Comparison of Median and Wiener Filter in Image De-noising

lemonadeviraginityAI and Robotics

Nov 6, 2013 (3 years and 9 months ago)

102 views

Comparison of Median and Wiener Filter in

Image
De
-
noising


M
r.Swaroopkrishna Nair
1
,
M
r.Mallikarjun Shivsharan
2

B
.E (E&TC)

SVERI’s
College of Engineering,

Pandharpur.,,Dist:Solapur.

413304

1
swaroopnair360@gmail.com,9860301795

2
shivsharan.arjun@gmail.
com,9970159924




Abstract



Noise is an inseparable part of any processing
technique. Image processing is no exception to
this. Noise may be added in the image during
capturing or during its transmission. Besides,
noise gets added at every point of the p
rocessing
procedure. It leads to the deterioration of the
quality of the image. Hence there is a necessity to
identify and remove the effects of these noises.
This paper aims at performance comparison of
various filters used for removal of different types
of noises.








INTRODUCTION



Image Processing
is an ever rising area
presently.

The

applications of image processing are wide
spread.
These

applications include the restoration,
enhancement and
s
torage

of the images on a large
scale. For

these purposes a large number of image
processing operations may be
needed. As

we know,
all processing operations are always affected by some
form of noise or
other. Hence

there is a need for
removal of
noise, pre
cisely

called Image De
-
Noising.







This process of Image De
-
Noising is performed using
different types of filters. This paper aims at a
comparative study of two filters viz. Median filter and
Wiener Filter
.

Classification

of Noise

The most commonly found noises are
:



Gaussian N
oise



Salt and pepper Noise



Speckle Noise



Periodic Noise



1. GAUSSIAN NOISE



It is also called as ‘Am
plifier Noise’ or ‘Normal
Noise
. This type of noise gets introduced into the
image due to the factors such as electronic circuit
noise and sensor noise due to poor illumination and/or
high temperature.


2.

SALT AND PEPPER NOISE



An image containing salt
-
and
-
pepper noise will
have dark pixels in bright regions and bright pixe
ls in
dark regions [4]. This type of noise can be caused by
dead pixels, analog
-
to
-
digital converter errors, bit
errors in transmission, etc. This can be eliminated in
large part by using dark frame subtraction and by
interpolating around dark/bright pixel
s.



3. SPECKLE NOISE



Speckle noise is a granular noise that inherently
exists in and degrades the quality of the active radar
and synthetic aperture radar (SAR) images. Speckle
noise in conventional radar results from random
fluctuations in the r
eturn 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 is caused by signals from elementary scatterers,
the gravity
-
capillary ripples, and manifests as a
pedesta
l image, beneath the image of the sea waves.

Classification of Filters:

The commonly used filters to reduce these noises are :



Spatial
domain
filters



Frequency domain filters

Most of the noises in nature could be removed by
Spatial filtering, i.e. by proce
ssing directly on the pixels
itself.

The frequency domain filtering is used for the
removal of some of the complex noises like periodic
noise.In this paper, we dicuss the performance
comparrison between two

very important spatial filters:
Median filter and

Wiener filter.

1.MEDIAN FILTER


This kind of filters replace the value of a pixel by
the median of the gray levels in the neighbo
u
rhood of
that pixel(the original value of the pixel is included in
the computation of the median).These are quite
popular
, because for certain types of random
noise(impulse noise)
, they provide excellent noise
reduction with less blurring. It forces the noise with
distinct gray levels to be more like their neighbours.
Larger clusters are affected considerably less.


2.W
IENER FILTER


It is a kind of adaptive filer. Wiener filtering
approach incorporates both the degradation function
and statistical characteristics of noise into restoration.
In this method, an estimate
F* of the uncorrupted
image F is obtained such tha
t the mean square error
between them is minimized.

In other words, the wiener
filter adapts itself to the corrupting function and the
kind of noise present in the image.



SIMULATION RESULTS

The Original Image is Flowertitlee image, adding three
types of

Noise (Gaussian noise, Speckle noise and
Salt & Pepper noise) and De
-
noised image using
Median filter and Wiener filter and comparisons among
them.



Fig.1.Original Image






















Fig .2.
Noisy image


(
Salt and Pepper
)




Fig .3.Noisy image

















































































































































Fig.4.Noisy image(Speckle Noise)


























































































































Fig.5.De
-
noising by Median filter



(for salt and pepper)




Fig.6.De
-
noising by Median filter



(for Gaussian noise)






















Fig.5.De
-
noising by Median filter











































































Fig.7.De
-
noising by Median filter





(for speckle noise)
































































































































































Fig.8.De
-
noising by wiener filter



(for salt and pepper)





















Fig.9.De
-
noising by wiener filter



(for Gaussian noise)


























































































































































Fig.10.De
-
noising by wiener f
ilter


(for speckle noise)



































CONCLUSION



We have used the image ‘a.jpg’ (fig.1), added three
types of noises (salt and pepper, Gaussian, speckle)
and de
-
noised them using the two filters in
consideration v
iz. Median and Wiener filters.From the
results, we conclude that:

(a)The performance of the Wiener Filter after de
-
noising for Speckle and Gaussian noisy image is
better than Median filter.

(b)The performance of the Median filter after de
-
noising for Salt
& Pepper noisy image is better than
Wiener filter.






REFERENCES


[1]International Journal of Computer Applications
(0975


8887) Volume 12


No.4, November 2010.

[2]

Wavelet domain image de
-
noising by thresholding
and Wiener filtering. Kazubek, M.
Signal Processing
Letters IEEE, Volume: 10, Issue: 11, Nov. 2003 265
Vol.3.

[3]Rafael. C. Gonzalez, Dgital Image Processing,
Edition III.

[4]

IEEE
Transactions on Image Processing, VOL. 17,
No. 1, January

2008.