IOP Conference Series: Materials

Science and Engineering

PAPER • OPEN ACCESS You may also like

- Fusing complementary images for
A framework for pavement crack detection and pavement cracking measurements
Ming Yao, Zuyun Zhao, Xun Yao et al.
classification - Measuring the impacts of a real-world
neighborhood-scale cool pavement
deployment on albedo and temperatures
To cite this article: A Habib and M M Kunt 2020 IOP Conf. Ser.: Mater. Sci. Eng. 800 012020 in Los Angeles
Joseph Ko, Hannah Schlaerth, Alexandra
Bruce et al.

- Crack identification for rigid pavements

using unmanned aerial vehicles
View the article online for updates and enhancements. Ahmet Bahaddin Ersoz, Onur Pekcan and
Turker Teke

This content was downloaded from IP address 194.27.73.86 on 14/05/2024 at 09:13

5th International Conference on New Advances in Civil Engineering (ICNACE 2019) IOP Publishing
IOP Conf. Series: Materials Science and Engineering 800 (2020) 012020 doi:10.1088/1757-899X/800/1/012020

A framework for pavement crack detection and classification

A Habib*, M M Kunt
Civil Engineering Department, Eastern Mediterranean University, Famagusta. North
Cyprus

*E-mail: [email protected]

Abstract. Pavement damage detection is indeed a very important process for the management
of roads. Nowadays, scholars are focusing on finding a simple and accurate way to detect road
cracks aiming to increase its life span and improve its safety and quality. However, due to some
factors including cost of implementation and level of experiences required small communities
and developing countries are yet unable to adapt this idea. Therefore, developing a feasible
method is becoming extremely important for improving the quality and safety of pavements.
Thus, this study is intended to propose a framework for pavement crack detection and
classification through the use of inexpensive set of sensors and the applications of artificial
neural networks. As part of the project, difficulties that are going to be faced by local agencies
in small communities will be highlighted and possible solutions will be suggested. In general,
this framework is expected to be helpful to government entities in developing plans and taking
actions toward providing a suitable pavement distress mitigation strategy.

1. Introduction
Deterioration of the pavement is the system through which distresses develop under the combined
impacts of road traffic and environmental factors. Serviceability, security and road quality are
significantly affected by deterioration of pavements. Because of its extensive use, roads deteriorate with
time and must, therefore, be maintained to ensure that safety, efficiency and durability requirements are
met. Common deterioration of pavement involves cracking, surface defects, deformation and structural
failure [1]. The most common types of cracking are fatigue cracking, longitudinal cracking, transverse
cracking, block cracking, slippage cracking, reflective cracking and edge cracking [2]. The combination
of fatigue and cracking block may be regarded as alligator cracking. These cracks are mainly caused by
high load applied or excess moisture. Therefore, it is important to continuously check the pavement
condition in order to provide a solid data for a pavement management system [3]. However, local
governments in small cities often lack experienced staff for effectively using a fully implemented
pavement management system [3]. Furthermore, available techniques for pavement crack detection are
considered to be very expensive especially for small countries such as the case of Turkish Republic of
North Cyprus (TRNC). For this reason, finding a way to reduce this cost and complexity is essentially
towards an effective management of pavements.
Thus, the aim of this paper is to propose a framework empowered by a low-cost tool that can be used
by small communities and to highlight the main difficulties and their solution that might be faced during
the implementing of this methods.

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
5th International Conference on New Advances in Civil Engineering (ICNACE 2019) IOP Publishing
IOP Conf. Series: Materials Science and Engineering 800 (2020) 012020 doi:10.1088/1757-899X/800/1/012020

2. Materials and methods

In this section, components of the framework for pavement damage detection empowered by a low-cost
tool is going to be discussed.

