利用改良式雷達干涉技術偵測地表微量變化之研究

Abstract

依經濟部水利署監測資料顯示，彰化和雲林等地區近年續有地層下陷現象，影響人民生活安全及土地運用效益甚鉅，值予關注。為有效監測地表微量變化情形，其中以「合成孔徑雷達干涉技術（InSAR）」具有涵蓋面積廣、觀測精度高及不受天氣限制等優勢，已迅速成為常用的大地測量方法之一，並進階運用雷達差分干涉技術（DInSAR）、永久散射體干涉技術（PSInSAR），以兩幅或多幅同一地區不同時間（位置）所拍攝的影像進行干涉處理，計算相位差值，萃取影像間細微變化情況，獲得大面積的地變資訊。 對於穩定雷達波散射體密度較低的區域，例如平原地區或植被多的地區，因固定建築物較少，利用PSInSAR技術進行點位選取時，可能會因PS（Persistent Scatterers）點位不夠而無法進行分析。因此，本研究選用2016至2017年歐洲太空總署的Sentinel-1A衛星IW模式雷達影像，以雲林土庫鎮作為實驗區域，並提出改良式雷達差分干涉技術，將影像進行PSInSAR運算後，挑選永久散射體候選點（PSC），另運用相關係數、KS檢定、CVM檢定及AD檢定等統計分析方法，找出雷達波反射較為一致的同質像素（SHP），萃取分布散射體候選點（DSC），藉由共同處理上述點位資訊，獲取此區地層下陷之微量變化及形變速度，適時提供有關單位參用並採取預警措施應處。

According to the investigation report issued by the Water Resources Agency, Ministry of Economic Affairs, there exists continuous subsidence in the Changhua and Yunlin areas of Taiwan. This adverse situation has greatly degraded the performance of local land utilization and the livelihood of inhabitants. In order to effectively address this issue, one must monitor the changes of ground surface elevation, and the solution nowadays is the Interferometric Synthetic Aperture Radar （InSAR）technique which has the advantage of broad area coverage, high measurement accuracy, and all-weather operations. The Differential Interferometric Synthetic Aperture Radar （DInSAR） and Persistent Scatterers InSAR（PSInSAR） techniques take two or more SAR images collected from different times and viewing angles over the same geographical location into the interferometric processing where the phase differences and small changes between images are extracted to render the information on geographical changes. However, for areas where the density of stabilized radar wave scatterers is low, such as flat terrain or lush vegetation, due to few fixed structures, the DInSAR and PSInSAR techniques may not function satisfactorily. To alleviate this limitation, the present study improves the differential SAR interferometry where the SAR images are first processed using PSInSAR technique to determine Persistent scatterer candidates （PSC）and then the images are further processed using statistical analyses such as Kolmogorov-Smirnov test, Cramer-von Mises test, Anderson-Darling test and correlation coefficient to acquire Distributed scatterer candidates （DSC）. By processing these scatterer candidates, the present study establishes ground subsidence displacement and velocity for the Tuku Township of Yunlin County using Sentinel-1A SAR satellite images of European Space Agency from 2016 to 2017.