See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/250801437

Object Detection using Circular Hough Transform

Article in American Journal of Applied Sciences · December 2005

DOI: 10.3844/ajassp.2005.1606.1609

CITATIONS READS
113 4,457

5 authors, including:

Haniza Yazid Puteh Saad

Universiti Malaysia Perlis Universiti Malaysia Perlis
96 PUBLICATIONS 1,252 CITATIONS 76 PUBLICATIONS 1,241 CITATIONS

SEE PROFILE SEE PROFILE

ali yeon md shakaff

Universiti Malaysia Perlis
230 PUBLICATIONS 3,015 CITATIONS

SEE PROFILE

All content following this page was uploaded by Puteh Saad on 09 July 2014.

The user has requested enhancement of the downloaded file.

American Journal of Applied Sciences 2 (12): 1606-1609, 2005
ISSN 1546-9239
© 2005 Science Publications

Object Detection using Circular Hough Transform

1
Mohamed Rizon, 2Haniza Yazid, 2Puteh Saad, 2Ali Yeon Md Shakaff, 2Abdul Rahman Saad
3,4
Masanori Sugisaka, 1Sazali Yaacob, 5M.Rozailan Mamat and 1M.Karthigayan
1
School of Mechatronics Engineering, Kolej Universiti Kejuruteraan Utara Malaysia
Jalan Kangar-Arau 02600 Jejawi, Perlis, Malaysia
2
School of Computer Engineering, Kolej Universiti Kejuruteraan Utara Malaysia
Jalan Kangar-Arau 02600 Jejawi, Perlis, Malaysia
3
Advanced Research Institute for Science and Enginering, Waseda University
3-4-1 Okubo, Shinjuku, Tokyo, 169-8555 Japan
4
Department of Electrical and Electronics Engineering, Faculty of Engineering
Oita University, 700 Dannoharu, Oita-shi, 870-1192 Japan
5
Terengganu Advanced Technical Institute (TATI), Jalan Panchor, Teluk Kalong
24000 Kemaman, Terengganu, Malaysia

Abstract: In this study we propose a new system to detect the object from an input image. The
proposed system first uses the separability filter proposed by Fukui and Yamaguchi (Trans. IEICE
Japan J80-D-II. 8, 2170-2177, 1997) to obtain the best object candidates and next, the system uses the
circular Hough transform (CHT) to detect the presence of circular shape. The main contribution of this
work consists of using together two different techniques in order to take advantages from the
peculiarity of each of them. As the results of the experiments, the object detection rate of the proposed
system was 96% for 25 images by moving the circle template every 20 pixels to right and down.

Key words: Separability filter, circular hough transform, edge detection, mathematical morphology

INTRODUCTION best object candidates and second, the CHT is

performed to detect the presence of circular shape. The
Object detection and recognition in noisy and separability filter was introduced by Fukui et al.[1] and
cluttered images is challenging problem in computer had been used in iris detection. Meanwhile the CHT is a
vision. The goal of this research is to identify objects in kind of Hough transform (HT) that can extract circular
an image (here we consider coconuts) by using two objects from an image. The CHT had been used in
techniques, separability filter and the CHT. Today, several researches in detecting fingertips position,
there is an increase need of coconuts. Due to the automatic ball recognition and iris detection for face
necessities of the coconuts, new method had to be recognition[2,3].
explored to assist the coconut gripping process. It will
help to pluck the coconut from a tree using image Proposed system: The image with 320x240 pixels has
processing techniques and it will be faster, easier and been used through out this study. PGM (portable gray
convenient than the manual plucking. There are several map) has been used as the input to the proposed
problems in detecting and recognizing the coconut in recognition process. Figure 1 shows the block diagram
the image. First, the target object is obscured due to the of the recognition process.
presence of the other object which can interfere with
recognition process such as the palm leaves. Second, Input Preprocessing Image
some of the objects are overlapping between each other segmentation
that make the recognition process challenging. Third,
the various object positions and finally, the images
itself contain noise that make the recognition process
Result
difficult without proper preprocessing and segmentation
process. Circular Hough Separability
The main contribution of this work consists of Transform filter
using together two different techniques in order to take
advantages from the peculiarity of each of them: first
the separability filter technique is used to obtain the Fig. 1: The coconut recognition process
Corresponding Author: Mohamed Rizon, School of Mechatronic Engineering, Kolej Universiti Kejuruteraan Utara Malaysia,
Jalan Kangar-Arau, 02600 Jejawi, Perlis, Malaysia, Tel: +60-4-9798335, Fax: +60-4-9798334
1606
 Am. J. Appl. Sci., 2 (12): 1606-1609, 2005

