應用NDVI時序軌跡分析於產製多時序崩塌圖資—以台灣南投地區為例;Generating multi-temporal landslide inventory using the NDVI temporal trajectory in Nantou County, Taiwan

NCU Institutional Repository > 工學院 > 國際永續發展在職專班 > 博碩士論文 > Item 987654321/86216

請使用永久網址來引用或連結此文件: http://ir.lib.ncu.edu.tw/handle/987654321/86216

題名:	應用NDVI時序軌跡分析於產製多時序崩塌圖資—以台灣南投地區為例;Generating multi-temporal landslide inventory using the NDVI temporal trajectory in Nantou County, Taiwan
作者:	阮成惟;Duy, Nguyen Thanh
貢獻者:	國際永續發展碩士在職專班
關鍵詞:	地震型崩塌;颱風型崩塌;時序軌跡;角變值分析
日期:	2021-08-12
上傳時間:	2021-12-07 12:17:47 (UTC+8)
出版者:	國立中央大學
摘要:	邊坡災害是許多國家必須面對的一個重要災害問題，而完整、準確和即時的崩塌目錄以及崩塌圖資對於崩塌相關的災害管理和預防可說是相當重要的資料。近年來，地球觀測衛星（EOS）已廣泛用於土地監測的工作上，包括崩塌的監測以及崩塌圖資之產出。現今應用遙測技術所產製之崩塌目錄以及圖資多為事件型的，也就是針對單一崩塌事件，並且主要是針對災害地區域進行崩塌圖資繪製，這使得崩塌圖資在多時、廣域的應用上時必須進行繁複的整合工作。本研究認為，使用衛星資料，並透過分析地表光譜軌跡的時間序列來檢視地表變遷可有效地、經濟地針對大範圍區域建立多時序的崩塌圖資。具體來說，本研究將基於Landtrendr（Landsat-based detection of Trends in Disturbance and Recovery）技術，使用包括 Landsat 和 SPOT 在內的多重衛星、多時序資料來檢測地震和颱風發生前後的地表標準化差異植生指數(Normalized Difference Vegetation Index, NDVI)，並應用NDVI之時間變化來檢視其角變值(Angular Chang, AC)之多時序軌跡用以進行崩塌判釋。本研究區位於台灣中部南投縣，該地區於1999年的集集地震中遭受嚴重的破壞，產生大量、大規模的崩塌事件，並且其後數年也遭受數個颱風襲擊，使得崩塌地在時間及空間上均有較大的變動性。本研究針對台灣中部南投縣一帶，1999年至2004年地震和颱風事件誘發的崩塌進行試驗，進行多時序崩塌目錄的建構，期望本研究提出之方法可以更有效、更快速地產制大範圍多時序之崩塌圖資，以精進邊坡災害之防治工作。;The landslide hazard is a crucial issue in many countries. A complete, accurate, and timely-provided landslide inventory is important to generate necessary information for hazard management and prevention. In recent years, the Earth Observation Satellites (EOS) have been widely used for land monitoring practice, including landslide detection and inventory preparation. However, most landslide mapping tasks are mainly event-based and focus on particular areas. In recent years, Landtrendr (Landsat-based detection of Trends in Disturbance and Recovery) is an approach process to perform the change of a land surface through time-series spectral trajectories. This study applies the analysis concept of the Landtrendr method in detecting temporal changes of earthquake- and typhoon-triggered landslides by using multiple satellite images, including Landsat and SPOT. The study area, Nantou County, located in central Taiwan, experienced high consequences of the Chi-Chi earthquake in 1999 and also suffered from seasonal typhoons year by year. Earthquake events with a magnitude over 0.6g of peak ground acceleration (PGA) and typhoon events with precipitation higher than 200mm/day are the main triggers that caused landslides in this area. The multi-temporal landslide mapping method, the angular change (AC) analysis, proposed in this study is based on NDVI (Normalized Difference Vegetation Index) temporal trajectory which uses EOS time-series data to identify the location, extent, and timing of landslides induced by the Chi-Chi earthquake and four typhoon events in Nantou County from 1999 to 2004. This study expects that the proposed method can contribute to the generation of multi-temporal landslide inventory more efficiently.
顯示於類別:	[國際永續發展在職專班] 博碩士論文

文件中的檔案:

檔案	描述	大小	格式	瀏覽次數
index.html		0Kb	HTML	183	檢視/開啟

在NCUIR中所有的資料項目都受到原著作權保護.

社群 sharing

資料載入中.....