Dalarna University's logo and link to the university's website

du.sePublications
Planned maintenance
A system upgrade is planned for 10/12-2024, at 12:00-13:00. During this time DiVA will be unavailable.
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • chicago-author-date
  • chicago-note-bibliography
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Automatic loose gravel condition detection using acoustic observations
Dalarna University, School of Information and Engineering.
Dalarna University, School of Information and Engineering.
2022 (English)Independent thesis Advanced level (degree of Master (One Year)), 10 credits / 15 HE creditsStudent thesis
Abstract [en]

Evaluation of the road's condition and state is essential for its upkeep, especially when discussing gravel roads, for the following reasons, among other. When loose gravel is not adequately maintained, it can pose a hazard to drivers, who can lose control of their vehicle and cause accidents. Current maintenance procedures are either laborious or time-consuming. Road agencies and institutions are on the lookout for more effective techniques. This study seeks to establish an automatic method for estimating loose gravel using acoustic observation. On gravelroads, recordings from a car's interior were evaluated and matched to the road's state. The first strategy examined road sections with a four-tier (multiclass) manual classification, based on their perceived condition of loose gravel, in accordance with the Swedish road administration authority’s guidelines. The second, examined two tier (binary) manual classification, distinguishing between roads with low and high maintenance needs. Sound features were extracted and processed for subsequentanalysis. Several supervised machine learning methods and algorithms, combined with selected data preprocessing strategies, were deployed. The performance of each strategy and model is determined by assessing and evaluating their classification accuracy along with other performance metrics. The SVM classifier had the best performance in classifying both multiclass as well as binary gravel road conditions. SVM achieved an accuracy of 57.8% when classifying on a four-tier scale and an accuracy of 82% when classifying on a two-tier scale. These results indicate some merits of using audio features as predictive features in the automatic classification of loose gravel conditions on gravel roads.

Place, publisher, year, edition, pages
2022.
Keywords [en]
Sound Classification, Supervised Machine Learning, Gravel roads assessment, SVM, Random Forest, Ensemble of bagged trees, Sound feature extraction
National Category
Computer and Information Sciences
Identifiers
URN: urn:nbn:se:du-41898OAI: oai:DiVA.org:du-41898DiVA, id: diva2:1682300
Subject / course
Microdata Analysis
Available from: 2022-07-08 Created: 2022-07-08 Last updated: 2022-07-11

Open Access in DiVA

fulltext(778 kB)83 downloads
File information
File name FULLTEXT01.pdfFile size 778 kBChecksum SHA-512
e2d016c75cbccf6646cc4477dbc8cc47c0c8c9afa2c57ea2f3bd13fad5a7f9f67f62bc680720bd8c29871be81cfc542dfaefdc0bf866e5f290158abefca7f645
Type fulltextMimetype application/pdf

By organisation
School of Information and Engineering
Computer and Information Sciences

Search outside of DiVA

GoogleGoogle Scholar
Total: 83 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

urn-nbn

Altmetric score

urn-nbn
Total: 289 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • chicago-author-date
  • chicago-note-bibliography
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf