CSE 307

Fundamentals of Image Processing

Lecture #7 Edge Detection

Prepared & Presented by Asst. Prof. Dr. Samsun M. BAŞARICI

 Objectives

 What is edge detection and why is it so important to

computer vision?

 What are the main edge detection techniques and how

well do they work?

 How can edge detection be performed in MATLAB?

 What is the Hough Transform and how can it be used to

post-process the results of an edge detection algorithm?
 Edge detection

Edge detection is a fundamental tool in image processing,

machine vision and computer vision, particularly in the areas
of feature detection and feature extraction.
The purpose of detecting sharp changes in image brightness
is to capture important changes in the image.
Image filters are commonly used for edge detection.
 Introduction

Image filtering allows us to apply various effects on

images.
Here we consider some 2D filters similar to the ones used
in Photoshop as Custom Filter.

This filter has no effect

on the image
0 0 0 
 0 1 0
 
0 0 0 
 Introduction (cont.)

Blurring is done by taking the average of the current pixel

and it’s 4 neighbors.

0.0 0.2 0.0

0.2 0.2 0.2
 
0.0 0.2 0.0
 Introduction (cont.)

Bluring with a 5x5 filter

0 0 1 0 0
0 1 1 1 0
 
1 1 1 1 1 
 
0 1 1 1 0
0 0 1 0 0
 Introduction (cont.)

Motion blur is achieved by blurring in only 1 direction.

1, 0, 0, 0, 0, 0, 0, 0, 0  +
0, 1, 0, 0, 0, 0, 0, 0, 0 
 
0, 0, 1, 0, 0, 0, 0, 0, 0 
 
0, 0, 0, 1, 0, 0, 0, 0, 0 
0, 0, 0, 0, 1, 0, 0, 0, 0 
 
0, 0, 0, 0, 0, 1, 0, 0, 0 
0, 0, 0, 0, 0, 0, 1, 0, 0 
 
0, 0, 0, 0, 0, 0, 0, 1, 0 
0, 0, 0, 0, 0, 0, 0, 0, 1 
 
 Introduction (cont.)

 A filter to find the horizontal edges .

+
 0, 0, 0, 0, 0 
 0, 0, 0, 0, 0 
 
 - 1, - 1, 2, 0, 0 
 
 0, 0, 0, 0, 0 
 0, 0, 0, 0, 0,
 Introduction (cont.)

 A filter that finds vertical edges.

+
 0, 0, - 1, 0, 0 
 0, 0, - 1, 0, 0
 
 0, 0, 4, 0, 0
 
 0, 0, - 1, 0, 0
 0, 0, - 1, 0, 0
 Introduction (cont.)

 A filter that is good at finding edges of 45°

+
- 1, 0, 0, 0, 0 
 0, - 2, 0, 0, 0 
 
 0, 0, 6, 0, 0 
 
 0, 0, 0, - 2, 0 
 0, 0, 0, 0, - 1,
 Introduction (cont.)

 A simple edge detection filter that detects edges in all

directions
+

- 1, - 1, - 1 
- 1, 8, - 1 
 
- 1, - 1, - 1 
 Introduction (cont.)
 To sharpen the image is very similar to finding edges, The result will be
a new image where the edges are enhanced, making it look sharper.
 For example, the center value of the edge detection filter from the
previous example, is incremented with 1. The result will be an image
with the same brightness as the original, but sharper

- 1, - 1, - 1 
- 1, 9, - 1 
 
- 1, - 1, - 1 
 Introduction (cont.)

 Another example of sharpen filter:

+
 - 1, - 1, - 1, - 1, - 1 
 - 1, 2, 2, 2, - 1 
 
 - 1, 2, 8, 2, - 1 
 
 - 1, 2, 2, 2, - 1 
 - 1, - 1, - 1, - 1, - 1 
 What is edge detection?
 Edge detection is a fundamental image processing
operation used in many computer vision solutions.
 Goal of edge detection algorithms: to find the most relevant
edges in an image or scene.
These edges should then be connected into meaningful lines
and boundaries, resulting in a segmented image containing
two or more regions.
Subsequent stages in a machine vision system will use the
segmented results for tasks such as object counting,
measuring, feature extraction, and classification.
 What is edge detection? (cont.)

 Biological vision: there is compelling evidence that the very

early stages of the human visual system (HVS) contain
edge-sensitive cells that respond strongly (i.e., exhibit a
higher firing rate) when presented with edges of certain
intensity and orientation.
Edge detection algorithms attempt to emulate an ability
present in the HVS.
Such an ability is, according to many vision theorists,
essential to all processing steps that take place afterward in
the HVS.
 What is edge detection? (cont.)

