Automatic detection of earthquake-induced ground failure effects through Faster R-CNN deep learning-based object detection using satellite images

Kemal Hacıefendioğlu (), Hasan Basri Başağa () and Gökhan Demir ()
Additional contact information
Kemal Hacıefendioğlu: Karadeniz Technical University
Hasan Basri Başağa: Karadeniz Technical University
Gökhan Demir: Ondokuz Mayıs University

Natural Hazards: Journal of the International Society for the Prevention and Mitigation of Natural Hazards, 2021, vol. 105, issue 1, No 20, 383-403

Abstract: Abstract The seismically induced ground failure is defined as any earthquake-generated process that leads to deformations within a soil medium, which in turn results in permanent horizontal or vertical displacements of the ground surface. As a result, relative movements occur on the ground and structures affected by these movements and thus they may be damaged. Determining earthquake-induced ground failure areas is important to carry out damage assessment studies more quickly and reliably and to prevent more destructive damages. Large earthquake-induced ground failure areas or limited access to the areas due to earthquake causes costly and unsafe fieldwork. Using satellite photographs, earthquake-induced ground failure areas can be easily and reliably detected and the fieldwork can be planned quickly. This study aimed at determining the postearthquake-induced ground failure areas and buildings or structures partially ruined (damaged) by using a deep learning-based object detection method, using Google Earth satellite images after an earthquake. The data set obtained after the earthquake occurred in the 2018 Palu region of Indonesia was used. This data set is divided into two parts for training and test areas. A descriptive approach is considered for detecting the earthquake-induced ground failure areas and damaged structures from collected images from Google Earth software using satellite photographs, using a pretrained Faster R-CNN. To demonstrate the effectiveness of the proposed method, the data set was first created with Google Earth Pro software and it was generated with 392 images for the earthquake-induced ground failure area and 223 images for the damaged area with a resolution of 1024 × 600 pixels. The analyses were carried out by taking into account different image scales. As a result of the analyses, it was concluded that the earthquake-induced ground failure effects (liquefied soil) and damaged structures can be detected to a large extent by using object detection-based deep learning methods.

Keywords: Satellite image; Liquefaction; Faster R-CNN; Deep learning; Object detection (search for similar items in EconPapers)
Date: 2021
References: View complete reference list from CitEc
Citations: View citations in EconPapers (2)

Downloads: (external link)
http://link.springer.com/10.1007/s11069-020-04315-y Abstract (text/html)
Access to the full text of the articles in this series is restricted.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:nathaz:v:105:y:2021:i:1:d:10.1007_s11069-020-04315-y

Ordering information: This journal article can be ordered from
http://www.springer.com/economics/journal/11069

DOI: 10.1007/s11069-020-04315-y

Access Statistics for this article

Natural Hazards: Journal of the International Society for the Prevention and Mitigation of Natural Hazards is currently edited by Thomas Glade, Tad S. Murty and Vladimír Schenk

More articles in Natural Hazards: Journal of the International Society for the Prevention and Mitigation of Natural Hazards from Springer, International Society for the Prevention and Mitigation of Natural Hazards
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().