2.1. Data collection

In this study a data collection circuit will be developed based on the use of Raspberry Pi Zero, Figure 1,
as the main computer that works for data acquisition and data storing. An acceleration sensor MPU6050,
Figure 2, with a data sampling rate of 60 Hz and range of ±8g will be used for measuring the vibration
in the vertical direction. Furthermore, a Global Positioning System (GPS) GY-NEO6MV2 sensor,
Figure 3, will be used for defining the location of the crack. In general, the total cost of the circuit and
the sensors is considered low-cost system which allows small communities to get benefit of this method
even at small budgets.
In order to collect the data, the device must be mounted on any vehicle or bicycle. Thereafter, the
Raspberry Pi Zero will work as a data acquisition system through a python coded program that will read
the output of the sensors and will store it in the memory card.

Figure 1. Raspberry Pi Zero

Figure 2. Chip of the MPU 6050 sensor

Figure 3. Chip of the GY-NEO6MV2 GPS sensor

2.2. Data Pre-Processing

In fact, the recorded data must undergo a pre-processing stage to reduce their noise emerged from the
uncontrollable environment in the pavement. The suggested method in this study is called the moving
average filter which is composed of taking the average of a certain number of outputs and subtract that
value from the readings of the sensor [4]. In addition, to removing the surrounding noise the suggested
method will also remove the readings caused by the Earth’s gravity as the observed accelerations are in
vertical direction.

2
5th International Conference on New Advances in Civil Engineering (ICNACE 2019) IOP Publishing
IOP Conf. Series: Materials Science and Engineering 800 (2020) 012020 doi:10.1088/1757-899X/800/1/012020

2.3. Data processing

This stage includes both the crack detection and classification. At the beginning, the collected data
mainly the vibration one will be divided into number of cases in which each case will represent a sudden
change in the vibration input. This is going to be done by selecting all the values that cross a specified
acceleration limit such as 10 m/s2. Thereafter, 15 readings before and 15 readings after each of these
values will be defined as a crack that will be classified separately.
Classifying each crack is going to be undertaken through the use of Artificial Neural Network (ANN)
Figure 4 because of its capability to learn and capture the functional relationships among the hard
description of data [5]. In order to develop the ANN system part of the recorded data will be used to
train the system and rest will be to verify and test it.
In fact, the use of an accelerometer provides an easy way for identifying the vibration behavior of
each change in the pavement texture such as cracks or speed humps. Furthermore, using a reliable ANN
system can help in classifying each of the texture changes into the correct crack type. Table 1 shows
examples of different types of changes in the pavement surface. It is clear that each one has unique
characteristics. For example, the alligator crack can be defined by consecutive fluctuation of a similar
peak values. On the other hand, the pothole can be observed by a sudden change in the vibration behavior
that have a single peak value, while the speed hump is identified by a negative peak directly followed
by a high value of a positive peak and its long duration. Using the same procedure, it is possible to
specify the characteristics of each crack. These characteristics can be the duration of the fluctuation in
vibration, the standard deviation in of the vibrations, and the number of positive and negative peak
values.

Figure 4. An artificial neural network [6].

3
5th International Conference on New Advances in Civil Engineering (ICNACE 2019) IOP Publishing
IOP Conf. Series: Materials Science and Engineering 800 (2020) 012020 doi:10.1088/1757-899X/800/1/012020

Table 1. Vibration behavior of various pavement texture changes.

Type Photo Vibration Behavior

Alligator
Crack

Pothole
Crack

Speed
Humps

2.4. GIS Database

The importance of any GIS is its capability in data modelling, which is a set of constructs for describing
and representing selected aspects of the real world in computer [7]. It is a procedure of designing and
developing a data system by taking all the information that would be needed to support various business
processes of the organisation [8]. Data model is created to depict the structure of the data handled in
information systems that is demonstrated in entity-relationship diagrams (ERD). Three basic styles of
data modelling, Figure 5, are generally used in practice today [9]:
 Conceptual data models: the process of constructing a model of the data, creation of ERD, used
in organization and independent of all physical considerations.
 Logical data models: the method of representing data architecture and organization in a
graphical way, mapping ERD to tables, without any regard to the physical data storage device.
 Physical data models: the internal schema database design.

