Tuan 5 Morphological Operation Principle

Image Processing
Robotic & Vision Lab RoVis

Counting Coins
• Counting coins is difficult because they touch each

other!
• Solution: Binarization and Erosion separates them!

19-Sep-19
Structuring Element (Kernel)
• Structuring Elements can have varying sizes
• Usually, element values are 0,1 and none(!)
• Structural Elements have an origin
• For thinning, other values are possible
• Empty spots in the Structuring Elements are don’t
care’s!
Box
Disc
19-Sep-19 Examples of stucturing elements

3

Erosion
Erosion is used for shrinking of element A by using

element B
With A and B as sets in Z 2 , the erosion of A by B, denoted A B,
defined
A B   z | ( B) Z  A
The set of all points z such that B, translated by z , is contained by A.
A B   z | ( B) Z  Ac  
Consider each foreground pixel in the input image

If the structuring element fits in, write a “1” at the origin of the
structuring element!

9/19/2019 4
Erosion

Example for Erosion
Input image 1 0 0 0 1 1 1 0 1 1
Structuring Element 1 1 1
Output Image 0

19-Sep-19 7
Example for Erosion
Input image 1 0 0 0 1 1 1 0 1 1
Output Image 0 0

19-Sep-19 8
Example for Erosion
Input image 1 0 0 0 1 1 1 0 1 1
Output Image 0 0 0

19-Sep-19 9
Example for Erosion
Input image 1 0 0 0 1 1 1 0 1 1
Output Image 0 0 0 0

19-Sep-19 10
Example for Erosion
Input image 1 0 0 0 1 1 1 0 1 1
Output Image 0 0 0 0 1

19-Sep-19 11
Example for Erosion
Input image 1 0 0 0 1 1 1 0 1 1
Output Image 0 0 0 0 1 0

19-Sep-19 12
Example for Erosion
Input image 1 0 0 0 1 1 1 0 1 1
Output Image 0 0 0 0 1 0 0

19-Sep-19 13
Example for Erosion
Input image 1 0 0 0 1 1 1 0 1 1
Output Image 0 0 0 0 1 0 0 0

19-Sep-19 14
Exercise
• Erode

9/19/2019 16
Example access matrix
for (int y = 0; y < output.rows; y++)
{
for (int x = 0; x < output.cols; x++)
{
Mat ROI = Mat output(Rect(Point(x, y), Size(3, 3))); //tao window
for (int i = 0; i < ROI.rows; i++)
{
for (int j = 0; j < ROI.cols; j++)
{
kiểm tra: ROI.at<uchar>(i, j)
}
}
}
}

Remember from previous Lecture:
GRAY LEVEL THRESHOLDING
Objects
Set threshold
here
9/19/2019 Introduction to Machine Vision
BINARY IMAGE
Problem here
How do we fill “missing pixels”?

9/19/2019 Introduction to Machine Vision
Dilation
• Dilation is used for expanding an element A by using structuring
element B
• Dilation of A by B and is defined by the following equation:
• This equation is based 0n obtaining the reflection 0f B

about its origin and shifting this reflection by z.
• The dilation of A by B is the set of all displacements z,
such that and A overlap by at least one element. Based
On this interpretation the equation of (9.2-1) can be
rewritten as:
Consider each pixel in the input image

If the structuring element touches the foreground image, write a “1” at the
origin of the structuring element!

void cv::erode (cv::InputArray src, // Input image
cv::OutputArray dst, // Result image
cv::InputArray element, // Structuring, a cv::Mat()
cv::Point anchor = cv::Point(-1,-1), // Location of anchor point
int iterations = 1, // Number of times to apply
int borderType = cv::BORDER_CONSTANT // Border extrapolation
const cv::Scalar& borderValue = cv::morphologyDefaultBorderValue()
);
Hàm getStructuringElement():
cv::getStructuringElement(int shape, Size ksize, Point anchor = Point(-1,-1));
Với:
 shape: Là hình dạng của phần tử cấu trúc. Ví dụ: MORPH_RECT. Bạn có thể tham
khảo các loại hình dạng của phần tử cấu trúc trong cv::MorphShapes.
 ksize: Là kích thước của ma tận phần tử cấu trúc.
 anchor: Là điểm neo của phần tử cấu trúc. Giá trị mặc định là (-1, -1).

