L3 - Image Noise and Image Restoration

Department of Mechatronics
Image Acquisition and Pre-processing
MACHINE VISION AND IMAGE PROCESSING

(MTE 4075)
Year/ Semester : IVth / VII th
Name of Faculty:
Dr. Umesh Sahu
(Assistant Professor)
Manipal Institute of Technology, Manipal, India
−1−
Syllabus
Image acquisition and pre-processing: [15]
Vision system components, Image acquisition and analysis, Image
formation - Sampling & Quantization, simple operations on image,
Image enhancement: Spatial Domain and frequency domain
methods, Image noise, image restoration , morphological
operations, Segmentation, image analysis, representation of regions,
Feature extraction, fundamentals of color image processing,
conversion of color spaces.

−2−
Content:
▪Image Restoration
–What is image restoration?
–Noise and images
–Noise models
–Noise removal using spatial domain filtering
–Periodic noise
–Noise removal using frequency domain filtering
R. C. Gonzalez, R. E. Woods. Digital Image Processing_3rd Edition (Chapter-5:

Page 333)
−3−
“It makes all the difference whether
one sees darkness through the light
or brightness through the shadows.”
David Lindsay
5
of
31
What is Image Restoration?
Image restoration attempts to restore
images that have been degraded
– Identify the degradation process and attempt
to reverse it
– Similar to image enhancement, but more
objective
6
of
31
Noise and Images
The sources of noise in digital
images arise during image
acquisition (digitization) and
transmission
– Imaging sensors can be
affected by ambient
conditions
– Interference can be added
to an image during
transmission
7
of
31
Noise Model
We can consider a noisy image to be
modelled as follows:
g ( x, y) = f ( x, y) +  ( x, y)
where f(x, y) is the original image pixel,
η(x, y) is the noise term and g(x, y) is the
resulting noisy pixel
If we can estimate the model the noise in an
image is based on this will help us to figure out
how to restore the image.
8
of
A Model of Image Degradation and
31
Restoration Process
9
of
31
Noise Models
There are many different
Images taken from Gonzalez & Woods, Digital Image Processing (2002)
Gaussian Rayleigh
models for the image
noise term η(x, y):
– Gaussian
• Most common model Erlang Exponential
– Rayleigh
– Erlang (gamma)
– Exponential Uniform
– Uniform Impulse
– Impulse
• Salt and pepper noise
10
of
31
Noise Example
The test pattern to the right is
ideal for demonstrating the

addition of noise
The following slides will show Image
the result of adding noise

based on various models to
this image Histogram to go here
Histogram
of
11
31
Gaussian
Rayleigh
Erlang
Example 5.1
Noise Example (cont…)
12 Try out !!
of
31
Noise Example (cont…) Write a function to add uniform,
gaussian, salt and pepper,
Rayleigh, exponential and
erlang noise and analyze its
histogram
Hint:
Refer page 112/344 (pp 148 of
Rafael C. Gonzalez - Digital
Image Processing Using MATLAB
(R)-Prentice Hall (2003)
Exponential Uniform Impulse

13
of
31
Application of Various Noise model
▪Gaussian noise: electronic noise and sensor
noise due to poor illumination and/or high
temperature
▪Rayleigh noise: range imaging
▪Exponential and gamma noise: laser imaging
▪Impulse noise: faulty switching in image
digitization process
▪Uniform density: Least descriptive of practical
situations
Content:
–Noise and images
–Noise models
–Periodic noise

