Book chapter
A Video-Based Crack Detection in Concrete Surfaces
Data Science in Engineering, Volume 9, pp 245-252
05 Oct 2021
Abstract
Damage signs in concrete elements and structures appear on the surface in the form of cracks. Precise detection of existing crack pattern provides a reliable basis for updating the structural parameters of damaged concrete elements and predicting future behavior. Several studies have developed crack detection methods based on image processing of cracked surfaces. However, these methods have several deficiencies, including sensitivity to noise. In some cases, the recorded videos during the occurrence of damage are available, which provides a set of images for one damage level. In this research, a methodology is developed to track crack formation taking advantage of video processing. Specifically, robust principal component analysis is utilized to detect new cracks. To evaluate the methodology, the test data of a one-third scale rectangular RC shear wall is used. Images and video of this specimen were captured while applying a cyclic loading protocol. To achieve a reliable detection, two video stabilization methods are applied to the videos. These methods are based on feature point matching and phased-based motion processing. The results show that the local maxima of Gini coefficients of frames indicate new crack formation.
Metrics
20 Record Views
Details
- Title
- A Video-Based Crack Detection in Concrete Surfaces
- Creators
- Hamed Momeni - New Mexico Institute of Mining and TechnologySina Basereh - University at Buffalo, State University of New YorkPinar Okumus - University at Buffalo, State University of New YorkArvin Ebrahimkhanlou - New Mexico Institute of Mining and Technology
- Publication Details
- Data Science in Engineering, Volume 9, pp 245-252
- Series
- Conference Proceedings of the Society for Experimental Mechanics Series
- Publisher
- Springer International Publishing; Cham
- Resource Type
- Book chapter
- Language
- English
- Academic Unit
- Civil, Architectural, and Environmental Engineering
- Scopus ID
- 2-s2.0-85117451787
- Other Identifier
- 991021890004404721