere constructively or destructively to produce stronger or weaker signals. When a SAR
image is formed by coherently processing the backscatter returns from successive radar
pulses, this effect causes a pixel-to-pixel variation in intensity, which manifests itself as a
salt-and-pepper granular pattern called speckle (or fading). These bright and dark pixels result
1

in a SAR image that fails to have a constant mean radiometric level in homogeneous areas
[7].
Radar speckle noise has a standard deviation linearly related to the mean and is often
modelled as a multiplicative process [8]. This means that the higher the signal strength the
higher the noise. As a result, more speckle noise is commonly present near brighter pixel
areas. The noise of single-look SAR amplitude imagery often has a Rayleigh distribution,
whereas that of single-look intensity imagery has a negative exponential distribution [9].
Multi-look SAR imagery usually follows a gamma distribution, assuming the looks are
independent.
The presence of speckle may decrease the utility of SAR imagery by reducing the
ability to detect ground targets and obscuring the recognition of spatial patterns [10].
Consequently, it not only complicates visual image interpretation, but also makes automated
digital image classification a difficult problem. Therefore, speckle noise in radar data must
often be reduced before the data can be used for further analysis or information extraction
[11]. Dozens of de-speckle filters have been in order to remove multiplicative speckle noise.
In general, no filter consistently outperforms others. Each filter has its unique strengths and
limitations [12].
Thus, the choice of which filter to use is dependent on the requirements of the specific
application and the characteristics of the dataset employed [13]. De-speckle filters with good
noise removal capabilities often tend to degrade the spatial and radiometric resolution of an
original image and cause the loss of image detail. This may be acceptable for applications
involving large scale image interpretation or mapping. However, in many cases where the
retention of the subtle structures of the image is important, the performance of noise
suppression must be balanced with the filters effectiveness in order to preserve fine detail.

1.1 OBJECTIVE
An image is often corrupted by noise since its acquisition or transmission. The goal of
de-noising is to remove the noise while retaining as much as possible the important signal
features of an image. The most critical part of developing a method for recovering a signal
from its noisy environment seems to be choosing a reasonable statistical (or analytic)
description of the physical phenomena underlying the data-formation process. The
availability of an accurate and reliable model of speckle noise formation is a prerequisite for
development of a valuable de-speckling algorithm.
2

Mathematically there are two basic models of Noise: additive noise and multiplicative
noise. Additive noise is systematic in nature and can be easily modeled and hence removed or
reduced easily, whereas multiplicative noise is image dependent, complex to model and hence
difficult to reduce. Speckle noise is multiplicative noise. The received image needs
processing before it can be used for further applications. The suppression of the speckle noise
is a necessary approach before classification, target detection and abstraction of other
quantitative information from the SAR image. Therefore, in order to facilitate further
information extraction process, it is crucial to preserve image sharpness and detail while
suppressing noise. Different noise models including additive and multiplicative types are
studied. The main challenge in digital image processing is to remove noise from the original
image. Hence, it is very essential to keep the data close to originality.
Selection of the de-noising algorithm is application dependent. Hence, it is necessary
to have knowledge about the noise present in the image so as to select the appropriate denoising algorithm. The proposed method reduces speckle noise in SAR images by retaining
the original content of these images. This method enhances the signal to noise ratio and
perceives the original features of the images. The output de-noised image is produced by
applying selective filter and the project will be implemented in hardware using FPGA kit.

1.2 MOTIVATION
SAR images play an important role in the observation and collection of the ground
information. SAR technology has resulted in marked improvements in the spatial resolution
images when observing a ground scene from aircraft or satellites. Speckle noise gives a
grainy appearance to SAR images. It reduces the image contrast which has a direct negative
effect on texture based analysis of the images. Speckle noise is one of the most critical
disturbances that alter the quality of SAR coherent images. Before using SAR images in
automatic target detection and recognition, the first step is to reduce the effect of speckle
noise. In this project, an improved method is proposed that reduces the speckle noise in SAR
images and satellite images, and retains the original content of these images.

1.3 BACKGROUND
Synthetic Aperture Radar (SAR) systems which are able to operate under all weather
conditions have a significant role in the observation and collection of the ground information.
The idea behind SAR is to synthesize the effect of large aperture by moving small aperture
radar along the flight path to obtain much finer resolution. While moving along the flight
path, the SAR system records reflected waves from the imaged surface at different instants.
Coherent processing of this reflected waves from the illuminated area of different range and
azimuth results in the formation of 2-D SAR images. However, such recorded SAR raw data
contain unwanted artifacts, which result in granular appearances in SAR image. Those
granular parts of SAR image are known as speckle, which is multiplicative in nature and
degrades the SAR image quality significantly, leading to loss of crucial information. Thus,
removal of noise without blurring the edges is an important problem in SAR imaging.
The theoretical description which involves with definitions and descriptions of
various topics used in the project are given as below:

1.3.1 SYNTHETIC APERTURE RADAR (SAR)

SAR is a form of radar which is used to create images of an object, such as a
landscape these images can be 2D or 3D representations of the object. SAR uses the motion
of the SAR antenna over a target region to provide finer spatial resolution when compared to
conventional beam-scanning radars. SAR imagery uses microwave radiation so that it can
illuminate the earth surface. It provides its own illumination. It is not affected by cloud cover
or radiation in solar illumination.
SAR technology has resulted in marked improvements in the spatial resolution images
when observing a ground scene from aircraft or satellites, and it can be used to estimate the
features like the dampness of the soil, the thickness of a forest, or the roughness of the sea.
Nevertheless, SAR images are contaminated by multiplicative noise, due to the coherence of
the radar wavelength, labelled as speckle noise, which results in an important reduction in the
efficiency of target detection and classification algorithms.
Remote sensing is the science and art of obtaining information about an object, area,
or phenomenon through the analysis of data acquired by a device that is not in contact with
the object, area, or phenomenon under investigation. Based on the wavelength in which the
system works, remote sensing is categorized into two different groups, i.e., optical and
4

microwave. Optical remote sensing uses visible and infrared waves whereas microwave
remote sensing uses radio waves.
As a microwave remote sensing, RADAR (Radio Detection And Ranging) sends out
pulses of microwave electromagnetic radiation and measures the strength as well as time
between the transmitted and reflected pulses to determine both the type of reflector and its
distance from the transmitter. Different pulse intervals, different wavelengths (which range
from less than 1 mm to 1 m), different geometry and polarizations can all be used to
determine the roughness, geometry and moisture content of the earth surface. During the past
two decades different satellites using RADAR sensors have been put into the orbit.
A SAR system illuminates a scene with microwaves and records both the amplitude
and the phase of the back-scattered radiation, making it a coherent imaging process. The
received signal is sampled and converted into a digital image as shown in Figure 1.1 (Source:
http://www.mathworks.com/matlabcentral/fileexchange/screenshots/22325/original.jpg)
In comparison to optical remote sensing, radar imaging has some advantages.

As an active system, it is a day and night data acquisition system.

Considering the behaviour of electromagnetic waves in the range of radar wavelength,
it can be seen that atmospheric characteristics such as cloud, light rain, haze, and
smoke has little effect on the capability of radar data acquisition system. This makes

radar as an all-weather remote sensing system.

As the radar signals partially penetrate into soil and vegetation canopy, in addition to
surface information, it can provide subsurface information too.
Inherent with all radar images is speckle noise which is nothing else but variation in

backscatter from inhomogeneous cells. Speckle noise gives a grainy appearance to radar
images. It reduces the image contrast which has a direct negative effect on texture based
analysis of the images. Meanwhile, speckle noise also changes the spatial statistics of the
underlying scene backscatter which in turn makes the classification of images a difficult task.
Obviously, it is seen that to interpret radar images correctly one has to reduce the effect of
speckle noise. However, as the speckle noise reduction process changes the image as well,
one should use proper filter to keep the image degradation minimum.
SAR images have wide applications in remote sensing and mapping of the surfaces of
both the Earth and other planets. SAR can also be implemented as inverse SAR by observing
5

a moving target over a substantial time with a stationary antenna. Speckle noise is one of the
most critical disturbances that alter the quality of SAR coherent images. Before using SAR
images in automatic target detection and recognition, the first step is to reduce the effect of
speckle noise.

