7 Morphological

Morphological Image Processing
n Binary dilation and erosion

n Set-theoretic interpretation
n Opening, closing, morphological edge detectors
n Hit-miss filter
n Morphological filters for gray-level images
n Cascading dilations and erosions
n Rank filters, median filters, majority filters
Digital Image Processing: Bernd Girod, © 2013-2018 Stanford University -- Morphological Image Processing 1
Binary image processing
n Binary images are common
l Intermediate abstraction in a gray-scale/color image analysis system
• Thresholding/segmentation
• Presence/absence of some image property
l Text and line graphics, document image processing
n Representation of individual pixels as 0 or 1, convention:
l foreground, object = 1 (white)
l background = 0 (black)
n Processing by logical functions is fast and simple
n Shift-invariant logical operations on binary images:
“morphological” image processing
n Morphological image processing has been generalized to
gray-level images via level sets
Shift-invariance
n Assume that digital images f [x,y] and g[x,y] have infinite support
[ x, y ] ∈{,−2,−1,0,1,2,} × {,−2,−1,0,1,2,}
. . . then, for all integers a and b
f [x, y ] Shift- g [x, y ]

invariant
system
f [x - a, y - b ] Shift- g [x - a, y - b ]
invariant
system
n Shift-invariance does not imply linearity (or vice versa).
Structuring element
n Neighborhood “window” operator
180 degree rotation “structuring element”
{ } {
W f ⎡⎣ x, y ⎤⎦ = f ⎡⎣ x − x′, y − y′ ⎤⎦ : ⎡⎣ x′, y′ ⎤⎦ ∈Π xy }
The “hat” notation indicates
upright structuring element
(i.e., not rotated!)
n Example structuring elements P xy :
5x5 square cross

x x
y y
Binary dilation (expanding foreground)
{ }
g ⎡⎣ x, y ⎤⎦ = OR ⎡⎣W f ⎡⎣ x, y ⎤⎦ ⎤⎦ := dilate ( f ,W )
y
f [x, y] P xy g[x, y]
Binary dilation with square structuring element
{ }
g ⎡⎣ x, y ⎤⎦ = OR ⎡⎣W f ⎡⎣ x, y ⎤⎦ ⎤⎦ := dilate ( f ,W )
n Expands the size of

1-valued objects
n Smoothes object
boundaries
n Closes holes and gaps
dilation with dilation with

Original (701x781)
3x3 structuring element 7x7 structuring element
Binary erosion (shrinking foreground)
y
f [x, y] P xy g[x, y]
Caveat: There is another definition of erosion in the literature, which flips the structuring element, as for dilation.
The Lagunita online videos use that alternative definition. Matlab function imerode uses the definition on this slide.
To the best of our knowledge, there is no such discrepancy defining dilation.
Binary erosion with square structuring element
n Shrinks the size of

1-valued objects
n Smoothes object
boundaries
n Removes peninsulas,
fingers, and small
objects
erosion with erosion with
Original (701x781)
3x3 structuring element 7x7 structuring element
Relationship between dilation and erosion
n Duality: erosion is dilation of the background
n But: erosion is not the inverse of dilation
(
f ⎡⎣ x, y ⎤⎦ ≠ erode dilate ( f ,W ) ,W )
≠ dilate ( erode ( f ,W ) ,W )
Example: blob separation/detection by erosion
Original binary image Erosion by 30x30 Erosion by 70x70 Erosion by 96x96

Circles (792x892) structuring element structuring element structuring element
Example: blob separation/detection by erosion
Original binary image Erosion by disk-shaped Erosion by disk-shaped Erosion by disk-shaped

Circles (792x892) structuring element structuring element structuring element
Diameter=15 Diameter=35 Diameter=48
Example: chain link fence hole detection
Original grayscale image Fence thresholded Erosion with 151x151

Fence (1023 x 1173) using Otsu’s method “cross” structuring element
Set-theoretic interpretation
Continuous (x,y).
n Set of object pixels
{( x, y ) : f ( x, y ) = 1}
Works for discrete [x,y]
F≡ in the same way.
n Background: complement of foreground set