Page 333)
− 14 −
15
of
31
Filtering to Remove Noise
We can use spatial filters of different kinds
to remove different kinds of noise
The arithmetic mean filter is a very simple
one and is calculated as follows:
ˆf ( x, y ) = 1
 g ( s, t )
mn ( s ,t )S xy
1/ 1/ 1/ This is implemented as the
9 9 9
1/ 1/ 1/
simple smoothing filter
9 9 9
Blurs the image to remove
1/ 1/ 1/
9 9 9 noise
16
of
31
Other Means
There are different kinds of mean filters all of
which exhibit slightly different behaviour:
– Geometric Mean
– Harmonic Mean
– Contraharmonic Mean
17
of
31
Other Means (cont…)
There are other variants on the mean which
can give different performance
Geometric Mean:
1
  mn
fˆ ( x, y ) =   g ( s, t )
( s ,t )S xy 
• Achieves similar smoothing to the
arithmetic mean, but tends to lose less
image detail
18
of
31
Harmonic Mean:
mn
fˆ ( x, y ) =
1

( s ,t )S xy g ( s, t )
• Works well for salt noise, but fails for
pepper noise
• Also does well for other kinds of noise
such as Gaussian noise
19
of
31
Contraharmonic Mean:
 g ( s, t )
( s ,t )S xy
Q +1
fˆ ( x, y ) =
 g ( s ,
( s ,t )S xy
t ) Q
Q is the order of the filter and adjusting its

value changes the filter’s behaviour
• Positive values of Q eliminate pepper noise
• Negative values of Q eliminate salt noise
20
of
31
Noise Removal Examples
Image
Original Corrupted
Image By Gaussian
Noise
After A 3*3 After A 3*3

Arithmetic Geometric
Mean Filter Mean Filter
Example 5.2
21
of
31
Noise Removal Examples (cont…)
Image
Corrupted
By Pepper
Noise
Result of
Filtering Above
With 3*3
Contraharmonic
Q=1.5
22
of
31
Image
Corrupted
By Salt
Noise
Result of
Filtering Above
With 3*3
Contraharmonic
Q=-1.5
23
of
31
Contraharmonic Filter: Here Be Dragons
Choosing the wrong value for Q when using

the contraharmonic filter can have drastic

results
24
of
31
Order Statistics Filters
Spatial filters that are based on ordering
the pixel values that make up the

neighbourhood operated on by the filter
Useful spatial filters include
– Median filter
– Max and min filter
– Midpoint filter
– Alpha trimmed mean filter
25
of
31
Median Filter
Median Filter:
fˆ ( x, y ) = median{g ( s, t )}
( s ,t )S xy
• Excellent at noise removal, without the

smoothing effects that can occur with
other smoothing filters
• Particularly good when salt and pepper
noise is present
26
of
31
Max and Min Filter
Max Filter:
fˆ ( x, y ) = max {g ( s, t )}
( s ,t )S xy
Min Filter:
fˆ ( x, y ) = min {g ( s, t )}
( s ,t )S xy
• Max filter is good for pepper noise and min

is good for salt noise
27
of
31
Midpoint Filter
Midpoint Filter:
ˆf ( x, y ) = 1  max {g ( s, t )} + min {g ( s, t )}
2 ( s ,t )S xy ( s ,t )S xy 
• Good for random Gaussian and uniform

