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

net/publication/260350631

Agar Wood Grade Determination System Using Image Processing Technique

Article · June 2011

CITATIONS READS
6 732

9 authors, including:

Azma Abdullah Tuty Asmawaty Abdul Kadir

Universiti Malaysia Pahang Universiti Malaysia Pahang
2 PUBLICATIONS   6 CITATIONS    21 PUBLICATIONS   65 CITATIONS   

SEE PROFILE SEE PROFILE

Jasni Mohamad Zain

Universiti Teknologi MARA
160 PUBLICATIONS   1,053 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Android Device Security View project

Processing and Power Efficiency of Mobile Devices through Computational Offloading View project

All content following this page was uploaded by Azma Abdullah on 14 January 2016.

The user has requested enhancement of the downloaded file.

Proceedings of the International Conference on C-12
Electrical Engineering and Informatics
Institut Teknologi Bandung, Indonesia June 17-19, 2007

Agar Wood Grade Determination System Using Image Processing

Technique
Azma Abdullah*, Nik Kamariah Nik Ismail, Tuty Asmawaty Abdul Kadir,
Jasni Md Zain, Nor Amizam Jusoh, Noorlin Mohd Ali

Faculty of Computer Systems & Software Engineering, Universiti Malaysia Pahang

Locked Bag 12, 25000 Kuantan, Pahang Malaysia
Phone: (60)-09-5492119; Fax: (60)-09-5492144
Email: [email protected]; [email protected];
[email protected]; [email protected]; [email protected];

Abstract

The aim of this study is to develop a prototype of Agar wood grade determination system using image processing technique. Agar wood
becomes one of the valuable forest products compared to others because of its quality that makes the wood is high in demand. It is also
famous because of its special aroma and used for many purposes such as ingredients in perfume, medical and others. The raw Agar wood is
sold in either as large pieces of block chips or powder form. Manually, the quality and price of Agar wood are determined by its shape, size,
color and weight. There is no such modern way in determining the grade of Agar wood except by using the traditional or manual methods
above. As a result, the process of Agar wood’s grade determination consumes time and performed ineffectively. Besides, many experts and
experienced employees are required to determine the grade of Agar wood. As a solution, we need a system that is able to detect the
percentage of the density of black color by using grayscale images of Agar wood as well as determining the grade based on the percentage.
The objective of this paper is to determine the appropriate image processing technique that is able to perform the detection and
determination tasks. The detail outcome and result from the experiments of the technique are discussed in this paper.

Introduction Species of Aquilaria and Gyrinops, the two important

gaharu producing genera are naturally distributed in at
Internationally, agarwood, otherwise known as eaglewood, least 12 countries: Bangladesh, Bhutan, Cambodia, India,
aloeswood or gaharu (the Indonesian name frequently Indonesia, Lao PRD, Malaysia, Myanmar, Philippines,
used by villagers in Papua New Guinea), is known for its Thailand, Vietnam and Papua New Guinea (4).
fragrant resinous wood (1) .
The agar wood can be sold in different forms. The forms
Agar wood has three principal uses: medicine, perfume of agar wood in trade include large pieces of tree-trunks
and incense. Agar wood has been used for medicinal and branches, chips and flakes of uniform quality, powder
purposes for thousands of years, and continues to be used and oil (5).
in Ayurvedic, Tibetan and traditional East Asian medicine.
The use of agar wood for perfumery extends back several The purpose of this project is to advance understanding
thousands of years, and is referenced, for example, in the about agar wood and the batter solution or technique to
Old Testament several times using the term ‘aloes’. Both detect grade agar wood using image processing with fast
agar wood smoke and oil are customarily used as perfume and effective. It is because agar wood has high value in
in the Middle East (2). industry and all the process to define grade in agar wood
still using manually method.
Agarwood incense is burned to produce a pleasant aroma,
its use ranging from a general perfume to an element of Problem Statement
important religious occasions. Irregular chunks of
agarwood, usually a few centimetres long and weighing
10-200 g, may be cut or broken into smaller pieces and Traditionally, gaharu industry has uses human expert to
then burned, usually in a specially made incense burner (3). identify the gaharu's grade. However, by only using
expert naked eyes to classify it have leads to several
Agarwood refers to species within four genera; Gyrinops, problems such as inconsistence grading results, only
Aetoxylon, Gongystylis and more commonly, Aquilaria worker who very experiences are able to classify it and
within the family Thymelaeaceae. The process of olio- the process of grading is very depending on an experts'
resin production is the tree's response to injury if its first opinion.
line of defence, formation of phloem callus tissue, is
inhibited from forming over the injury (1). Beside that, the process that depending human expert are
______________________________________________ always have a problem especially in managing resources.
* Responsible author. E-mail: [email protected] For example, if a worker is retired, company needs to
make another investment to train a new expert in
classifying gaharu's grade. Moreover, traditional method

ISBN 978-979-16338-0-2 427

Proceedings of the International Conference on C-12
Electrical Engineering and Informatics
Institut Teknologi Bandung, Indonesia June 17-19, 2007

need a huge number of worker and time to made a process Where xi, i=1, 2…n with n is assumed as odd number that
done. produced by n = 2n + 1. Then, the arrangement of pixels
value and the neighbor for x is like the equation (1).
Approach and Methods
X 1 < X 2 < .... < X n (1).
Figure 1 shows the methodology in this research. There
are four phases involved; image acquisition, pre-
The median value is determined by the middle value. The
processing, segmentation and feature analysis.
main function of median filtering is to ensure that nearby
Image Acquisition
pixels have the value that is almost the same and no big
different between them (6).

