| 摘要: | 地質災害發生頻率的增加，加上氣候變遷的惡化，對越南構成了重大的國家關注。越南被視為兩種主要地質災害——地面沉降和山崩——廣泛發生的熱點地區。這些災害通常在對基礎設施造成顯著影響並導致人員傷亡之前，往往不被察覺。雖然地面觀測在監測這些災害方面效果不佳，但整合衛星數據提供了一種可能更高效的解決方案。使用衛星數據的主要問題在於其實際應用和準確性。本論文的目標是展示衛星數據在檢測和管理人為地質災害方面的實際應用案例。通過適當的處理，衛星測量可以實現高精度的全球應用。此外，將地面觀測與衛星分析相結合，提供比單獨依賴任一方法更全面的研究。本論文展示了兩個利用衛星數據研究越南地質災害的案例。第二章使用干涉合成孔徑雷達（InSAR）技術研究河內市的長期沉降。我們處理了2007年至2018年間的多顆SAR衛星影像，提供連續的地面位移觀測。位移的時間序列顯示，郊區的顯著沉降與自1990年以來的地下水枯竭密切相關。二十多年來，地下水儲存的迅速減少，結合水文地質層的逐漸壓實，導致了延遲的沉降。為了改進沉降預測，我們應用一維擴散模型來確定這些層的特性，從而提高我們預測的準確性。此外，我們還探討了導致東南亞城市沉降的因素，以解決這一問題。在第三章中，我們重點評估了一個露天礦場的潛在地質災害。首先，我們利用攝影測量立體技術生成了多時期的光學衛星影像數字高程模型，揭示了廣寧省 Nui Beo 礦場二十年來的高度變化。隨後，我們發現礦場的顯著景觀改變，通過2015年至2021年的InSAR觀測，導致了附近住宅區的沉降。在露天礦場中，InSAR時間序列中水相關信號和位移趨勢的混合，對識別邊坡穩定性構成挑戰。為了解決這一問題，我們對信號進行分解以獲取校正後的位移時間序列。我們應用反速度法對校正後的位移進行評估，以評估邊坡不穩定性並估算潛在失敗的時間。最終，我們確定了應進行風險管理和預警監測的熱點地區。;The increasing frequency of geological hazards, worsened by climate change, poses a major national concern for Vietnam. The country stands out as a hotspot for widespread occurrences of two dominant geohazards: land subsidence and landslides. These hazards often remain unnoticed until they trigger pronounced impacts on infrastructure and result in casualties. While ground observation proves ineffective in monitoring these hazards, integrating satellite-based data offers a potentially more efficient solution. The primary concerns with utilizing satellite data are its practical application and accuracy levels. The objective of this thesis is to demonstrate practical use cases of satellite data in the detection and management of anthropogenic geohazards. With appropriate processing, satellite measurement can achieve high accuracy for practical applications worldwide. Furthermore, combining ground observations with satellite analysis offers a more comprehensive study than relying solely on either method. This thesis presents two cases demonstrating the utilization of satellite data to study geohazards in Vietnam. Chapter 2 examines the long-term subsidence in Hanoi city using Interferometric Synthetic Aperture Radar (InSAR) technology. We process multi-SAR satellites imagery to provide a continuous observation of ground displacement between 2007 and 2018. The time series of displacement reveals the significant subsidence in suburban strongly correlated with the depletion of groundwater since 1990. Over two decades, the rapid loss of groundwater storage, combined with the gradual compaction of hydrogeological layers, has resulted in delayed subsidence. To improve the subsidence prediction, we apply a one-dimensional diffusion model to determine properties of these layers, thereby enhancing the accuracy our predictions. Furthermore, we explore factors contributing to sinking cities in Southeast Asia to address the issue. In Chapter 3, we focus on assessing potential geohazards at an open-pit mining site. Initially, we employ photogrammetric stereo techniques to generate multi-temporal Digital Elevation Models from optical satellite images, unveiling two decades of elevation changes with the emergence of manmade slopes at the Nui Beo mining site in Quang Ninh province. Subsequently, we discover that significant landscape alterations at the mining site have led to subsidence in nearby residential areas, as detected through InSAR between 2015 and 2021. In the open-pit, the mixing of water-related signals and displacement trends in InSAR time series poses a challenge for identifying slope stability. To address this, we decompose the signal to retrieve corrected displacement time series. We apply the inverse velocity method on corrected displacement to assess slope instability and estimate the timing of potential failure. Ultimately, we identify hotspots that should be monitored for risk management and warning purposes. |
