Building damage assessment using improved combined adjustment and automatic building corners extraction method

ABSTRACT

Accurately estimating the collapse height of buildings during earthquakes holds great significance. Acquiring precise height information of buildings using dual-phase high-resolution stereo satellite images (HRSSIs) requires high-precision ground control points, which are difficult to obtain in earthquake-stricken areas. Spaceborne laser data has been validated to significantly improve the vertical accuracy of HRSSIs. Therefore, this paper first proposes the application of a combined stereo adjustment model in building collapse estimation and thoroughly verifies its feasibility. Firstly, the laser altimetry points (LAPs) acquired by GF-7 and ICESat-1 serve as elevation control points (ECPs), and the tie points from HRSSIs were used as constraint to conduct combined adjustment to improve the consistency of stereoscopic positioning accuracy between HRSSIs before and after the earthquake. Subsequently, a footprints-to-conners, coarse-to-fine, automatic building corners extraction method was performed. Avoiding the time-consuming generation of large-scale DSM while also improving the accuracy of building collapse assessment. Finally, according to the height change of building corners calculated from HRSSIs, building damage assessment was completed. According to the above methods, independent experiments were carried out respectively for the Oaxaca earthquake in Mexico in 2020 and the Kahramanmaras earthquake in Türkiye in 2023. GF-7 HRSSIs was used to verify the accuracy and stability of the proposed method for 3D assessment of building damage. The results indicate that the RMSE of elevation positioning accuracy has been greatly reduced in both regions, ranging from 10.29 m to 0.95 m and from 6.67 m to 0.99 m, respectively. This accuracy guarantees the feasibility of detecting the building floor-level collapse. And 2541 building corner points were extracted from two disaster areas, with correct corner points accounting for 85.18%. A total of 520 buildings were correctly extracted corner points, with an accuracy rate of 94.12% for assessing the degree of damage.

KEYWORDS:

• Building damage assessment
• GF-7
• laser altimetry
• block adjustment
• building corners extraction

Acknowledgements

We would like to thank the Land Satellite Remote Sensing Application Center of the Ministry of Natural Resources of China for providing the GF-7 HRSSIs of the earthquake area in Mexico and Turkey before and after the earthquake.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

ICESat-1 laser points data used in this study were acquired in Earth Data’s official website(https://nsidc.org/data/icesat).

The Gaofen-7 datasets that support the findings of this study are available from the Land Satellite Remote Sensing Application Center of the Ministry of Natural Resources of China.

The Real images of urban buildings in Turkey’s Kahramanmaras are from Google Street View.

Additional information

Funding

The work was supported in part by the National Key R&D Program of China under Grant 2018YFB0505400 and National Natural Science Foundation of China under Grant 42241164 and 41871325.