 Edge detection is a hard image processing problem.

 Most edge detection solutions exhibit limited performance in

the presence of images containing real-world scenes.

 It is common to precede the edge detection stage with pre-

processing operations such as noise reduction and
illumination correction
 Basic concepts
 Edge: a boundary between two image regions having distinct
characteristics according to some feature (e.g., gray level, color, or
texture).
 In grayscale 2D images: a sharp variation of the intensity function
across a portion of the image.
 Basic concepts (cont.)

 Edge detection methods usually rely on calculations of the

first or second derivative along the intensity profile.
The magnitude of the first derivative can be used to detect the
presence of an edge at a certain point in the image.
The sign of second derivative can be used to determine whether
a pixel lies on the dark or bright side of an edge.
Moreover, the zero crossing between its positive and negative
peaks can be used to locate the center of thick edges.
Basic concepts (cont.)
 First and
second
derivative
 Basic concepts (cont.)

 The
influence of
noise
 Basic concepts (cont.)

 The process of edge detection consists of three

main steps:
Noise reduction
Detection of edge points
Edge localization

 In MATLAB: edge
 Background: Derivatives (remember?)

First-order derivative
f
 f '( x)  f ( x  1)  f ( x)
x

Second-order derivative
2
 f
2
 f ( x  1)  f ( x  1)  2 f ( x)
x
Characteristics of First and Second Order Derivatives

First-order derivatives generally produce thicker edges in

image
Second-order derivatives have a stronger response to fine
detail, such as thin lines, isolated points, and noise
Second-order derivatives produce a double-edge
response(negative to positive or positive to negative at
ramp and step transition in intensity
The sign of the second derivative can be used to
determine whether a transition into an edge is from light
to dark or dark to light
 Characteristics of First and Second Order
Derivatives (cont.)
To compute the first and second drivatives at every pixel
location is to use spatial filters.
The response of the mask at the center point of the region
is computed as

R  1z1  2 z2    9 z9
 Detection of Isolated Points

The Laplacian
2 2
2  f  f
 f ( x, y )  2  2
x y
 f ( x  1, y )  f ( x  1, y )  f ( x, y  1)  f ( x, y  1)
4 f ( x, y )

1 if | R( x, y ) | T 9
R   wk zk
g ( x, y )  
0 otherwise k 1
 First-order derivative edge detection

 Digital approximations of the first-order derivative

 Roberts:
First-order derivative edge detection (cont.)

 Prewitt operators:
First-order derivative edge detection (cont.)

 Sobel operators:
First-order derivative edge detection (cont.)

 Prewitt operator
J = imread('bw_scissors.png');
Jt = imcomplement(J);
Jt = Jt .* 0.5;
Ju = imcomplement(Jt);
Jd = im2double(Ju);
hy = fspecial('prewitt')
hx = hy'
J8 = imfilter(Jd,hy);
J7 = imfilter(Jd,hx);
Jx = mat2gray(J7);
Jy = mat2gray(J8);
imshow(Jy), figure, imshow(Jx)
J10 = abs(J7) + abs(J8);
Jxy = mat2gray(J10);
figure, imshow(Jd), figure, imshow(Jxy)
First-order derivative edge detection (cont.)

 Prewitt operator
First-order derivative edge detection (cont.)

 Sobel operator
J = imread(’Figure14_05_a.png');
Jd = im2double(J);
hy = fspecial('sobel')
hx = hy'
J1 = imfilter(Jd,hy);
J2 = imfilter(Jd,hx);
imshow(J1)
J1c = imcomplement(J1); figure, imshow(J1c)
J2c = imcomplement(J2); figure, imshow(J2c)
J3 = abs(J1) + abs(J2);
J3c = imcomplement(J3); figure, imshow(J3c)
 First-order derivative edge detection (cont.)
 Sobel operator
 First-order derivative edge detection (cont.)

 Kirsch compass masks:

First-order derivative edge detection (cont.)

 Robinson compass masks:

First-order derivative edge detection (cont.)

 Sobel operator with threshold (example)

J = imread('Figure14_05_a.png');
thresh = 0;
BW0 = edge(J,'sobel',thresh);
imshow(imcomplement(BW0))
thresh = 0.05;
BW1 = edge(J,'sobel',thresh);
figure, imshow(imcomplement(BW1))
thresh = 0.2;
BW2 = edge(J,'sobel',thresh);
figure, imshow(imcomplement(BW2))
[BW,thresh] = edge(J,'sobel');
thresh
figure, imshow(imcomplement(BW)))
 First-order derivative edge detection (cont.)