Figure 1.1 Original SAR image

1.3.2 SPECKLE NOISE

Speckle is a granular noise that inherently exists in and degrades the quality of the
active radar, synthetic aperture radar (SAR) and medical ultrasound images. The vast
majority of surfaces, synthetic or natural, are extremely rough on the scale of the wavelength.
Images obtained from these surfaces by coherent imaging systems such as laser, SAR, and
ultrasound suffer from a common phenomenon called speckle. Speckle, in both cases, is
primarily due to the interference of the returning wave at the transducer aperture. The
scattered signals add coherently; that is, they add constructively and destructively depending
on the relative phases of each scattered waveform. Speckle noise results from these patterns
of constructive and destructive interference shown as bright and dark dots in the image.
Speckle noise in conventional radar results from random fluctuations in the return
signal from an object that is no bigger than a single image-processing element. It increases
the mean grey level of a local area. Speckle noise in SAR is generally serious, causing

difficulties for image interpretation. It is caused by coherent processing of backscattered

signals from multiple distributed targets.
Speckle noise is commonly observed in radar (microwave or millimetre wave)
sensing systems. Like the light from a laser, the waves emitted by active sensors travel in
phase and interact minimally on their way to the target area. After interaction with the target
area, these waves are no longer in phase because of the different distances they travel from
targets, or single versus multiple bounce scattering. Once out of phase, radar waves can
interact to produce light and dark pixels known as speckle noise. Speckle noise in radar data
is assumed to have multiplicative error model and must be reduced before the data can be
utilized otherwise the noise is incorporated into and degrades the image quality (See Figure
1.2). Ideally, speckle noise in radar images must be completely removed. However, in
practice it cannot be reduced completely, only significant amount of speckle noise can be
reduced. Reducing the effect of speckle noise permits both better discrimination of scene
targets and easier automatic image segmentation.
Speckle noise can be reduced by multi-look processing or spatial filtering. While
multi-looking process is usually done during data acquisition stage, speckle reduction by
spatial filtering is performed on the image after it is acquired. No matter which method is
used to reduce the effect of speckle noise, the ideal speckle reduction method preserves
radiometric information, the edges between different areas and spatial signal variability, i.e.,
textural information.

Figure 1.2 SAR image corrupted with speckle noise

1.3.3 SAVITZKYGOLAY FILTER

A SavitzkyGolay filter is a digital filter that can be applied to a set of digital data
points for the purpose of smoothing the data, that is, to increase the signal-to-noise ratio
without greatly distorting the signal. This is achieved, in a process known as convolution, by
fitting successive sub-sets of adjacent data points with a low-degree polynomial by the
method of linear least squares. When the data points are equally spaced an analytical solution
to the least-squares equations can be found, in the form of a single set of "convolution
coefficients" that can be applied to all data sub-sets, to give estimates of the smoothed signal
(or derivatives of the smoothed signal), at the central point of each sub-set.
Savitzky-Golay smoothing filters (also called digital smoothing polynomial filters or
least-squares smoothing filters) are typically used to "smooth out" a noisy signal whose
frequency span (without noise) is large. In this type of application, Savitzky-Golay smoothing
filters perform much better than standard averaging FIR filters, which tend to filter out a
significant portion of the signal's high frequency content along with the noise. Although
Savitzky-Golay filters are more effective at preserving the pertinent high frequency
components of the signal, they are less successful than standard averaging FIR filters at
rejecting noise. Savitzky-Golay filters are optimal in the sense that they minimize the leastsquares error in fitting a polynomial to frames of noisy data.

1.3.4 MEDIAN FILTER

The spatial filters are categorized into two different groups, i.e., non-adaptive and
adaptive. Non-adaptive filters take the parameters of the whole image signal into
consideration and leave out the local properties of the terrain backscatter or the nature of the
sensor. These kinds of filters are not appropriate for non-stationary scene signal. Fast Fourier
Transform (FFT) is an example of such filters. On the other hand, adaptive filters
accommodate changes in local properties of the terrain backscatter as well as the nature of the
sensor. In these types of filters, the speckle noise is considered as being stationary but the
changes in the mean backscatters due to changes in the type of target are taken into
consideration. Adaptive filters reduce speckles while preserving the edges (sharp contrast
variation). These filters modify the image based on statistics extracted from the local
environment of each pixel. Adaptive filter varies the contrast stretch for each pixel depending
upon the Digital Number (DN) values in the surrounding moving kernel. Obviously, a filter
that adapts the stretch to the region of interest (the area within the moving kernel) would
8

produce a better enhancement. Mean, median, Lee-Sigma, Local-Region, Lee, Gamma MAP,
Frost are examples of such filters.
In signal processing, it is often desirable to be able to perform some kind of noise
reduction on an image or signal. The median filter is a nonlinear digital filtering technique,
often used to remove noise. Such noise reduction is a typical pre-processing step to improve
the results of later processing (for example, edge detection on an image). Median filtering is
very widely used in digital image processing because, under certain conditions, it preserves
edges while removing noise.
All smoothing techniques are effective at removing noise in smooth patches or
smooth regions of a signal, but adversely affect edges. At the same time as reducing the noise
in a signal, it is important to preserve the edges. Edges are of critical importance to the visual
appearance of images.
Speckle filtering consists of moving a kernel over each pixel in the image and
applying a mathematical calculation using the pixel values under the kernel and replacing the
central pixel with the calculated value. By applying the filter a smoothing effect is achieved
and the visual appearance of the speckle is reduced.
In speckle noise and salt and pepper noise (impulsive noise), Median filter is
particularly effective. Because of this, median filtering is very widely used in digital image
processing.

1.3.5 WAVELET TRANSFORM

Non-decimated 2-D wavelet transform is used to decompose a signal or an image
using non-decimated wavelet analysis. This kind of redundant, translation-invariant transform
is especially useful for de-noising, which is one of the most important wavelet applications.
This type of analysis has already been developed for stationary wavelet transform (SWT)
functions (1-D stationary wavelet functions). This transform however has a serious limitation:
the signal length must be a power of 2 and the periodised extension mode must be used for
the underlying DWT.
Recently, the use of wavelet transform has led to significant advances in image
analysis. The main reason for the use of multi-scale processing is the fact that many natural
signals, when decomposed into wavelet bases are significantly simplified and can be
9

modelled by known distributions. Besides, wavelet decomposition is able to separate noise

and signal at different scales and orientations. Therefore, the original signal at any scale and
direction can be recovered and useful details are not lost. The following is the process where
the wavelet packet decomposition takes place over 3 levels as shown in the figure 1.3, where
g[n] is the low-pass approximation coefficients, h[n] is the high-pass detail coefficients.

Figure 1.3 Wavelet Packet decomposition over 3 levels

1.3.6 WAVELET TRANSFORM VS. FOURIER TRANSFORM

Fourier analysis is based on an indefinitely long cosine wave of a specific frequency
whereas, wavelet Analysis is based on a short duration wavelet of a specific centre frequency.
Wavelet transform is most appropriate for non-stationary signals. The wavelets are localised
in both time and frequency whereas the standard Fourier transform is only localised in
frequency. In addition, the Wavelet Transform provides a multi-resolution system which is
useful in several applications like image communications and image data base.

1.3.7 LITERATURE SURVEY

To remove speckle noise from SAR images, dozens of de-speckle filters have been in
order. In general, no filter consistently outperforms others. Each filter has its unique strengths
and limitations. Thus, the choice of which filter to use is dependent on the requirements of
the specific application and the characteristics of the dataset employed. De-speckle filters
with good noise removal capabilities often tend to degrade the spatial and radiometric
resolution of an original image and cause the loss of image detail. This may be acceptable for
10

applications involving large scale image interpretation or mapping. However, in many cases
where the retention of the subtle structures of the image is important, the performance of
noise suppression must be balanced with the filters effectiveness in order to preserve fine
detail.
The following literature survey suggests different sources of theoretical study and the
subjects they deal with. A conclusion will be reached at the end of this literature survey with
which the project will be implemented.
1.