Pre-processing
Image Threshold
Threshold is used to perform the segmentation process. In
Segmentation this phase Otsu’s method is used in order to enhance the
intensities of foreground and background.
Feature Analysis
Figure 1: Methodology for Agar Wood • Segmentation
Basically the segmentation process is done to find regions
that represent object or meaningful parts of object. In this
• Image Acquisition
phase clustering technique by row and column is used.
In this phase, 240 x 240 images of Agar Wood are
obtained using 7.2 mega pixels Sony digital camera with
4X optical zooming capability. About 74 images have
been captured for several categories and saved in JPEG Start Reading
numbers for each pixel value, total(m) = 0;
format. This image was categories into Grade A, B, C, D, for value of pixel not more than 150
and E. Figure 2 shows examples of the images. for each row & column
check the values of pixel
if threshold == 0
total(m)=total(m) + 1
else no calculation for total(p) will done
End Reading
Figure 3: Algorithm Applied

Figure 3 shows the algorithm applied in this technique.

Grade A Grade C This algorithm will calculate the numbers of pixels for
each grey scale value (from 0 to 150). This process
continued with categorizing each grey scale value into 5
ranges as it refers to the grade of Agar Wood (refer to
table 1).

Table 1: Range of Agar Wood

Ranges of grey Agar wood
Grade D Grade E scale value Gred
Figure 2: Images of Agar Wood 0-25 A
26-31 B
• Pre - Processing 32-44 C
Pre-processing phase used to improve an image’s contrast 45-60 D
and brightness characteristic, reduce its noise content, or 61-150 E
sharpen its details. There are two (2) processes of image
enhancement which are image filtering and image
threshold. Before accomplish these processes, color image • Features Analysis
has been converted into grey scale level. Feature analysis produced after the segmentation process
to detect grade of Agar Wood. Percentage of grey scale
Image Filtering value range will calculated during this phase. The higher
After testing several types of filtering techniques, percentage of grey scale value range, the better grade of
researchers find the most suitable technique to filter an Agar Wood.
image is using median filtering. This technique produces
by the sequence of arranged data to get the middle value.

ISBN 978-979-16338-0-2 428

Proceedings of the International Conference on C-12
Electrical Engineering and Informatics
Institut Teknologi Bandung, Indonesia June 17-19, 2007

Results References

In evaluating the proposed grade determination system for (1). B. Gunn, P. Stevens, M. Singadan, L. Sunari and P. Chatterton.
(2004). Eaglewood in Papua New Guinea. Resource Management
agar wood, 74 images of agar wood were tested to obtain in Asia-Pacific Program, Paper No. 51. Canberra, Australia.
the grade. The grade of agar wood was identifying by (2). Chakrabarty, K., Kumar, A. and Menon, V. (1994). Trade in
expect using manually technique. Table 2 shows the result Agarwood. TRAFFIC India and WWF-India, New Delhi. 51pp.
from expect. (3). Heuveling van Beek, H. and Phillips, D. (1999). Agarwood: Trade
and CITES Implementation in Southeast Asia. Unpublished report
prepared for TRAFFIC Southeast Asia, Malaysia.
Table 2: Grade of Agar Wood (4). Barden, A., Noorainie Awang Anak, Muliken, T., and Song, M.
Grade Total Images 2001. Heart of the Matter: Gaharu and Trade and CITES
A 4 implementation for Aquilaria malaccensis. TRAFFIC
(5). Zich, F. and Compton, J. (2002): The Final Frontier: Towards
B 0 Sustainable Management of Papua New. Guinea’s Agarwood
C 20 Resource.
D 20 (6). Zalili Musa (2003) Sistem Pengesanan Kerosakan Pada Jalinan
E 30 Tekstill, 1st Edition, Skudai Johor.

After the images testing to the prototype, the result is

show in table 3. The grade of agar wood has been
dividing into five (5) ranges of pixel. The ranges are 0-25
(grade A), 26-31 (grade B), 32-44 (grade C), 45-63 (grade
D) and 64-150 (grade E).

Table 3: Overall Result Data

Grade Range
0-25 26-31 32-44 45-60 61-150
A 2 0 1 1 0
B 0 0 0 0 0
C 2 0 16 2 0
D 0 0 3 14 3
E 0 0 0 0 30

Based on result show in table 4, only grade A was 50%

below from the true of agar wood grade result. That
means, this prototype system is suitable for the lower
grade of Agar wood.

Grade A must be detected by expert naked eyes to classify

it. The grade of agar wood is determined by its color (the
darker it is, the more valuable) and intensity of its
fragrance (5).

Table 4: Accurate and In-accurate Data

Grade A B C D E
Accurate 2 0 16 14 30
Data
In 2 0 4 6 0
Accurate
Data
Percentage 50% 100% 80% 70% 100%

Conclusion

We proposed a prototype that can detect a grade of agar

wood. The purpose is to develop a prototype of Agar
wood grade determination system using image processing
technique. The experimental results demonstrate that the
precision of detection a grade of agar wood is close to
80%.

ISBN 978-979-16338-0-2 429

View publication stats