The input image has to come across several steps Separability filter: The separability filter is a template-
before the proposed technique is performed. The image based method. The proposed technique will place the
had to be enhanced using histogram equalization in template of Fig. 2 at each point (xi,yi) and then
preprocessing step, edge detection and mathematical compute the separability between the two regions R1
morphology in segmentation step. Histogram and R2 in the template with size r.
equalization tends to increase the contrast of the image
and produced a better result for region-based feature
extraction. Then, the image segmentation is carried out. 1/4r
The edge detection and mathematical morphology had
been employed to perform segmentation. The edge
detection process is very important as the edge
information is required for the CHT technique. Various r
edge detection methods have been applied for different
application. Among them, Canny edge detector has
been employed to the image. Canny gives thin edge
compared to Sobel[4]. Beside that, the opening and
closing morphology are employed to minimize the
points to apply the template matching. The closing
tends to close small gaps and removes much of the
white pixel noise, giving a fairly clean image. Opening Fig. 2: The circle template to detect the object
on the other hand tends to open small gaps or spaces candidate
between touching objects in an image. The closing has
Equation 3 is used to measure the separability
been applied then followed by the opening. The closing
between the two regions.
‘close’ small gaps and make the region is clearly seen
N
even though it consists of noise. Then, the opening took
( )
2
place. The opening will eliminate some of the unwanted A= I ( xi , yi ) − Pm (1)
information. i =1

( ) ( )
After the segmentation process, both separability 2 2
B = n1 P1 − Pm + n2 P2 − Pm (2)
filter and CHT techniques are applied. The process of
separability filter and the CHT technique are described B
later.
η= (3)
A
The CHT is applied to image after the separability where
technique had been employed. When using the CHT, Nk (k=1,2) : number of pixels in Rk ;
the radius r is known in advanced because of the N=n1+n2;
object’s shape is peculiar feature by each other, so the
Pk (k=1,2) : the average intensity in Rk ;
radius r can be treated as known parameter. In this
work, the radii for the object largeness have been set to Pm : the average intensity in the union of R1 and R2 ;
30, 40 and 50 pixels. I ( xi , yi ) : the intensity values of pixels ( xi , yi ) in the
In this experiment, the object of interest is the
coconut throughout this research. Coconut is known in union of R1 and R2 .
scientific as coco nucifera and is a member of Family The circles that give the local maxima of the
Arecaceae (palm family). Coconut is an important plant separability η are selected as object candidates. The
in the lives and economies of people in South East Asia points (x i i obtained after the opening morphology in
,y )
like Burma, Indonesia, Philippines and Malaysia. In the segmentation process is used as the input to the
Malaysia, coconut is planted either for personal use or separability filter. The range of radius for the R2 region
commercialize. There are varieties of commercial is set to rL, rL+10, …, rU . The template of Fig. 2 will
coconut that had been planted in Malaysia such as move every 20 pixels from left to the right and 20
Malayan Tall (MT), Malayan yellow Dwarf (MYD), pixels from top to down as shown in Fig. 3.
Malayan Green Dwarf (MGD), Rennel Tall, Malayan
Red Dwarf (MRD) and Pandan. Coconut provides Circular hough transform (CHT): The Hough
almost all necessities of life like food, drink, oil, transform (HT) and several modified versions have
medicine, timber, thatch, mats, fuel and domestic been recognized as robust techniques for curve
utensils. For these, it has been called the “tree of detection. This method can detect object even polluted
heaven” and “tree of life”. A 40 year old palm typically by noise. The CHT was sketched by Duda et al.[5]. The
attains a height of 20-22m and an 80 year old palm may
CHT is one of the modified versions of the HT. The
attain a height of 35-40m. Due to the increasing usage
CHT aims to find circular patterns within an image. The
of coconut, new method had to be explored to assist the
coconut gripping process. CHT is used to transform a set of feature points in the
1607
 Am. J. Appl. Sci., 2 (12): 1606-1609, 2005