4
5th International Conference on New Advances in Civil Engineering (ICNACE 2019) IOP Publishing
IOP Conf. Series: Materials Science and Engineering 800 (2020) 012020 doi:10.1088/1757-899X/800/1/012020

In this stage, each crack will be defined by its type and location on a designed geodatabase that composed
of a digitized roads network such as the one shown in Figure 7.

Figure 5. Levels of abstraction in data modelling [10]

Figure 6. An example of a digital road network

3. Proposed Framework

3.1. Pavement Crack Detection and Classification Framework

As mentioned previously, this paper is proposing a framework that can be used for pavement crack
detection and classification.
The framework proposed in this study, as depicted in Figure 7, is composed of four main stages
which are data collection, their pre-processing and processing and storing them in a GIS database for
possible mitigation plans.

5
5th International Conference on New Advances in Civil Engineering (ICNACE 2019) IOP Publishing
IOP Conf. Series: Materials Science and Engineering 800 (2020) 012020 doi:10.1088/1757-899X/800/1/012020

Figure 7. General framework for pavement crack detection and classification

3.2. Challenges for Small Communities

In fact, the major difficulty that is expected to be faced by small communities is the lack of an accurate
GIS database. Previous study [11] have illustrated how a GIS map can be collected, edited, and verified
from other map sources such as locally maintained road ones. Furthermore, another challenge would be
the cost of continuous data collection which can be overcome by licensing local agencies who run
vehicle or bicycle renting services to work as a data provider for the government by introducing this
low-cost device to their business.

4. Conclusion
This study contributes to the pavement damage detection and classification in small communities with
the developed framework. As a part of this study, a proposed framework that can be implemented with
a low-cost tool mounted to a vehicle or a bike was discussed. Furthermore, some of the challenges that
are expected to be faced by these communities were mentioned and possible solutions were suggested.

References
[1] Aziz, M. M., Mustafa, K. H., Hainin, M. R., Zaini, A. A., & Warid, M. N. (2016). A Review on
Crack Relief Layer (crl) in Roads and Highways. Jurnal Teknologi, 45-52.
[2] Adlinge, S. S., & Gupta, A. K. (2013). Pavement deterioration and its causes. International Journal
of Innovative Research and Development, 2(4), 437-450.
[3] Kunt, M. M. (2008). The Benefits of Introducing a Pavement Management System in Small Cities:
The Case of the City of Famagusta. Washington D.C.: Transportation Research Board 87th
Annual Meeting.
[4] Smith, S. W. (1997). The scientist and engineer's guide to digital signal processing.

6
5th International Conference on New Advances in Civil Engineering (ICNACE 2019) IOP Publishing
IOP Conf. Series: Materials Science and Engineering 800 (2020) 012020 doi:10.1088/1757-899X/800/1/012020

[5] Gunawan, F. (2018, 11 1). Detecting road damages by using gyroscope sensor. ICIC Express
Letters, 12(11), 1089-1098.
[6] Daniel, G. (2013). Principles of artificial neural networks . World Scientific.
[7] Longley, P. A., Goodchild, M. F., Maguire, D. J., & Rhind, D. W. (2005). Geographic information
systems and science. John Wiley & Sons.
[8] Ponniah, P. (2007). Data modeling fundamentals: a practical guide for IT professionals. John Wiley
& Sons.
[9] Singh, S. P., & Singh, P. (2014). Modeling a geo-spatial database for managing travelers' demand.
International Journal of Database Management Systems.
[10] Akinyemi, F. O. (1999). Database design and development for census mapping in Nigeria.
Cartography in the Service of Government, Nigeria, Lagos: Nigerian Cartographic
Association, 191-202
[11] He, Z., Qin, X., Wang, H., & Comes, C. (2017). Implementing practical pavement management
systems for small communities: a south Dakota case study. Public Works Management &
Policy, 22(4), 378-391.