ISSN 2347 - 3983

Ibthaj Nasr Asmaeil et al., International JournalVolume

of Emerging TrendsFebruary
10. No.2, 2022Research, 10(2), February 2022, 128 – 132
in Engineering

International Journal of Emerging Trends in Engineering Research

Available Online at http://www.warse.org/IJETER/static/pdf/file/ijeter171022022.pdf
https://doi.org/10.30534/ijeter/2022/171022022

Currency Recognition System Using Image Processing:

Libyan Banknote as a Case Study
Ibthaj Nasr Asmaeil1, A. Albagul2, Abdussalam Ali Ahmed3
College of Electronics, Bani Walid, Libya1
The Libyan Center for Engineering Research and Information Technology Bani Walid, Libya2
Mechanical and Industrial Engineering Department, Bani Waleed University, Bani Waleed/Libya 3

Received Date : January 05, 2022 Accepted Date : January 30, 2022 Published Date : February 07, 2022

the creation form database. Suggested approach as for the

ABSTRACT recognition of the Libyan currency, it is discussed in the third
part of this paper.
Banknote recognition in the field of image processing is used
to distinguish between banknotes of different countries and Table I: Libyan currencies
determine their values. One of the problems that emerged with
this process is the recognization of the banknote. The
recognition operations for Libyan currency denominations
comprise six stages: image capturing, preprocessing, target
region extraction, number segmentation, currency value
recognition, and output with screen displaying and voice. The
currency recognition system is implemented by comparing the
similarity of each banknote value template via image
processing steps in the Matlab environment. To prove the
effectiveness of the proposed system, the number of banknote
images with two types (new and medium) from each of the four
types of denominations were tried.

Key words: Currency recognition, Image processing, Libyan

currency, Matlab.

1. INTRODUCTION 2. LIBYAN CURRENCY SPECIFICATION

Recognition of a currency is generally known as a
classification Denomination banknote that mean value Features of the Libyan currency
Currency note from different countries. Rating and calculation There are many features in the Libyan currency; the Libyan
all banknotes on hand will be boring, long and error prone. Bank decided it. Figure 1. Give an idea about the features of
Therefore, it is an automated and reliable currency the currency denomination 20 dinars.
Recognition and counting are good options for transactions
Handle different types of tasks such as selling merchandise to
the public Transportation, gas stations, banks, etc. Until now,
there Many works have been developed to identify the single
currency (currency for a specific country) for example
Bangladeshi currency [1], Mexican currency [2], US dollar
[3], Indian currency [4] Although a lot of research is working
to identify currencies For different countries, there is little
work done on paper Getting to know the Libyan currency The
benefit of this work is the design and development of a
framework Four types of. Recognizable Libyan currency
denominations that are currently circulating in everyday life
(5, 10, 20, 50). The first table shows the four types of The Figure 1: Security Features 20 Dinar.
Libyan banknotes used in this work. In addition to organizing
the rest, the sections are as follows: The second section is for
128
Ibthaj Nasr Asmaeil et al., International Journal of Emerging Trends in Engineering Research, 10(2), February 2022, 128 – 132

The Read Serial Prefix and Artistic impression

The height and width of the currency Another parameter that we can use is reading text and words
Currency notes have distinct widths calculation of diagonal which is unique to each denomination of currency note. Serial
values (See the following formula) Due to the high circulation Prefix (4) and text which describes the Artistic impression (2)
of currency notes, boundaries of notes are folded or damaged on currency image (Figure 1) are some of the unique
easily and the size of the note will get change. Hence, characters that can be used.
considering only the width and height of the notes, are not
enough to identify currency notes 3. PROPOSED SYSTEM
The system proposed here works here on the image of the
(1) R : the diagonal of the paper currency obtained by the scanner. The algorithm
currency implemented here is as follows
w : the width of the currency 1. Get a picture of the paper currency from a scanner.
h : the height of the currency 2. The obtained image is an RGB image and is now converted
to a grayscale image.
Average the color values of the currency 3. Edge detection of full gray scale image.
Each pixel has a red component, a green component, and a 4. The features of the paper currency properties will be
blue component called RGB value so, given an image, we can cropped and split.
calculate the average value of red, blue and green according to 5. After the split, the characteristics of the banknotes are
the formula that gives values very close to each denomination extracted.
of banknotes. The color contrast on the note of the captured 6. The intensity of each feature is calculated.
banknote would give fair accurate results for new banknotes. 7. If the condition is met, the currency denomination will be
Old currency notes, due to frequent circulation and excessive recognized.
use, the colors of some categories of notes tend to overlap each
other. Then use RGB only Values to distinguish or identify In this system, the characteristics of banknotes that are used by
observations will not be sufficient. the common people are used Differentiate between different
denominations of banknotes that were discussed in the
previous section Figure 3 given below describes flowchart for
currency Recognition.