Reduction of Speckle Noise and Image Enhancement of Images Using Filtering

Technique, authored by Milindkumar V. Sarode and Prashant R. Deshmukh
It proposes filtering techniques for the removal of speckle noise from the digital
images. Quantitative measures are done by using signal to noise ration and noise level
is measured by the standard deviation.
The main purpose of the noise reduction technique is to remove speckle noise by
retaining the important feature of the images. This paper offers some idea about
various noise reduction techniques. SAR imagery uses microwave radiation so that it
can illuminate the earth surface. SAR provides its own illumination. It is not affected
by cloud cover or radiation in solar illumination. It also discusses about model of
speckle noise and noise in ultrasound images as well as noise in SAR images.
This paper introduced a novel method which reduces speckle noise in ultrasound
images and SAR images, retaining the original content of these images. This method
enhances the Signal to Noise ratio and perceives the original features of the images.
This paper introduced a Speckle noise reduction model for Ultrasound Sound images
as well as Synthetic Aperture Radar (SAR) imagery. Both models preserve the
appearances of structured regions. In case of Ultrasound Images, Texture and organ
surfaces have been enhanced. The performance of the algorithm has been tested using
visual performance measures. Many of the methods are failure to remove speckle
noise present in the Ultrasound images, since the information about the variance of
the noise may not be able to identify by the methods. Introduced model automatically
collect the information about the noise variance. Performance of the Speckle noise
reduction model for Synthetic Aperture Radar (SAR) imagery is well as compared to
other filters. Histogram results shows very closed equivalency in between SAR
original images and SAR de-noised i.e. enhanced images.

2. An Improved SAR Image Speckle Reduction Algorithm of Wavelet Threshold,

authored by Chen Yi Tian and Li Bin Bing
11

It proposes an improved SAR image speckle reduction algorithm of wavelet

threshold. This method aims at overcoming the shortages of the discontinuity of the
hard threshold function in the traditional wavelet threshold noise reduction method
and the permanent bias between the estimated wavelet coefficient and composed
wavelet coefficient in soft threshold function. Also on the basis of Garrote threshold
function, it constructs a new self-adaption threshold function, which adds in 3 shape
adjustment factors to make the function own the characteristic of shape visual
adjustment. In MATLAB simulation experiment, X-band SAR river image shot by
Scan SAR sensor of Radarsat-1satellite verifies the authentic capability of the
improved algorithm to SAR image speckle reduction. Compared with the noise
reduction effect of the traditional threshold function, the new threshold function is
more effective in SAR image speckle noise reduction and has an excellent edge
preservation capability.
The wavelet transform with multi-scale frequency domain and superior edge
preservation is widely used in the noise reduction of SAR image. Through further
study, the shortage of the noise reduction of wavelet threshold is easy to find, namely
there is a discontinuity in hard threshold function and the estimated variance may be
larger; while there exists a permanent bias between the estimated wavelet coefficient
and composed wavelet coefficient in soft threshold function. The newly proposed
Garrote function and half soft threshold function can balance the soft and hard
threshold function and combine their advantage, but the structure is relative fixed and
it overly depends on threshold, lacks differential of higher order and decreases the
function and flexibility of the threshold. This thesis proposes an improved algorithm
by combining with Garrote threshold function. Compared with the traditional
threshold function, the new threshold function with excellent mathematical
characteristics has a simpler expression and is easy to calculate. At the same time, the
three adjustable factors of the new non-linear threshold function make the signal selfadaption noise reduction possible. The simulation experiment verifies the new
threshold function is more effective in SAR image speckle reduction and has an
excellent edge preservation capability.
This thesis proposes an improved wavelet threshold function based on Garrote
threshold function and add three shape adjustment factors, which makes it a flexible
threshold function and can adjust the threshold function to the ideal shape according
to the needs. Using this function can consider the two important factors during the
12

noise reduction: threshold and threshold function. The experiment result indicates that
the speckle algorithm based on wavelet threshold in this thesis is simple and effective
and can remarkably improve SAR image visual effect.
3. Effects and performance of speckle noise reduction filters on active radar and
SAR images, authored by M. Mansourpour, M.A. Rajabi and J.A.R. Blais
It focuses on reduction of speckle noise to increase the quality of radar coherent
images. Image variances or speckle is a granular noise that inherently exists in and
degrades the quality of the active radar and SAR images. Before using active radar
and SAR images, the very first step is to reduce the effect of Speckle noise. Most of
speckle reduction techniques have been studied by researchers; however, there is no
comprehensive method that takes all the constraints into consideration. Filtering is
one of the common methods which is used to reduce the speckle noises. This paper
compares six different speckle reduction filters quantitatively using both simulated
and real images. The results have been presented by filtered images, statistical tables
and diagrams. Finally, the best filter has been recommended based on the statistical
and experimental results.
It also discusses about speckle noise, its reduction and speckle filtering. By applying
the filter a smoothing effect is achieved and the visual appearance of the speckle is
reduced. It explains different filters like Mean filter, Median filter, Lee-Sigma and Lee
filters.
This paper reviewed the effect of applying six different adaptive filters on a simulated
image as well as a real SAR imagery. To test the effect and performance of the filters,
as the original and noisy images were available for the simulated image, one can use
Mean Square Error, Signal to Noise ratio and parameter (which shows the edge
preserving strength of the filters). These three measures are able to evaluate the
performance of filters quantitatively when both the original as well as noisy image are
available. A good filter shows lower Mean Square Error, higher Signal to Noise Ratio,
and a closer to one.
However, in cases when the original image is not available, one can use only two
parameters of Mean and Standard Deviation of noisy and filtered images to perform
the filter assessment. A good filter has a lower difference between Means of the
original and filtered images while preserving a low Standard Deviation for the filtered
image.
In both simulated and real images it is seen that regardless of the kernel size, Mean,
Median and Local Region filters perform poorly. This sounds a reasonable result as
13

these filters do not take all the statistical characteristics of the image into
consideration. In case of simulated imagery it is seen that the Gamma-MAP filter with
a 3x3 kernel has the lowest MSE and highest SNR and in compare to other filters
with the same kernel. However, Frost filter with a 7x7 kernel has the lowest MSE and
highest SNR. The numerical results show that Gamma-MAP filter performs much
better for preserving the edge information.
In the case of real SAR imagery, the Gamma-MAP, Frost, and Lee filters with a 5x5
kernel show better results as the differences of their Means from the Mean of original
image is low while they all have low Standard Deviation.
4. Analysis of Speckle Noise Reduction in Synthetic Aperture Radar Images,
authored by Pranali A. Hatwar and Dr. Heena R. Kher
It proposes that Synthetic aperture radar (SAR) is an active imaging system that can
achieve high resolutions both in range and azimuth. Speckle noise is one of the most
critical disturbances that alter the quality of SAR coherent images. Before using SAR
images in automatic target detection and recognition, the first step is to reduce the
effect of speckle noise. In this paper, filtering techniques have been used on the input
image and statistical parameters are calculated for the output images obtained from all
filters for performance measurement and these are implemented in MATLAB.
This paper discusses about the adaptive methods concern for eliminating speckle
noise in SAR images. It gives a brief review of SAR de-speckling methods that are
present in literature and the adaptive de-speckling filters and also algorithm for
speckle noise reduction is discussed for SAR images. Along with these, estimation of
statistical parameter for performance measurement is explained. An experimental
result for applying these filters on some real SAR images and compares it with
existing speckle removal methods.
In this paper, the performance of the algorithm has been tested using statistical
parameter measures. Many of the methods are failures to remove speckle noise
present in the Synthetic Aperture Radar (SAR) images, since the information about
the variance of the noise may not be able to identify by the methods. The Performance
of the Speckle noise reduction model for Synthetic Aperture Radar is well as
compared to other filters. Lee filters smoothest the image data, without removing
edges or sharp features in the images. Enhanced lee filter and enhanced frost filter
divide the image into areas of three classes which are homogeneous area,
heterogeneous area and isolated point. In the homogeneous area in which speckle
eliminated by averaging the multi-look processing, in the heterogeneous area in which
14