Example for Dilation
Input image 1 0 0 0 1 1 1 0 1 1
Output Image 1

19-Sep-19 23
Input image 1 0 0 0 1 1 1 0 1 1
Output Image 1 0

19-Sep-19 24
Input image 1 0 0 0 1 1 1 0 1 1
Output Image 1 0 1

19-Sep-19 25
Input image 1 0 0 0 1 1 1 0 1 1
Output Image 1 0 1 1

19-Sep-19 26
Input image 1 0 0 0 1 1 1 0 1 1
Output Image 1 0 1 1 1

19-Sep-19 27
Input image 1 0 0 0 1 1 1 0 1 1
Output Image 1 0 1 1 1 1

19-Sep-19 28
Input image 1 0 0 0 1 1 1 0 1 1
Output Image 1 0 1 1 1 1 1

19-Sep-19 29
Input image 1 0 0 0 1 1 1 0 1 1
Output Image 1 0 1 1 1 1 1 1

19-Sep-19 30
Dilation

Dilation : Bridging gaps
33
Exercise

9/19/2019 34
Mat image, dst;
image = imread("demcoin2.jpg",1);
// Create a structuring element
int erosion_size = 6;
Mat element = getStructuringElement(cv::MORPH_ERODE,
cv::Size(2 * erosion_size + 1, 2 * erosion_size + 1),
cv::Point(erosion_size, erosion_size));
// Apply erosion or dilation on the image
erode(image, dst, element);
namedWindow("Display window", CV_WINDOW_AUTOSIZE);
imshow("Display window", image);
namedWindow("Result window", CV_WINDOW_AUTOSIZE);
imshow("Result window", dst);
waitKey(0);
return 0;

Useful
• erosion
• removal of structures of certain shape and size, given by
SE
• Dilation
• filling of holes of certain shape and size, given by SE
36
Exercise
• A) erode by hand
• B) erode by programing do not use erode function
• C) using erode function to compare
• D) dilate by hand
• E) dilate by programing do not use dilate function
• F) using dilate function to compare

Opening
Erosion followed by dilation, denoted ∘
A  B  ( A  B )  B
• First – erode A by B, and then dilate the result by B
First – erode A by B, and then dilate the result by B
• eliminates protrusions
• breaks necks
• smoothes contour
In other words, opening is the unification of all B objects
Entirely Contained in A
38
Opening
• Structuring element: 3x3 square

19-Sep-19 41
Opening Example
• Opening with a 11 pixel diameter disc

19-Sep-19 42
Opening Example
• 3x9 and 9x3 Structuring Element
3*9
9*3

19-Sep-19 43
Opening on Gray Value Images
• 5x5 square structuring element

19-Sep-19 44
Use Opening for Separating Blobs
• Use large structuring element that fits into the big

blobs
• Structuring Element: 11 pixel disc

19-Sep-19 45
Exercise opening

Closing
dilation followed by erosion, denoted •
A  B  ( A  B)  B
First – dilate A by B, and then erode the result by B
• smooth contour
• fuse narrow breaks and long thin gulfs
• eliminate small holes
• fill gaps in the contour
In other words, closing is the group of points, which the
intersection of object B around them with object A – is not
empty
47
Closing
• Structuring element: 3x3 square

19-Sep-19 50
Closing Example
• Closing operation with a 22 pixel disc

• Closes small holes in the foreground

19-Sep-19 51
Closing Example 1
1. Threshold
2. Closing with disc of size 20
19-Sep-19
Thresholded closed
52
A) open by hand
B) open by programing do not use morphologyex function
C) using morphologyex function to compare
D) close by hand
E) close by programing do not use morphologyex function
F) using close function to compare

Using opening to remove the line

Closing Gray Value Images
• 5x5 square structuring element

19-Sep-19 55
Hàm cv::morphologyEx().
cv::morphologyEx( InputArray src, OutputArray dst, int op, InputArray kernel, Point anchor = Point(-1,-1
), int iterations = 1, int borderType = BORDER_CONSTANT, const Scalar& borderValue = morphologyDe
faultBorderValue());
Với
 src: Là ảnh ban đầu:
 dst: Là ảnh sau khi xử lý hình thái học.
 op: Là loại hình thái học, ví dụ: MORPH_ERODE. Bạn có thể tham khảo các loại phép toán hình thái
học trong cv::MorphTypes.
 kernel: Là phần tử cấu trúc. Bạn hoàn toàn có thể khởi tạo và sử dụng bằng cách sử dụng hàm getStruc
turingElement().
 anchor: Là điểm neo của phần tử cấu trúc kernel. Giá trị mặc định là (-1, -1).
 iterations: Số lần lặp đi lặp lại các phép toán hình thái học lên ảnh. Ví dụ: Bạn càng để phép toán giãn
nở ảnh (Dialation) thì ảnh càng giãn nở nhiều.
 borderType: BORDER_CONSTANT
 borderValue: Là giới hạn biên của những điểm ảnh (Pixel) nằm ngoài kích thước của ảnh trong quá t
rình tính toán.

