| 講演要旨(和文) | | 土砂災害の発生危険度を高精度に評価することは, 効率的な整備を行うための重要な課題である. しかし, 一般に土砂災害の発生要因は多様かつ, 複雑であり, 高精度な危険度評価を実現することは難しい. そこで本研究では, 奈良県全域において2007年に取得された斜面データをケーススタディーとして, 機械学習の一手法であるSVMを用いた災害危険度の総合評価（大石ら, 2007）を試みた. その結果, 落石・崩壊を対象災害として, 従来よりも高精度な危険度評価を可能にした. 特に, 全斜面に一律の基準で評価が行われている斜面データに, 地質区分データを加えることで, 広範囲な対象地域において高精度な危険度評価が実現できることを明らかにした. |
|
|
| | 講演要旨(英文) | | Evaluation of landslide risk with high accuracy is an important issue to maintain the stability of slopes efficiently. However, it is difficult to evaluate the landslide risk with high accuracy because the causes of landslides are generally diverse and complex. In this study, we conducted the comprehensive evaluation of the landslide risk by using the machine learning technique, called SVM (Ohishi et al. 2007), based on the slope investigation data which were acquired in Nara Prefecture in 2007. As a result, we were able to evaluate the landslide risk with higher accuracy than the conventional method. In particular, it becomes clear that it is possible to evaluate the landslide risk with high accuracy in a large area, adding the geological classification data to the slope investigation data, whose evaluation was carried out in the standard uniform for all slopes. |
|
|
|
| |
|