International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169

Volume: 3 Issue: 2 563 - 566

_______________________________________________________________________________________________
Circular Hough Transform For Detecting And Measuring Circles of Object

Mr. Chetan S. More1 Prof. Mrs. Anita P. Patil2

Electronics and Tele Communication Department Electronics Department
Dr. J. J. Magdum College Of Engineering Dr. J. J. Magdum College Of Engineering
Jaysingpur, India Jaysingpur, India
E-mail: [email protected] E-mail: [email protected]

Abstract- This article presents a robust method for detecting and measuring the circles of industrial object like water pump pulley(WPP-44),
images based on circular Hough transform. The software of the application is based on detecting and measuring the circles surrounding of that
object. The objective of this application is the recognition of circular shapes in an image. This paper presents efficient implementation of Image
Processing technique for measurement of dimension of the object used in Turbine system. The proposed work makes use of the Matlab software
for the said work..At First an segmentation or edge detection technique is used for finding the edges in the input image, then the characteristic
points of circles are determined, after this process the pattern of the circles is extracted, and then with the help of other extraction parameter of
the diameter of object(WPP-44).

Keywords- Hough Transform, Circle Hough Transform, Detection, Measurement, Edges, Water Pump Pulley.
__________________________________________________*****_________________________________________________

I. INTRODUCTION (CHT)[5].The Hough transform can be described as a

Nowadays, in industry various methods are used for quality transformation of a point in the x,y-plane to the parameter
assessment of final product. For measuring various circles in space. The parameter space is defined according to the shape
any product different producer are used. Likely for Water of the object of interest. A straight line passing through the
Pump Pulley - 44 (WPP-44) there are many circles present at points (x1, y1) and (x2,y2 ) can be in the x,y-plan to be described
top side of an object. by:

y  ax  b
Therefore, industry has to balance the demand of high
quality production and low prices may ways, such as to (1)
increase the capacity and production technology. The process This is the equation for a straight line in the Cartesian
of quality assessment production is complex. Due to its coordinate system, where a and b represent the parameters of
complexity, there are failures object to occur in each the line. The Hough transform for lines does not use this
production process. The failure object is found from the defect representation of lines, since lines perpendicular to the x-axis
of object which cannot be repaired and thus wastes the will have an a-value of infinity[9]. This will force the
production cost. The Failure Analysis (FA) is to identify the parameter space a,b to have infinite size. Instead a line is
failure of defect on media and classify the pattern of defect. represented by its normal which can be represented by an
The FA helps technicians to access problems rapidly, which angel θ and a length ρ .

ρ  xcos(θ )  ysin(θ )
decreases the waste and the production cost. Consequently,
this study is to introduce a method of Detection and (2)
measurement of these circles present in WPP-44 based on the The parameter space can now spanned by θ and ρ , where θ
Circular Hough Transform and Image . will have a finite size, depending on the resolution used for θ .
The distance to the line ρ will have a maximum size of two
times the diagonal length of the image.
II. HOUGH TRASFORM
A commonly faced problem in computer vision is
III. CIRCULAR HOUGH TRASFORM
determining the location, number or orientation of a particular
object in an image. One problem could, for instance, be to The circle is actually simpler to represent in parameter space,
determine the straight roads on an aerial photo, this problem compared to the line, since the parameters of the circle can be
can be solved using Hough transform for lines. Often the directly transfer to the parameter space. The equation of a
objects of interest have other shapes than lines, it could be circle is
parables, circles or ellipses or any other arbitrary shape. The
general Hough transform can be used on any kind of shape, r2 = (x − a)2 + ( y − b)2 (3)
although the complexity of the transformation increases with As it can be seen the circle got three parameters, r, a and b.
the number of parameters needed to describing the shape. In Where a and b are the center of the circle in the x and y
the following we will look at the Circular Hough Transform direction respectively and where r is the radius. The

