Automatic classification of buildings using post-earthquake single-epoch photogrammetric 3D point cloud data for rapid assessment: a case study of the February 6, 2023 Kahramanmaraş earthquakes in Türkiye
JOURNAL OF CIVIL STRUCTURAL HEALTH MONITORING, cilt.16, sa.6, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 16 Sayı: 6
- Basım Tarihi: 2026
- Doi Numarası: 10.1007/s13349-026-01143-9
- Dergi Adı: JOURNAL OF CIVIL STRUCTURAL HEALTH MONITORING
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Compendex, INSPEC, Materials Science & Engineering Collection (ProQuest), Technology Collection (ProQuest)
- Karadeniz Teknik Üniversitesi Adresli: Evet
Özet
The rapid and accurate acquisition of information on earthquake-damaged structures is of critical importance for reducing loss of life. Although remote sensing data are widely employed for mapping affected regions after earthquakes, precise assessment results are required for rapid response and economic loss estimation. Therefore, in post-disaster detailed damage assessment, 3D photogrammetric point cloud data obtained from unmanned aerial vehicle imagery are commonly utilized due to their high resolution and rapid data acquisition capabilities. In the proposed study, a fast and effective method has been developed to identify buildings damaged following the Kahramanmara & scedil; earthquakes on February 6, 2023; by using 3D photogrammetric point cloud data acquired from a single post-earthquake flight combined with cadastral maps. The buildings were automatically classified into four different damage types within a short time. The developed algorithm enables the detection of intact, slightly damaged, damaged and destroyed buildings by analyzing 3D geometric changes for 695 buildings located in Hatay province, with the entire process being performed fully automatically in approximately 10 min, achieving an accuracy rate of 98.71%. These findings highlight the effectiveness of the proposed approach in identifying destroyed, slightly damaged, damaged, and intact buildings after a disaster. Furthermore, the proposed method is not influenced by radiometric changes nor dependent on training data, which indicates its potential to provide similar accuracy rates across different regions. Additionally, by leveraging 3D geometric changes, it offers the capability to automatically generate a four-class building damage map with high accuracy in a very short time.