International Conference on Smart Computing and Electronic Enterprise.
(ICSCEE2018) ©2018 IEEE
Analysis and Improvements on Current Pothole
Detection Techniques
Sumit Srivastava Ayush Sharma Harsh Balot
School of Computing & IT Graduate Student, IT Graduate Student, IT
Manipal Univeristy Jaipur Manipal University Jaipur Manipal University Jaipur
India India India
[email protected] [email protected] [email protected] Abstract—India recorded at least 4,80,652 accidents in 2016,
leading to 1,50,785 deaths which were caused by speeding,
overtaking, drunk driving, speed breakers and potholes. Potholes are
State wise reported deaths
not just structural failure in a road surface, here in India these days
is a major cause of road side causalities. A constant detection and 367
repair in proper time can not only result in ensure road surface RAJASTHAN 440
quality but can also save many lives. The proposal describes one such 481
road maintenance system which uses Basic Ultrasonic Sensors, ANDHRA PRADESH 497
Raspberry Pi and A Mobile Phone with Internet Capabilities which 597
BIHAR 659
are connected to an Internet-of-Things platform over the Internet. In
783
addition to providing a generic Internet-of-Things based platform, 1244
MADHYA PRADESH
the proposed solution brings objective real time data about the state 1410
of the roads in a particular region which can be sent to the local UTTAR PRADESH 3428
authorities to take further actions upon.
Keywords— pothole, cloud, IoT , ultrasonic 0 1000 2000 3000 4000
I. INTRODUCTION Fig. 2: The state wise data compiled by the ministry of road transport
Understanding the condition of paved surface is extremely From the above data we can see Pothole-ridden roads have
helpful for road users as a result with the supply of such data road taken 11,386 lives across India in the last four years, which
users will avoid or use caution of what's coming ahead. interprets into roughly seven deaths each day.
These dangerous road conditions are a distraction for all the
In addition, the information regarding the road condition may commuters, hence detection of these potholes plays a major role
be applied in higher cognitive process and strategic designing of in fixing them on time and can prevent many road accidents. For
the Govt. authorities. pothole detection many approaches are preexisting but each of
them has their own drawbacks. This paper analyses the related
work and proposes an efficient solution which may prove to be a
Road accident in India due to Potholes great help to the commuters.
In this paper the Related work and their shot comings is
4000 3416
3039 presented in Section 2, System requirements and Methodology
3000 2324 and Implementation of our proposed solution is discussed in
2067 Section 3, followed by the conclusion in Section 4.
2000
II. RELATED WORK AND ANALYSIS
1000 The existing research work in this domain of pothole detection
us divided into three categories on the bases of technologies that
0
are used Fig. 3 show the categorization of the previous work. In
2013 2014 2015 2016
this section we will analyze each of these approach and give their
Fig. 1: The data compiled by the ministry of road transport drawbacks.
Authorized licensed use limited to: BMS College of Engineering. Downloaded on October 28,2024 at 17:02:47 UTC from IEEE Xplore. Restrictions apply.
International Conference on Smart Computing and Electronic Enterprise. (ICSCEE2018) ©2018 IEEE
reduced the equipment’s cost but still failed to reduce the
computational power and error rate for analyzing the
potholes.
C. Vibration based detection:
Yu and Yu [9] were the first one to perform a real time analysis
on the detected pothole with the use of an accelerometer and an
oscilloscope connected with a laptop. Major advantage of this
system was its portability and its cost effectiveness. Another
approach that Erikson et al. [10] used was using a three axis
acceleration sensor and a GPS device deployed on laptop called
Pothole Patrol (P2) which was placed inside the car. P2 used the
gathered data and uploaded that to the central server. The major
Fig. 3: Approaches for pothole detection drawback of these vibration based methods was they are providing
wrong results and not able to distinguish between a pothole, a
A. Vision Based Detection: speed breaker or a manhole.
This research work was presented by Koch and Brilakis [1].
they proposed the use of 2D images and texture-comparison The novel approach provided in this paper tries its best to solve
implemented in MATLAB utilizing the Image Processing the issues marked in bold that are faced by the predecessors.