The Blind Recognition (BR) currency

Above-mentioned features will not always give accurate
results for used currencies. Hence additional methods need to
be used to identify currency notes. One solution to identified
issue is to get the Blind Recognition value (Figure 2). Image
processing methods like Feature extraction and Edge
Detection have been used for this purpose. Similarly, several
image processing filters has been used without effecting the
basic features of an image

Figure 3: The architecture of proposed system.

1. Image Acquisition: Performing image acquisition in image

processing is always the first step in the workflow
sequence because, without an image, no processing is
possible. After the image has been obtained, various
methods of processing can be applied to the image to
perform the many different vision tasks. There are various
ways to acquire image such as with the help of camera or
scanner. Acquired image should retain all the features [5].
The following figure shows the image acquisition of
Libyan banknote from the denomination of fifty dinars.

Figure 2: Blind Recognition Circles

129
Ibthaj Nasr Asmaeil et al., International Journal of Emerging Trends in Engineering Research, 10(2), February 2022, 128 – 132

Figure 4: Input Image. Figure 6: Binary Image.

2. Preprocessing: serves as the front end for recognition 2.3. Noise filtering: Various noises can appear in digital
system, which transforms the coarse image data into a images during capturing and transmission due to
suitable format. In the preprocessing of the images, the environmental condition and the quality of
following process steps are needed: gray scale converting, sensors. Noise may affect segmentation and
image binarization, noise filtering, morphological pattern matching. Hence, we want to eliminate
operation, and black area extraction for all types of salt and pepper noise and we work with
currency images. non-linear filter known as median filter. They are
2.1. Gray scale conversion: Images from your camera also called order-statistics filter, because their
or scanner are usually in the form of 24-bit true response is based on the ordering or ranking of
color images. Therefore, the image is first the pixels contained within the mask. It
converted to grayscale and resized as shown in preserves edges while removing unwanted noise
figure 5 that will be processed more easily and [7]. In our system 5x5 size of mask is used.
efficiently for successive steps. 3. Edge Detection: Because of circulation, edges of the
currency notes are usually worn and torn when they get
old. Therefore, to overcome the problem of recognizing
noisy currency notes, need to apply a suitable
pre-processing step. Edge detection is a fundamental tool
in image processing and computer vision, particularly in
the areas of feature detection and feature extraction,
which aim at identifying points of a digital image. Edge
detection reflects sharp intensity changes in colors of the
image as shown in figure 7. There are many ways to
perform edge detection. Gradient and Laplacian are the
mostly used methods. Since there are lots of objects and
shapes on Bank notes, which are unique to each
denomination, the use of edge detection techniques will
be helpful to success the research in terms of accuracy.

Figure 5: Gray Scale Image.

2.2. Binarization: The grayscale image is further

processed into a binary image as illustrated in
figure 6 that will reduce storage space and fast
process speed. For this process, Otsu’s method
[6] is used. Otsu’s method finds the threshold
level of image and then the calculated threshold
that minimizes the variance within the classes,
i.e. black and white. Each black pixel is assigned
value to 0 and each white pixel is assigned value
to 1. Figure 7: Edge Detection Image.

130
Ibthaj Nasr Asmaeil et al., International Journal of Emerging Trends in Engineering Research, 10(2), February 2022, 128 – 132

4. Image Segmentation: Image segmentation such illustrated Table 2: Statistics for Accuracy and Reliability.
in the figure below is the process of partitioning a digital
image into multiple Segments (sets of pixels, also known
as super pixels). The goal of segmentation is to simplify
and/or change the representation of an image into
something that is more meaningful and easier to analyze.