speckle are reduced while preserving the texture and preserving texture in isolated
point target. Kuan, Enhanced Kuan and Gamma MAP filter reduces speckle while
preserving edges in SAR images.
5. Wavelet Based Techniques for Speckle Noise Reduction in Ultrasound Images,
authored by Nishtha Attlas and Dr.Sheifali Gupta
It presents a review of various techniques for reduction of speckle noise in ultrasound
images. Speckle Noise is one of the most prominent noises seen in the ultrasound
images and corrupts the visual quality of the image for further processing being
multiplicative in nature. This paper demonstrates wavelet based techniques for
improving visual image quality in ultrasound images and de-noising. With the help of
variable window technique and region based processing, discrete wavelet transform
technique provided better noise rejection in ultrasound images by removing the
speckle noise.
This paper presents various techniques for removing speckle noise from images and
image enhancement by thresholding using various spatial domain filters and Wavelet
analysis on the corrupted images have been discussed. Latest domain in the field of
Image de-noising and compression is using wavelet analysis. Multi-resolution image
analysis using wavelets is the latest modification in the field of image enhancement
and de-noising. Wavelet analysis represents the next logical step: a windowing
technique with variable-sized regions. Wavelet analysis allows the use of long time
intervals where we want more precise low-frequency information, and shorter regions
where we want high-frequency information. Speckle Noise is the high frequency
content in the ultrasound images and can be easily removed using wavelet based
thresholding technique.
The algorithm follows a quantization approach that divides the input image in 4 filter
coefficients, and then performs further quantization on the lower order filter or
window of the previous step. This quantization depends upon the decomposition
levels and maximum numbers of decomposition levels to be entered are 3 for Disvrete
Wavelet Transform (DWT).
6. Speckle Noise Reduction in SAR Imaging using Lattice Filters Based Sub band
Decomposition, authored by Gkhan Karasakal and I. Erer
It presented a new speckle reduction algorithm based on lattice filters for SAR
imaging. In the new method, the sub-band decomposition of the speckled image is
performed using lattice filters. The noisy image is decomposed into sub band images
15

using high-pass and low-pass filters having lattice structure, then a threshold value is
estimated according to noise variance in each sub and soft-thresholding is applied on
the sub band images. The de-speckled image is obtained from the thresholded sub
band images using the inverse lattice filter. The proposed speckle reduction method is
applied to SAR images. The performance of the proposed method has also been
compared with median filtering, and discrete and stationary wavelet transform based
speckle reduction methods. Results show that the proposed method may be used
efficiently for speckle noise reduction in SAR images. It also discusses brief
information about the sub-band decomposition and reconstruction based on lattice
filters.
In this work, a new speckle reduction method which uses lattice filters based sub-band
decomposition is presented. The performance of the proposed method has also been
compared with median filtering, and discrete and stationary wavelet transform based
speckle reduction methods. Results show that the proposed method may be an
alternative to the widely used wavelet transform based methods for speckle noise
reduction in SAR images.
7. Wavelet-Based De-speckling of Synthetic Aperture Radar Images Using
Adaptive and Mean Filters, authored by Syed Musharaf Ali, Muhammad
Younus Javed, and Naveed Sarfraz Khattak
It introduced new wavelet based algorithm for speckle reduction of synthetic aperture
radar images, which uses combination of undecimated wavelet transformation, wiener
filter (which is an adaptive filter) and mean filter. Furthermore instead of using
existing thresholding techniques such as sure shrinkage, Bayesian shrinkage,
universal thresholding, normal thresholding, visu thresholding, soft and hard
thresholding, we use brute force thresholding, which iteratively run the whole
algorithm for each possible candidate value of threshold and saves each result in array
and finally selects the value for threshold that gives best possible results. That is why
it is slow as compared to existing thresholding techniques but gives best results under
the given algorithm for speckle reduction.
Previously most of the researchers use discrete wavelet transformation for reduction
of speckle. Drawback of discrete wavelet transformation is that it is not translation
invariant. That means it will lose lots of important coefficients during translation from
original signal to sub bands. In order to solve this problem and to save the
coefficients, derivate form of discrete wavelet transformation is used called
16

undecimated wavelet transformation. Basic idea is that it does not lose any
coefficients, all coefficients remain intact. That is why it is also called redundant
wavelet transformation. It requires more storage space and need more time for
computation. Whether discrete or undecimated wavelet is used, biggest problem is the
selection of optimal thresholding .Some researchers use wavelet based hard or soft
thresholding.Other thresholding techniques were also used such as VisuShrink,
SureShrink, OracleShrink, OracleThresh, NormalShrink, BayesShrink, Thresholding
Neural Network (TNN) etc. But there are some draw backs of using those thresholds.
Therefore a new technique for speckle reduction of SAR images was implement that
use the combination of undecimated wavelet transformation, wiener filter and mean
filter. For thresholding brute force thresholding was used which always gives best
results as compared to above mention thresholding techniques.
This paper proposed a new wavelet based method for de-speckling of SAR images
that uses combination of undecimated wavelet transformation, wiener filter and mean
filter. Moreover instead of using existing thresholding techniques, it uses brute force
thresholding that always outperforms the other existing techniques but only drawback
is that it is slow.
8. SAR images speckle noise reduction by wavelet thresholding, authored by
Punya Thitimajshima
It describes a method 0f speckle reduction in Synthetic Aperture Radar (SAR) images
based on the wavelet transform. To deduce the problem of filtering the multiplicative
noise to the case of an additive noise. The wavelet decomposition is performed On the
1ogairithm of the image gray levels. A threshold value is estimated according to the
noise variance and used for the soft-thresholding performed on all the high frequency
sub images. The filtered logarithmic image is then obtained by reconstruction from
the thresholded coefficients. The exponential function of this reconstructed image
gives the final filtered image. Experimental results on SAR images showed that the
proposed method results in a significant noise removal.
9. Subspace-Based Technique for Speckle Noise Reduction in SAR Images,
authored by Norashikin Yahya, Nidal S. Kamel and Aamir Saeed Malik
It proposed the Image-subspace-based approach for speckle noise removal from
synthetic aperture radar (SAR) images. The underlying principle is to apply
homomorphic framework in order to convert multiplicative speckle noise into additive
and then to decompose the vector space of the noisy image into signal and noise
17

subspaces. Enhancement is performed by nulling the noise subspace and estimating

the clean image from the remaining signal subspace. Linear estimator minimizing
image distortion while maintaining the residual noise energy below some given
threshold is used to estimate the clean image. Experiments are carried out using
synthetically generated data set with controlled statistics and real SAR image of
Selangor area in Malaysia. The performance of the proposed technique is compared
with Lee and homomorphic wavelet in terms of noise variance reduction and
preservation of radiometric edges. The results indicate moderate noise reduction by
the proposed filter in comparison to Lee but with a significantly less blurry effect and
a comparable performance in terms of noise reduction to wavelet but with less
artifacts. The results also show better preservation of edges, texture, and point targets
by the proposed filter than both Lee and wavelet and less required computational
time.
10. Speckle Noise Reduction in SAR Imagery Using a Local Adaptive Median
Filter, authored by Fang Qiu, Judith Berglund, John R. Jensen, Pathik
Thakkar and Dianwei Ren
It developed a local adaptive median filter that uses local statistics to detect SAR
speckle noise and to replace it with a local median value. The performance of the
local adaptive median filter was evaluated using RADARSAT and JERS-1 datasets
based on both visual assessment and a number of numerical measures. In comparison
with established filters, it was found that the local adaptive median filter
outperformed others in achieving the best balance between speckle suppression and
image detail preservation.
The performance of the proposed filter was evaluated using a number of quantitative
and qualitative criteria. These included the Speckle Suppression Index (SSI), Speckle
Image Statistical Analysis (SISA), Edge Enhancing Index (EEI), Feature Preserving
Index (FPI), and Image Detail Preserving Coefficient (IDPC), in addition to visual
assessment. Two different kinds of SAR datasets, RADARSAT Fine Beam Mode
SAR and JERS-1 SAR data were used for evaluation purposes. In comparison to
many established adaptive filters, including the Lee filter, Frost filter, Kuan filter,
Gamma (MAP) filter, the local sigma filter, and the Lee-sigma filter, it was found that
the proposed filter not only achieved comparable effectiveness in speckle noise
reduction, but also demonstrated superior capability in structure and detail
preservation. In conclusion, the proposed filter is considered to be the best filter to
18

accommodate the tradeoff between sufficient speckle suppression and effective detail
retention among those evaluated in this study.
It is necessary to have knowledge about the noise present in the image so as to select
the appropriate de-noising algorithm. From the above literature study, various filtering
techniques like frost, mean, median, Lee filters were studied [12] and the Median filter is
chosen as the best filter for the rejection of the speckle noise in SAR images. The proposed
method reduces speckle noise in SAR images by retaining the original content of these
images. This method enhances the signal to noise ratio and perceives the original features of
the images. The output de-noised image is produced by applying different smoothening
techniques like wavelet transform and different filters like SavitzkyGolay filter and Median
filter.

