Towards automated UAV assisted bridge inspections using photogrammetry and image processing techniques

Habeene Habeenzu, Patrick McGetrick, David Hester

Research output: Contribution to conferencePaperpeer-review

9 Downloads (Pure)

Abstract

At the heart of addressing bridge condition challenges are bridge inspections. The main activity during a bridge inspection is close, arm’s length visual inspection of the entire bridge structure. During this process, all defects such as cracks, spalls and material degradation are manually recorded on the bridge itself and on inspection forms. Where access is difficult such as where a safe working platform cannot be mounted under bridge decks, or on high bridges, expensive underbridge equipment is required which when used results in expensive lane closures. Furthermore, visual inspections have been shown to lack consistency from inspector to inspector and can be unreliable. Technological solutions such as using drones with digital cameras combined with post-processing of images using digital image processing and photogrammetry techniques can potentially assist bridge inspectors in the provision of reliable information on structure geometry, inventory and structure condition, supplementing traditional methods. This information can also be packaged in easy to understand 2D or 3D formats making it more straightforward for bridge owners to make timely decisions about allocating bridge maintenance funds. This paper investigates the use of digital image processing and photogrammetry techniques to detect and annotate 3D models of cracked concrete specimen obtained using drones and presents the results of laboratory tests.
Original languageEnglish
Publication statusPublished - 27 Aug 2020
EventCERI 2020 - Cork University, Cork, Ireland
Duration: 27 Aug 202028 Aug 2020

Conference

ConferenceCERI 2020
Country/TerritoryIreland
CityCork
Period27/08/202028/08/2020

Fingerprint

Dive into the research topics of 'Towards automated UAV assisted bridge inspections using photogrammetry and image processing techniques'. Together they form a unique fingerprint.

Cite this