Fc ≡ {( x, y ) : f ( x, y ) = 0}
n Dilation is Minkowski set addition
G = F ⊕ Π xy Commutative and associative!
= {( x + p , y + p ) : ( x, y ) ∈F,( p , p ) ∈Π }
x y x y xy
=  F+ p , p
( )
( )
px , p y ∈Π xy
x y
translation of F by vector px , p y ( )
Set-theoretic interpretation: dilation
F
x
y P xy
G = F ⊕ Π xy
= {( x + p , y + p ) : ( x, y ) ∈ F,( p , p ) ∈Π }
x y x y xy
=  F+
( p ,p )
( p , p )∈Π
x y xy
x y
Set-theoretic interpretation: erosion
F x
y P xy
G
Not commutative!
Not associative!
Minkowski set subtraction
Opening and closing
n Goal: smoothing without size change
Open filter
( )
n
open ( f ,W ) = dilate erode ( f ,W ) ,W

close ( f ,W ) = erode ( dilate ( f ,W ) ,W )
n Close filter
n Open filter and close filter are biased

l Open filter removes small 1-regions
l Close filter removes small 0-regions
l Bias is often desired for enhancement or detection!
n Unbiased size-preserving smoothers
(
close − open ( f ,W ) = close open ( f ,W ) ,W )
open − close ( f ,W ) = open ( close ( f ,W ) ,W )
n close-open and open-close are duals, but not inverses of each other.
Small hole removal by closing
Original binary mask Dilation Closing 10x10 Difference to original mask

10x10
Morphological edge detectors
f éë x, y ùû (
dilate f ,W ¹ f ) (
erode f ,W ¹ f ) ( ) (
dilate f ,W ¹ erode f ,W )
Recognition by erosion
Binary image f
1400
2000
Structuring
44
element W
34
Structuring
44
element W
34
Binary image f
1400
2000
Structuring
62
element W
18
Structuring
62
element W
18
Hit-miss filter
Binary image f
1400
2000
Structuring Structuring
62
62
element V element W
18 18
Hit-miss filter
Structuring
62
element V
18
Structuring
62
element W
18
Morphological filters for gray-level images
n Threshold sets of a gray-level image f [x,y] (aka level sets)

( ) {
Τθ f ⎡⎣ x, y ⎤⎦ = ⎡⎣ x, y ⎤⎦ : f ⎡⎣ x, y ⎤⎦ ≥ θ , } − ∞ < θ < +∞
n Reconstruction of original image from threshold sets
{ (
f ⎡⎣ x, y ⎤⎦ = sup θ : ⎡⎣ x, y ⎤⎦ ∈Τθ f ⎡⎣ x, y ⎤⎦ )}
n Idea of morphological operators for multi-level
(or continuous-amplitude) signals
l Decompose into threshold sets
l Apply binary morphological operator to each threshold set
l Reconstruct via supremum operation
lGray-level operators thus obtained: flat operators
è Flat morphological operators and thresholding are commutative
Dilation/erosion for gray-level images
n Explicit decomposition into threshold sets not required in practice
n Flat dilation operator: local maximum over window W
{ {
g ⎡⎣ x, y ⎤⎦ = max W f ⎡⎣ x, y ⎤⎦ }} := dilate( f ,W )
n Flat erosion operator: local minimum over window W
n Binary dilation/erosion operators contained as special case
1-d illustration of erosion and dilation
Structuring element length =
——Dilation
——Erosion
——Original
Structuring element:
horizontal line
Amplitude
Sample no.
Image example
Original 394 x 305 Dilation 10x10 square Erosion 10x10 square
Flat dilation with different structuring elements
Original Diamond Disk 20 degree line
9 points 2 horizontal lines
Example: counting coins
thresholded 1 connected component
Original 20 connected
components
dilation thresholded after dilation
Example: chain link fence hole detection
Original grayscale image Flat erosion with 151x151 Binarized by Thresholding

Fence (1023 x 1173) “cross” structuring element
Morphological edge detector
original f dilation g g− f |g-f

g − f| thresholded
Beyond flat morphological operators
n General dilation operator
{ } {
g ⎡⎣ x, y ⎤⎦ = sup f ⎡⎣ x − α , y − β ⎤⎦ + w ⎡⎣α , β ⎤⎦ = sup w ⎡⎣ x − α , y − β ⎤⎦ + f ⎡⎣α , β ⎤⎦
α ,β α ,β
}
n Like linear convolution, with sup replacing summation, addition replacing multiplication
n Dilation with “unit impulse”
⎧0 α =β =0
d ⎡⎣α , β ⎤⎦ = ⎨
⎩−∞ else
does not change input signal:
{
f ⎡⎣ x, y ⎤⎦ = sup f ⎡⎣ x − α , y − β ⎤⎦ + d ⎡⎣α , β ⎤⎦
α ,β
}
Flat dilation as a special case
n Find w éëa , b ùû such that
{ }
f ⎡⎣ x, y ⎤⎦ = sup f ⎡⎣ x − α , y − β ⎤⎦ + w ⎡⎣α , β ⎤⎦ = dilate ( f ,W )
α ,β
n Answer:
⎧⎪0 ⎡⎣α , β ⎤⎦ ∈Π xy
w ⎡⎣α , β ⎤⎦ = ⎨
⎪⎩−∞ else
n Hence, write in general
{
g !" x, y #$ = sup f !" x − α , y − β #$ + w !"α , β #$
α ,β
}
= dilate ( f , w) = dilate ( w, f )
General erosion for gray-level images
n General erosion operator
n Dual of dilation
n Flat erosion contained as a special case
Cascaded dilations
f dilate ( f , w1 ) dilate ( f , w1 , w2 )
dilate ⎡⎣ dilate ( f , w1 ) , w2 ⎤⎦ = dilate ( f , w)
where w = dilate ( w1 , w2 )
Cascaded erosions
n Cascaded erosions can be lumped into single erosion
n New structuring element (SE) is not the erosion of one SE

