Deep Neural Networks Based Approach for Pothole Detection

Anup Pandey; Rahat Iqbal; Saad  Amin; Tomasz Maniak; Vasile Palade; Charalampos Karyotis

doi:10.1109/ICSPIS53734.2021.9652434

Deep Neural Networks Based Approach for Pothole Detection

Anup Pandey, Rahat Iqbal, Saad Amin, Tomasz Maniak, Vasile Palade, Charalampos Karyotis

Research Centre for Computational Science and Mathematical Modelling

Research output: Chapter in Book/Report/Conference proceeding › Chapter

9 Citations (Scopus)

Abstract

Potholes on the road cause inconvenience to commuters and delay in delivering products and services. The current pothole detection methods require manual inspection of roads using custom sensors installed on specially adapted vehicles. The procedure is time-consuming and labour intensive. Internet of Things (IoT) is an emerging technology that has the potential to provide an efficient and cost-effective solution to road pothole detection. This paper proposes a novel Convolution Neural Networks (CNN) based approach for pothole detection. The approach fuses imagery and sensory data to perform pothole detection. In the experimental studies performed, the proposed approach was able to achieve 87.20% precision, 92.7%recall and 89.9% F1-Score.

Original language	English
Title of host publication	2021 4th International Conference on Signal Processing and Information Security, ICSPIS 2021
Publisher	IEEE
Pages	1-4
Number of pages	4
ISBN (Electronic)	9781665437967
ISBN (Print)	978-1-6654-3797-4
DOIs	https://doi.org/10.1109/ICSPIS53734.2021.9652434
Publication status	Published - 27 Dec 2021
Event	4th International Conference on Signal Processing and Information Security - Dubai, United Arab Emirates Duration: 24 Nov 2021 → 25 Nov 2021

Conference

Conference	4th International Conference on Signal Processing and Information Security
Abbreviated title	ICSPIS
Country/Territory	United Arab Emirates
City	Dubai
Period	24/11/21 → 25/11/21

Bibliographical note

Publisher Copyright:
© 2021 IEEE.

Keywords

1D-CNN
2D-CNN
Accelerometer
Crowdsource Data
Convolution Neural Networks
IoT
Machine Learning
Pothole Detection

Access to Document

10.1109/ICSPIS53734.2021.9652434

Cite this

Pandey, A, Iqbal, R, Amin, S, Maniak, T, Palade, V & Karyotis, C 2021, Deep Neural Networks Based Approach for Pothole Detection. in 2021 4th International Conference on Signal Processing and Information Security, ICSPIS 2021. IEEE, pp. 1-4, 4th International Conference on Signal Processing and Information Security , Dubai, United Arab Emirates, 24/11/21. https://doi.org/10.1109/ICSPIS53734.2021.9652434

@inbook{655a754722bc4d56b80acb244e246487,

title = "Deep Neural Networks Based Approach for Pothole Detection",

abstract = "Potholes on the road cause inconvenience to commuters and delay in delivering products and services. The current pothole detection methods require manual inspection of roads using custom sensors installed on specially adapted vehicles. The procedure is time-consuming and labour intensive. Internet of Things (IoT) is an emerging technology that has the potential to provide an efficient and cost-effective solution to road pothole detection. This paper proposes a novel Convolution Neural Networks (CNN) based approach for pothole detection. The approach fuses imagery and sensory data to perform pothole detection. In the experimental studies performed, the proposed approach was able to achieve 87.20% precision, 92.7%recall and 89.9% F1-Score.",

keywords = "1D-CNN, 2D-CNN, Accelerometer, Crowdsource Data, Convolution Neural Networks, IoT, Machine Learning, Pothole Detection",

author = "Anup Pandey and Rahat Iqbal and Saad Amin and Tomasz Maniak and Vasile Palade and Charalampos Karyotis",

note = "Publisher Copyright: {\textcopyright} 2021 IEEE.; 4th International Conference on Signal Processing and Information Security , ICSPIS ; Conference date: 24-11-2021 Through 25-11-2021",

year = "2021",

month = dec,

day = "27",

doi = "10.1109/ICSPIS53734.2021.9652434",

language = "English",

isbn = "978-1-6654-3797-4",

pages = "1--4",

booktitle = "2021 4th International Conference on Signal Processing and Information Security, ICSPIS 2021",

publisher = "IEEE",

address = "United States",

}

TY - CHAP

T1 - Deep Neural Networks Based Approach for Pothole Detection

AU - Pandey, Anup

AU - Iqbal, Rahat

AU - Amin, Saad

AU - Maniak, Tomasz

AU - Palade, Vasile

AU - Karyotis, Charalampos

PY - 2021/12/27

Y1 - 2021/12/27

N2 - Potholes on the road cause inconvenience to commuters and delay in delivering products and services. The current pothole detection methods require manual inspection of roads using custom sensors installed on specially adapted vehicles. The procedure is time-consuming and labour intensive. Internet of Things (IoT) is an emerging technology that has the potential to provide an efficient and cost-effective solution to road pothole detection. This paper proposes a novel Convolution Neural Networks (CNN) based approach for pothole detection. The approach fuses imagery and sensory data to perform pothole detection. In the experimental studies performed, the proposed approach was able to achieve 87.20% precision, 92.7%recall and 89.9% F1-Score.

AB - Potholes on the road cause inconvenience to commuters and delay in delivering products and services. The current pothole detection methods require manual inspection of roads using custom sensors installed on specially adapted vehicles. The procedure is time-consuming and labour intensive. Internet of Things (IoT) is an emerging technology that has the potential to provide an efficient and cost-effective solution to road pothole detection. This paper proposes a novel Convolution Neural Networks (CNN) based approach for pothole detection. The approach fuses imagery and sensory data to perform pothole detection. In the experimental studies performed, the proposed approach was able to achieve 87.20% precision, 92.7%recall and 89.9% F1-Score.

KW - 1D-CNN

KW - 2D-CNN

KW - Accelerometer

KW - Crowdsource Data

KW - Convolution Neural Networks

KW - IoT

KW - Machine Learning

KW - Pothole Detection

UR - http://www.scopus.com/inward/record.url?scp=85124421431&partnerID=8YFLogxK

U2 - 10.1109/ICSPIS53734.2021.9652434

DO - 10.1109/ICSPIS53734.2021.9652434

M3 - Chapter

SN - 978-1-6654-3797-4

SP - 1

EP - 4

BT - 2021 4th International Conference on Signal Processing and Information Security, ICSPIS 2021

PB - IEEE

T2 - 4th International Conference on Signal Processing and Information Security

Y2 - 24 November 2021 through 25 November 2021

ER -

Deep Neural Networks Based Approach for Pothole Detection

Abstract

Conference

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this