noise
28
of
31
Alpha-Trimmed Mean Filter
Alpha-Trimmed Mean Filter:
ˆf ( x, y ) = 1
 g r ( s, t )
mn − d ( s ,t )S xy
We can delete the d/2 lowest and d/2 highest
grey levels
So gr(s, t) represents the remaining mn – d
pixels
29
of
31
Noise Removal Examples
Image Result of 1
Corrupted Pass With A
By Salt And 3*3 Median
Pepper Noise Filter
Result of 2 Result of 3
Passes With Passes With
A 3*3 Median A 3*3 Median
Filter Filter
Example 5.3
30
of
31
Image Image
Corrupted Corrupted
By Pepper By Salt
Noise Noise
Result Of Result Of
Filtering Filtering
Above Above
With A 3*3 With A 3*3
Max Filter Min Filter
Example 5.3
31
of
31
Image Image Further
Corrupted Corrupted
By Uniform By Salt and
Noise Pepper Noise
Filtered By Filtered By
5*5 Arithmetic 5*5 Geometric
Mean Filter Mean Filter
Filtered By Filtered By
5*5 Median 5*5 Alpha-Trimmed
Filter Mean Filter
Example 5.3
32
of
31
Adaptive Filters
• The filters discussed so far are applied to
an entire image without any regard for how
image characteristics vary from one point
to another
• The behaviour of adaptive filters changes
depending on the characteristics of the
image inside the filter region
• We will take a look at the adaptive
median filter
33
of
31
Adaptive Median Filtering
• The median filter performs relatively well on
impulse noise as long as the spatial density
of the impulse noise is not large
• The adaptive median filter can handle much
more spatially dense impulse noise, and
also performs some smoothing for non-
impulse noise.
• The key insight in the adaptive median filter
is that the filter size changes depending on
the characteristics of the image
34
of
31
Adaptive Median Filtering (cont…)
Remember that filtering looks at each
original pixel image in turn and generates a
new filtered pixel
First examine the following notation:
– zmin = minimum grey level in Sxy
– zmax = maximum grey level in Sxy
– zmed = median of grey levels in Sxy
– zxy = grey level at coordinates (x, y)
– Smax =maximum allowed size of Sxy
35
of
31
Level A: A1 = zmed – zmin
A2 = zmed – zmax
If A1 > 0 and A2 < 0, Go to level B
Else increase the window size
If window size ≤ repeat Smax level A
Else output zmed
Level B: B1 = zxy – zmin
B2 = zxy – zmax
If B1 > 0 and B2 < 0, output zxy
Else output zmed
36
of
31
The key to understanding the algorithm is to
remember that the adaptive median filter
has three purposes:
– Remove impulse noise
– Provide smoothing of other noise
– Reduce distortion
37
of
31
Adaptive Filtering Example
Image corrupted by salt Result of filtering with a 7 Result of adaptive

and pepper noise with * 7 median filter median filtering with i = 7
probabilities Pa = Pb=0.25
Example 5.5
Content:
–Noise and images
–Noise models
–Periodic noise

Page 333)
− 38 −
39
of
31
Periodic Noise
• Typically arises due to
electrical or electromagnetic
interference
• Gives rise to regular noise
patterns in an image
• Frequency domain
techniques in the Fourier
domain are most effective at
removing periodic noise
40
of
31
Band Reject Filters
Removing periodic noise form an image involves
removing a particular range of frequencies from that
image
Band reject filters can be used for this purpose
An ideal band reject filter is given as follows:
 W
1 if D(u , v)  D0 − 2
 W W
H (u, v) = 0 if D0 −  D(u, v)  D0 +
 2 2
1 if D(u , v)  D0 + W
 2
41
of
31
Band Reject Filters (cont…)
The ideal band reject filter is shown below,
along with Butterworth and Gaussian

versions of the filter
Ideal Band Butterworth Gaussian

Reject Filter Band Reject Band Reject
Filter (of order 1) Filter
42
of
31
Band Reject Filter Example
Image corrupted by Fourier spectrum of
sinusoidal noise corrupted image
Butterworth band Filtered image

reject filter
43
of
Periodic Noise Reduction by Frequency
31 Domain Filtering
2) Bandpass Filters
Hbp(u,v)=1-Hbr(u,v)
It helps to isolate the noise pattern.

44
of
31 Domain Filtering
3) Notch Filters
0 if D1 (u , v )  D 0 or D 2 (u, v )  D 0
3-1) Ideal : H(u, v ) = 
1 otherwise
1
 2
  