Toolbox to detect the pothole but the drawback was that they were Section 3 provides brief implementation of the system proposed in
dependent on still images or 1 frame of the image which was very paper.
difficult to get from a moving car and cannot predict more than
one pothole. Buza et al. [2] recently proposed an approach which
uses spatial clustering to detect rough surfaces and potholes on the III. METHODOLOGY AND IMPLEMENTATION
road. To overcome the problem of analyzing one single frame at a The presented approach provides a user friendly standalone
time a more refined approach was used by Koch et al. i.e. use of cloud connected mobile application which alerts drivers about the
video and vision tracking onto multiple number of image frames upcoming potholes in his route and also an affordable low cost
to detect magnitude of potholes in the path. Here the drawback device which can be installed on any vehicle for the collection of
was the use of costly equipment, video cannot be captured form a the data about the potholes. This data can be used by other
vehicle moving at high speeds and the computation time that is application users as well as further compiled and sent to authorities
required for model training. to take a fast track action against the pothole prone region.
B. 3D Reconstruction:
For 3D Reconstruction of the potholes on the road there are
three already existing approaches each having their respective
drawbacks.
1) Projecting lasers onto a surface and using the reflected
pulse to create an accurate model in digital form. Chang et al. [3]
used these lasers to project 3D Cloud Points (Grid Based
Processing) with their elevations, although the results were
accurate but the approach was costly to implement. To improve
upon this Li et al. [4] introduced a much more cost effective
solution which used infrared laser line projectors, a digital camera
and a multi-view coplanar scheme for calibrating the lasers to plot
more feature points in the cameras point of view.
2) Wang [5] and Hou et al. [6] instead of using laser
projection used stereo vison which used an array of 2 to 4 digital
cameras (according to the size of the road under observation)
which captured the series of 2D images that can be further Fig 4: Information flow in the presented approach
combined to form a 3D surface model with the help of geometric
modeling. The drawback was the high computation required in the Implementation:
3D reconstruction of road surface and the feature point detection. As discussed in Section 2 all the existing approaches are
3) Joubert et al. [7] and Moazzam et al. [8] further tried to having some issues. The approach prescribed has tried to
reduce the cost of the equipment by using Xbox Kinect sensor and solve these mentioned issues and provides a cost efficient
its high speed camera for pothole detection. Through Kinect they solution for the problem of automated pothole detection.
collected depth images and calculated the approximate volume
using trapezoidal rule on area depth curves that are generated
through the Kinect collect images. This approach greatly
Authorized licensed use limited to: BMS College of Engineering. Downloaded on October 28,2024 at 17:02:47 UTC from IEEE Xplore. Restrictions apply.
International Conference on Smart Computing and Electronic Enterprise. (ICSCEE2018) ©2018 IEEE
Instead of using a 3D reconstructed image which is the data value and store them on a new table. This trigger solves
computationally heavy and resource hungry this approach uses the error issue which was present in vibration based approach.
simple ultrasonic modules which are mounted onto a test vehicle
and can detect the distance changes between the road and the All the given previous approaches only did the job of
moving vehicle in real time. collecting the information about the potholes and building a
dataset[6,8], but none of the approaches presented us with an end
user application which can be beneficial for public transportation
and commuters. The prescribed approach also provides end user
solution by the use of a Multi-Platform (Apple® iOS® or
Google™ Android®) application which show a push notification
to the user as soon as the user enter the vicinity of nearby potholes.
Fig. 6 shows when the user starts the application and selects the
Fig 5: Working of Ultrasonic sensor module to detect potholes start and end points of trip, the data between those points is fetched
from the crowd sourced cloud based database. The application
The ultrasonic sensor does this by firing a high-frequency plots Point Markers for the potholes present on the Google Maps
sound pulse from one end and the other end receives the reflected and in background also checks for nearby potholes in user’s radius
echo, then the time taken between the two pulses can be calculated. of 500m. If user is in the range application shoots a push
Now we know the speed of sound is 343m/s and sensor provides notification, warning user about the nearby potholes.
time so form the equation:
Distance =Speed/Time In case of formation of sudden road caves which occur mostly
Distance is returned to the Raspberry Pi. If the value is greater after heavy rainfall user can manually click a photo of the pit
than (Pothole/Manhole) or less than (Speed Breaker) a particular formed from the mobile app and notify other app users. This
limit (Distance between the car and the ground on a plane surface process is also controlled by crowd sourcing if many reports form
a single vicinity is captured then only nearby users are notified of
the pothole ahead.
Furthermore, the crowd source data with the use of data
visualization techniques can be presented to National Highway
Authorities, State Transport Authorities, and Public Works
Departments for the better up keeping of the roads to avoid motor
vehicle collisions and traffic congestions.