According to the results in the table (Table 2), system has

correctly identified the blind recognition value of 10 out of 10
notes of each denomination, where the maximum success rate
has a value of 100%. The value of 8 out of 10 banknotes of
each denomination, where the minimum success rate is 80%.
According to the results in the table (Table 2), system has
correctly read the serial prefix of 12 out of 12notes of each
denomination, where the minimum success rate has a value of
100%.The value of 9 out of 10 banknotes of each
denomination, where the minimum success rate is 90%.
However, the success rate and rate are low. The main reason
Figure 8: Segmentation Image for this is being traded the edges and borders of the coin are
easily bent, torn or frayed due to high circulation of the
5. Feature Extraction: Feature extraction is a special form of currency in the open. The reason is that there is writing on the
dimensional reduction. When the input data to an currency, which causes distortion and change of its shape.
algorithm is too large to be processed and it is suspected
to be very redundant then the input data will be 5. CONCLUSION
transformed into a reduced representation set of features. Currency denomination detection was implemented using
Transforming the input data into the set of features is image processing using MATLAB. Currency features
called feature extraction. If the features extracted are observation such as color value, size value for observation,
carefully chosen it is expected that the features set will blind recognition value and artistic impression were
extract the relevant information from the input data in extracted. The process is started from image acquisition to
order to perform the desired task using this reduced calculating the intensity of each extracted feature. The system
representation instead of the full size input [5]. can extract features even if the banknote has scribbles. The
processing algorithm here works suitable for the newly
4. RESULTS AND DISCUSSION introduced class 5, 10, 20 and 50. The system not only
In this section, the results are obtained after performing the determines the denomination of the banknote, but also gives
morphological image processing process. The results are the result in the form of an audio output.
displayed in a GUI made in MATLAB as shown in figure 9
that displays the extracted features (color value, size value for REFERENCES
observation, blind recognition value,and artistic impression.)
of a given currency note and those features extracted are used 1. Rafael C. Gonzalez, Richard E. Woods. “Digital
in an intelligent banknote recognition system. The accuracy Image Processing 2nd edition” Pearson Education,
and reliability of the system were measured by individually New York, 2001
considering the results obtained for serial prefix, blind 2. K. Debnath, S. Ahmed and Md. Shahjahan (2010). “A
recognition value and technical impression. paper currency recognition system using Negatively
Correlated Neural Network Ensemble”, Journal of
Multimedia.
3. W. Yan, J. Chambers and A. Garhwal (2015). “An
Empirical Approach for Currency Identification”,
Multimedia Tools and Applications.
4. M. Aoba, T. Kikuchi and Y. Takefuji (2003). “Euro
banknote recognition system using a three-layered
perceptron and RBF networks”. Transactions on
Mathematical Modeling and Its Applications.
5. M.Deborah and Soniya Prathap “Detection of Fake
currency using Image Processing”.
IJISETInternational Journal of Innovative Science,
Figure 9: Currency Recognition system via Matlab. Engineering & Technology, Vol. 1, Issue 10, 2014.
6. N. Otsu, “A Threshold Selection Method from
Gray-Level Histograms,” IEEE Transactions on

131
Ibthaj Nasr Asmaeil et al., International Journal of Emerging Trends in Engineering Research, 10(2), February 2022, 128 – 132