563
IJRITCC | February 2015, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 3 Issue: 2 563 - 566
_______________________________________________________________________________________________
parametric representation of the circle is same size as the parameter space. In this way, we sweep over
every edge point in the input image drawing circles with the
x = a + r cos(θ ) desired radii and incrementing the values in our accumulator.
(4) When every edge point and every desired radius is used, we
y = b + r sin(θ ) can turn our attention to the accumulator[9]. The accumulator
will now contain numbers corresponding to the number of
Thus the parameter space for a circle will belong to R3 where circles passing through the individual coordinates. Thus the
as the line only belonged to R2 . As the number of parameters highest numbers (selected in an intelligent way, in relation to
needed to describe the shape increases, as well as the the radius) correspond to the center of the circles in the
dimension of the parameter space R increases, so do the image[4][6].
complexity of the Hough transform[1]. Multiple circles with the same radius can be found with the
same technique. The center points are represented as red cells
Therefore, the Hough transform in general is only considered in the parameter space drawing. Overlap of circles can cause
for simple shapes with parameters belonging to R2 or at most spurious centers to be found, such as at the blue cell[7]. The
R3. In order to simplify the parametric representation of the circles can be removed by matching to circles in the original
circle, the radius can be held as a constant or limited to number image.
of known radii.

Fig. 1 The parameter space used for CHT Fig. 3 Each point in geometric space (left) generates a
circle in parameter space (right). The circles in parameter
The process of finding circles in an image using CHT is:First space intersect at the (a, b) that is the center in geometric
we find all edges in the image. This step has nothing to do space[10]
with Hough Transform and any edge detection technique of
your desire can be used. It could be Canny, Sobel or
Morphological operations[3][2]. IV. CIRCLE DETECTION ALGORITHM
Circular Hough Transform function detects multiple circles
in an image. The image contains holes or circles or disks
whose centers can be in or out of the image.
Syntax:
circles = houghcircle(im, minR, maxR); or
circles = houghcircle(im, minR, maxR, ratio); or
circles = houghcircle(im, minR, maxR, ratio, delta);

Inputs:
im: input image (gray or color)
Fig. 2 Circular HT from the x,y-space (left) to the parameter
minR: minimal radius (unit: pixels)
space (right), for a constant radius[2]
maxR: maximal radius (unit: pixels)
ratio: minimal number of detected edge pixels of a circle to
At each edge point we draw a circle with center in the point
the circle perimeter (0<ratio<1, default:0.3)
with the desired radius. This circle is drawn in the parameter
delta: the maximal difference between two circles for them
space, such that our x axis is the a-value and the y axis is the b
to be considered as the same one (default: 12); e.g.,
value while the z axis is the radii. At the coordinates, which
circle1=[x1 y1 R1],circle2=[x2 y2 R2], delta = |x1-x2|+|y1-
belong to the perimeter of the drawn circle, we increment the
y2|+|R1-R2|.
value in our accumulator matrix, which essentially has the
Output:
564
IJRITCC | February 2015, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 3 Issue: 2 563 - 566
_______________________________________________________________________________________________
circles: n-by-4 array of n circles; each circle is represented by
[cx cy R c], where (cx cy), R, and c are the center coordinate,
radius, and pixel count, respectively.
Steps:
1.First of all we check the input parameter.
2.If the number of input is ‘4’, in that case , delta=12, because
Each element in [cx cy R] may deviate 4 pixels approx.
3.If the number of input is ‘3’, then ratio=0.3 ,which is the 1/3
of the perimeter and delta=12.
4.If the number of input is not ‘5’ ,at that time program will
show a message that” Require at least 3 input arguments”
5.If minR<0 or maxR<0 or minR>maxR or ratio<0 or ratio>1
Figure 4: Input Image
or delta<0, the it will show this message
'Required: 0<minR, 0<maxR, minR<=maxR, 0<ratio<1, and
A threshold value was defined to represent the boundary of
0<delta'
the circle pattern by applying edge detection technique. Then,
the maximum values from the layer were compared with the
• Turn the colour image into gray image.
threshold to identify for the circle pattern.
• Create a 3D Hough array. The first two dimensions
specify the coordinates of the circle centers (x,y), and the
third specifies the radii(r).
• Detect edge pixels using Sobel edge detector. Reset
the lower and/or upper thresholds to balance between the
performance and detection quality.
• For an edge pixel (px,py), the locations of its
corresponding possible circle centers are within the area
(px-maxR:px+maxR, py-maxR:py+maxR).

• Create the grid [0:maxR2, 0:maxR2], and then

compute the distances between the center and all the grid
points to form a radius map (Rmap), followed by clearing Figure 5: Shows the segment of top view detection
out-of-range radii . Where : maxR2= 2 * max R.
• For each edge pixel, increment the corresponding By using circular hough transform implantation the
elements in the Hough array. detection of cirles are done. Figure 6 demonstrated circle
• Collect circles pattern depicting the peak value voted from the accumulator
Format: [cx cy R c] array, that the defect was the circle pattern.
Clear pixel count < 2*pi*R*ratio
• Delete similar circles.
• Draw circles on the original image.
• After drawing circle on image than using circular hough
transform(CHT) , the radius of circles is find out.

