篇名 | 整合遙測特徵指標與自組織聚類神經網路模式於崩塌地萃取之研究―以莫拉克風災之阿里山溪集水區為例 |
---|---|
卷期 | 6:4 |
作者 | 林文賜 、 林文賜 、 黃碧慧 |
頁次 | 185-189 |
關鍵字 | 崩塌地 、 遙測特徵指標 、 自組織聚類神經網路 、 Landslide 、 Landslide Feature Indices 、 Self-Organizing Map |
出刊日期 | 201110 |
九十八年八月八日莫拉克颱風(又稱為八八水災)造成中南部及東部地區嚴重山崩、地滑、土石流、淹水及村毀等災情,本研究係以遭受莫拉克颱風災害之阿里山溪集水區為範圍,蒐集研究地區之集水區資料、空間數值資料及風災前後SPOT 衛星影像資料,發展萃取崩塌地影像之遙測光譜特徵指標及二層式自組織聚類神經網路萃取崩塌區位。分析結果顯示,第一層SOM 模式為粗分類,將影像分類為精確崩塌區、模糊崩塌區及非崩塌區;再將模糊崩塌區透過第二層SOM 模式(細分類),可分類為精確崩塌區及非崩塌區,合併二層之分類成果即可獲得全區崩塌地;研究地區經二層SOM 模式萃取,其崩塌面積為619.91 公頃,準確度評估之Kappa 值為0.9766,顯示本研究發展模式可供崩塌地評估及坡地防災之參考。
Morakot typhoon, occurred on August 8, 2009, causedserious damages such as villages and bridges destroyed by landslides,debris flow and flood hazards in south-central and eastern Taiwan. TheAlisan creek watershed, struck by Morakot typhoon, was chosen asstudied area for assessing the landslide sites. In this study, watersheddisasters, spatial digital data and SPOT images before and after typhoonwere collected, and two landslide feature indices and two-layerself-organizing map neural network were developed to extract accuratelandslides. The analyzed results indicate that the image after typhoon canbe classified as precise landslide area (PLA), fuzzy landslide area (FLA)and non-landslide area (NLA) by the first-layer SOM extraction (calledthe coarse classification). The FLA can be isolated as PLA and NLA bythe second-layer SOM extraction (called the fine classification). Thecompleted accurate landslide sites can be obtained by merging first- andsecond-layer PLAs. By two-layer SOM extraction, there are 619.91 ha oflandslide area with high Kappa value 0.9766 extracted at studied site. Itshows the developed model can be used for landslide assessment andslopeland disaster prevention.