IV. CONCLUSION
In this paper the authors explorers and give analysis on various
pre exiting approaches for detecting the pavement potholes. The
authors also suggest the issues in the predecessors work [3, 4, 6,
8, 9] and tries to give a novel approach with the assumption to
eliminate the high equipment cost, high computation power and
error in the data collection. Authors propose a demo of a very low
cost embedded system, installed on public transports for crowd
sourcing the relevant pothole data along with the location
coordinate. The demo also includes a multi-platform (Apple®
iOS® or Google™ Android®) easy to operate user friendly
application (with off-the-shelf open source packages further
reducing backend cost) which uses this crowd sourced data to
warn user about the nearby potholes.
Although we can conclude that a simple model presented may be
sufficient to estimate the road quality and avoiding traffic
congestion and collisions, but there are still many issues that are
Fig. 6: End-user Android application work flow
to be dealt with in our future works to make the approach more
efficient.
Which can be conFig.d at the run time of the application) then Pi
sends a message to the android device which in turn replies the
In future, we envisage to improve and extend this approach to
GPS location coordinates of the detected ditch to the Pi Board.
accommodate with simulation results and with information about
Then location coordinates and the depth measurement of the
the quality of road by detecting the vibration data we get from
pothole are then send to an Amazon Relational Database Service
accelerometer sensor module.
instance. For the accuracy of data, a trigger which is fired in every
24 hours is created on the RDS instance which takes the mean of
Authorized licensed use limited to: BMS College of Engineering. Downloaded on October 28,2024 at 17:02:47 UTC from IEEE Xplore. Restrictions apply.
International Conference on Smart Computing and Electronic Enterprise. (ICSCEE2018) ©2018 IEEE
REFERENCES [11] Seung-Ki Ryu, Taehyeong Kim: Review and Analysis of Pothole Detection
Methods, Journal of Emerging Trends in Computing and Information
[1] Christian Koch, Ioannis Brilakis:Advanced Engineering Informatics archive Sciences, Vol. 5, No. 8 August 2014
Volume 25 Issue 3, August 2011, Pages 507-515
[2] E. Buza S. Omanovic and A. Huseinovic: Pothole Detection with Image
Processing and Spectral Clustering. In 2nd International Conference on
Information Technology and Computer Networks, Pages 48–53, 2013.
[3] K. T. Chang, J. R. Chang and J. K. Liu: Detection of Pavement Distresses
Using 3D Laser Scanning Technology, International Conference on
Computing in Civil Engineering 2005
[4] Li, Q., Yao, M., Yao, X and Xu, B. (2009): A real-time 3D Scanning System
for pavement distortion inspection, measurement science and
technology,Pages 15702-15709.
[5] K. C. P. Wang: Challenges and feasibility for comprehensive automated
survey of pavement conditions, In 8th International Conference on
Applications of Advanced Technologies in Transportation Engineering
(2004), Pages 531-536
[6] Z. Hou, K. C. P. Wang, and W. Gong: Experimentation of 3D pavement
imaging through stereovision, In International Conference on Transportation
Engineering (2007), Pages 376-381
[7] D. Joubert, A. Tyatyantsi, J. Mphahlehle, and V.Manchidi: Pothole tagging
system, In 4th Robotics and Mechatronics Conference of South Africa(2011),
Pages 1-4
[8] I. Moazzam, K. Kamal, S. Mathavan, S. Usman, and M. Rahman: Metrology
and visualization of potholes using the Microsoft Kinect sensor, In 16th
International IEEE Annual Conference on Intelligent Transportation
Systems(2013), Pages 1284-1291
[9] B. X. Yu and X. Yu: “Vibration-based system for pavement condition
evaluation,” In 9th International Conference on Applications of Advanced
Technology in Transportation, August 2006, Pages 183–189.
[10] J. Eriksson, L. Girod, B. Hull, R. Newton, S. Madden, and H. Balakrishnan:
“The pothole patrol: using a mobile sensor network for road surface
monitoring”, In 6th International Conference on Mobile Systems,
Applications, and Services (MobiSys '08), June 2008, Pages 29–39
Authorized licensed use limited to: BMS College of Engineering. Downloaded on October 28,2024 at 17:02:47 UTC from IEEE Xplore. Restrictions apply.
