[H-DS07] 地すべりおよび関連現象

[HDS07-P11] Detection of spectral characteristics of vegetation recovery at landslide scars using hyperspectral images

*齋藤 仁¹、内山 庄一郎²、手代木 功基³、伊藤 千尋⁴、早川 裕弌⁵ (1.名古屋大学 大学院環境学研究科、2.防災科研、3.金沢大学、4.九州大学、5.北海道大学)

キーワード：Landslides、Hyperspectral camera、Vegetation recovery

Landslides are important natural hazards that cause damage to society. Landslides destroy vegetation on hillslopes by removal, resulting in unstable land followed by unprecedented surface erosion for many years. In this regard, vegetation recovery at landslide scars are important for stabilizing the land surface and ecological systems on mountain slopes. However, analysis of these characteristics from traditional optical satellite and aerial images is difficult due to the limitation of low spatial resolution and broad-band multispectral data. Recently, narrow-band spectral data from hyperspectral cameras improves the measuring of geological characteristics and detailed vegetation features. The purpose of this study is to detect the spectral characteristics of vegetation recovery at landslide scars using hyperspectral images. We preliminarily measured landslide scars and vegetation using small hyperspectral cameras at Aso Volcano, Japan. The study area has frequently experienced co-seismic and rainfall-induced landslide events and is characterized by the typical grass vegetation landscape. As the results, recovered grass vegetation at landslide scars were characterized by relatively low reflectance at near-infrared bands although NDVI (Normalized difference vegetation index) has little differences from the recovered vegetation and the vegetation at non-landslide slopes. The preliminary results showed a significant difference in the reflectance of geology at the landslide slip surface. Further studies are necessary to validate the spectral characteristics in other landslide sites.