The system be down for regular maintenance on April 3rd, 2024 from 8:00-10:00am.

Show simple item record

dc.contributor.advisorMiller, Gregory Ren_US
dc.contributor.authorLattanzi, Daviden_US
dc.date.accessioned2013-07-25T17:54:13Z
dc.date.available2013-07-25T17:54:13Z
dc.date.issued2013-07-25
dc.date.submitted2013en_US
dc.identifier.otherLattanzi_washington_0250E_11891.pdfen_US
dc.identifier.urihttp://hdl.handle.net/1773/23578
dc.descriptionThesis (Ph.D.)--University of Washington, 2013en_US
dc.description.abstractCivil inspection images are notoriously difficult to interpret for those who were not present during an inspection, and so they have historically had limited value in ongoing condition monitoring applications. Within the last decade the advent of inexpensive digital imaging has compounded this problem by making it possible and common to capture and report numerous images, creating information overload that can easily overwhelm engineers and structure owners. More recently, developments in electronics manufacturing and robotics have made automated structural inspection an emergent technology for civil engineers, which inevitably is leading to even more digital image data that must be handled. This dissertation presents a systematic approach to capturing, processing, and representing inspection image data such that the inherent value of these increasingly large data sets can be realized. Using a data pipeline combining automated image capture, contextualized 3D visualization, and robust computational imaging and regression techniques, the goal is to allow engineers to view inspection images in their 3D spatial context, while aiding them through enhanced damage detection routines, all while helping to minimize field inspection disruptions through the use of robotic imaging. In the development of this pipeline and an associated prototype implementation, several key challenges have been addressed: (i) automated systems capable of comprehensive field imaging; (ii) 3D reconstruction algorithms which provide accurate, photorealistic image interpretations; (iii) robust computer vision algorithms suitable for field applications; and (iv) nonlinear regression models which correlate the relationships between extracted image information and structural performance. The results of prototype testing show that, given due consideration to the inherently large data sets that robots produce, systematic imaging can enable entirely new ways of visualizing and interacting with inspection informationen_US
dc.format.mimetypeapplication/pdfen_US
dc.language.isoen_USen_US
dc.rightsCopyright is held by the individual authors.en_US
dc.subjecthealth-monitoring; image analysis; inspection; robotics; scene reconstructionen_US
dc.subject.otherCivil engineeringen_US
dc.subject.othercivil engineeringen_US
dc.titleA Computational Framework for Next-Generation Inspection Imagingen_US
dc.typeThesisen_US
dc.embargo.termsNo embargoen_US


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record