以遙測方式獲取土地資訊具有涵蓋面積廣、花費較低廉以及資料易於取得等優勢，且隨著太空科技之發展，繞行固定軌道之衛星更具長期監測之能力。合成孔徑雷達影像係藉由雷達天線主動發射並接收其地表反射訊號來探求地表資訊，除可穿透雲、雨、濛、霧外，亦具有全天候監測能力。 但同一地區不同時間所拍攝之兩幅影像，由於拍攝間隔時間長、且基線長度不一，導致差分合成孔徑雷達干涉技術於應用在台灣這種地貌變化大且多植被的地方成果不佳。故本研究藉由偵測出自然界中的永久散射體於影像中的點位，進而得到該點位的長時間的地形變化量。 而過去永久散射體雷達干涉技術所使用的估算模型，是將整幅影像的視角與斜距以影像的中心像元之數值為代表，這種模型對觀測地區落於影像中心的情況是合理的簡化；但是台灣地區較多人口的區域是落於影像的邊緣地帶，採取中心像元之數值的方式就顯得不恰當。故本論文改用適應性模型依據各個像點給值，使得成果能更加精確。 論文成果發現，使用適應性模型對於觀測的成果確實有所改善，若是能夠改良候選點的挑選方式，使其在大範圍的呈現上能夠更密集，對於大範圍的變遷偵測之研究會更有幫助。 Repeat-pass satellite Interferometric SAR(InSAR) is potentially a unique tool for low cost precise Digital Terrain Model(DTM) generation and large-coverage surface deformation monitoring. Differential InSAR(DInSAR) may be used to monitor centimeter-scale surface displacements over large geographic extents. However, the method has limitations set by temporal and geometric decorrelation and atmospheric inhomogeneities. We present a Persistent Scatterer(PS) InSAR method for analyzing episodic crustal deformation in non-urban environments. This method for identifying PS pixels in a series of interferograms is based primarily on coherence index and finds low-amplitude pixels with phase stability that are not identified by the existing amplitude-based algorithm. The model which considers linear velocity deformation and DTM error usually using a centre target’s range distance. This method is fit for a test area nearby a centre area, but it’s not suitable for Taiwan’s urban. Taiwan’s urban usually locate in side of SAR images, that let a centre area’s value see not suitable. We chose adaptable range distance instead the fixed value. Using this method, we identified a density of 4.1 PS/km2. Even for longer than 500 m baselines, the PSInSAR technique can help build land subsidence maps better than differential InSAR does.