Ví dụ 4.13: Sử dụng opencv để thực hiện các phép toán tử sau:
Opening Closing Tophat Gradient
Blackhat
createTrackbar("Operator:\n 0: Opening - 1: Closing \n 2: Gradient - 3: Top Hat \n 4: Black Hat",
window_name, &morph_operator, max_operator, Morphology_Operations);
/// Create Trackbar to select kernel type
createTrackbar("Element:\n 0: Rect - 1: Cross - 2: Ellipse", window_name, &morph_elem, max_elem,
Morphology_Operations);
/// Create Trackbar to choose kernel size
createTrackbar("Kernel size:\n 2n +1", window_name, &morph_size, max_kernel_size,
Morphology_Operations);
/// Default start
Morphology_Operations(0, 0);
waitKey(0);
return 0;
}
/**
* @function Morphology_Operations
*/
void Morphology_Operations(int, void*)
{
// Since MORPH_X : 2,3,4,5 and 6
int operation = morph_operator + 2;
Mat element = getStructuringElement(morph_elem, Size(2 * morph_size + 1, 2 * morph_size +

1), Point(morph_size, morph_size));
/// Apply the specified morphology operation
morphologyEx(src, dst, operation, element);
imshow(window_name, dst);
}
Useful: open & close
60
Some Basic Morphological Algorithms (1)
• Boundary Extraction
The boundary of a set A, can be obtained by first eroding A
by B and then performing the set difference between A and
its erosion.
 ( A)  A   A  B 

9/19/2019 62
Example 1

9/19/2019 63
Boubdary extracion

Example 2

9/19/2019 65
int erosion_size = 1;
Mat element = getStructuringElement(cv::MORPH_ERODE,
cv::Size(2 * erosion_size + 1, 2 * erosion_size + 1),
cv::Point(erosion_size, erosion_size));
// Apply erosion or dilation on the image
erode(image, dst, element);
cv::subtract(image, dst, dst);

• Hole Filling
A hole may be defined as a background region surrounded
by a connected border of foreground pixels.
Let A denote a set whose elements are 8-connected

boundaries, each boundary enclosing a background region
(i.e., a hole). Given a point in each hole, the objective is to
fill all the holes with 1s.

9/19/2019 67
• Hole Filling
1. Forming an array X0 of 0s (the same size as the array
containing A), except the locations in X0 corresponding to
the given point in each hole, which we set to 1.
c
2. Xk = (Xk-1 + B) A k=1,2,3,…
Stop the iteration if Xk = Xk-1

9/19/2019 68
Example

9/19/2019 69
Hold filling

9/19/2019 70
9/19/2019 71
Ví dụ 4.21: Tô đầy vùng các ký tự có lỗ trống
cv::Mat mask;
image_thresh.copyTo(mask);
for (int i = 0; i < mask.cols; i++) {
if (mask.at<char>(0, i) == 0) {
cv::floodFill(mask, cv::Point(i, 0), 255, 0, 10, 10);
}
if (mask.at<char>(mask.rows - 1, i) == 0) {
cv::floodFill(mask, cv::Point(i, mask.rows - 1), 255, 0, 10, 10);
}
}
for (int i = 0; i < mask.rows; i++) {
if (mask.at<char>(i, 0) == 0) {
cv::floodFill(mask, cv::Point(0, i), 255, 0, 10, 10);
}
if (mask.at<char>(i, mask.cols - 1) == 0) {
cv::floodFill(mask, cv::Point(mask.cols - 1, i), 255, 0, 10, 10);
}
}

HIT and MISS

9/19/2019 73
Hit-and-miss Transform
• Used to look for particular patterns of foreground

and background pixels
• Very simple object recognition
• All other morphological operations can be derived
from it!!
• Input:
• Binary Image
• Structuring Element, containing 0s and 1s!!

74
Hit-and-miss Transform
• Example for a Hit-and-miss Structuring Element
• Contains 0s, 1s and don’t care’s.
• Usually a “1” at the origin!