CHAPTER 2
PROJECT DESCRIPTION AND GOALS
Several different methods are used to eliminate speckle noise, based upon different
mathematical models of the phenomenon. One method, for example, employs multiple-look
processing, averaging out the speckle noise by taking several "looks" at a target in a single
radar sweep. The average is the incoherent average of the looks.
Second method uses adaptive and non-adaptive filters on the signal processing (where
adaptive filters adapt their weightings across the image to the speckle level, and non-adaptive
filters apply the same weightings uniformly across the entire image). Such filtering also
eliminates actual image information as well, in particular high-frequency information, and the
applicability of filtering and the choice of filter type involves trade-offs. Adaptive speckle
19

filtering is better at preserving edges and detail in high-texture areas (such as forests or urban
areas). Non-adaptive filtering is simpler to implement, and requires less computational power.
There are two forms of non-adaptive speckle filtering: one based on the mean and one
based upon the median (within a given rectangular area of pixels in the image). The latter is
better at preserving edges whilst eliminating noise spikes, than the former is. There are many
forms of adaptive speckle filtering, including the Lee filter, the Frost filter, and the Refined
Gamma Maximum-A-Posteriori (RGMAP) filter. They all rely upon three fundamental
assumptions in their mathematical models:

Speckle noise in SAR is a multiplicative noise, i.e. it is in direct proportion to the

local grey level in any area.

The signal and the noise are statistically independent of each other.

The sample mean and variance of a single pixel are equal to the mean and variance of
the local area that is centred on that pixel.

The Lee filter converts the multiplicative model into an additive one, thereby reducing the
problem of dealing with speckle noise to a known tractable case.
In this case, Median filter and wavelet transform are used since they are the best
methods to remove speckle noise and to produce de-noised image without affecting the
original image much, which is proved by the literature study.
The goal is to write the Verilog code for the removal of speckle noise in SAR images
and to implement it in the FPGA hardware kit. The signal to noise ratio is enhanced and the
original features of the images are perceived.
The report deals with following chapters:
Chapter 3: Technical specification Deals with software and hardware modules.
Chapter 4: Design approach and details Block diagram, flowchart, hardware interfacing
details, codes and standards, constraints, alternatives and tradeoffs.
Chapter 5: Schedule, Tasks and Milestones Project completion details
Chapter 6: Project demonstration Output details and figures
Chapter 7: Cost analysis
Chapter 8: Summary

20

CHAPTER 3
TECHNICAL SPECIFICATION

The project contains both software and hardware modules which contains various technical
aspects. They are explained in the following sections.

3.1 SOFTWARE MODULE

Verilog HDL is a Hardware Description Language. A Hardware Description Language
is a language used to describe a digital system, for example, a computer or component of a
computer. One may describe a digital system at several levels. For example, HDL might
describe the layout of the wires, resistors and transistors on an Integrated Circuit chip. It
might describe the logical gates and flip flops in a digital system. The higher level HDL

21

describes the registers and the transfers of vectors of information between registers. This is
called the Register Transfer Level. Verilog supports all of these levels.
The software implementation is carried out using Xilinx software with the Verilog code.
3.1.1 Comparator
A comparator is a device that compares two voltages or currents and indicates a larger digital
signal. It has two analog input terminals V+, V- and one binary digital output Vo.
The output is

A digital comparator or magnitude comparator is a hardware electronic device that takes two
numbers as input in binary form and determines whether one number is greater than, less than
or equal to the other number.
There are two main types of Digital Comparator, they are:

Identity Comparator an Identity Comparator is a digital comparator that has only

one output terminal for when A = B either HIGH A = B = 1 or LOW A = B = 0

Magnitude Comparator a Magnitude Comparator is a digital comparator which has
three output terminals, one each for equality, A = B greater than, A > B and less than
A<B

The purpose of a Digital Comparator is to compare a set of variables or unknown numbers.

The following table shows the 1-bit comparator truth table (Table 3.1).
Table 3.1 1-bit comparator truth table
Inputs

Outputs

A>B

A=B

A<B

22

A Magnitude Comparator is a type of digital comparator that has three output terminals, one
each for equality, A = B greater than, A > B and less than A < B (see Figure 3.1). The digital
comparator accomplishes this using several logic gates that operate on the principles of
Boolean algebra.

Figure 3.1 n-bit Magnitude Comparator

A comparator consists of a specialized high-gain differential amplifier. They are commonly
used in devices that measure and digitize analog signals, such as analog-to-digital converters
(ADCs), as well as relaxation oscillators (see figure 3.2).

23

Figure 3.2 8-bit comparator

Digital or Binary Comparators are made up from standard AND, NOR and NOT gates that
compare the digital signals present at their input terminals and produce an output depending
upon the condition of those inputs.

3.2 VERILOG
Verilog, standardized as IEEE 1364, is a hardware description language (HDL) used to model
electronic systems. It is most commonly used in the design and verification of digital circuits
at the register-transfer level of abstraction. It is also used in the verification of analog circuits
and mixed-signal circuits.
Hardware description languages such as Verilog differ from software programming languages
because they include ways of describing the propagation time and signal strengths
(sensitivity). There are two types of assignment operators; a blocking assignment (=), and a
non-blocking (<=) assignment. The non-blocking assignment allows designers to describe a
state-machine update without needing to declare and use temporary storage variables. Since
24

these concepts are part of Verilog's language semantics, designers could quickly write
descriptions of large circuits in a relatively compact and concise form. At the time of
Verilog's introduction (1984), Verilog represented a tremendous productivity improvement for
circuit designers who were already using graphical schematic capture software and specially
written software programs to document and simulate electronic circuits.
The designers of Verilog wanted a language with syntax similar to the C programming
language, which was already widely used in engineering software development. Like C,
Verilog is case-sensitive and has a basic pre-processor (though less sophisticated than that of
ANSI C/C++). Its control flow keywords (if/else, for, while, case, etc.) are equivalent, and its
operator precedence is compatible with C. Syntactic differences include: required bit-widths
for variable declarations, demarcation of procedural blocks (Verilog uses begin/end instead of
curly braces {}), and many other minor differences. Verilog requires that variables be given a
definite size. In C these sizes are assumed from the 'type' of the variable (for instance an
integer type may be 8 bits).
A Verilog design consists of a hierarchy of modules. Modules encapsulate design hierarchy,
and communicate with other modules through a set of declared input, output, and
bidirectional ports. Internally, a module can contain any combination of the following:
net/variable declarations (wire, reg, integer, etc.), concurrent and sequential statement blocks,
and instances of other modules (sub-hierarchies). Sequential statements are placed inside a
begin/end block and executed in sequential order within the block. However, the blocks
themselves are executed concurrently, making Verilog a dataflow language.
Verilog's concept of 'wire' consists of both signal values (4-state: "1, 0, floating, undefined")
and signal strengths (strong, weak, etc.). This system allows abstract modeling of shared
signal lines, where multiple sources drive a common net. When a wire has multiple drivers,
the wire's (readable) value is resolved by a function of the source drivers and their strengths.
A subset of statements in the Verilog language is synthesizable. Verilog modules that conform
to a synthesizable coding style, known as RTL (register-transfer level), can be physically
realized by synthesis software. Synthesis software algorithmically transforms the (abstract)
Verilog source into a netlist, a logically equivalent description consisting only of elementary
logic primitives (AND, OR, NOT, flip-flops, etc.) that are available in a specific FPGA or
VLSI technology. Further manipulations to the netlist ultimately lead to a circuit fabrication

