A Review on Edge Detection Algorithms in Digital Image Processing Applications

International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 5 Issue: 10 69 – 75
_______________________________________________________________________________________________
69
IJRITCC | October 2017, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
A Review on Edge Detection Algorithms in Digital Image Processing
Applications
1
R V Ramana, 2
T V Rathnam,3
A Sankar Reddy
1, 2, 3
Assistant Professor
1
SVCE, 2, 3
AITS, JNTUA University
1
rvramana.r@gmail.com
2
isit.rathnam@gmail.com
3
sankar551@gmail.com
Abstract— Edge detection is one of the major step in Image segmentation, image enhancement, image detection and recognition applications.
The main goal of edge detection is that to localize the variation in the intensity of an image to identify the phenomena of physical properties
which produced by the capturing device. An edge might be characterized as a set of neighborhood pixels that forms a boundary between two
different regions. Detecting the edges is an essential technique for segmenting the image in to various regions based on their discontinuity in the
pixels. Edge detection has very important applications in image processing and computer vison. It is broadly used technique and quick feature
extraction technique hence used in various feature extraction and feature detection techniques. There exists several methods in the literature for
edge detection such as Canny, Prewitt, Sobel, Maar Hildrith, Robert etc. In this paper we have studied and compared Prewitt, Sobel, and Canny
detection operators. Our experimental study shows that the canny operator is giving better results for different kinds of images and has numerous
advantages than the other operators such as the nature of adaptive, works better for noisy images and providing the sharp edges with low
probability of false detection edges.
Index Terms: edges, Sobel, Prewitt, Canny edge detection Operators.
__________________________________________________*****_________________________________________________
I.INTRODUCTION
Image enhancement is considered as standout amongst the
most essential techniques in the image processing
applications. The primary aim of enhancing the image is to
improve the quality and to increase the visual appearance of
a picture to get a better representation than the original
image for automated image processing. It is important to
upgrade the contrast and to remove the noise present in the
image to enhance the image quality. The goal of image
enhancement is to process the one image so that the result is
apparently better than the original image for several
applications such as edge detection, pattern and various
object detection and recognition.
Edge detection is one of the major step in image
segmentation, image enhancement, image compression etc.
The goal of edge detection is to identify the intensity regions
that are having a slighter or more discontinuity with the
neighborhood. It is widely used technique and easy to
extract the features in a given input image. The results of the
edge detection must be reliable and accurate hence the
efficiency of the final result of the sub sequent image
processing operations depends on this. To meet this
requirement the edge detection algorithms should provide
the complete and significant information about the image
derivatives. However, differentiation of an image is
sensitive to various sources of noise, i.e., electronic,
semantic and discretization or quantification effects. Hence
to regularize the differentiation, the image must be
smoothed. But there are undesirable effects associated with
smoothing, i.e., loss of information and displacement of
prominent structures in the image plane. Hence it is difficult
to design a general edge detection algorithm which performs
well in many contexts and captures the requirements of
subsequent processing stages. Consequently, over the
history of digital image processing a variety of edge
detectors have been devised which differ in their purpose
and their mathematical & algorithmic properties. This
chapter makes a survey of some of the popular edge
detection techniques.
Edge detection is basically image segmentation technique,
divides spatial domain, on which the image is defined, into
meaningful parts or regions. Edges characterize boundaries
and are therefore a problem of fundamental importance in
image processing [1]. Edges typically occur on the boundary
between two different regions in an image. Edge detection
allows user to observe those features of an image where
there is a more or less abrupt change in gray level or texture
indicating the end of one region in the image and the
beginning of another. It finds practical applications in
medical imaging, computer guided surgery diagnosis, locate
object in satellite images, face recognition, and finger print
recognition ,automatic traffic controlling systems, study of
anatomical structure etc. Many edge detection techniques
have been developed for extracting edges from digital
images[2] .Gradient based classical operators like Robert,
Prewitt, Sobel were initially used for edge detection but
they did not give sharp edges and were highly sensitive to
noise image .Laplacian based Marr Hildrith operators also

