DC 欄位 值 語言 DC.contributor 遙測科技碩士學位學程 zh_TW DC.creator 羅子翔 zh_TW DC.creator Zi-Xiang Lo en_US dc.date.accessioned 2020-8-20T07:39:07Z dc.date.available 2020-8-20T07:39:07Z dc.date.issued 2020 dc.identifier.uri http://ir.lib.ncu.edu.tw:88/thesis/view_etd.asp?URN=107022005 dc.contributor.department 遙測科技碩士學位學程 zh_TW DC.description 國立中央大學 zh_TW DC.description National Central University en_US dc.description.abstract 氣候變遷，導致極端降雨的事件頻傳，使得大規模崩塌的發生頻率越來越高，藉由建置災後崩塌目錄，做為日後崩塌潛勢分析與災害風險管理的基礎。近年來，經常使用光學衛星影像進行大範圍的崩塌測繪，有效節省成本並提升效率，除了視覺判釋外，常態化差異植生指標 (Normalized Difference Vegetation Index, NDVI) 及其他衛星衍生的光譜指標也常被用於崩塌裸露地偵測。根據受災面積和破壞深度，可將崩塌分為深層崩塌和淺層崩塌兩種類型。深層崩塌是指滑動面深入新鮮岩體中之滑動現象，因此可以在地表觀察到母岩露頭。深層崩塌對於環境具有較大的影響，這是因為其大量的沉積物透過破壞材料輸送造成嚴重破壞以及重大社會經濟損失的機會更大。因此深層崩塌的辨識對崩塌災害評估具有重要意義。但是目前尚未有用於區分這兩種崩塌類型的特定衛星衍生指標。本研究的主要目的為建立一個指標，該指標稱為母岩裸露指標 (Bedrock Exposure Index, BEI)，以識別山坡裸露地上母岩露頭的狀況，判斷崩塌之破壞面是否為深入岩體，進而能區判該崩塌是否為深層崩塌。母岩裸露指標建立方法為首先以光譜輻射儀 (FieldSpec®3) 進行野外採樣來分析崩塌地及鄰近區域之母岩、土壤 (岩屑)、植被和水體的光譜連續性與吸收特徵，以及檢視 EO-1 Hyperion 之歷史高光譜影像。由上述的光譜特徵可以發現，在可見光 (Visible Spectrum; VIS; 380-750 nm) 區間 和短波紅外光 (Short-Wave Infrared; SWIR; 850-2500 nm) 反射率之間的偏差，以更好地區分裸露土地 (即崩塌) 的母岩露頭和土壤 (岩屑)，本研究便藉此特性來建立母岩裸露指標 (BEI)。 本研究提出之BEI之值域在 -1〜1 之間，透過分析發現其中水體主要為BEI<0(負值); 母岩露頭為0〜0.2; 土壤岩屑為0.2〜0.5; 植被則為BEI> 0.5。另外，由於Sentinel-2 多光譜影像的多光譜儀 (MultiSpectral Instrument, MSI) 可以感知 BEI 計算所需之光譜波段，其波段2、波段11的波長與本研究中選定的波長 518 nm 與 1648 nm 相近，因此本研究認為亦適合用於 BEI 之計算與應用。本研究亦比較了BEI 和其他可用於偵測裸露地的相關指標，如 NDVI、NDWI(Normalized difference water index)、EVI(Enhanced vegetation index)等，比較結果發現，相較於其他指標BEI在區分母岩與土壤 (岩屑)具有更好的光譜分離性，顯示出 BEI 在判釋不同類型崩塌的優勢以及應用於崩塌監測的潛力。 zh_TW dc.description.abstract Landslide mapping for a wide region using optical satellite imagery has become a popular method due to its relatively low cost and high efficiency. In addition to manually visual interpretation, the Normalized Difference Vegetation Index (NDVI) and other satellite-derived indices have been applied for detecting bare soils which can be considered as a proxy of landslide scars. However, according to the affected area and the depth of the failure, the landslide, by simplification, can be categorized into deep-seated and shallow ones. The deep-seated landslide is defined as the slip plane penetrates the fresh rock mass and therefore bedrock exposure (outcrop) can be observed on the surface. The deep-seated landslide generally has a higher impact on the environment due to its considerable amount of sediments delivered form failure materials and a higher chance to causes severe damages with significant socio-economic loss. Therefore, the identification of deep-seated landslide is important to landslide hazard assessment. So far, however, no particular satellite-derived index has been developed for the discrimination of the two landslide types. This study aims to develop an index, the Bedrock Exposure Index (BEI), for identifying the condition of bedrock exposure over hilly bare lands. The development of BEI was conducted by performing field sampling with radiometer and analyzing the hyperspectral signatures of bedrock, soil regolith, vegetation, and water body using EO-1 Hyperion imagery. The BEI adopts the reflectance deviation between Visible Spectrum (VIS) and the Short-Wave Infrared (SWIR) to better distinguish the bedrock exposure and soil regolith over the bare land surface (i.e. landslide) Specifically, the bedrock exposure can be identified when BEI is within 0~0.2. Meanwhile, Sentinel-2 multispectral imagery is considered suitable for BEI application, thanks to its MultiSpectral Instrument (MSI) sensing all required spectrums for BEI calculation. en_US DC.subject 母岩裸露指標 zh_TW DC.subject 深層崩塌 zh_TW DC.subject EO-1 Hyperion zh_TW DC.subject 土壤岩屑 zh_TW DC.subject Sentinel-2 zh_TW DC.subject Bedrock Exposure Index en_US DC.subject Deep-seated landslides en_US DC.subject EO-1 Hyperion en_US DC.subject Soil regolith en_US DC.subject Sentinel-2 en_US DC.title 母岩裸露指標之建立並應用於崩塌判釋與監測 zh_TW dc.language.iso zh-TW zh-TW DC.title Developing Bedrock Exposure Index for Landslide Mapping and Monitoring en_US DC.type 博碩士論文 zh_TW DC.type thesis en_US DC.publisher National Central University en_US