Deakin University
Browse

File(s) under permanent embargo

Water leaks detection based on thermal images

conference contribution
posted on 2018-01-01, 00:00 authored by Cristiane Penteado, Yuri Olivatti, Guilherme Lopes, Paulo Rodrigues, Rodrigo Filev MaiaRodrigo Filev Maia, Plinio T Aquino
In this paper, a digital image processing methodology based on thermal images to detect and locate water leak in underground pipes will be presented. With the advent of smart cities, some countries began to face this problem with technologies, such as Smart Water Grids. However, with this proposal it is intended to contribute with the smart cities context but with a nondestructive and less complex system. Also, it is proposed in this paper the use of the q-sigmoid function that is an alternative to pre-processing step to digital image processing. The potential of this function will be validated as method to contrast enhancement and to highlight regions of interest. The proposal was applied in thermal images captured from the soil surface with an underground water leak. Such conditions were proposed into a laboratory, using an ideal model with sandy soil. The results obtained from this experiment were promising, since it was able to detect leaks at early stages of the experiment, which suggest a potential use of the methodology and the proposed preprocessing technique to detect water leaks.

History

Pagination

1-8

Location

Kansas City, Missouri

Start date

2018-09-16

End date

2018-09-19

ISBN-13

9781538659595

ISBN-10

153865959X

Language

eng

Publication classification

E1.1 Full written paper - refereed

Copyright notice

2018, IEEE

Editor/Contributor(s)

[Unknown]

Title of proceedings

ISC2 : 2018 IEEE International Smart Cities Conference

Event

IEEE International Smart Cities. Conference (4th : 2018 : Kansas City, Mo.)

Publisher

IEEE

Place of publication

Piscataway, N.J.

Usage metrics

    Research Publications

    Categories

    No categories selected

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC