Journal article
Automated Rust-Defect Detection of a Steel Bridge Using Aerial Multispectral Imagery
Journal of infrastructure systems, v 25(2)
01 Jun 2019
Featured in Collection : UN Sustainable Development Goals @ Drexel
Abstract
AbstractComputer vision methods have the potential to detect rust defects in steel components of bridges. However, direct use of images collected by aerial means to identify such defects is currently difficult because of obstructions caused by other objects in the image field of view. In this context, an automated rust-defect-determination method that leverages aerial imagery, including both visible and infrared images, is presented in this investigation. The proposed method consists of three steps. The first step deals with image registration for which a binary information method is proposed to match the infrared images to their visible counterparts. In the second step, bridge components are retrieved from the captured images via automated segmentation obtained by fusion of visible and infrared images. Finally, rusted regions are identified in YCbCr colorspace, and a rust percentage is calculated. Experimental results obtained by aerial images collected on a real operating structure demonstrate that the proposed methodology can directly use the original captured images and can be successfully applied to real-world scenarios.
Metrics
Details
- Title
- Automated Rust-Defect Detection of a Steel Bridge Using Aerial Multispectral Imagery
- Creators
- Yundong Li - North China University of TechnologyAntonios Kontsos - Drexel UniversityIvan Bartoli - Drexel University
- Publication Details
- Journal of infrastructure systems, v 25(2)
- Publisher
- American Society of Civil Engineers
- Resource Type
- Journal article
- Language
- English
- Academic Unit
- Civil, Architectural, and Environmental Engineering; Mechanical Engineering and Mechanics
- Web of Science ID
- WOS:000464578100018
- Scopus ID
- 2-s2.0-85063580617
- Other Identifier
- 991019169421404721
UN Sustainable Development Goals (SDGs)
This publication has contributed to the advancement of the following goals:
InCites Highlights
Data related to this publication, from InCites Benchmarking & Analytics tool:
- Collaboration types
- Domestic collaboration
- International collaboration
- Web of Science research areas
- Engineering, Civil