22
D1 (u, v ) =  u − − u 0  +  v − − v 0  
M N

 2   2  
1
 2
  
22
D1 (u, v ) =  u − + u 0  +  v − + v 0  
M N

 2   2  
The center of the frequency rectangle has been shifted to the point
(M/2,N/2)
45
of
31 Domain Filtering
H(u , v ) =
1
3-2) Butterworth  
n
D02
of order n : 1+  
 D1 (u, v ) D 2 (u , v ) 
1  D1 (u , v ) D 2 (u , v ) 
−  
3-3) Gaussian : H(u, v ) = 1 − e
2  2
D0 
46
of
31 Domain Filtering
47
of
31 Domain Filtering
Example 5.8
48
of
31 Domain Filtering
4) Optimum Notch Filters
f (x , y ) = g (x , y ) − w (x , y ) (x , y )
(x , y ) = F H(u, v )G (u, v )
−1
g (x , y )(x , y ) − g (x , y ) (x , y )
w (x , y ) =
 (x , y ) −  (x , y )
2 2
To obtain w(x,y) the goal is to minimize the variance in the

neighborhood of x,y in the image.
49
of
31 Domain Filtering
50
of
31 Domain Filtering
Example 5.9
51
of
31 Domain Filtering
52
of
31 Domain Filtering
Image Noise, Image Restoration : MATLAB
Imnoise(): Add noise to image
J = imnoise(I,'gaussian')
J = imnoise(I,'gaussian',m)
J = imnoise(I,'gaussian',m,var_gauss)
J = imnoise(I,'localvar',var_local)
MATLAB
J = imnoise(I,'localvar',intensity_map,var_local)
DEMO
J = imnoise(I,'poisson')
J = imnoise(I,'salt & pepper')
J = imnoise(I,'salt & pepper',d)
J = imnoise(I,'speckle')
J = imnoise(I,'speckle',var_speckle)
https://in.mathworks.com/help/images/ref/imnoise.html#:~:text=example-
,J%20%3D%20imnoise(%20I%20%2C'salt%20%26%20pepper'%2C,mean%200%20and%20variance%200.05.

− 78 −
Practice Problem
Gonzalez, Rafael C., and Richard E. Woods. "Digital Image Processing,
New Jersey.", 2007. 3rd edition.
Example 5.1 to 5.20

− 79 −
Summary
▪We have looked at:

❑Image Restoration
•Introduction
•Various noise models
•Gaussian, Rayleigh, Erlang, Exponential, uniform,
impulse
•Image restoration using spatial domain filtering
•Mean, Order statistics, Adaptive
•Periodic Noise
•Image restoration using frequency domain filtering End of
Lecture-3

− 80 −
Thank you
for your attention

− 81 −

L3 - Image Noise and Image Restoration

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

L3 - Image Noise and Image Restoration

Uploaded by

Copyright:

Available Formats

Department of Mechatronics

Image Acquisition and Pre-processing

MACHINE VISION AND IMAGE PROCESSING

Year/ Semester : IVth / VII th

Manipal Institute of Technology, Manipal, India

R. C. Gonzalez, R. E. Woods. Digital Image Processing_3rd Edition (Chapter-5:

ideal for demonstrating the

The following slides will show Image

the result of adding noise

Exponential Uniform Impulse

R. C. Gonzalez, R. E. Woods. Digital Image Processing_3rd Edition (Chapter-5:

Q is the order of the filter and adjusting its

After A 3*3 After A 3*3

Choosing the wrong value for Q when using

the contraharmonic filter can have drastic

the pixel values that make up the

• Excellent at noise removal, without the

• Max filter is good for pepper noise and min

• Good for random Gaussian and uniform

Image corrupted by salt Result of filtering with a 7 Result of adaptive

R. C. Gonzalez, R. E. Woods. Digital Image Processing_3rd Edition (Chapter-5:

along with Butterworth and Gaussian

Ideal Band Butterworth Gaussian

sinusoidal noise corrupted image

Butterworth band Filtered image

It helps to isolate the noise pattern.

4) Optimum Notch Filters

To obtain w(x,y) the goal is to minimize the variance in the

Imnoise(): Add noise to image

Manipal Institute of Technology, Manipal, India

Example 5.1 to 5.20

Manipal Institute of Technology, Manipal, India

▪We have looked at:

Manipal Institute of Technology, Manipal, India

Manipal Institute of Technology, Manipal, India

You might also like

After A 33 After A 33