Chapter 09

Digital Image Processing
Chapter 9:
Morphological Image Processing
What are Morphological Operations?
Morphological operations come from the word “morphing”

in Biology which means “changing a shape”.
Morphing
Image morphological operations are used to manipulate

object shapes such as thinning, thickening, and filling.
Binary morphological operations are derived from

set operations.
Basic Set Operations
Concept of a set in binary image morphology:
Each set may represent one object. Each pixel (x,y) has
its status: belong to a set or not belong to a set.
(Images from Rafael C. Gonzalez and Richard E.

Wood, Digital Image Processing, 2nd Edition.
Translation and Reflection Operations
Translation Reflection
Az  c c  a  z, for a  A Bˆ  w w  b, for b  B
B
A
z = (z1,z2)
(A)z B̂
Logical Operations*
*For binary images only

Dilation Operations
 
A  B  z Bˆ z  A   
 = Empty set
Dilate means “extend”
A = Object to be dilated
B = Structuring element

Dilation Operations (cont.)
Reflection B̂
Structuring
Element (B)
Original image (A) Intersect pixel Center pixel

Dilation Operations (cont.)
Result of Dilation Boundary of the “center pixels”


where B̂ z intersects A
Example: Application of Dilation
“Repair” broken characters

Erosion Operation
A B  z B z  A
Erosion means “trim”
A = Object to be eroded
B = Structuring element

Erosion Operations (cont.)
Structuring
Element (B)
Original image (A) Intersect pixel Center pixel

Erosion Operations (cont.)
Result of Erosion Boundary of the “center pixels”

where B is inside A
Example: Application of Dilation and Erosion
Remove small objects such as noise

Duality Between Dilation and Erosion
( A B )  A  Bˆ
c c
where c = complement
Proof:
( A B )  z B z  A
c c
 z B z  A   
c c
 z B z  Ac   
 A  Bˆ
c
Opening Operation
A  B  ( A B)  B
or
A  B  B z B z  A
= Combination of all parts of A that can completely contain B
Opening eliminates narrow and small details and corners.

Example of Opening
(Images from Rafael C.

Gonzalez and Richard E.
Wood, Digital Image
Processing, 2nd Edition.
Closing Operation
A  B  (A  B) B
Closing fills narrow gaps and notches

Example of Closing

Wood, Digital Image
Duality Between Opening and Closing
A  B c

 Ac  Bˆ 
Properties Opening
1. A  B  A
2. If C  D then C  B  D  B
3. A  B   B  A  B
Properties Closing
1. A  A  B
2. If C  D then C  B  D  B
3. A  B   B  A  B
Idem potent property: can’t change any more
Example: Application of Morphological Operations
Finger print
enhancement

Wood, Digital Image
Hit-or-Miss Transformation
A * X  A 
X   Ac (W  X ) 
where X = shape to be detected
W = window that can contain X

Hit-or-Miss Transformation (cont.)
A * B  A 
B   Ac (W  X ) 

Wood, Digital Image
Boundary Extraction
β(A)  A  A B
Original Boundary
image
Wood, Digital Image
Region Filling
X k  X k 1  B   A
c
where X0 = seed pixel p
Original Results of region filling

image

Extraction of Connected Components
X k  X k 1  B   A where X0 = seed pixel p

Example: Extraction of Connected Components
X-ray image
of bones
Thresholded
image
Connected
components
Wood, Digital Image
Convex Hull
Convex hull has no concave part.
Convex hull
4
Algorithm: C ( A)   D D i  X conv
i
i
where
i 1
X ki  X k 1 * B i  A, i  1,2,3,4

Example: Convex Hull

Wood, Digital Image
Thinning
A  B  A  ( A * B)
 A  ( A * B)c
A  B ((...(( A  B1 )  B 2 )...)  B n )

Example: Thinning
Make an object
thinner.
Wood, Digital Image
Thickening
A . B  A  ( A * B)
A . B ((...(( A . B ) . B )...) . B )
1 2 n
Make an object thicker

Skeletons
Dot lines are

skeletons of this
structure

Skeletons (cont.)
K
S ( A)   Sk ( A)
k 0
with Sk ( A)  ( A kB)  ( A kB)  B
where ( A kB)  (...( A B ) B ) ...) B

k times
and K  maxk A kB    
Skeletons

Wood, Digital Image
Pruning
X 1  A  B = thinning
8
X 2   ( X 1 * B k ) = finding end points
k 1
X 3  ( X 2  H )  A = dilation at end points
X 4  X 1  X 3 = Pruned result

Example: Pruning
Original
image
After
Thinning End
3 times points
Dilation
of end Pruned
points result
Summary of Binary Morphological Operations
(Tables from Rafael C.

Wood, Digital Image
Summary of Binary Morphological Operations (cont.)