25

blueprint (such as a photo mask set for an ASIC or a bitstream file for an FPGA). Verilog can
be used at different levels of abstraction.
3.2.1 System level
Verilog is not ideally suited for abstract system-level simulation, prior to the hardwaresoftware split. This is to some extent addressed by System Verilog. Unlike VHDL, which has
support for user-defined types and overloaded operators which allow the designer to abstract
his work into the domain of the problem, Verilog restricts the designer to working with predefined system functions and tasks for stochastic simulation and can be used for modelling
performance, throughput and queuing but only in so far as those built-in language features
allow. Designers occasionally use the stochastic level of abstraction for this phase of the
design process.
3.2.2 Digital
Verilog is suitable for use today in the digital hardware design process, from functional
simulation, manual design and logic synthesis down to gate-level simulation. Verilog tools
provide an integrated design environment in this area.
Verilog is also suited for specialized implementation-level design verification tools such as
fault simulation, switch level simulation and worst case timing simulation. Verilog can be
used to simulate gate level fan-out loading effects and routing delays through the import of
SDF files.
The RTL level of abstraction is used for functional simulation prior to synthesis. The gate
level of abstraction exists post-synthesis but this level of abstraction is not often created by
the designer, it is a level of abstraction adopted by the EDA tools (synthesis and timing
analysis, for example).
3.2.3 Analog
Because of Verilog's flexibility as a programming language, it has been stretched to handle
analog simulation in limited cases. There is a draft standard Verilog-AMS that addresses
analog and mixed signal simulation.
3.2.4 Synthesizing Verilog

26

Synthesis is a broad term often used to describe very different tools. Synthesis can include
silicon compilers and function generators used by ASIC vendors to produce regular RAM and
ROM type structures. Synthesis in the context of this tutorial refers to generating random
logic structures from Verilog descriptions. This is best suited to gate arrays and
programmable devices such FPGAs. Synthesis is not a panacea! It is vital to tackle High
Level Design using Verilog with realistic expectations of synthesis.
The definition of Verilog for simulation is cast in stone and enshrined in the Language
Reference Manual. Other tools which use Verilog, such as synthesis, will make their own
interpretation of the Verilog language. There is an IEEE standard for Verilog synthesis (IEEE
Std. 1364.1-2002) but no vendor adheres strictly to it.
It is not sufficient that the Verilog is functionally correct; it must be written in such a way that
it directs the synthesis tool to generate good hardware, and moreover, the Verilog must be
matched to the idiosyncrasies of the particular synthesis tool being used. Some of these
idiosyncrasies will be tackled in this Verilog tutorial.
There are currently three kinds of synthesis:

behavioural synthesis
high-level synthesis
RTL synthesis

There is some overlap between these three synthesis domains. The essence of RTL code is
that operations described in Verilog are tied to particular clock cycles. The synthesised netlist
exhibits the same clock-by-clock cycle behaviour, allowing the RTL test bench to be easily
re-used for gate-level simulation.

3.3 MATLAB
MATLAB (matrix laboratory) is a multi-paradigm numerical computing environment and
fourth-generation programming language which is developed by Math Works. MATLAB
allows matrix manipulations, plotting of functions and data, implementation of algorithms,
creation of user interfaces, and interfacing with programs written in other languages,
including C, C++, Java, FORTRAN and Python.
Although MATLAB is intended primarily for numerical computing, an optional toolbox uses
the MuPAD symbolic engine, allowing access to symbolic computing capabilities. An
27

additional package, Simulink, adds graphical multi-domain simulation and Model-Based

Design for dynamic and embedded systems.
MATLAB is a high-level language and interactive environment for numerical computation,
visualization, and programming. Using MATLAB, one can analyse data, develop algorithms,
and create models and applications. The language, tools, and built-in math functions enable
you to explore multiple approaches and reach a solution faster than with spread sheets or
traditional programming languages, such as C/C++ or Java. One can use MATLAB for a
range of applications, including signal processing and communications, image and video
processing, control systems, test and measurement, computational finance, and computational
biology. More engineers and scientists in industry and academia use MATLAB, the language
of technical computing.

3.4 HARDWARE MODULE

The hardware part is implemented in Altera DE2-115 FPGA board (Figure 3.3- Source:
http://www.terasic.com.tw/cgi-bin/page/archive.pl?Language=English&No=502).
Its specifications are as follows:
3.4.1 Altera DE2-115 (Education and Development board)
The DE2 board has many features that allow the user to implement a wide range of designed
circuits, from simple circuits to various multimedia projects.
The following hardware is provided on the DE2 board:

Altera Cyclone IV EP4CE115 device

Altera Serial Configuration device - EPCS16
USB Blaster (on board) for programming and user API control; both JTAG and Active

Serial (AS) programming modes are supported

8-Mbyte SDRAM
4-Mbyte Flash memory (1-Mbyte on some boards)
4 pushbutton switches
18 toggle switches
18 red user LEDs
9 green user LEDs
50-MHz oscillator and 27-MHz oscillator for clock sources
VGA DAC (10-bit high-speed triple DACs) with VGA-out connector
10/100 Ethernet Controller with a connector
28

USB Host/Slave Controller with USB type A and type B connectors

RS-232 transceiver and 9-pin connector
PS/2 mouse/keyboard connector
IR Receiver
2 SMA connectors for external clock input/output
One 40-pin Expansion Header with diode protection
One High Speed Mezzanine Card (HSMC) connector
16*2 LCD module

In addition to these hardware features, the DE2-115 board has software support for standard
I/O interfaces and a control panel facility for accessing various components. Also, the
software is provided for supporting a number of demonstrations that illustrate the advanced
capabilities of the DE2-115 board.

Figure 3.3 Altera DE2-115 (Education and Development board)

3.4.2 Block Diagram of the DE2-115 Board and its specifications
To provide maximum flexibility, all connections are made through the Cyclone IV E FPGA
device. Thus, the user can configure the FPGA to implement any system design (Figure 3.4).

29

Fig.3.4 Block Diagram of the DE2-115 Board

Following is more detailed information about the blocks:
3.4.2.1 FPGA device

Cyclone IV EP4CE115F29 device

114,480 Les
432 M9K memory blocks
3,888 Kbits embedded memory
4 PLLs

3.4.2.2 FPGA configuration

JTAG and AS mode configuration

EPCS64 serial configuration device
On-board USB Blaster circuitry

3.4.2.3 Memory devices

128MB (32Mx32bit) SDRAM

2MB (1Mx16) SRAM
8MB (4Mx16) Flash with 8-bit mode
32Kb EEPROM

30

3.4.2.4 Connectors

Two Ethernet 10/100/1000 Mbps ports

High Speed Mezzanine Card (HSMC)
Configurable I/O standards (voltage levels:3.3/2.5/1.8/1.5V)
USB type A and B
o Provide host and device controllers compliant with USB 2.0
o Support data transfer at full-speed and low-speed
o PC driver available
40-pin expansion port
o Configurable I/O standards (voltage levels:3.3/2.5/1.8/1.5V)
o VGA-out connector
o VGA DAC (high speed triple DACs)
o DB9 serial connector for RS-232 port with flow control
PS/2 mouse/keyboard

3.4.2.5 Display

16x2 LCD module

3.4.2.6 Switches and indicators

18 slide switches and 4 push-buttons switches

18 red and 9 green LEDs
Eight 7-segment displays

3.4.2.7 Power

Desktop DC input
Switching and step-down regulators LM3150MH

CHAPTER 4
DESIGN APPROACH AND DETAILS

In this chapter, details about engineering design and hardware implementation are discussed.
Also, various aspects like codes and standards, constraints, alternatives, trade-offs and
problems faced during the course of project are discussed.

