2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS | 2021
Detection and Volume Estimation of Large-Scale Landslide in Abe Barek, Afghanistan Using Nonlinear Mapping of DEMs
Abstract
Evaluations of the extent and the volume of the displaced materials are vital for post-disaster management activities. In this study, we present a nonlinear geometric correction technique for identifying landslide affected areas and estimating the volume by using pre-and post-event digital elevation models (DEMs) generated from high-resolution stereo pair satellite imagery of the Abe Barek landslide, which was a devastating landslide that occurred On May 2, 2014, in the northern part of Afghanistan. The proposed method consists of shifting vector generation in subareas of the DEMs, consensus operations, and interpolation of the shifting vectors. The quality assessment confirmed that the proposed method outperformed simple DEMs of difference technique by eliminating a large scale of geometric errors in unaffected areas. We estimated the landslide volume as 1.06 x 106 m3 from the pre-and post-event DEMs corrected by the proposed method. Finally, the building damage assessment on the landslide-affected area has been carried out in terms of the deposition depths obtained in this analysis.