Wood, Digital Image

Wood, Digital Image

Wood, Digital Image
Basic Types of Structuring Elements
x = don’t care

Gray-Scale Dilation
1-D Case f  b  maxf ( s  x )  b( x ) | ( s  x )  D f and x  Db 
2-D Case
f  b  maxf ( s  x, t  y )  b( x, y ) | ( s  x ), (t  y )  D f ; ( x, y )  Db 

Gray-Scale Dilation (cont.)
Original image Reflection
Subimage of B
+
Max
Moving
window
Structuring
element B
Note: B can be any shape and

subimage must have
the same shape as
reflection of B. Output image
Gray-Scale Erosion
1-D Case f  b  minf ( s  x )  b( x ) | ( s  x )  D f and x  Db 
2-D Case
f  b  minf ( s  x, t  y )  b( x, y ) | ( s  x ), (t  y )  D f ; ( x, y )  Db 

Gray-Scale Erosion (cont.)
Original image
Subimage B
-
Min
Moving
window
Structuring
element B
Note: B can be any shape and

subimage must have
the same shape as B.
Output image
Example: Gray-Scale Dilation and Erosion
Original image After dilation
Darker Brighter
After erosion (Images from Rafael C. Gonzalez and Richard E.

Gray-Scale Opening
f  b  ( f  b)  b
Opening cuts peaks

Gray-Scale Closing
f  b  ( f  b)  b
Closing fills valleys

Example: Gray-Scale Opening and Closing
Original image After opening After closing
Reduce dark
Reduce white
objects
objects
Gray-scale Morphological Smoothing
Smoothing: Perform opening followed by closing
Original image After smoothing

Morphological Gradient
g  ( f  b)  ( f  b)
Original image Morphological Gradient

Top-Hat Transformation
h  f  ( f  b)
Original image Results of top-hat transform

Example: Texture Segmentation Application
Small
blob
Original image Segmented result

Large blob
Algorithm:
1. Perform closing on the image by using successively larger
structuring elements until small blobs are vanished.
2. Perform opening to join large blobs together
3. Perform intensity thresholding
Example: Granulometry
Objective: to count the number of particles of each size
Method:
1. Perform opening using structuring elements of increasing size
2. Compute the difference between the original image and the result
after each opening operation
3. The differenced image obtained in Step 2 are normalized and used
to construct the size-distribution graph.
Size distribution
Original image graph (Images from Rafael C. Gonzalez and Richard E.
Morphological Watershads



Gradient Image
Original P Surface of P
image
P at edges look
like mountain ridges.



Convex Hull


Chapter 09

Uploaded by

Document Information

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

Chapter 09

Uploaded by

Copyright:

Available Formats

Digital Image Processing

Morphological operations come from the word “morphing”

Image morphological operations are used to manipulate

Binary morphological operations are derived from

(Images from Rafael C. Gonzalez and Richard E.

*For binary images only

(Images from Rafael C. Gonzalez and Richard E.

Original image (A) Intersect pixel Center pixel

Result of Dilation Boundary of the “center pixels”

“Repair” broken characters

Erosion means “trim”

(Images from Rafael C. Gonzalez and Richard E.

Original image (A) Intersect pixel Center pixel

Result of Erosion Boundary of the “center pixels”

Remove small objects such as noise

(Images from Rafael C. Gonzalez and Richard E.

Opening eliminates narrow and small details and corners.

(Images from Rafael C.

Closing fills narrow gaps and notches

(Images from Rafael C. Gonzalez and Richard E.

(Images from Rafael C.

(Images from Rafael C.

(Images from Rafael C. Gonzalez and Richard E.

(Images from Rafael C.

where X0 = seed pixel p

Original Results of region filling

(Images from Rafael C. Gonzalez and Richard E.

X k  X k 1  B   A where X0 = seed pixel p

(Images from Rafael C. Gonzalez and Richard E.

(Images from Rafael C. Gonzalez and Richard E.

(Images from Rafael C.

A  B ((...(( A  B1 )  B 2 )...)  B n )

(Images from Rafael C. Gonzalez and Richard E.

Make an object thicker

Dot lines are

(Images from Rafael C. Gonzalez and Richard E.

with Sk ( A)  ( A kB)  ( A kB)  B

where ( A kB)  (...( A B ) B ) ...) B

(Images from Rafael C.

X 3  ( X 2  H )  A = dilation at end points

(Images from Rafael C. Gonzalez and Richard E.

(Tables from Rafael C.

(Tables from Rafael C.

(Tables from Rafael C.

(Tables from Rafael C.

(Images from Rafael C. Gonzalez and Richard E.

(Images from Rafael C. Gonzalez and Richard E.

Note: B can be any shape and

(Images from Rafael C. Gonzalez and Richard E.

Note: B can be any shape and

After erosion (Images from Rafael C. Gonzalez and Richard E.

Opening cuts peaks

(Images from Rafael C. Gonzalez and Richard E.

Closing fills valleys

Original image After opening After closing

Original image After smoothing

(Images from Rafael C. Gonzalez and Richard E.

Original image Morphological Gradient

(Images from Rafael C. Gonzalez and Richard E.

Original image Results of top-hat transform

(Images from Rafael C. Gonzalez and Richard E.

Original image Segmented result

(Images from Rafael C. Gonzalez and Richard E.