4.1 HARDWARE DESIGN

31

Figure 4.1 describes about the hardware design of the project. The Verilog code is written in
Quartus II software by giving 5 numbers each of 3 bit as input. The code is programmed in
the Altera DE2-115 FPGA kit by connecting it to the PC using USB cable. The program is
stored in SDRAM.
Power
supply

Quartus II
Software
(PC)

SDRAM
USB
Cable

Red LED lights

Slide Switches
Altera DE2-115

Figure 4.1 Hardware design

After assigning pins, compiling the code and interfacing hardware kit to the computer, the kit
can be programmed to get the required output which is, arranging the given input values in
ascending order in order to find the median value. The inputs are given using slide switches
and the outputs are shown using the red LEDs. The hardware implementation cannot be done
for 256*256 values since there are only 18 inputs available in the hardware kit.

4.2 HARDWARE IMPLEMENTATION

The hardware part is implemented in Altera DE2-115 FPGA kit, using the slide switches as
inputs and the red LEDs as outputs.
4.2.1 Switches
There are also 18 slide switches on the DE2-115 board (See Figure 4.2). These switches are
not debounced, and are assumed for use as level-sensitive data inputs to a circuit. Each switch
is connected directly to a pin on the Cyclone IV E FPGA. When the switch is in the DOWN

32

position (closest to the edge of the board), it provides a low logic level to the FPGA, and
when the switch is in the UP position it provides a high logic level.
The inputs are 3 bit- 5 numbers and a reset clock, which are assigned to 16 switches.
A list of the pin names on the Cyclone IV E FPGA that are connected to the slide switches is
given in the section 4.2.3.

Figure 4.2 Connections between the slide switches and Cyclone IV E FPGA

4.2.2 Light Emitting Diodes (LEDs)

There are 27 user-controllable LEDs on the DE2-115 board. Eighteen red LEDs are situated
above the 18 Slide switches, and eight green LEDs are found above the push-button switches
(the 9th green LED is in the middle of the 7-segment displays). Each LED is driven directly
by a pin on the Cyclone IV E FPGA; driving its associated pin to a high logic level turns the
LED on, and driving the pin low turns it off. Figure 4.3 shows the connections between LEDs
and Cyclone IV E FPGA.

33

The outputs are 3 bit- 5 numbers, which are assigned to 15 LEDs in order to get the required
order. The pins used to connect to the LEDs are listed under the section 4.2.3.

Figure 4.3 Connections between the LEDs and Cyclone IV E FPGA

4.2.3 Pin assignment
The input switches which are named as SW[0] to SW[17], have the following pin numbers in
order to assign them to the required inputs using Quartus-II software pin planner tool.

SW[0]
SW[1]
SW[2]
SW[3]
SW[4]
SW[5]
SW[6]
SW[7]
SW[8]
SW[9]
SW[10]
SW[11]
SW[12]
SW[13]
SW[14]
SW[15]
SW[16]
SW[17]

PIN_AB28
PIN_AC28
PIN_AC27
PIN_AD27
PIN_AB27
PIN_AC26
PIN_AD26
PIN_AB26
PIN_AC25
PIN_AB25
PIN_AC24
PIN_AB24
PIN_AB23
PIN_AA24
PIN_AA23
PIN_AA22
PIN_Y24
PIN_Y23

The output red LEDs which are named as LEDR[0] to LEDR[17], have the following pin
numbers in order to assign them to the required outputs.
34

LEDR[0]
LEDR[1]
LEDR[2]
LEDR[3]
LEDR[4]
LEDR[5]
LEDR[6]
LEDR[7]
LEDR[8]
LEDR[9]
LEDR[10]
LEDR[11]
LEDR[12]
LEDR[13]
LEDR[14]
LEDR[15]
LEDR[16]
LEDR[17]

PIN_G19
PIN_F19
PIN_E19
PIN_F21
PIN_F18
PIN_E18
PIN_J19
PIN_H19
PIN_J17
PIN_G17
PIN_J15
PIN_H16
PIN_J16
PIN_H17
PIN_F15
PIN_G15
PIN_G16
PIN_H15

The output is checked on the Altera DE2-115 kit using LEDs as indicators.

4.3 SOFTWARE IMPLEMENTATION

The filtering process of SAR image to remove speckle noise is implemented in software
using MATLAB and Verilog codes. In MATLAB, the output de-speckled image can be
compared to the images which were produced in the intermediate processes of smoothening
and filtering using Median filter. A Verilog code is also written to remove the speckle noise
from an image of 256*256 resolution by taking its pixel values as input. It produces output
pixel values after applying Median filter.
4.3.1 Flowchart
Firstly, MATLAB code is written in order to check the selected median filter in software. The
implementation process is shown in the flowchart (Figure 4.4).

35

Figure 4.4 Design Procedure of Speckle Noise Reduction in MATLAB

The SAR image is taken as input and pre-processing is done. Then, speckle noise is added to
the input image. Different filtering and smoothening techniques like wavelet noise
thresholding algorithm and Median filter are applied. The output image obtained is a
denoised SAR image in which most of the noise is reduced.

4.4 CODES AND STANDARDS

4.4.1 IEEE 1801-2009 Unified Power Format (UPF) standard
Addressing static or leakage-based power consumption requires new techniques and
standards that fall outside the scope of traditional HDL-based flows. The IEEE Standard
1801-2009, based on Accelleras Unified Power Format (UPF), allows designers to describe
low power design intent and improve the way complex integrated circuits can be designed,
36

verified and implemented. This world-wide open standard for low power from IEEE permits
all EDA tool providers to implement advanced tool features that enable the design of lowpower ICs. Starting at the Register Transfer Level (RTL) and progressing into the detailed
levels of implementation and verification, UPF facilitates an interoperable, multi-vendor tool
flow and ensures consistency throughout the design process.
Advantages of UPF:

The UPF file is the input to several tools (e.g., simulation, synthesis, formal
verification, and place-and-route tools).

Synthesis tools can read the RTL/UPF design input files and produce a net list.

The UPF file may be reused without change later in the tool flow.

A UPF specification can be included with the other deliverables of

intellectual-property (IP) blocks and reused along with the other delivered IP
files.

The same standard can be used in a multi-vendor tool flow

4.4.2 USB 2.0 - Universal serial Bus- Standard to interface the FPGA board to the computer.
4.4.3 IEEE 1364.2001- Verilog coding standard

4.5 CONSTRAINTS, ALTERNATIVES AND TRADEOFFS

4.5.1 Realistic design constraints
Filter selection: Since most of the methods of speckle noise reduction involves in different
types of filters which is dependent on the application, it is difficult to select filters which can
give best de-speckled output from the given input image.
Implementing image processing in hardware: Since the hardware kit (Altera DE2-115
FPGA) consists of only 18 inputs, the 256*256 pixel resolution image cannot be
implemented.
4.5.2 Significant trade-off

37

In this design, there is a trade-off between order of the filter and area occupied by the design.
To increase the order of the filter for improving the smoothness of the image, the compromise
is made with the increase in area.
4.5.3 Problems faced during the course of project:

At first, the code was written in MATLAB and then tried to convert it into Verilog
code using code converter. But, it was not possible because of the command imread,
which reads the pixel values of the input image. The error shown was
buildInstrumentedMex. Later, it was removed and values of pixels of the image
were given as input directly but the same error appeared again.

So, an attempt was made to convert the same code into C++ but very few call
functions are converting and the rest are not converting because the code is not
standalone. Thus the attempts of conversion were terminated.

At last, code was written directly in Verilog format using Xilinx software and later
implemented in the FPGA board.

The proposal of the project is to display the output through Video Graphics Array
(VGA) module but configuring VGA module is not possible because frequency of
operation is not matched with the display device. So, Phased Locked Loop (PLL)
cannot be configured.

CHAPTER 5
Schedule, Tasks and Milestones
The whole project is divided into following tasks:

Understanding SAR images and speckle noise

Studying various de-noising techniques

Understanding different filters like median filter etc.

Studying wavelet transform and threshold algorithm

38

Writing Verilog code in Xilinx software

Design for hardware implementation

Implementing software code in the FPGA kit

Evaluating input and output values

There are three major milestones as well as above mentioned smaller tasks that must be
achieved in order to reach the milestones. The three milestones are:

Literature study

Writing software code

Hardware implementation and verification

Tasks will be split up among group members according to each members level of expertise
or comfort. Each task will have a leader who is responsible for completion of that task.
However, the other group members are expected to provide assistance if needed. This way, an
engineer can concentrate on a task but still get help if needed.
A Gantt chart outlining important tasks and schedule can be seen in Appendix.

