合成孔徑雷達干涉法 (InSAR) 已被廣泛應用到檢測地表變形，如：地震、地層下陷和火山活動。永久散射體雷達差分干涉法 (PSInSAR) 是多時序合成孔徑雷達干涉法去測量來自地面特徵物的穩定反射相位，例如：建築物、橋樑或裸岩。然而，每個合成孔徑雷達 (SAR) 都有其特定的雷達波長、空間解析度、時間解析度和影像模式。此外，散射和穿透特性取決於與目標物間的交互作用。因此，本研究提出了一個全新概念來整合 Sentinel-1 和 TerraSAR-X PSInSAR 結果，並結合它們各自在陸地觀測中的優勢。此研究進行了兩個案例來探討所提出方法在不同情況下的可行性：雲林縣因抽取地下水而造成的地層下陷以及臺北盆地由自然以及人為活動所造成的地表活動。在研究結果中：雲林地區在 2014 年至 2016 年中，累積下陷量達到視衛星方向(Line-Of-Sight)上的-51 mm，而與 GNSS 站相比的均方根誤差(RMSE)為 6.2 mm。第二件案例分析顯示可以觀察到時間序列上沿山腳斷層的移動還有輕微下陷的臺北盆地。與 GNSS資料驗證可得到，相關係數(R)為 0.86，且RMSE為 6 mm。因此，我們可以得出下列結論：藉由兩種不同的雷達影像可以提高時間解析度以及利用兩者雷達影像的優點可以得到更佳的觀察地表變形的能力，從而修正了S1 PSInSAR高估的累積變形量且符合現地資料的趨勢，並保留了更完善的地表特徵。;Interferometric Synthetic Aperture Radar (InSAR) has been widely used to detect surface deformation such as earthquake, subsidence, and volcanic activities. Persistent Scatterer InSAR (PSInSAR) is one of the multi-temporal InSAR techniques to measure stable phase reflected by ground features, such as buildings, bridges, or other bright targets in radar wavelength. However, each SAR mission has its specified wavelength, resolution, revisit, and acquisition mode. Moreover, the scattering and penetration characteristics depend on the interaction with the target. Hence, this study proposes a novel concept to merge Sentinel-1 and TerraSAR-X PSInSAR results to leverage their individual advantages in land observations. Two experiments are conducted to explore the performance of the proposed method under different scenarios: Yunlin County with land subsidence and Taipei Basin with both natural and anthropogenic deformation. In our results, Yunlin area has a maximum cumulative subsidence reaching -51 mm in 2014 ~ 2016 in Line-Of-Sight (LOS) direction, while the root-mean-square error (RMSE) as compared against GNSS stations is 6.2 mm. Also, in Taipei Basin we observe a movement along the Shanchiao fault in the time series and slight settlement in the central of Taipei Basin. The correlation is 0.86 and RMSE is 6 mm in GNSS validation. We conclude that the mid-resolution SAR images calibrated by some temporally interleaved high-resolution images have better capability to observe land deformation that revising the overestimated S1 PSInSAR result, which shows a good agreement with on-site data, and to preserve detailed patterns.