V. EXPERIMENTAL RESULTS
The following Figure 4 is used as a sample image which is
taken from camera, that image act as a real time image which
is stored in a temporary image. This image will detect and
measure parameter of WPP-44 by applying circular hough
transform technique..The sample images then were resized to Fig.6 : The implementation of Circular Hough Transform
365x352 pixels for a fast image processing response. Technique and detection of circles (bore)

Using circular hough transform the circles are detected of

water pump pulley and also the centre point of circle is
determined.
565
IJRITCC | February 2015, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 3 Issue: 2 563 - 566
_______________________________________________________________________________________________
VI. CONCLUSION
This study was to introduce an algorithm for circle pattern
detection and measurement of WPP-44 object based on the
Hough Transform method. The experiment demonstrated ous
a satisfying result. The developed algorithm was advantage
for the parameter measurement and the failure analysis of
WPP-44 production process, such as help to access problems
rapidly, decrease waste, and reduce the production cost of
WPP-44 industry.

REFERENCES
Figure7: Final Output [1] Daniel C. H. Schleicher and Bernhard G. Zagar "Image
Processing and Concentric Ellipse Fitting to Estimate the
To get clear image of circles and to calculate diameter of Ellipticity of Steel Coils" in intechopen Journal October 11,
2009.
object, histogram is applied to Figure7. By using histogram
[2] Rafael C. Gonzalez and Richard E. Woods. Digital Image
accuracy of parameter measurement is increased.
Processing. Prentice Hall, 2007. ISBN 0-13-168728-x.
[3] Carolyn Kimme, Dana Ballard, and Jack Sklansky. Finding
circles by an array of accumulators. Communication of the
ACM, Volume 18, Number 2, February 1975.
[4] Bryan S. Morse. Lecture 15: Segmentation (edge based,
hough transform). Brigham Young University: Lecture
Notes, 2000.
[5] Sirisak Liangwongsan, Boonraung Marungsri, Ratchadaporn
Oonsivilai, Anant Oonsivilai "Extracted Circle Hough
Transform and Circle Defect Detection Algorithm" World
Academy of Science, Engineering and Technology 60 2011.
[6] Shylaja S S,K N Balasubramanya Murthy, S Natarajan
Nischith,Muthuraj R,Ajay S.Feed Forward Neuarl Network
Base Eye Localization and Recognition Using Hough
Transform. IJACSA Vol. 2,No.3,pp104-109, March 2011
Figure8: Histogram of Final OutPut [7] Mohamed Rizon, Haniza Yazid "Object Detection using
Circular Hough Transform" in Am. J. Appl. Sci., 2 (12):
The experiment showed that the developed circle detection 1606-1609, 2005.
algorithm was able to correctly detect and measure the circles [8] Duda RO, Hart PE, “Use of the Hough Transf orm To Detect
of sample with maximum accuracy. Lines and Curves In Pictures”, In the Communications of
ACM, pp. 11–15, 1972.
TABLE I. EXPERIMENTAL RESULTS [9] Kimme, C., D. Ballard and J. Sklansky, “Finding Circles by
an Array of Accumulators”, In the Communications of AC
M, pp. 120–122
Sr. Name Standard Calculated Accuracy
No. Readings Readings in % [10] Mohamed Roushdy “Detecting Coins with Different Radii
based on Hough Transform in Noisy and Deformed Image”
1 Total Diameter 69.85mm 67.74mm 96.979
In the proceedings of GVIP Journal, Volume 7, Issue 1,
2 Outer 33.15mm 32.55 mm 98.19 April, 2007.
Diameter of
bore
3 Centre bore 15.86mm 14.88 mm 93.82
4 Thread Holes
I Hole1 7.98mm 7.1 mm 87.88
II Hole2 7.98mm 7.8 mm 97.74
III Hole3 7.98mm 7.85 mm 98.37
IV Hole4 7.98mm 7.7 mm 97.5

566
IJRITCC | February 2015, Available @ http://www.ijritcc.org
_______________________________________________________________________________________