Detection of Sediment Disaster Due to the 2000 Tokai Heavy Rainfall Using High Resolution Satellite Images

This paper describes detection of the sediment disaster due to the 2000 Tokai Heavy Rainfall in Japan, using IKONOS Pan-sharpen images. In most case of steep mountainside in Japan, slope failures are often covered by leaves in the forests. In this study, a detector based on region growing algorithm is proposed for high resolution satellite images to estimate the shadow area of the slope failure. The magnitude and the morphology of slope failures are calculated by the proposed method. On the other hand, the boundary of sub-compartment shown in forest planning maps was digitized, and the forest database was developed for the attributes such as forest management type shown in the forest register of Gifu Prefecture in Japan. The detected results of slope failures are stored in the database, and they are compared with the attribute data for the characteristic analysis. The GIS model was applied to detection of heavy rainfall induced slope failures in the upstream of the Yahagi River triggered by the 2000 Tokai Heavy Rainfall. The conclusion is shown as follows: (1) Most of sliding failures can be detected precisely using threshold value. The shaded relief processing was effective to detect the debris flow along a valley, and the accuracy was improved. (2) The quantitative theory analysis for slope failures and the database concerning the topography, geological features and forests is effective to create the hazard assessment map. This result is useful for planning of erosion control dams.