_______________________________________________________________________________________________
70
_______________________________________________________________________________________
suffers from two limitations : high probability of detecting
false edges and the localization error may be severe at
curved edges but algorithm proposed by John F. Canny in
1986 is considered as the ideal edge detection algorithm for
images that are corrupted with noise. Canny's aim was to
discover the optimal edge detection algorithm which reduces
the probability of detecting false edge, and gives sharp edges
[3].
There are mainly three stages in the edge detection process
i.e. Filtering, Enhancement and detection. These techniques
are explained as follows. Images are often corrupted by
random variations in intensity values, called as noise. Some
common types of noises are salt and pepper noise, impulse
noise and Gaussian noise. For example, salt and pepper
noise contains random occurrences of both black and white
intensity values. These noises should be removed for
effectual edge detection. Hence, filters are used for
removing the noise. More filtering to reduce noise results in
a loss of edge strength and hence there is always a trade-off
between edge strength and noise reduction. In order to
facilitate the detection of edges, it is essential to determine
changes in intensity in the neighborhood of a point
[Enhancement emphasizes pixels where there is a significant
change in local intensity values and is usually performed by
computing the gradient magnitude.
Figure 1. Example masks in Horizontal and Vertical direction
Many points in an image have a nonzero value for the
gradient, and not all of these points are edges for a particular
application. Therefore, some method should be used to
determine which points are edge points. Frequently,
thresholding is used for edge detection.
In recent years, digital images plays an important role in
different applications such as restorations and
enhancements, image transmission and coding, color
processing, remote sensing, robot vision, hybrid techniques,
facsimile, pattern recognition, registration techniques, multi-
dimensional image processing, video processing, high
resolution display, high quality color representation and
super high definition applications. Image segmentation
process is important for these digital image processing
applications because the raw images are captured by the
digital camera or mobile camera with nonessential
background scenes or noise [4]. The elimination of
background images and noise is quite important to get
accurate results. This process is a low level image
engineering process which converts the raw images into
segments are pixels or objects. These pixels are converted
into vectors and analyzed or tested with the any one of the
image segmentation process. The removal of noise from the
images is performed using de- noising techniques such as
filtering, enhancement, detection and localization for
identifying the edges. These edges are analyzed with the
help of mid and high – level image engineering processing
methods. The adequate edges are identified by diverse edge
detection techniques in several image processing
applications such as object recognition, motion analysis,
pattern recognition, computer- guided surgery, finger print
recognition, automatic traffic controlling systems,
anatomical structure and image processing [5]. Detecting the
edges from noisy images or corrupted images is difficult in
nature. In the past two decades‟ several edge detection
techniques or algorithms are proposed, based on that the
effective edges are evaluated or analyzed. The ultimate
reason behind in these methods to restrict the false detection
in the edges, edge localization and computational time. In
this, canny optimal detection algorithm aims to discover the
optimal edge which reduces the probability of detecting
false edges, and gives sharp edges [6].
II.RELATED STUDY
Image segmentation is the method to simplify the digital
image into segments or pixels which are easier to analyze
and identify the effective edges in a complex image. The
image engineering processes are sub divided into low level,
mid-level, high level. In low-level engineering process, the
raw image is taken as an input and the noises are eliminated.
These raw images are transformed into pixels. The pixel is a
collection of discrete cells in the particular image. The
characteristics of color and texture are found as similarity in
a pixel. Generally, the raw images are taken as a colored
image which is needed to be modify into grayscale or black
and white images, because the edges can be detected using
the pixels. In the Mid –level, the output is presented in the
form of attributes like edges, contours, and the identity of
individual objects. In the High- level engineering process,
involves making sense, of a recognized object in the image
analysis and it performs the cognitive functions associated
with computer vision.
A. Edge Types
An edge in an image is a significant local change in the
image intensity, usually associated with a discontinuity in
either the image intensity or the first derivative of the image
intensity. Discontinuities in the image intensity can be either
„step edge‟, where the image intensity abruptly changes
from one value on one side of the discontinuity to a different
value on the opposite side, or „line edge‟, where the image
intensity abruptly changes value but then returns to the
starting value within some short distance. Because of low
frequency components or the smoothing introduced by most
sensing devices, sharp discontinuities, like step and line