19-Sep-19 76
The Hit-or-Miss
Transformation
if B denotes the set composed of

D and its background,the match
(or set of matches) of B in A,
denoted A  B,
A * B   A  D    Ac  W  D  
B   B1 , B2 
B1 : object
B2 : background
A  B   A  B1   ( Ac  B2 )

9/19/2019 79
The Hit-or-Miss Transformation

Exercise hit and fit
in this case, we are looking for a pattern in which the central pixel belongs to the background
while the north, south, east, and west pixels belong to the foreground. The rest of pixels in the
neighbourhood can be of any kind, we don't care about them. Now, let's apply this kernel to an
input image:

Hàm cv::morphologyEx().
cv::morphologyEx( InputArray src, OutputArray dst, int op, InputArray kernel, Point anchor = Point(-1,-1
), int iterations = 1, int borderType = BORDER_CONSTANT, const Scalar& borderValue = morphologyDe
faultBorderValue());
Với
 src: Là ảnh ban đầu:
 dst: Là ảnh sau khi xử lý hình thái học.
 op: Là loại hình thái học, ví dụ: MORPH_ERODE. Bạn có thể tham khảo các loại phép toán hình thái
học trong cv::MorphTypes.
 kernel: Là phần tử cấu trúc. Bạn hoàn toàn có thể khởi tạo và sử dụng bằng cách sử dụng hàm getStruc
turingElement().
 anchor: Là điểm neo của phần tử cấu trúc kernel. Giá trị mặc định là (-1, -1).
 iterations: Số lần lặp đi lặp lại các phép toán hình thái học lên ảnh. Ví dụ: Bạn càng để phép toán giãn
nở ảnh (Dialation) thì ảnh càng giãn nở nhiều.
 borderType: BORDER_CONSTANT
 borderValue: Là giới hạn biên của những điểm ảnh (Pixel) nằm ngoài kích thước của ảnh trong quá t
rình tính toán.

#include <opencv2/core.hpp>
#include <opencv2/imgproc.hpp>
#include <opencv2/highgui.hpp>
using namespace cv;
int main(){
Mat input_image = (Mat_<uchar>(8, 8) <<
0, 0, 0, 0, 0, 0, 0, 0,
0, 255, 255, 255, 0, 0, 0, 255,
0, 255, 255, 255, 0, 0, 0, 0,
0, 255, 255, 255, 0, 255, 0, 0,
0, 0, 255, 0, 0, 0, 0, 0,
0, 0, 255, 0, 0, 255, 255, 0,
0, 255, 0, 255, 0, 0, 255, 0,
0, 255, 255, 255, 0, 0, 0, 0);
Mat kernel = (Mat_<int>(3, 3) <<
0, 1, 0,
1, -1, 1,
0, 1, 0);
Mat output_image;
morphologyEx(input_image, output_image, MORPH_HITMISS, kernel);
const int rate = 50;
kernel = (kernel + 1) * 127;
kernel.convertTo(kernel, CV_8U);
resize(kernel, kernel, Size(), rate, rate, INTER_NEAREST);
imshow("kernel", kernel);
moveWindow("kernel", 0, 0);
resize(input_image, input_image, Size(), rate, rate, INTER_NEAREST);
imshow("Original", input_image);
moveWindow("Original", 0, 200);
resize(output_image, output_image, Size(), rate, rate, INTER_NEAREST);
imshow("Hit or Miss", output_image);
moveWindow("Hit or Miss", 500, 200);
waitKey(0);

return 0;
}
Kernel and output result for Kernel and output result for finding left
finding top-right corners end points

Basic
THINNING
9/19/2019 86
Thinning
1. Used to remove selected foreground pixels from

binary images
2. After edge detection, lines are often thicker than

one pixel.
3. Thinning can be used to thin those line to one

pixel width.

19-Sep-19 87
Definition of Thinning
• Let B be a kernel and A be an image
with 0-1=0!!
• If foreground and background fit the structuring
element exactly, then the pixel at the origin of the
SE is set to 0
• Note that the value of the SE at the origin is 1 or

don’t care!
19-Sep-19 88
• A more useful expression for thinning A symmetrically is

based on a sequence of structuring elements:
  
B  B1
, B 2
, B 3
,..., B n

where B i is a rotated version of B i -1
The thinning of A by a sequence of structuring element {B}

A  {B}  ((...(( A  B1 )  B 2 )...)  B n )

9/19/2019 89
9/19/2019 90
Ví dụ 4.18: Sử dụng kernel sau để làm mỏng hình
Kernel:

// seRight90
Mat seRight90 = getStructuringElement(MORPH_RECT, Size(3, 3));
data = { { { 0, 1, 0 },{ 1, 1, 0 },{ 0, 0, 0 } } };
seRight90 = Mat(3, 3, CV_8UC1, &data);
cout << "seRight90 = " << endl << " " << seRight90 << endl << endl;
Mat dum1 = dst_bin.clone();

morphologyEx(dst_bin, dum1, MORPH_HITMISS, seLeft0);
Mat dum3;
morphologyEx(dum1, dum3, MORPH_HITMISS, seRight0);
Mat dum5;
morphologyEx(dum3, dum5, MORPH_HITMISS, seLeft90);
Mat dum7;
morphologyEx(dum5, dum7, MORPH_HITMISS, seRight90);
cv::imshow("ketqua", dum7 * 255);

cout << endl;
cv::waitKey();
return 0;

Basic
THICKENING

9/19/2019 93
49/19/2019
• Thickening:
The thickening is defined by the expression

A B  A   A* B
The thickening of A by a sequence of structuring element {B}

A {B}  ((...(( A B1 ) B 2 )...) Bn )
In practice, the usual procedure is to thin the background of the set

and then complement the result.

59/19/2019

Example Thinning
We use two Hit-and-

miss Transforms

19-Sep-19 96
Skeleton
K
Gọi: S ( A)   S k ( A)
k 0
K  max{k | A  kB  };
Với
Sk ( A)  ( A  kB)  ( A  kB) B
Với B là kernel hay structure elements

( A kB )  ((..(( A B ) B ) ...) B)

Skeleton Finding via Distance Transform
• Distance transform: find the distance from

the nearest boundary for each point
- initialization
x0 (m, n)  x(m, n)
- iteration
xk (m, n)  xo (m, n)  min{xk 1 (i, j ) : d (m, n; i, j )  1}

98
Euclide Distance

Skeleton Finding via Distance Transform
• Skeleton is the set of points whose distance

from the nearest boundary is locally maximum
{(m, n) : xk (m, n)  xk (i, j ), d (m, n; i, j )  1}

102
Numerical Example
11111 11111 11111 1 1

11111 12221 12221 2 2
11111 12221 12321 3
11111 12221 12221 2 2
11111 11111 11111 1 1
x0(m,n) x1(m,n) x3(m,n) local skeleton
maximum

103
Image Example
original skeleton
MATLAB code: y=bwmorph(x,’skel’,inf);

104
Image Example (Con’t)
Binary fingerprint image Skeleton image

105
Thickening
• Used to grow selected regions of foreground pixels

• E.g. applications like approximation of convex hull

19-Sep-19 106
Definition Thickening
• Let B be a kernel and A be an image
with 1+1=1
• If foreground and background match exactly the SE,
then set the pixel at its origin to 1!
• Note that the value of the SE at the origin is 0 or
don’t care!

19-Sep-19 107
Example Thickening

19-Sep-19 108
Ví dụ 4.20: làm dày hình sau sử dụng các kernel
Kernel:
Hình 4.44 Hình và Kernel cho ví dụ 4.20

for (int iii = 0; iii < 10; iii++)
{
filter2D(dst_bin, dum, dst_bin.depth(), seLeft0);
filter2D(dum, dum, dst_bin.depth(), seRight0);
filter2D(dum, dum, dst_bin.depth(), seLeft90);
cv::bitwise_and(dum, dum, dum);
}

Tuan 5 Morphological Operation Principle

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

Tuan 5 Morphological Operation Principle

Uploaded by

Copyright:

Available Formats

Image Processing

Robotic & Vision Lab RoVis

• Counting coins is difficult because they touch each

Robotic & Vision Lab RoVis

19-Sep-19 Examples of stucturing elements

Robotic & Vision Lab RoVis

Erosion is used for shrinking of element A by using

Consider each foreground pixel in the input image

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

How do we fill “missing pixels”?

• This equation is based 0n obtaining the reflection 0f B

Consider each pixel in the input image

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Erosion followed by dilation, denoted ∘

• Structuring element: 3x3 square

Robotic & Vision Lab RoVis

• Opening with a 11 pixel diameter disc

Robotic & Vision Lab RoVis

• 3x9 and 9x3 Structuring Element

Robotic & Vision Lab RoVis

• 5x5 square structuring element

Robotic & Vision Lab RoVis

• Use large structuring element that fits into the big

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

dilation followed by erosion, denoted •

• Structuring element: 3x3 square

Robotic & Vision Lab RoVis

• Closing operation with a 22 pixel disc

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

• 5x5 square structuring element

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Opening Closing Tophat Gradient

Mat element = getStructuringElement(morph_elem, Size(2 * morph_size + 1, 2 * morph_size +

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis

Robotic & Vision Lab RoVis