Sytem, Man, and Cybernetics,” 1979, Vol. 9, pp. 21. A. Douma, G. Şengül, F. G. Ibrahim Salem and A.
62-66. Ali Ahmed, "Applying the Histogram of Oriented
7. R. C. Gonzalez, R.E. Woods, “Digital Image Gradients to Recognize Arabic Letters," 2021 IEEE
Processing,” Addison Wesley, 1998. 1st International Maghreb Meeting of the Conference
8. F. Garcia-Lamont, J. Cervantes, A. Lopez (2012). on Sciences and Techniques of Automatic Control
“Recognition of Mexican Banknotes via their colour and Computer Engineering MI-STA, 2021, pp.
and texture features”, Expert Systems with 350-355,doi:
Applications. 10.1109/MI-STA52233.2021.9464499.
9. A. Ahmadi, S. Omatu, T. Kosaka and T. Fujinaka 22. V. Abburu, S. Gupta, S. R. Rimitha, M. Mulimani
(2004). “A reliable method for classification of bank and S. G. Koolagudi, "Currency recognition system
notes using artificial neural networks”. Artificial Life using image processing," 2017 Tenth International
and Robotics, Conference on Contemporary Computing (IC3),
10. T. Kagehiro, H. Nagayoshi and H. Sako (2006). “A 2017, pp. 1-6, doi: 10.1109/IC3.2017.8284300.
hierarchical classification method for US bank 23. G. V. Rajan, D. M. Panicker, N. E. Chacko, J. Mohan
notes”. Transactions on Information and Systems. and V. K. Kavitha, "An Extensive Study on Currency
11. L. Georgieva, T. Dimitrova, N. Angelov (2005). Recognition System Using Image Processing," 2018
“RGB and HSV Colour Models in Colour Conference on Emerging Devices and Smart Systems
Identification of Digital Traumas Images”, (ICEDSS), 2018, pp. 228-230, doi:
International Conference on Computer Systems and 10.1109/ICEDSS.2018.8544310.
Technologies. 24. K. Gautam, "Indian Currency Detection using Image
12. Y. Liu, L. Zheng and Y. Huang (2014). “Haar-SVM Recognition Technique," 2020 International
for Real-Time Banknotes Recognition”. Journal of Conference on Computer Science, Engineering and
Information and Computational Sciences. Applications (ICCSEA), 2020, pp. 1-5, doi:
13. Litjens G, Sánchez CI, Timofeeva N, Hermsen M, 10.1109/ICCSEA49143.2020.9132955.
Nagtegaal I, Kovacs I, et al.” Deep learning as a tool 25. J. Guo, Y. Zhao and A. Cai, "A reliable method for
for increased accuracy and efficiency of paper currency recognition based on LBP," 2010 2nd
histopathological diagnosis”. Sci Rep. (2016). IEEE InternationalConference on Network
14. F. G. Ibrahim Salem, R. Hassanpour, A. A. Ahmed Infrastructure and Digital Content, 2010, pp.
and A. Douma, "Detection of Suspicious Activities of 359-363, doi: 10.1109/ICNIDC.2010.5657978.
Human from Surveillance Videos," 2021 IEEE 1st 26. J. Akter, M. K. Hossen and M. S. A. Chowdhury,
International Maghreb Meeting of the Conference on "Bangladeshi Paper Currency Recognition System
Sciences and Techniques of Automatic Control and Using Supervised Learning," 2018 International
Computer Engineering MI-STA, 2021, pp. 794-801, Conference on Computer, Communication,
doi: 10.1109/MI-STA52233.2021.9464477. Chemical, Material and Electronic Engineering
15. Goodfellow I, Bengio Y, Courville A.” Deep (IC4ME2), 2018, pp. 1-4, doi:
Learning. Cambridge”, MS: MITPress (2016). 10.1109/IC4ME2.2018.8465595.
16. Fernandes AFA, Dórea JRR, Fitzgerald R, Herring
W, Rosa GJM. “A novelautomated system to acquire
biometric and morphological measurements and
predict body weight of pigs via 3D computer vision. J
Anim Sci”. (2019)
17. Boyle WS, Smith GE. “Buried channel charge
coupled devices”. [(accessed May 17, 2019].
18. A. Douma, A. A. Ahmed, G. Şengül, J. Santhosh, O.
S. M. Jomah and F. G. Ibrahim Salem, "A
comparison of Pattern Recognition Approaches for
Recognizing Handwriting in Arabic Letters," 2021
IEEE 1st International Maghreb Meeting of the
Conference on Sciences and Techniques of
Automatic Control and Computer Engineering
MI-STA, 2021, pp. 818-824, doi:
10.1109/MI-STA52233.2021.9464483.
19. Lister M. (1995).” The photographic image in digital
culture”. London: Routledge.
20. Sussman, D., L., Tang, M., Priebe, C., E.,“Consistent
Latent Position Estimation and Vertex Classification
for Random Dot Product Graphs”, IEEE transactions
on pattern analysis and machine intelligence, (2014).

132