_______________________________________________________________________________________________
71
_______________________________________________________________________________________
edges, rarely exist in real.Because of low frequency
components or the smoothing introduced by most sensing
devices, sharp discontinuities, like step and line edges,
rarely exist in realimages. „Step edges‟ become „ramp
edges‟ and „line edges‟ become „roof edges‟, where intensity
changes are not instantaneous but occur over a finite
distance. Illustrations of these edge shapes are shown in
Figure 2.
Figure 2.Edge types a. step edge b. Ramp edge c. line edge d. roof
edge.
B. Edge Detection
Edge detection is a fundamental step in image segmentation,
image recognition, image enhancement, restoration, image
registration, image compression, and so on. The goal of edge
detection is to localize the variations in the intensity of an
image and to identify the physical phenomena which
produce them. Edge detection is widely used in image
processing as it is a quick and easy way of extracting most
of the important features in an image. Edge detection must
be efficient and reliable because the validity, efficiency and
possibility of the completion of subsequent processing
stages rely on it. To fulfill this requirement, edge detection
should provide all significant information about the image
including image derivatives. However, differentiation of an
image is sensitive to various sources of noise, i.e.,
electronic, semantic and discretization or quantification
effects. Hence to regularize the differentiation, the image
must be smoothed. But there are undesirable effects
associated with smoothing, i.e., loss of information and
displacement of prominent structures in the image plane.
Hence it is difficult to design a general edge detection
algorithm which performs well in many contexts and
captures the requirements of subsequent processing stages.
Consequently, over the history of digital image processing a
variety of edge detectors have been devised which differ in
their purpose and their mathematical & algorithmic
properties. This chapter makes a survey of some of the
popular edge detection techniques.Edge detection contain
three steps namely, filtering, enhancement and detection.
Filtering
Images are often corrupted by random variations in intensity
values, called as noise. Some common types of noises are
salt and pepper noise, impulse noise and Gaussian noise. For
example, salt and pepper noise contains random occurrences
of both black and white intensity values. These noises
should be removed for effectual edge detection. Hence,
filters are used for removing the noise. More filtering to
reduce noise results in a loss of edge strength andhence there
is always a trade-off between edge strength and noise
reduction.
Enhancement
In order to facilitate the detection of edges, it is essential to
determine changes in intensity in the neighborhood of a
point. Enhancement emphasizes pixels where there is a
significant change in local intensity values and is usually
performed by computing the gradient magnitude.
Detection
Many points in an image have a nonzero value for the
gradient, and not all of these points are edges for a particular
application. Therefore, some method should be used to
determine which points are edge points. Frequently,
thresholding is used for edge detection.
C. Techniques used
Edge detection is the name for a set of mathematical
methods which aim at identifying points in a digital image at
which the image brightness changes sharply or, more
formally, has discontinuities. The points at which image
brightness changes sharply are typically organized into a set
of curved line segments termed edges. In this paper three
edge detection methods are used to find the better
performance on various images. The list of edge detection
operators are shown in Figure 3. In this paper we have
studied Sobel, Prewitt, Canny Edge detection operators.
Figure 3. Types of edge detection operators
Most of the edge detection algorithms in the literature will
read an input image and performed convolution operation
with mask. Later the gradients are computed in both
directions.i.e. Horizontal and vertical directions. Once the
gradients are computed a thresholding is used to remove the
pixels value above the specified threshold. If the pixel is
greater than the specified threshold it is considered as an

_______________________________________________________________________________________________
72
_______________________________________________________________________________________
edge otherwise it is not an edge. The working flow of a
generic edge detection algorithm is shown in Figure 4.
Figure 4. General flow of a conventional edge detection algorithm.
i) Sobel Edge Detection: In digital images, the approximate
partial derivation in gradient is computed by the Sobel
operator. In terms of computations, the edge is based on the
edge convolving with the integer, separable and small
valued filter in vertical and horizontal directions
.Mathematically, the approximations of the derivative can be
calculated by using two 3*3 kernels which are convolved
with the original image.
A 3X3 kernel is used for enhancement. Maximum of edges
are identified with respect to perpendicular angle. In this,
one kernel is allowed to rotate with 90 degrees and another
one kernel is stay on this position. The raw image is taken as
an input. It detects two types of edges like Horizontal edges
(Sy) and Vertical Edges (Sx). The Magnitude (M) is
calculated by adding the partial derivatives of Sx and Sy.
The threshold is added with the magnitude to get the output
image. The Magnitude of the gradients (M) are calculated
with Sx, Sy indicates vertical and horizontal positions. The
Sx, Sy partial derivations are given below [7]
M= 𝑆𝑥2 + 𝑆𝑦2
The angle of orientation of the edge is given by:
ɵ= arctan (S𝑦/S𝑥)
Where ɵ is angle to find the direction.
Sx = (a2+Ca3+a4) – (a0+Ca7+a6)
Sy= (a0+Ca1+a2) – (a6+Ca5+a4)
ii) Prewitt Edge Detection: The maximum responses which
are directly from the kernel are obtained by the use of
Prewitt Edge Detector. The Prewitt edge operator or
detectors are used for the measurement of two components
i.e. horizontal edge components and vertical edge
components. These two components (vertical and
horizontal) are used different kernels.
Prewitt Operator:
Prewitt operator is quite similar to the Sobel Operator with
the difference of the C value is 1.
The main advantage of this technique is to provide a better
performance on horizontal and vertical edges in the images
and higher responses for noisy images. The operator should
have the following properties: one is
i) Both negative and positive values should be in the all
convolution masks.
ii) The final results should be zero when sum is obtained.
It is widely used technique to evaluate the magnitude (M) of
the edges.
The Sx, Sy partial derivations are given [8]
M= 𝑆𝑥2 + 𝑆𝑦2
Sx = (a2+Ca3+a4) – (a0+Ca7+a6)
Sy= (a0+Ca1+a2) – (a6+Ca5+a4)
Where the threshold value C=1 and a0, a1, a2, a3, a4, a5, a6,
a7 are masks.
The original image pixel 12 is taken and the magnitude is
calculated with Sx mask and the same procedures are
followed to calculate the Sy mask. The final value of Sy
mask is 20.
iii) Canny Edge Detection: The Canny Edge detection is
introduced by John Canny (1983). Canny edge detection
technique is one of the standard edge detection techniques. It
is used many of the newer algorithms that have been
developed. The noise can be reduced and suppression can be
minimum are the stages of canny algorithms which are used
in images.
This algorithm is focuses to separate the background noise
from complex image and to find the effective edges for
optimized solutions. It uses first derivative of an image. It is
used to measures the mean square distance, error edge map
and signals to noise ratio. The improved canny edge
detection algorithm provides the better optimal solutions
with respect to noisy images [9]. Canny edge detection steps
are shown in Figure.

