- 學位論文
差分干涉雷達地變形方向模糊的最佳分解
Optimal Resolving Directional Ambiguity in DInSAR Deformation Mapping
摘要
干涉雷達利用感測器收發訊號得到相位差對於地球表面資訊的萃取是很有用的技術。然而，有待討論的問題是單純用此技術得到的位移結果是相對於感測器的斜距方向，並非當地的垂直或水平位移，但若結合多觀測姿態就可以解決垂直和水平方向模糊的問題。 本研究提出一個基於加權最小平方的最佳求解法，結合昇、降的軌道模式求解三維地變形。而此研究方法和過去不同的是求解結果為三維方向位移量並非斜距方向位移量，除了能直接了解當地地變形情形外和地變形中心外，同時也便於與水準測量或GPS量測的結果來做直接比較。假設所有的測量無論是昇或降軌道模式是不相互影響的、各個資料面互相獨立，但在不同的相位值和同調性是相關的，於此最佳線性估計的求解方法便可實現。為了進一步的驗證由ALOS- PALSAR多組的干涉結果在求解地變形結果是可靠的，必須與精密水準測量的結果比較。 在良好的情況下使用ALOS/PALSAR資料經由本研究所提出之方法可以獲得準確的三維地變形量，靈敏度約10公分(L波段)。結合昇、降模式求解結果指出了在2007年到2010年的地變形主要是下陷的情形，而最大下陷速率是每年6.7公分(在最大下陷中心-溪湖)。在未來建議結合多衛星多觀測模式(多入射角、多方位方向)、長的和短的時間間距對地層下陷確實做長期監測。
並列摘要
Radar interferometry is a powerful technique for extracting information about the Earth’s surface using the phase difference between the signals arriving at the sensor antenna during repeated observations from the same platform. However, the problematic issue is that since radar senses and measures in slant range, it does not tell whether the displacement comes from horizontal or vertical direction, namely, a directional ambiguity inherently exists. A method based on weighted least squares is proposed to resolve it by integrating both ascending and descending orbits in order to determine whether the deformation is of uplift type or of subsidence. The resolved deformation pattern into three directions is useful to understand the movement mechanism and also to facilitate the direct comparison with ground leveling or measurements. Assuming that all the measurements, whether descending or ascending, are uncorrelated and independent but with different phase variances depending on the coherence, best linear unbiased estimator may be derived. To better confirm the subsidence map derived from series ALOS-PALSAR image data takes, comparisons with the precision leveling survey were made. Excellent agreement was obtained on both the spatial pattern and scale at centimeter resolution. Both ascending and descending mapping results indicate that the maximum subsidence rate in that period as about 6.7 cm/yr (at an epicenter). It was suggested that reliable deformation estimate is contingent upon the combination of long-term and short-term observations using dual-beam SAR image data sets preferably at different inclination angles.