CHAPTER 6
PROJECT DEMONSTRATION
In MATLAB implementation, SAR image is taken as input (Figure 6.1) and speckle noise is
added to it (Figure 6.2).

39

Figure 6.1 Input image

Figure 6.2 Speckle noise added image

Then, SavitzkyGolay filter is applied to the noise added image. The resulting output image
looks as shown in the Figure 6.3.

Figure 6.3 SavitzkyGolay filtered image

And then, Median filter is applied to it in order to remove the noise and to smooth the image.
The resulting output image looks as shown in Figure 6.4.

40

Figure 6.4 Median filtered image

After applying different smoothening techniques such as wavelet thresholding algorithm, the
output de-speckled image is produced and can be compared with the speckled image. The
comparison is shown in the Figure 6.5.

Figure 6.5 Comparison of Speckled and De-speckled output images

The code for 8-bit 8 numbers is written in Xilinx software. Eight random values are taken as
input and checked for ascending order using Verilog code. The output is shown in the Figure
6.6.

41

Figure 6.6 Output of Verilog software code of 8-bit

Then the code for 256*256 pixel values is written in Verilog and Median filter is applied to it.
The result is shown in Figure 6.7

Figure 6.7 Output of Verilog software code

The hardware part is implemented in Altera DE2-115 and the inputs are given using slide
switches and the outputs are shown using Red LED lights (Figure 6.8).
The inputs given are 101, 010, 111, 100 and 001
42

The program sorts the given inputs in ascending order.

The outputs are 001, 010, 100, 101 and 111.

Figure 6.8 Hardware output on Altera DE2-115 FPGA kit

CHAPTER 7
COST ANALYSIS
43

The process of removal of speckle noise in SAR images is a complicated one. The existing
methods are not much efficient in doing so since it involves not only in removal of noise but
also in preserving the image details. The proposed method can work efficiently by achieving
high SNR.
The cost analysis of the project is as follows:
Table 7.1 Cost analysis
1
2

Altera DE2-115 (Education and Development board)

USB cable for Altera board
Total cost

$280 or 17420
740
18160

CHAPTER 8
SUMMARY

44

SAR images are useful sources of information for roughness, geometry, and moisture
content of the Earth surface. As an active, day/night, and all-weather remote sensing system,
SAR images can provide us information from both surface and subsurface of the Earth.
Inherent with SAR images is speckle noise which gives a grainy appearance to the
images. Speckle noise reduces the image contrast and has a negative effect on texture based
analysis. Moreover, as speckle noise changes the spatial statistics of the images, it makes the
classification process a difficult task to do. All of these show that to get information out of
SAR images one should first remove/reduce the effect of speckle noise.
The presence of speckle noise as grainy salt-and-pepper patterns in a SAR image
often obscures the underlying image content, reduces the interpretability of the image, and
complicates digital image processing. To take full advantage of active SAR data, the
reduction of speckle is often desired. Substantially reducing speckle noise while effectively
preserving image detail are two important speckle suppression considerations. In many
applications, it is the balance between these two considerations that determines the success of
a speckle suppression filter. For this reason, a new local adaptive median filter for speckle
suppression was proposed. It attempted to integrate the advantages of the median filter in
structure retention.
Generally there are two techniques of removing/reducing speckle noise, i.e., multilook process and spatial filtering. Multi-look process is used at the data acquisition stage
while spatial filtering is used after the data is stored. No matter which method is used to
reduce/remove the speckle noise, they should preserve radiometric information, edge
information and last but not least, spatial resolution. These are the conditions that any speckle
noise reduction technique should meet.
Spatial filters are mainly categorized into two general groups, i.e., non-adaptive and
adaptive filters. Non-adaptive filters are those which neglect the local properties of the terrain
backscatter or nature of sensor. However, adaptive filters accommodate the change in the
local properties of the terrain backscatter or the nature of sensor.
The use of filter in Digital Image Processing improves the image to a great extent.
The model preserves the appearances of structured regions. In case of SAR Images, Texture
and land surfaces have been enhanced. The performance of the algorithm has been tested
using statistical parameter measures. Many of the methods are failures to remove speckle
noise present in the SAR images, since the information about the variance of the noise may
not be able to identify by the methods. The Performance of the selected filter for speckle
45

noise reduction is far better when compared to the other filters mentioned in the literature
study.
The performance of the proposed filter was evaluated using a number of quantitative
and qualitative criteria. These included the Signal to Noise Ratio (SNR) and Peak Signal to
Noise Ratio (PSNR).
Speckle noise reduction model for Synthetic Aperture Radar (SAR) imagery is
introduced. It preserves the appearances of structured regions. Performance of the Speckle
noise reduction model for SAR imagery is well as compared to other filters. The results show
much closed equivalency in between SAR original images and SAR de-noised i.e. enhanced
images.

REFERENCES

46

[1] Milindkumar V. Sarode, Prashant R. Deshmukh Reduction of Speckle Noise and

Image Enhancement of Images Using Filtering Technique IJoAT, Vol 2, No 1
(January 2011).
[2] Chen Yi Tian, Li Bin Bing An Improved SAR Image Speckle Reduction Algorithm
of Wavelet Threshold 2012 IEEE, 978-1-4673-0875-5/12.
[3] T.Vigneswaran, J.Selva Kumar, Dr.P.Subbarami Reddy Design of digital FIR filter
using dynamic distributed arithmetic algorithm with improved table look up scheme
for residue number system, IET-UK International Conference on Information and
Communication Technology in Electrical Sciences (ICTES 2007) pp. 606-610
[4] M. Mansourpour, M.A. Rajabi, J.A.R. Blais Effects and performance of speckle
noise reduction filters on active radar and SAR images
[5] Pranali A. Hatwar, Dr. Heena R. Kher Analysis of Speckle Noise Reduction in
Synthetic Aperture Radar Images, International Journal of Engineering Research &
Technology (IJERT) ISSN: 2278-0181, Vol. 4 Issue 01, January-2015.
[6] Nishtha Attlas, Dr.Sheifali Gupta Wavelet Based Techniques for Speckle Noise
Reduction in Ultrasound Images, Int. Journal of Engineering Research and
Applications ISSN: 2248-9622, Vol. 4, Issue 2 (Version 1), February 2014, pp.508513.
[7] Gkhan Karasakal, I. Erer, Speckle Noise Reduction in SAR Imaging using Lattice
Filters Based Subband Decomposition, 1-4244-1212-9/07/$25.00 2007 IEEE.
[8] Syed Musharaf Ali, Muhammad Younus Javed, and Naveed Sarfraz Khattak,
Wavelet-Based De-speckling of Synthetic Aperture Radar Images Using Adaptive
and Mean Filters, International Scholarly and Scientific Research & Innovation 1(7)
2007.
[9] Punya Thitimajshima, SAR images speckle noise reduction by wavelet
thresholding, 1999 IEEE, 0-7803-5761-2/99.

[10] Norashikin Yahya, Nidal S. Kamel, Aamir Saeed Malik, Subspace-Based Technique
for Speckle Noise Reduction in SAR Images, IEEE transactions on geoscience and
remote sensing, Vol. 52, No. 10, October 2014.

47

[11] Fang Qiu, Judith Berglund, John R. Jensen, Pathik Thakkar and Dianwei Ren,
Speckle Noise Reduction in SAR Imagery Using a Local Adaptive Median Filter,
GIScience and Remote Sensing, 2004, 41, No. 3, pp. 244-266.
[12] M. Mansourpour, M.A. Rajabi, J.A.R. Blais, Effects and performance of speckle
noise reduction filters on active radar and SAR images, Dept. of Geomatics Eng.,
University of Calgary, Calgary, Alberta, T2N 1N4, Canada.
[13] Wei Zhang, Jinzhong Yang, Le Yu, Comparisons of speckle noise filtering methods
on high resolution SAR image, 2010 IEEE, CCTAE, 978-1-4244-6947-5/10.

48

APPENDIX
Gantt chart

49