應用高解析度SAR影像於大型浮頂式油槽油位動態監測

High-Resolution SAR-Based Monitoring of Fuel-Level Dynamics in Large Floating-Roof Oil Tanks

合成孔徑雷達具備全天候、晝夜可觀之特性，能在不利光照與氣象條件下提供監測燃油基礎設施的可靠替代資訊。本文建立高解析度TerraSAR-X影像量化大型浮頂式油槽油位動態之操作框架。使用X波段聚束模式影像萃取油槽特徵散射弧，推估油槽直徑、槽壁高與浮頂高，並以次像素精度進行量測，同時構建專用的誤差傳播模型以評估幾何參數之不確定性。研究分析2022年12月至2023年1月之多時序TerraSAR-X影像，重建浮頂高的時間變化並換算對應體積變動。蒐集中油大林煉油廠之地面紀錄進行對照驗證，結果顯示兩者高度一致；即便在地面儀器失效或資料中斷時，該框架仍能有效辨識浮頂高度之時序變化。綜上所述，本研究以單一且穩定之感測來源，將可物理詮釋的SAR散射特徵連結至可量化之幾何反演，提供可靠、可重複且近即時油位監測能力，對於能源設施之持續觀測與能源管理應用具實務價值。

Synthetic Aperture Radar imagery enables all-weather, day-and-night observation and thus provides a robust alternative to optical imagery for monitoring petroleum infrastructure. This study establishes an operational framework to quantify fuel-level dynamics in large floating-roof oil tanks using multi-temporal, high-resolution TerraSAR-X data. Characteristic backscattering arcs are extracted from X-band spotlight images to derive tank diameter, wall height, and floating-roof elevation with sub-pixel precision, and a dedicated error-propagation model is developed to assess uncertainty in these geometric estimates. Multi-date TerraSAR-X acquisitions collected between December 2022 and January 2023 are analyzed to retrieve temporal variations in floating-roof height and convert them to volumetric change. Validation against CPC refinery records shows close agreement, and the method remains effective even when field instruments are unavailable. By linking physically interpretable SAR signatures to quantitative geometric retrievals from a single, stable sensor, the proposed framework delivers reliable, repeatable, and near–real-time assessments of fuel storage, offering a practical basis for continuous monitoring and energy-management applications.