2020-02-062024-05-14https://scholars.lib.ntu.edu.tw/handle/123456789/660286摘要：山崩為台灣常見的自然災害之一，若能快速辨識山崩事件位置，以及實施坡地監測，便能提供重要資訊用於救災與防災。在台灣偵測山崩與地滑，卻因為山區交通不易以及光學影像易被雲霧遮蔽有實際上的困難。除此之外即便在已知山崩區域，現地崩移速率的監測亦有實際的困難。因此本研究目標將利用雷達影像不受雲層覆蓋影響之特性，針對台灣山區的崩塌地進行監測。本研究的目的在於建立合成孔徑雷達強度影像之時間序列，進而透過空間上的統計分析系統地分類出強度與環境因子之間的關係，探討多時序強度影像監測崩塌地的可行性以及崩塌地本身的活動特性。因此本研究將使用多時段的雷達差分干涉技術和多時段的雷達強度分析技術，分析歐洲太空總署拍攝之Sentinel-1資料，並以日本JAXA發射之ALOS與ALOS2雷達資料為輔，針對台灣近年主要的崩塌地進行監測，並對山坡地崩塌潛勢進行評估，最後利用光學影像進行檢核。<br> Abstract: Landslides are common natural hazards in Taiwan, and it is crucial to provide useful information for hazard mitigation by rapidly identifying landslide location and monitoring landslide process. In Taiwan, inaccessibility in mountainous areas and cloud shadow for optical images cause landslide detection difficult. In addition, in-situ monitoring at landslides is also somewhat difficult to conduct. Therefore, this study aims to use synthetic aperture radar (SAR) images, which are able to see through clouds, . The purpose of this study is to establish time series of SAR intensity images and to systematically analyze the relationship between environmental factors and intensity through spatial statistics. We will then discuss possibility of monitoring landslides by using intensity time series and discuss their moving characteristics. This research will use multi-temporal InSAR and intensity analysis techniques for Sentinel-1 images launched by European Space Agency (ESA) and ALOS and ALOS-2 images launched by Japan Aerospace Exploration Agency (JAXA). We will analyze major landslides in recent years and evaluate landslide potential with validation by using optical images.SARintensitycoherenceInSAR合成孔徑雷達強度同調性干涉合成孔徑雷達