 Sobel operator with threshold (example)

 Second-order derivative edge detection

 Problems with Laplacian:

It generates “double edges”, i.e., positive and
negative values for each edge.
It is extremely sensitive to noise.

 In MATLAB:
h = fspecial('laplacian',0);
J = edge(I,'zerocross',t,h);
Second-order derivative edge detection (cont.)

 Laplacian and zero-cross

J = imread('Figure14_05_a.png');
h = fspecial('laplacian',0)
J1 = edge(J,'zerocross',0,h);
J2 = edge(J,'zerocross',[],h);
figure, imshow(imcomplement(J1));
figure, imshow(imcomplement(J2));
K = imnoise(J,'gaussian',0,0.0001);
K1 = edge(K,'zerocross',0,h);
K2 = edge(K,'zerocross',[],h);
figure, imshow(imcomplement(K1));
figure, imshow(imcomplement(K2));
Second-order derivative edge detection (cont.)

 Laplacian and
zero-cross
 Second-order derivative edge detection (cont.)

 Laplacian of Gaussian (LoG):

Works by smoothing the image with a Gaussian
low-pass filter, and then applying a Laplacian
edge detector to the result.
The LoG filter can sometimes be approximated
by taking the differences of two Gaussians of
different widths, in a method known as
Difference of Gaussians (DoG).
 Second-order derivative edge detection (cont.)

 Laplacian of Gaussian (LoG) transfer function

(Mexican hat)
 Second-order derivative edge detection (cont.)

 Laplacian of Gaussian (LoG) transfer function (2D

cross-section of 3D plot) and mask (kernel)
Second-order derivative edge detection (cont.)

 Laplacian of Gaussian (LoG):

J = imread('Figure14_05_a.png');
J1 = edge(J,'log');
imshow(imcomplement(J1))
J2 = edge(J,'log',[],1);
figure, imshow(imcomplement(J2))
J3 = edge(J,'log',[],3);
figure, imshow(imcomplement(J3))
 Second-order derivative edge detection (cont.)

Laplacian
of Gaussian
(LoG)
 Marr-Hildreth edge detector
x2  y 2

G ( x, y )  e ,  : space constant.
2 2

The
Laplacian of Gaussian (LoG)
Marr- 2  2 G ( x, y )  2G ( x, y )
 G ( x, y )  2

Hildreth x y 2
edge
   x  2 2     y  x  2y 
2 2 2 2
x y
detector   2e    2 e 2 
x    y   
2 x2  y 2 2 x2  y 2
x 1   2 2
y 1   2 2
  4  2 e   4  2 e
     
2 2 2 x2  y 2
 x  y    2 2
 4  e
  
Marr-Hildreth edge detector (cont.)
 Marr-Hildreth Algorithm

1. Filter the input image with an nxn Gaussian lowpass

filter. N is the smallest odd integer greater than or
equal to 6

2. Compute the Laplacian of the image resulting from

step1

3. Find the zero crossing of the image from step 2

g ( x, y )   G ( x, y ) f ( x, y ) 
2
Marr-Hildreth edge detector (cont.)
 The Canny edge detector
 Optimal for step edges corrupted by white
noise.

 The Objective

1. Low error rate

The edges detected must be as close as possible to the true
edge

2. Edge points should be well localized

The edges located must be as close as possible to the true
edges

3. Single edge point response

The number of local maxima around the true edge should be
minimum
 The Canny edge detector (cont.)

1. The input image is smoothed using a Gaussian lowpass filter (LPF),

with a specified value of σ.
2. The local gradient (intensity and direction) is computed for each
point in the smoothed image.
3. The edge points at the output of step 2 result in wide ridges. The
algorithm thins those ridges, leaving only the pixels at the top of each
ridge (non-maximal suppression).
4. The ridge pixels are then thresholded using two thresholds, Tlow and
Thigh: ridge pixels with value greater than Thigh are considered strong
edge pixels; ridge pixels with values between Tlow and Thigh are said to
be weak pixels. This process is known as hysteresis thresholding.
5. The algorithm performs edge linking, aggregating weak pixels that
are 8-connected to the strong pixels.

In MATLAB: J = edge(I, 'canny', T, sigma);

 The Canny edge detector (cont.)

Example 14.5 (PIVPUM):

J = imread('twin_towers_skybridge_gray.png');
BW = edge(J,'canny');
[BW, t] = edge(J,'canny');
imwrite(imcomplement(BW),'twin_towers_canny_default.png');
sigma = 0.5
BW = edge(J,'canny',t,sigma);
imwrite(imcomplement(BW),'twin_towers_canny_sigma_half.png');
sigma = 2
BW = edge(J,'canny',t,sigma);
imwrite(imcomplement(BW),'twin_towers_canny_sigma_two.png');
t = [0.01 0.1]
BW = edge(J,'canny',t);
imwrite(imcomplement(BW),'twin_towers_canny_mod_thresh.png');
 The Canny edge detector (cont.)