EXPERIMENTAL RESULTS

We made experiments using 25 images to evaluate

the performance of the proposed system. The lower
bound rL and the upper bound rU on the radius of object
in the input images were set to 30 and 50, respectively.
20 pixels right Table 1 shows the results of the experiments.
The highest success rate of the proposed system to
detect the object was 96% by moving the circle
template every 20 pixels to the right and 20 pixels down
20 pixels down
and the success rate was 80% even for the worst
Fig. 3: Template moving moving as shown in Table 1.
Figure 5 and 6 show examples of the images for
image space into a set of accumulated votes in a which the proposed system gave successful results.
parameter space. Then, for each feature point, votes are The error occurred when the captured image was far
accumulated in an accumulator array for all parameter away from the coconut palm. If the images ware taken
combinations. The array elements that contain the far, the coconut image tends to be small and difficult to
highest number of votes indicate the presence of the detect as shown in Fig. 7.
shape. A circle pattern is described by equation 4
Table 1: The success rates of each template moving
( x p − x0 ) + ( y p − y0 )
2 2
= r2 (4) No Template moving [pixel] Success rate (%)
1 10 88
where x0 and y0 are the coordinates of the center and r 2 15 80
is the radius of the circle. An example of conventional 3 20 96
CHT is shown in Fig. 4. 3 25 80
4 30 80

Fig. 4: The contribution of the edge points to the

accumulator space

The black circles indicate a set edge points within Fig. 5: Successful detection of single coconut
the image. Each edge point contributes a circle of radius
R to an output accumulator space indicated by the grey
circles. The output accumulator space has a peak where
these contributed circles overlap at the center of the
original circle. Modification to the CHT has been
widely implemented to either increase the detection rate
or reduce its computational complexity[6-8]. In this
work, the edge orientation information is used to
increase the CHT performance. This method was first
proposed by Kimme et al.[6]. The use of the edge
orientation information limits the possible positions of
the center for each edge point. Using this method, only
an arc needs to be plotted perpendicular to the edge
orientation at a distance R from the edge point.
Fig. 6: Successful detection of single coconut

1608
 Am. J. Appl. Sci., 2 (12): 1606-1609, 2005

ACKNOWLEDGEMENT

The authors wish to thank Kolej Universiti

Kejuruteraan Utara Malaysia for their supports
throughout this research.

REFERENCES

1. Fukui, K. and O. Yamaguchi, 1997. Facial feature

points extraction method based on combination of
shape extraction and pattern matching. Trans.
IEICE Japan J80-D-II. 8: 2170-2177.
2. D’Orazio, T., C.Guaragnella, M. Leo and A.
Fig. 7: Failure detection of coconut
Distante, 2004. A new algorithm for ball
CONCLUSION recognition using circle Hough transform and
neural classifier. Pattern Recognition, 37: 393-408.
In this research the coconut location can be 3. Kawaguchi, T., D. Hidaka and M. Rizon, 2000.
detected but some constraint had been exposed. The Detection of eyes from human faces by Hough
separability filter tends to detect the possible transform and separability filter. Proc. of IEEE
coordinates of the coconut. Then, using the CHT, the ICRP, 1: 49-52.
obtained coordinates that are not considered as circle 4. Yazid, H., M. Rizon, P Saad, A.Y.M. Shakaff, S.
will be eliminated. The main constraint is the size of the Yaacob, A.R.M. Saad and M. Sugisaka, 2005. An
coconut. Template is used to get the radius of the approach of coconuts detection using edge
coconut largeness. information. Proc. of Intl. Advanced Technology
We made experiments using 25 images to evaluate Cong., CD-ROM.
the performance of the proposed system. As the results 5. Duda, R.O. and P.E Hart, 1972. Use of the Hough
of the experiments, the highest success rate of the transformation to detect lines and curves in picture.
proposed system to detect object was 96% by moving Commun. ACM, pp: 11-15.
the circle template every 20 pixels to the right and 20
6. Kimme, C., D. Ballard and J. Sklansky, 1975.
pixels down and the success rate was 80% even for the
Finding circles by an array of accumulators. Proc.
worst moving.
ACM, 18: 120-122.
The problem arises if the coconut image is
captured far away and the image tends to be small. 7. Ioannou, D., W. Duda and F. Laine, 1999. Circle
Beside that, some of the coconuts are in a bunch and recognition through a 2D Hough Transform and
overlaps between each other. These make the detection radius histogramming. Image and Vision
method difficult in carrying out the task. In this Computing, 17: 15-26.
research, automatic coconut recognition can be used to 8. Guil, N. and E.L. Zapata, 1997. Lower order circle
assist the coconut gripping process. Future work in and ellipse hough transform. Pattern Recognition,
progress includes detection of several coconuts in a 30: 1729-1744.
bunch and overlapping between each other.

1609

View publication stats