請用此 Handle URI 來引用此文件: http://hdl.handle.net/11455/47273
標題: 考慮時間空間尺度之坡地災害及受災潛勢研究-子計畫:多時期遙測影像應用於坡地災害監測與潛勢分析之研究
Application of Multi-Date Remote Sensing Images to Monitoring and Potential Assessment of Landslides
作者: 楊明德
關鍵字: 環保工程
應用研究
摘要: 88風災所挾帶的豪雨,造成台灣史上最大的崩塌事件,在南投地區則以陳有蘭溪流域之坡地崩塌為最。崩塌區域後續之植生復育及防災工作極為重要,鑑於崩塌區位分布遼闊且零散,利用遙測衛星影像能快速地監測、評估大範圍崩塌區位之植生復育及變遷情形,針對植生復育進行監測可做為坡地災害即時評估與水土保持設施規劃設計之用。除瞭解崩塌地植生復育及變遷狀況外,崩塌地所造成之土方量推估亦是重點之一,隨颱風挾帶之暴雨常會擴大崩塌區域且將大量土砂帶至下游段,造成嚴重土砂災害,對下游民眾之生命財產造成重大威脅。本研究計畫為整合型之三年計畫主要係整合遙測衛星影像資訊及地理資訊系統圖資,以自動化分類方式萃取集水區崩塌區位,分析植生復育率變遷及各項崩塌因子與崩塌規模之關係,藉以間接推算各地層之崩塌機率;另以植生覆蓋指數及崩塌區位分析,量化分析影響崩塌區位植生復育之影響因子以及集水區崩塌地之治理順序,以利作最有效整治計畫。另將提供植生復育區位資訊,再配合三維立體模型,可作為集水區崩塌區位植生復育監測依據與治理評估用。最後發展坡地危害保全對象風險評估機制,以便災前找出高險區域而加以整治或加強預警,達到減災避災的目的。此三年計畫之主要工作內容分述如下:(一)第一年度工作:搜集國內外坡地災害相關文獻,並以陳有蘭溪集水區為對象,利用風災前後期SPOT衛星影像以各種分類方式,進行崩塌區位判釋,再以88風災後之衛星影像進行崩塌區位植生復育率分析,針對近15年陳有蘭溪集水區之坡地變化進行輻射校正以利多時期監測,並對坡地災害保全區域進行分析與評估。(二)第二年度工作:以頭坑溪集水區為研究試區,利用監督式以及非監督式分類法,發展崩塌地自動辨識方法。利用地理資訊系統以及多變量統計分析影響植生復育率之環境因子,研究崩塌地與雨量、坡度、集水區面積與流出土方量之關係,及這些因子對於植生復育率之影響。建立類神經網路進行崩塌潛勢分析,並進一步進行植群演化推估與崩塌地區後續發展預測模式。(三)第三年度工作:利用前兩年之成果針對陳有蘭溪流域作為研究試區,探討88風災後之植生復育狀況,坡地崩塌潛勢區域圖,利用所發展之類神經網路進行崩塌潛勢分析,並提出植群復育率整合性評估與長期監測分析,以期達成有效掌握崩塌地變遷的相關資訊,提供相關單位決策之輔助。利用攝影測量技術產生即時三維模型,並針對研究試區之崩塌地進行三維監測,以期達成有效掌握崩塌地變遷的相關資訊。並配合保全對象之空間資訊與崩塌潛勢之分析,發展坡地危害保全對象風險評估機制,以便災前找出高險區域而加以整治或加強預警,提供相關單位決策之輔助達到減災避災的目的。
The geology of Taiwan mostly is weak shale and slate mostly that causes a great potential of mud flow occurrence and soil scour after heavy rainfalls, especially after Chi-Chi earthquake in 1999. During the typhoon season in summer and autumn, landslides due to heavy rainfalls often cause severe disasters and even great economic losses, such as the catastrophic event caused by Typhoon Morakot. Monitoring and investigation of landslide has become an essential task to deal with in Taiwan as well as other countries with a great percentage of mountainous area.This proposal expresses a three-year successive research project. At the first year, worldwide literatures about applying various remote sensing techniques to landslide monitoring and investigation will be collected and reviewed. Landslide area in Chenyulan River watershed will be investigated and analyzed by the multi-temporal satellite images (such as SPOT and FORMOSAT-II) in the recent 15 years.At the second year, an automatic landslide identifying classifier and a vegetation recovery model of landslide in Chenyulan River watershed will be developed. Furthermore, potential landslide area could also be estimated by analyzing the vegetation recovery rate, landslide sites, and affecting factors. Back-propagation neural network will be adopted to reveal the relationship between rainfalls and landslides.At the third year, the results of vegetation recovery and change detection of landslide area in Chenyulan River watershed will be discussed. The risk assessment and management of the landslides will also be established and processed. A systematic process of monitoring and assessment of landslide area will also be established for providing the mitigation strategies in landslide potential area.
URI: http://hdl.handle.net/11455/47273
其他識別: NSC99-2625-M005-008-MY3
顯示於類別:土木工程學系所

文件中的檔案:
沒有與此文件相關的檔案。


在 DSpace 系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。