_______________________________________________________________________________________________
73
_______________________________________________________________________________________
Figure 5. Working flow of canny edge detection algorithm.
IV.RESULTS & DISCUSSION
This section shows our works on the Dark Channel Prior and
the edge detection algorithms. Our work has done on
OpenCV environment using Python. In Figure 6 we have
taken an image with frontal face and in Figure 7 we
considered the image taken in outdoor i.e. building and the
results for various edge detection algorithms are shown.
a b
c d
e f
Figure 6.Comparison of edge detection results.
Figure 6.a is the input image, Figure 6.b is the Prewit
operator result and Figure 6.c and Figure 6.d are the results
of Sobel operator in x direction and y direction and the
Figure 6.e is the Laplacian result and Figure 6.d is the canny
edge detection result. The same results are considered for
outdoor image in Figure 7.
a b
c
d
e f
Figure 7. Comparison of edge detection results.

_______________________________________________________________________________________________
74
_______________________________________________________________________________________
Table 1. Summary of the edge detection algorithm.
S.no Edge
Detection
Operator
Sensitivity Method Merit Demerit
1 Sobel
Medium
Edges are detected
based on the
perpendicular angle
Simple and
smooth edges are
detected
In accurate average
results with respire
to complex images.
2 Prewitt Gives good results of
horizontal and
vertical edges
3 Canny High Used to eliminate the
noise and to
find the effective
edges
Better detection,
Adaptive,
Good
Localization.
False Zero crossing
PARAMETERS FOR EVALUATION
Correlation Coefficient
The correlation coefficient a concept from statistics is a
measure of how the trends in the predicted values follow
trends in past actual values. It is a measure of how well the
predicted values from a forecast model "fit" with the real-
life data. The correlation coefficient is a number between 0
and 1. As the strength of the relationship between the
predicted values and actual values increases so does the
correlation coefficient. A perfect fit gives a coefficient of
1.0. Thus the higher the correlation coefficient the better
[12].
Edge Detection Performance
The performance of the edge detection algorithm is
computed based on determining the false edges i.e. those
edges are not actual edges but the algorithm determined as
the edges and finding the missed edges i.e. the those edges
basically edges but that are not included in the result. The
mean and square error determine the true edges. Finally the
tolerance of the edges are also considered. The first two
parameters are about to the edge detection and the rest of the
two parameter denotes the edge localization. These
Performance Ratio (PR), Miss Count (MC), Peak Signal to
Noise Ratio (PNSR).
PR =
𝑇𝑟𝑢𝑒 𝑒𝑑𝑔𝑒𝑠
𝐹𝑎𝑙𝑠𝑒 𝑒𝑑𝑔𝑒𝑠
FoM=
1
max ⁡(IA,I1) ⁡
1
1+𝑑∝2
IA
𝑖−1
To compare image compression quality MSE and PSNR is
used [10][13].The PSNR and MSE are measured by the
following equation.
MSE=
𝐼1 m,n −𝐼2 m,n
𝑀∗𝑁
Where M and N are rows and column of an image 𝐼1 is
considered as an original raw image and 𝐼2 is a considered
as a detected output image [11].
PSNR=10log10
𝑅2
𝑀𝑆𝐸
V.CONCLUSION
In this paper, we have reviewed the edge detection
algorithms such as Sobel, Prewitt, and Canny edge
detection. Among these techniques each algorithms are
suitable for a specific application. But the canny edge
detection is giving the standout results in most of the
applications. As our experimental results also shows that the
canny algorithm is giving the best results for the tested input
images. The other edge detection algorithm such as
Robert‟s, Kirsch edge detection algorithms are to be
analyzed in the future work and needed to apply the same
techniques for poor contrast images as our future work.
REFERENCES
[1] James Clerk Maxwell, 1868 DIGITAL IMAGE
PROCESSING Mathematical and Computational Methods.
[2] R .Gonzalez and R. Woods, Digital Image Processing,
Addison Wesley, 1992, pp 414 - 428.
[3] S. Sridhar, Oxford university publication. , Digital Image
Processing.
[4] Beant K, Anil G. Mathematical morphological edge
detection for remote sensing images. IEEE Transactions.
2011; 324–7.
[5] Ireyuwa EI. Comparison of edge detection techniques in
image processing techniques. International Journal of
Information Technology and Electrical Engineering. 2013;
25–9.
[6] K. Bala Krishnan*, Shiva Prakash Ranga and Nageswara
Guptha “A Survey on Different Edge Detection
Techniques for Image Segmentation”, Indian Journal of
Science and Technology, Vol 10(4), DOI:
10.17485/ijst/2017/v10i4/108963 January 2017
[7] Khaire PA, Thakur NV. A fuzzy set approach for edge
detection.International Journal of Image Processing.
2012;403–12.
[8] Lakshmi SJ, Prasanna. Image segmentation using
variousedge detection techniques. International Journal of
Emerging Technologies in Engineering Research. 2016;
129–32.

