Smart Automated Fault Detection for Improved Road Maintenance Planning in Australia
Edited book (chapter)
Chapter Title | Smart Automated Fault Detection for Improved Road Maintenance Planning in Australia |
---|---|
Book Chapter Category | Edited book (chapter) |
ERA Publisher ID | 2797 |
Book Title | Recent Advances in Structural Health Monitoring Research in Australia |
Authors | Brisolin, Jacob (Author), Nguyen, Andy (Author), Ullah, Fahim (Author), Bourke, Allan (Author), Khuc, Tung (Author) and Nsanabo, Vivien (Author) |
Editors | Guan, Hong, Chan, Tommy H. T. and Li, Jianchun |
Page Range | 245-278 |
Series | Civil Engineering and Architecture |
Chapter Number | 8 |
Number of Pages | 34 |
Year | 2022 |
Publisher | Nova Science Publishers |
Place of Publication | New York, United States |
ISBN | 9781685077419 |
9781685076092 | |
Digital Object Identifier (DOI) | https://doi.org/10.52305/QHVI3457 |
Web Address (URL) | https://novapublishers.com/shop/recent-advances-in-structural-health-monitoring-research-in-australia/ |
Abstract | All assets deteriorate with time requiring regular maintenance and management. Asset management attempts to monitor and react to this deterioration to deliver the expected levels of service. However, current road asset management techniques are highly reactive and unable to thoroughly analyse the holistic environment and associated effects. The massive footprints of road networks make checking every road segment an impossible task for the asset inspectors. This chapter addresses this issue in the context of an Australian local government area (LGA) by using commonly available technologies. The technologies in focus include geographical mapping systems and mobile phones with remote access to cloud databases and image capturing capability. The inspectors and assessors can automatically detect road faults by integrating these tools coupled with image processing and artificial intelligence (AI). This is possible using image submissions by public members taken from smartphones, enabling a smart and cost-effective automated road fault detection system. In this chapter, we collected 863 photos of road surfaces across an LGA using a mobile photo capturing application. These photos were subdivided to construct a database of 12,598 images containing pavement cracking and unrelated objects visible from the road reserves. Next, the collected images were subjected to classification using deep learning for automated fault detection. The results of this chapter show that implementing this fault detection procedure into a common engineering software such as MATLAB achieved a validation accuracy of 95%. This was independent of the technical background or programming knowledge of the road users as data collectors. Finally, an integration framework to incorporate this smart tool into a standard asset management framework has been provided to improve current road asset maintenance planning and management practices in Australia. |
Keywords | Asset Management, Automated Fault Detection, Road Fault, Artificial Intelligence, Deep Learning, Mobile Application, Road Users |
ANZSRC Field of Research 2020 | 400510. Structural engineering |
Public Notes | Files associated with this item cannot be displayed due to copyright restrictions. |
Byline Affiliations | University of Southern Queensland |
Hanoi University of Civil Engineering, Vietnam | |
Brisbane City Council, Australia | |
Institution of Origin | University of Southern Queensland |
https://research.usq.edu.au/item/q7w7q/smart-automated-fault-detection-for-improved-road-maintenance-planning-in-australia
132
total views5
total downloads10
views this month0
downloads this month