A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Towards Automated Detection of Cracked Concrete
When inspecting bridges, unmanned aerial vehicles or drones can provide eyes that can be used even without traffic closures, which is a significant issue for bridge owners. Their impact on direct and indirect costs can be even more valuable when inspecting high bridges with large spans. If a drone carries high definition lenses with a long focal length that can capture images of cracked concrete from a distance, and if the process of automated detection of cracked concrete can be applied, its scope for monitoring the progress of damaged bridges can be significantly extended. This article presents the concept and some results of a study on using drones for such a purpose. The investigations were initially performed under controlled laboratory conditions to simulate cracks in the reinforced concrete block. The analysis showed that in a photogrammetry 3D model, cracks of 0.2 mm and wider could be measured with reasonable accuracy. Crack detection efficiency was demonstrated on a concrete sample using deep convolutional neural networks, and an 80% accuracy rate was achieved. The validation procedure applied only a small number of real-world images; therefore, its performance, as also stated by other studies, can be additionally improved when a larger dataset is considered.
Towards Automated Detection of Cracked Concrete
When inspecting bridges, unmanned aerial vehicles or drones can provide eyes that can be used even without traffic closures, which is a significant issue for bridge owners. Their impact on direct and indirect costs can be even more valuable when inspecting high bridges with large spans. If a drone carries high definition lenses with a long focal length that can capture images of cracked concrete from a distance, and if the process of automated detection of cracked concrete can be applied, its scope for monitoring the progress of damaged bridges can be significantly extended. This article presents the concept and some results of a study on using drones for such a purpose. The investigations were initially performed under controlled laboratory conditions to simulate cracks in the reinforced concrete block. The analysis showed that in a photogrammetry 3D model, cracks of 0.2 mm and wider could be measured with reasonable accuracy. Crack detection efficiency was demonstrated on a concrete sample using deep convolutional neural networks, and an 80% accuracy rate was achieved. The validation procedure applied only a small number of real-world images; therefore, its performance, as also stated by other studies, can be additionally improved when a larger dataset is considered.
Towards Automated Detection of Cracked Concrete
Lecture Notes in Civil Engineering
Pellegrino, Carlo (editor) / Faleschini, Flora (editor) / Zanini, Mariano Angelo (editor) / Matos, José C. (editor) / Casas, Joan R. (editor) / Strauss, Alfred (editor) / Žnidarič, Aleš (author) / Kreslin, Maja (author) / Anžlin, Andrej (author) / Krivic, Andraž (author)
International Conference of the European Association on Quality Control of Bridges and Structures ; 2021 ; Padua, Italy
Proceedings of the 1st Conference of the European Association on Quality Control of Bridges and Structures ; Chapter: 147 ; 1294-1300
2021-12-12
7 pages
Article/Chapter (Book)
Electronic Resource
English
Permeability, of cracked concrete
| British Library Online Contents | 1999
Permeability of cracked concrete
| Springer Verlag | 1999
Permeability of cracked concrete
| Online Contents | 1999
Chloride Diffusivity in Cracked Concrete
| British Library Conference Proceedings | 2011