Example
14.5:
 Edge linking and boundary detection

 Goal of edge detection: to produce an image containing

only the edges of the original image.
 However, due to the many technical challenges (noise,
shadows, occlusion, etc,), most edge detection algorithms
will output an image containing fragmented edges.
 Additional processing is needed to turn fragmented edge
segments into useful lines and object boundaries.
 Hough transform: a global method for edge linking and
boundary detection.
 The Hough transform (cont.)

 A mathematical method designed to find lines in images.

It can be used for linking the results of edge detection,
turning potentially sparse, broken, or isolated edges into
useful lines that correspond to the actual edges in the image.
 The Hough transform (cont.)
 Let (x,y) be the coordinates of a point in a binary image (containing
thresholded edge detection results).
 The Hough transform stores in an accumulator array all pairs (a,b)
that satisfy the equation y = ax+ b. The (a,b) array is called the
transform array.
 Example:, the point (x,y) = (1,3) in the input image will result in the equation b
= -a + 3, which can be plotted as a line that represents all pairs (a,b) that satisfy
this equation.
 The Hough transform (cont.)
 Since each point in the image will map to a line in the transform domain,
repeating the process for other points will result in many intersecting lines,
one per point.
 The meaning of two or more lines intersecting in the transform domain is
that the points to which they correspond are aligned in the image.
 The points with the greatest number of intersections in the transform
domain correspond to the longest lines in the image.
 The Hough transform (cont.)
 Coordinate conversion
Describing lines using the equation y =
ax + b (where a represents the gradient)
poses a problem, though, since vertical
lines have infinite gradient.
 This limitation can be circumvented by
using the normal representation of a line,
which consists of two parameters: ρ and θ.
 In this new representation, vertical lines
will have θ = 0.
 It is common to allow ρ to have negative
values, therefore restricting θ to the range
−90◦ < θ ≤ 90◦.
 The Hough transform (cont.)
 Under the new set of coordinates, the Hough transform
can be implemented as follows:
1. Create a 2D array corresponding to a discrete set of values
for ρ and θ. Each element in this array is often referred to
as an accumulator cell.
2. For each pixel (x,y) in the image and for each chosen value
of θ, compute x cos θ + y sin θ and write the result in the
corresponding position (ρ, θ) in the accumulator array.
3. The highest values in the (ρ, θ) array will correspond to the
most relevant lines in the image.
 The Hough transform (cont.)
 In MATLAB:
The IPT contains a function for Hough transform
calculations, hough, which takes a binary image as an
input parameter, and returns the corresponding Hough
transform matrix and the arrays of ρ and θ values over
which the Hough transform was calculated.
Optionally, the resolution of the discretized 2D array for
both ρ and θ can be specified as additional parameters.
 The Hough transform (cont.)

 Example 14.6 PIVPUM)

BW = imread('Figure14_16_a.png');
imshow(BW)

[H,T,R] = ...
hough(BW,'RhoResolution',0.5,'ThetaResolution',0.5);

figure, imshow(imadjust(mat2gray(H)),'XData', ...

T,'YData',R, 'InitialMagnification','fit');

xlabel('\theta'), ylabel('\rho');
axis on, axis normal, hold on;
colormap(gray);
 The Hough transform (cont.)
 Example 14.6
 The Hough transform (cont.)

 In MATLAB:
The IPT also includes two useful companion functions for
exploring and plotting the results of Hough Transform
calculations:
 houghpeaks: identifies the k most salient peaks in the Hough
transform results, where k is passed as a parameter.
 houghlines: draws the lines associated with the highest peaks
on top of the original image.
 The Hough transform (cont.)

 Example 14.7
(PIVPUM)
 Summary

You should now know

What edge detection is and its importance in computer
vision

The main edge detection techniques and how well they

work

How you can perform edge detection in MATLAB

What the Hough Transform is and how it can be used to

post-process the results of an edge detection algorithm
 References
 Rafael C. Gonzalez, Richard E. Woods, “Digital Image
Processing, 4th Ed.”; Pearson, 2018, (DIP)

 Rafael C. Gonzalez, Richard E. Woods, Steven L. Eddins,

“Digital Image Processing Using MATLAB”; Pearson,
2004, (DIPUM)

 Oge Marques, “Practical Image and Video Processing

Using MATLAB”; Wiley, 2011, (PIVPUM)