_______________________________________________________________________________________________
75
_______________________________________________________________________________________
[9] Liu X, Dawn Z, Xu WWE. Image edge deduction algorithm
based on improved wavelet transform. International Journal
of Signal Processing, Image Processing and Pattern
Recognition.2016; 9(4):435–42.
[10] Pradyna J, Usha J. Sobel edge detection implementation
using FPGA. International Journal of Science and
Research. 2015; 4(7):439–42.
[11] Ramn M, Aggarwal H. Study and comparison of various
image edge detection techniques. International Journal of
Image Processing. 2009; 3(1):1–12.
[12] Sos Agaian, Ali Almuntashri “Noise-Resilient Edge
Detection Algorithm for Brain MRI Images”, IEEE, 31st
Annual International Conference of the IEEE EMBS
Minneapolis, Minnesota, USA, September 2-6, 2009.978-1-
4244-3296-7/09/$25.00 ©2009 IEEE.
[13] Jianjum Zhao, Heng Yu, Xiaoguang Gu and Sheng Wang.
“The Edge Detection of River model Based on Self-
adaptive Canny Algorithm And Connected Domain
Segmentation” IEEE,Proceedings of the 8th World
Congress on Intelligent Control and Automation July 6-9
2010, Jinan, China, 978-14244-6712-9/10/$26.00 ©2010
IEEE
Authors Profile:
I, R Venkata Ramana is currently working as Assistant Professor
at the department of Computer Science and Engineering in SVCE-
Tirupati, JNTUA. He received his master‟s degree from JNTUA in
2012 and Bachelor‟s degree in 2008. He has 5 years of experience
and has taught various subjects in computer science stream and
organized various national conferences and workshops in the
organization. His research interests are Knowledge Engineering,
Image mining, Image Analytics, Recommender Systems. He also
published various National and International Journals.
Mr. T V Rathnam is currently working as Assistant Professor in
the department of Computer Science and Engineering in AITS-
Tirupati, JNTUA. He received his master‟s degree from JNTUA in
2012 and Bachelor‟s degree in 2008. He has 5 years of experience
and has taught various subjects in computer science stream and
organized various national conferences and workshops in the
organization. His research interests are Computer Architecture,
Computer networks, Big data Analytics.0
Mr. A. Sankar Reddy is currently working as Assistant Professor
in the department of Computer Science and Engineering in AITS-
Tirupathi, JNTUA. He received his master‟s degree from JNTUA
University in 2014 and Bachelor‟s degree from JNTUH in 2007.
He has 4 years of teaching experience and taught various subjects
in computer science stream and organized various national
conferences and workshops in the organization. His research
interests are cloud computing, computer vision and Computer
Graphics and knowledge engineering.

A Review on Edge Detection Algorithms in Digital Image Processing Applications

More Related Content

What's hot (17)

Similar to A Review on Edge Detection Algorithms in Digital Image Processing Applications (20)

More from rahulmonikasharma (20)

Recently uploaded (20)

A Review on Edge Detection Algorithms in Digital Image Processing Applications