Citation

Khow, Zu Jun and Tan, Yi Fei and Abdul Karim, Hezerul and Abdul Rashid, Hairul Azhar (2024) Comparative study of deep learning models for sewer pipes anomaly detection. In: MULTIMEDIA UNIVERSITY ENGINEERING CONFERENCE 2023 (MECON2023), 26–28 July 2023, Virtual Conference.

Full text not available from this repository.

Abstract

This research aims to evaluate the effectiveness of different existing vision-based deep-learning models for the inspection of sewer pipes. The inspection of the sewer pipes is critically essential because sewer damage can lead to several extreme situations which include but are not limited to property loss, environmental pollution, sewer line system collapse that will result in a flood, etc. On that account, a sewer inspection process is necessary. Most of the sewer inspections, currently, adopted by the people are still using human vision to evaluate the condition of a sewer pipe’s interior, which is easily subject to human error. Thus, this research proposed a vision-based Artificial lntelligence (Al) for inspecting the sewer pipes’ interior anomaly to assist the decision-making process as well as increase operational efficiency. The images used in this research are obtained from an open-source data set consisting of images depicting both defective and non­ defective sewer pipes interiors. A few deep learning models were trained by using ResNet, DenseNet, MobileNet and You Only Look Once (YOLO). The experiments showed that the accuracy of models is encouraging and promising.

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	Environmental pollution
Subjects:	T Technology > TD Environmental technology. Sanitary engineering
Divisions:	Faculty of Engineering (FOE)
Depositing User:	Ms Nurul Iqtiani Ahmad
Date Deposited:	03 Dec 2024 02:09
Last Modified:	03 Dec 2024 02:09
URII:	http://shdl.mmu.edu.my/id/eprint/13162

Downloads

Edit (login required)