by the other, but dilation.
Fast dilation and erosion
n Idea: build larger dilation and erosion operators by cascading simple, small
operators
n Example: binary erosion by 11x11 window
Erosion with … Erosion with
1x3 window 1x3 window
Erosion with
5 stages
11x11 window
Erosion with … Erosion with
3x1 window 3x1 window
5 stages
120 AND per pixel 2x10 = 20 AND per pixel
Rank filters
n Generalisation of flat dilation/erosion: in lieu of min or max value in window,
use the p-th ranked value
n Increases robustness against noise
n Best-known example: median filter for noise reduction
n Concept useful for both gray-level and binary images
n All rank filters are commutative with thresholding
Median filter
n Gray-level median filter
{ }
g ⎡⎣ x, y ⎤⎦ = median ⎡⎣W f ⎡⎣ x, y ⎤⎦ ⎤⎦ := median ( f ,W )
n Binary images: majority filter
{ }
g ⎡⎣ x, y ⎤⎦ = MAJ ⎡⎣W f ⎡⎣ x, y ⎤⎦ ⎤⎦ := majority ( f ,W )
n Self-duality
median ( f ,W ) = − ⎡⎣ median ( − f ,W ) ⎤⎦
(
majority ( f ,W ) = NOT ⎡⎣ majority NOT ⎡⎣ f ⎤⎦ ,W ⎤⎦ )
Majority filter: example
Binary image with

3x3 majority filter 20% ‘Salt&Pepper’ noise 3x3 majority filter
5% ‘Salt&Pepper’ noise
Median filter: example
Original image 5% ‘Salt&Pepper’ noise 3x3 median filtering 7x7 median filtering
Example: non-uniform lighting compensation
Original image Dilation (local max) Rank filter

1632x1216 pixels 61x61 structuring element 10st brightest pixel
61x61 structuring element
Example: non-uniform lighting compensation
Background – original image After global thresholding

7 Morphological

Uploaded by

Copyright:

Available Formats

7 Morphological

Uploaded by

Document Information

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

7 Morphological

Uploaded by

Copyright:

Available Formats

Morphological Image Processing

n Binary dilation and erosion

f [x, y ] Shift- g [x, y ]

n Shift-invariance does not imply linearity (or vice versa).

n Example structuring elements P xy :

5x5 square cross

n Expands the size of

dilation with dilation with

n Shrinks the size of

n Duality: erosion is dilation of the background

n But: erosion is not the inverse of dilation

Original binary image Erosion by 30x30 Erosion by 70x70 Erosion by 96x96

Original binary image Erosion by disk-shaped Erosion by disk-shaped Erosion by disk-shaped

Original grayscale image Fence thresholded Erosion with 151x151

n Background: complement of foreground set

open ( f ,W ) = dilate erode ( f ,W ) ,W

n Open filter and close filter are biased

Original binary mask Dilation Closing 10x10 Difference to original mask

n Threshold sets of a gray-level image f [x,y] (aka level sets)

n Binary dilation/erosion operators contained as special case

Original 394 x 305 Dilation 10x10 square Erosion 10x10 square

Original Diamond Disk 20 degree line

9 points 2 horizontal lines

thresholded 1 connected component

dilation thresholded after dilation

Original grayscale image Flat erosion with 151x151 Binarized by Thresholding

original f dilation g g− f |g-f

n Hence, write in general

n Flat erosion contained as a special case

dilate ⎡⎣ dilate ( f , w1 ) , w2 ⎤⎦ = dilate ( f , w)

n New structuring element (SE) is not the erosion of one SE

120 AND per pixel 2x10 = 20 AND per pixel

Binary image with

Original image Dilation (local max) Rank filter

Background – original image After global thresholding

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.